JP4366826B2 - Engine exhaust purification system - Google Patents

Description

【００４８】
ここで、上記推定イオウ付着量Ｑｓｓは、ＮＯｘ触媒１７に実質的にＮＯｘがまったく吸蔵されなくなるときの付着量を１とし、まったく付着していないときの付着量をゼロとした場合におけるゼロから１までの値（割合）として上記式に代入される。
【００４９】
そして、ステップＳ３で、その修正ＮＯｘ吸蔵量Ｑｎｏｘｓが所定の吸蔵量Ｑｎｏｘ１以上となれば、ステップＳ４で、排ガスの空燃比を上記のように少なくともリーン運転時の空燃比よりもリッチ側にシフトする（例えば図３に例示したように理論空燃比とする。あるいはそれ以上にリッチとする）。
【００５０】
そして、ステップＳ５で、このリッチ化時間が所定時間ｔ２を経過したときに、ステップＳ６で、リッチ化を終了し、空燃比を元のリーンの状態に復帰させる。この場合、上記処理時間ｔ２は、リッチ化の度合との関係で上記所定ＮＯｘ吸蔵量Ｑｎｏｘ１が全て還元放出される時間に予め実験的に設定される。
【００５１】
ＮＯｘ触媒１７に吸蔵されたＮＯｘは、このように定期的あるいは不定期的に放出処理されるほか、運転者の運転操作に応じて、例えば、図３に鎖線ウ、エで示したように、運転領域がリーン運転領域Ａから理論空燃比運転領域Ｃに切り換わったときやリッチ運転領域Ｂに切り換わったときにもＮＯｘ触媒１７から放出され、該触媒１７のＮＯｘ吸蔵能力が回復する。
【００５２】
さらに、吸蔵ＮＯｘは、後述するイオウ被毒解消制御や該制御で用いられる推定イオウ付着量Ｑｓｓの補正制御においても空燃比がリッチ化されて同様にＮＯｘ触媒１７から放出される。
［ＮＯｘ触媒のイオウ被毒解消制御］
図５はＮＯｘ触媒１７のイオウ被毒解消制御の具体的動作の一例を示すタイムチャートである。ＮＯｘ触媒１７はバリウムを含有するので、燃料中のイオウ成分が触媒１７に付着するというイオウ被毒の問題がある。これをそのまま放置するとＮＯｘ浄化能力が低下するから、付着したイオウを放出させて触媒１７の浄化能力を回復させるための処理、すなわちイオウ被毒解消のためのＮＯｘ触媒１７の昇温が行なわれる。
【００５３】
このＮＯｘ触媒１７の昇温は、燃料の分割噴射や点火時期の遅角により排ガス温度を上昇させることで達成される。イオウ被毒は空燃比がリーンのときほど進行しやすく、リッチになるほど遅くなる。例えば、図中鎖線カで示したように、空燃比が理論空燃比のときはイオウ被毒がほとんど停滞し、図中鎖線キで示したように、空燃比が理論空燃比よりリッチになると却ってイオウの放出が始まる。したがって、このイオウ被毒解消制御では空燃比をリッチ化したうえでＮＯｘ触媒１７を昇温してイオウの放出を促進する。
【００５４】
また、リーン状態でのイオウの付着速度はＮＯｘの吸蔵速度に比べてはるかに遅い。したがって、このイオウ被毒解消制御は前述のＮＯｘ浄化制御に比べて実行される頻度が少なく、例えばこのイオウ被毒解消制御が一回行われる間にＮＯｘ浄化制御は何度も繰り返し行われる。図５と図３のタイムチャートでは時間間隔は必ずしも同一には表わしていない。
【００５５】
図６はイオウ被毒解消制御の具体的動作の一例を示すメインフローチャートである。このプログラムはイグニッションスイッチがオンである期間中、例えば、燃料噴射量制御、燃料噴射時期制御、点火時期制御等と無関係に繰り返し実行される。
【００５６】
まず、ステップＳ１１で、イオウ被毒時間Ｔｓを計測する。このイオウ被毒時間Ｔｓは、ＮＯｘ触媒１７のイオウ被毒がほぼ完全に解消した時刻からの経過時間である。すなわち、イオウ被毒がほぼ完全に解消したのち初めてこのステップＳ１１に進んだときは、イオウ被毒時間Ｔｓはあらためてゼロから起算され、それ以外のときは以前からの計測が続行される。
【００５７】
次に、ステップＳ１２で、リーン運転の連続継続時間Ｔｌｎを計測する。すなわち、図７に示すように、ステップＳ４１でリーン運転か否かを判定し、リーン運転のときはステップＳ４２でリーン継続時間Ｔｌｎを計測する一方、リーン運転でないときはステップＳ４３でリーン継続時間Ｔｌｎをリセットする。ただし、リーン継続時間Ｔｌｎをリセットしたときは、それまでステップＳ４２で計測したリーン継続時間Ｔｌｎの最後の値をメモリに格納する。これにより、リーン運転がＮＯｘ浄化制御によって中断したり、運転者の運転操作に応じて中断したり、あるいは後述する推定イオウ付着量Ｑｓｓの補正制御によって中断したときは、それまで行なわれていたリーン運転の連続継続時間Ｔｌｎがメモリに残存する。
【００５８】
メインフローに戻り、次に、ステップＳ１３で、ＮＯｘ触媒１７へのイオウ付着量Ｑｓｓ、すなわちＮＯｘ触媒１７のイオウ被毒状態を推定する。このイオウ付着量Ｑｓｓの推定は、図８のフローチャートに従って行われ、ステップＳ５１で、各種のデータを読み込んだうえで、ステップＳ５２で、前回イオウ付着量Ｑｓｓを推定してから今回イオウ付着量Ｑｓｓを推定するまでの間にインジェクタ６から噴射された燃料の量（燃料供給量）Ｑｆを算出する。
【００５９】
次に、ステップＳ５３で、上記燃料供給量Ｑｆに基づいて、前回イオウ付着量Ｑｓｓを推定してから今回イオウ付着量Ｑｓｓを推定するまでの間に増加したイオウ付着量（すなわち、単位時間当たりにＮＯｘ触媒１７に付着したイオウ付着量の瞬時値）の基本値（基本イオウ増加量）Ｑｓａを推定する。ここで、この基本イオウ増加量Ｑｓａは、図９に示すように、イオウの発生源である燃料の供給量Ｑｆにほぼ比例するように推定される。
【００６０】
次に、ステップＳ５４で、数２に従って、上記基本イオウ増加量Ｑｓａを後述する推定イオウ付着量Ｑｓｓの補正制御で算定される燃料のイオウ含有量予測値Ｋｆｓｓに基いて補正することにより補正基本イオウ増加量Ｑｓｂを算出する。
【００６１】
【数２】

【００６２】
ここで、上記予測値Ｋｆｓｓは、いままで考慮していた燃料のイオウ含有量を１とした場合の係数（割合）として上記式に代入される。すなわち、燃料のイオウ含有量にはバラツキがあるから、単に燃料供給量Ｑｆに基いて推定した上記の基本イオウ増加量Ｑｓａには実際のイオウ付着量との間にずれがあるので、そのずれを是正してイオウ付着量の推定精度の向上を図るのである。
【００６３】
なお、例えば同じ燃料を使っていてイオウ含有量が変わらなければ、この燃料イオウ含有量予測値（係数）Ｋｆｓｓはほぼ１となる。この予測値Ｋｆｓｓの算定については推定イオウ付着量の補正制御で詳しく述べる。
【００６４】
次に、ステップＳ５５〜Ｓ５７で、上記補正基本イオウ増加量Ｑｓｂに対する補正係数Ｋ１，Ｋ２，Ｋ３をそれぞれ設定する。すなわち、ステップＳ５５では排気温センサ２７で検出された排ガス温度Ｔｍｐに基づいて第１補正係数Ｋ１を、ステップＳ５６では上記ステップＳ４２で計測されたリーン運転継続時間Ｔｌｎに基づいて第２補正係数Ｋ２を、そして、ステップＳ５７では推定イオウ付着量の前回値（既イオウ付着量）Ｑｓｓ［ｉ−１］に基づいて第３補正係数Ｋ３を設定する。
【００６５】
ここで、第１補正係数Ｋ１は、図１０に示すように、所定の排ガス温度Ｔｍｐａをピークにそれより高くなってもまた低くなっても小さい値に設定される。また、第２補正係数Ｋ２は、図１１に示すように、所定のリーン運転継続時間Ｔｌｎａをピークにそれより長くなってもまた短くなっても小さい値に設定される。また、第３補正係数Ｋ３は、図１２に示すように、所定の既イオウ付着量Ｑｓｓａ［ｉ−１］より多くなったときに小さい値に設定される。
【００６６】
特に第２補正係数Ｋ２が上記のような特性であるのはおよそ次のような理由による。すなわち、リーン運転継続時間Ｔｌｎが長いときは、図１３に示すように、ＮＯｘ触媒１７に用いられているバリウムにすでに多量のＮＯｘやイオウが付着している。そのため、新規にイオウがバリウムに付着し難くなり、単位時間あたりのイオウ付着量が減少する。
【００６７】
また、リーン継続時間Ｔｌｎが短いときは、図１４に示すように、イオウとバリウムとの接触時間が短くなり、これらの間に強固な結合が生成し難くなる。そのため、次に理論空燃比運転やリッチ運転に切り換わったときに、イオウがＮＯｘ同様放出されやすくなり、やはり単位時間あたりのイオウ付着量が減少する。
【００６８】
つまり、時間の経過とともに累積のイオウ付着量Ｑｓｓは増加するにしても、例えば、同じ５分間のリーン運転であっても、１分間のリーン運転を５回行ったときに比べて、５分間のリーン運転（図１１に示す所定のリーン運転継続時間Ｔｌｎａに相当する）を１回行ったときのほうが、イオウ付着量の累積総量が多くなる。また、例えば、同じ２０分間のリーン運転であっても、２０分間のリーン運転を１回行ったときに比べて、５分間のリーン運転（同じく図１１に示す所定のリーン運転継続時間Ｔｌｎａに相当する）を４回行ったときのほうが、やはりイオウ付着量の累積総量が多くなる。
【００６９】
したがって、リーン運転を連続して行う時間を、上記所定時間Ｔｌｎａを避けてそれより短くするかあるいは長くすることが、ＮＯｘ触媒１７にイオウを付着させ難くする観点からは有利となる。例えば、前述のＮＯｘ浄化制御において吸蔵能力回復のための空燃比のリッチ化を定期的に行う場合に（図３における符号ア）、その周期Ｔ１を上記所定時間Ｔｌｎａより短く設定するとよい。
【００７０】
図８のサブフローに戻り、次に、ステップＳ５８で、数３に従って、補正基本イオウ増加量Ｑｓｂに上記第１〜第３補正係数Ｋ１，Ｋ２，Ｋ３を乗算することにより補正イオウ増加量Ｑｓｃを算出する。
【００７１】
【数３】

【００７２】
そして、ステップＳ５９で、数４に従って、上記補正イオウ増加量Ｑｓｃを既イオウ付着量Ｑｓｓ［ｉ−１］に加算することにより今回の推定イオウ付着量Ｑｓｓを算出する。
【００７３】
【数４】

【００７４】
なお、イオウ被毒は、排ガスが先に流入するＮＯｘ触媒１７の上流部分から優先して始まる。つまり、ＮＯｘ触媒１７にはイオウ成分が一様には付着せず、イオウ成分は排ガスの通過経路に沿って偏って付着する。それゆえ、イオウ付着量Ｑｓｓを推定するにあたり、ＮＯｘ触媒１７を排ガスの通過経路に沿って一般にｎ個のブロックに分割して考え、各ブロック毎にイオウ付着量Ｑｓｓ［ｊ］（ｊ＝１〜分割数ｎ）を推定して、その総和（Ｑｓｓ［１］＋…＋Ｑｓｓ［ｎ］）を触媒１７全体のイオウ付着量Ｑｓｓとすることができる。この場合、上流側のブロックのイオウ付着量Ｑｓｓ［ｊ］は下流側のブロックのそれに比べて多く推定される。あるいは、各ブロック毎に推定したイオウ付着量Ｑｓｓ［ｊ］の平均値（（Ｑｓｓ［１］＋…＋Ｑｓｓ［ｎ］）／ｎ）を触媒１７のイオウ付着量Ｑｓｓを代表する値として取り扱うこともできる。このような考え方については後のステップＳ２１のイオウ残存量Ｑｚｓの推定でさらに詳しく述べる。
【００７５】
メインフローに戻り、次に、ステップＳ１４で、フラグＦが１にセットされているか否かを判定する。このフラグＦは、要するに、ＮＯｘ触媒１７がイオウ被毒を解消すべき状態にあるかどうかの指標となるフラグであり、解消すべき状態にあるときは１にセットされ、ないときはゼロにリセットされる。
【００７６】
そして、このステップＳ１４で、フラグＦが１にセットされていないときは、ステップＳ１５において、ステップＳ１３で推定したイオウ付着量Ｑｓｓが予め設定された判定基準量Ｑｓ１以上か否かを判定し、判定基準量Ｑｓ１以上のときは、ステップＳ１６で、上記フラグＦを１にセットしたうえで、ステップＳ１７以下に進む。これに対し、ステップＳ１５で、推定イオウ付着量Ｑｓｓが判定基準量Ｑｓ１未満のときは、ステップＳ１２に戻る。また、ステップＳ１４で、フラグＦがすでに１にセットされているときは、ステップＳ１５，Ｓ１６をスキップして、直ちにステップＳ１７以下に進む。
【００７７】
ここで、上記判定基準量Ｑｓ１は、例えば、ＮＯｘ触媒１７のＮＯｘ浄化能力が８０％にまで低下するときのイオウの付着量等に設定される。すなわち、上記フラグＦは、ＮＯｘ触媒１７のＮＯｘ浄化能力がＮＯｘエミッションに影響を及ぼすほどに低下するぐらいの量のイオウ成分が該触媒１７に付着しているかどうかを表示するイオウ除去要求フラグである。
【００７８】
次に、ステップＳ１７では、ＮＯｘ触媒１７に対するイオウ放出処理実行時の排ガス温度Ｔｍｐの標準的な目標値である第１目標温度Ｔｍｐ１を設定する。この第１目標温度Ｔ１は、ステップＳ１１で計測したイオウ被毒時間Ｔｓに基づいて設定される。その場合に、図１５に示すように、第１目標温度Ｔｍｐ１は、例えば６５０℃と７００〜７５０℃との間の温度に設定され、イオウ被毒時間Ｔｓが長いときは、短いときに比べて、高い温度に設定される。また、第１目標温度Ｔｍｐ１を、さらに、ステップＳ１３で推定したイオウ付着量Ｑｓｓに基づいて補正してもよい。その場合も、図１５に準じて、第１目標温度Ｔｍｐ１は、推定イオウ付着量Ｑｓｓが多いときは、少ないときに比べて、高い温度に補正される。
【００７９】
次に、ステップＳ１８で、イオウ放出処理実行時の排ガス温度Ｔｍｐの最終目標値Ｔｍｐｓを設定する。すなわち、上記ステップＳ１７で設定した第１目標温度Ｔｍｐ１と、後述するステップＳ２９で設定した第２目標温度Ｔｍｐ２とを比較し、高いほうを最終的に目標排ガス温度Ｔｍｐｓとして選択するのである。
【００８０】
なお、第２目標温度Ｔｍｐ２は第１目標温度Ｔｍｐ１よりも一般に高い値に設定される。また、第１目標温度Ｔｍｐ１が常に設定されるのに対し、第２目標温度Ｔｍｐ２は常に設定されるとは限らない。つまり、第２目標温度Ｔｍｐ２が設定されていないときは第１目標温度Ｔｍｐ１が最終目標温度Ｔｍｐｓとされ、第２目標温度Ｔｍｐ２が設定されたときは第２目標温度Ｔｍｐ２が最終目標温度Ｔｍｐｓとされる傾向にある。
【００８１】
なお、ステップＳ２０のイオウ放出処理では、この最終目標排気ガス温度Ｔｍｐｓが実現するように、燃料の分割噴射や点火時期のリタードが行なわれ、その結果、ＮＯｘ触媒１７に付着したイオウ成分が除去される。その場合に、ステップＳ２０のイオウ放出処理の開始時における一定期間だけ、正規に設定された上記の最終目標排ガス温度Ｔｍｐｓよりも所定温度高い温度を最終目標温度としてもよい。これにより、排ガス温度Ｔｍｐの立ち上がりが促進され、イオウ成分が放出処理開始後速やかに除去され始める。図１５に示す鎖線タは、例えば第１目標温度Ｔｍｐ１が最終目標排ガス温度Ｔｍｐｓに選択された場合において、そのようにイオウ放出処理の開始時一定期間だけ採用する高い目標温度を例示するものである。
【００８２】
次いで、ステップＳ１９で、イオウ放出実行許可条件が満足されているか否かの判定を行う。ここでは、車速Ｖがイオウ放出処理を実行しても不具合のない所定車速Ｖ１以上であるか否かが判定される。
【００８３】
イオウ放出実行許可条件を車速で設定したのは、低車速時にイオウ放出処理のために分割噴射や点火時期のリタードを行っても、排ガス温度Ｔｍｐがイオウ放出可能温度である目標温度Ｔｍｐｓ（例えばこの実施の形態でいうと、図１５よりＴｍｐｓ≧６５０℃）まで上昇せず、効率のよいイオウ放出処理が実現しないこと、低車速時にイオウ放出処理のために分割噴射や点火時期のリタードを行うと、エンジン出力が過度に不安定化することなどの理由による。
【００８４】
しかし、イオウ放出実行許可条件は、これに限られず、一般に、イオウ放出処理を実行することにより何らかの不具合が随伴する、あるいは随伴する不具合が相対的に大きくなるような状況を排除する目的で他のパラメータを用いて設定してもよい。
【００８５】
そして、車速Ｖが上記所定車速Ｖ１以上であると判定した場合は、ステップＳ２０に進んでイオウ放出処理を実行する。一方、車速Ｖが上記所定車速Ｖ１以上でないと判定した場合には、イオウ放出処理を実行せずにステップＳ２８に進む。
【００８６】
ステップＳ２０のイオウ放出処理は、図１６に示すフローチャートに従って行なわれる。まず、ステップＳ６１で、現在用いている図２に示す通常時の目標空燃比マップを、例えば図１７に示すようなイオウ放出処理時の目標空燃比マップに切り換える。ここで、このイオウ放出処理時の目標空燃比マップにおいては、全運転領域が理論空燃比領域Ｃとされている。すなわち、現在行なわれているリーン運転を禁止して強制的に理論空燃比運転とすると共に、燃料カット（Ｆ／Ｃ）を禁止するのである。理論空燃比運転とすることにより、排ガス中のＣＯ，ＨＣ濃度が高くなり、ＮＯｘ触媒１７に付着したイオウ成分が放出され易い環境が生成される。また、燃料カットを禁止することにより、燃料が常に供給され、燃料噴射制御を利用したイオウ放出処理が安定して遂行される。
【００８７】
次いで、ステップＳ６２で、現在の運転状態が分割噴射領域にあるか否かを判定する。つまり、現在のエンジン１の運転状態が、イオウ放出のための排ガスの昇温を燃料の分割噴射によって行う運転領域内にあるかどうかを判定するのである。ここで、分割噴射領域は、図１７のイオウ放出処理時のマップにおいて、中回転中負荷領域（斜線を施した全部分）に設定されている。
【００８８】
分割噴射領域である場合は、ステップＳ６３，Ｓ６４で、燃料の後期噴射量Ｑｆｋｓおよび後期噴射時期θｆｋｓを設定する。ここで、後期噴射とは、図２の通常時マップと同様、燃料を圧縮行程中に噴射することであり、燃料を吸気行程中に噴射する前期噴射に比べて燃料の気化霧化が進まず、未燃成分が増加する。
【００８９】
ステップＳ６３での後期噴射量Ｑｆｋｓの設定は、数５に従って、吸入空気量などから別途定められる全燃料噴射量Ｑｆｔｓに、後期噴射量係数Ｋ４，Ｋ５（いずれも１未満の値）を乗算することにより行なわれる。
【００９０】
【数５】

【００９１】
ここで、第４の補正係数である第１後期噴射量係数Ｋ４は、図１８に示すように、目標排ガス温度Ｔｍｐｓ（イオウ放出処理の開始時一定期間は目標排ガス温度Ｔｍｐｓを所定温度だけ高くする場合も含む）に基いて定められ、目標温度Ｔｍｐｓが高いほど大きな値に設定される。また、第５の補正係数である第２後期噴射量係数Ｋ５は、図１９に示すように、実排ガス温度Ｔｍｐに基いて定められ、実排ガス温度Ｔｍｐが低いほど大きな値に設定される。したがって、目標排ガス温度Ｔｍｐｓが高いほど、また、実排ガス温度Ｔｍｐが低いほど、後期噴射量Ｑｆｋｓが多くなる。その結果、燃料の未燃成分が一層増加し、ＮＯｘ触媒１７がより速やかに昇温される。
【００９２】
一方、ステップＳ６４では、後期噴射時期θｆｋｓは、エンジン負荷が低いときほどより遅くなるように設定される。したがって、低負荷時ほど、後期噴射時期θｆｋｓが遅くなり、燃料の気化霧化がなお一層進まず、未燃成分がより増加して、ＮＯｘ触媒１７がより速やかに昇温される。
【００９３】
このようにして設定された後期噴射量Ｑｆｋｓ及び後期噴射時期θｆｋｓはインジェクタ６に制御信号として出力され、エンジン１に対する燃料噴射量及び燃料噴射時期の制御に用いられる。
【００９４】
次いで、ステップＳ６５で、リタード（遅角）制御領域か否かを判定する。つまり、現在のエンジン１の運転領域が、イオウ放出のための排ガスの昇温を点火時期のリタードによって行う運転領域内にあるかどうかを判定するのである。ここで、リタード制御領域は、図１７のイオウ放出処理時のマップにおいて、分割噴射領域のうちの低負荷側の領域（ラインＤより下の密に斜線を施した部分）に重ねて設定されている。すなわち、このリタード制御領域では、燃料の分割噴射と点火時期のリタードとの両方が行なわれる。
【００９５】
リタード制御領域である場合は、ステップＳ６６で、リタード量を設定する。特に、このステップＳ６６では、リタード量は、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓに収束するようにフィードバック制御される。
【００９６】
ステップＳ６６のリタード量の設定およびそのフィードバック制御は図２０に示すフローチャートに従って行なわれる。まず、ステップＳ７１で、排気温センサ２７で検出される実排ガス温度Ｔｍｐを読み込んだうえで、ステップＳ７２で、リタード制御の実行開始時か否かを判定する。
【００９７】
そして、リタード制御の開始時の最初の一回だけステップ７３に進み、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓ以上であるか否かを判定する。その結果、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓ以上のときは、ステップＳ７４で、基本リタード量として、予め設定された高温用リタード量を設定する。一方、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓ以上でないときには、ステップＳ７５で、基本リタード量として、予め設定された低温用リタード量を設定する。
【００９８】
ここで、低温用リタード量は、高温用リタード量に比べて、大きなリタード量に設定されている。これにより、排ガス温度Ｔｍｐが低いときは、高いときに比べて、より程度の大きい昇温が図られる。
【００９９】
一方、リタード制御の開始時でないとき、つまりリタード制御がすでに開始しているときは、ステップＳ７６に直接進み、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓ以上であるか否かを判定する。その結果、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓ以上のときは、さらにステップＳ７７で、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓよりヒステリシスの増分ΔＴｍｐだけ高い温度（Ｔｍｐｓ＋ΔＴｍｐ）以上であるか否かを判定する。そして、そうであるときには、ステップＳ７８で、リタード量を所定量だけ減量する。
【０１００】
一方、ステップＳ７７で、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓ以上で、ヒステリシスの増分ΔＴｍｐだけ高い温度（Ｔｍｐｓ＋ΔＴｍｐ）未満であるときには、リタード量をそのまま維持する。
【０１０１】
さらに、ステップＳ７６で、実排ガス温度Ｔｍｐが目標温度Ｔｍｐｓ以上でないと判定されたときは、ステップＳ７９で、リタード量を所定量だけ増量する。
【０１０２】
このようにして設定された点火時期のリタード量は点火プラグ５の点火回路３５に制御信号として出力され、エンジン１に対する点火時期の制御に用いられる。
【０１０３】
以上により、図１６のイオウ放出処理のフローチャートおよび図１７のイオウ放出処理時の空燃比マップから明らかなように、エンジン１の運転状態が中回転中負荷領域のうちの高負荷側にある場合（ステップＳ６２で肯定的判定およびステップＳ６５で否定的判定の場合）は、排ガスの昇温のために分割噴射のみ行われる。また、エンジン１の運転状態が中回転中負荷領域のうちの低負荷側にある場合（ステップＳ６２で肯定的判定およびステップＳ６５でも肯定的判定の場合）は、分割噴射と点火時期のリタードとが併せて行われる。
【０１０４】
しかし、エンジン１の運転状態がもともと中回転中負荷領域にない場合（ステップＳ６２で否定的判定の場合）には、分割噴射も点火時期のリタードも行なわれない。つまり、低回転低負荷時、および高回転高負荷時には、排ガスの昇温が実質的に行なわれないのである。
【０１０５】
これは、低回転低負荷領域では、排ガス温度Ｔｍｐがもともと低く、排ガスの昇温を行ってもイオウ放出処理実行可能な目標温度Ｔｍｐｓまで上昇しない可能性があると共に、排ガスの昇温を行うとエンジン１の出力状態が不安定となる可能性があるから、そのような無駄を回避する目的である。
【０１０６】
一方、高回転高負荷領域では、排ガス温度Ｔｍｐがもともと高く、わざわざ排ガスの昇温を行わなくてもイオウ放出が行なわれる可能性があると共に、排ガスの昇温を行うとＮＯｘ吸収材ひいてはＮＯｘ触媒１７が過度に高温となって損傷する可能性があるから、やはりそのような無駄を回避する目的である。
【０１０７】
そして、このような対策を講じることにより、分割噴射やリタードを用いた排ガスの昇温に随伴して発生する燃費の悪化やトルクの低下などの不具合もまた必要最小限に抑制される。
【０１０８】
さらに、エンジン１の運転状態が中回転中負荷領域にあり、排ガスの昇温を行う場合においても、基本的には、排ガスの昇温を分割噴射で行い、排ガス温度が相対的に低く、より大きな程度に昇温する必要のある低負荷時においてのみ、リタードも併せて行うから、リタードによるトルクの低下の不具合がやはり必要最小限に抑制される。
【０１０９】
メインフローに戻り、次いで、ステップＳ２１で、ＮＯｘ触媒１７に残存しているイオウ残存量Ｑｚｓを推定する。このイオウ残存量Ｑｚｓの推定は、図２１のフローチャートに従って行なわれ、まず、ステップＳ８１で、排ガス温度Ｔｍｐがイオウ放出可能温度の最低値（例えばこの実施の形態でいうと６５０℃）以上であるか否かを判定し、その結果、排ガス温度Ｔｍｐが６５０℃未満のときは、ステップＳ８２でイオウ放出処理時間Ｔｒをリセットする一方、６５０℃以上のときは、ステップＳ８３でイオウ放出処理時間Ｔｒを計測する。
【０１１０】
つまり、上記ステップＳ１７で、図１５を用いて第１目標温度Ｔｍｐ１を設定したときと同様に、排ガス温度ＴｍｐひいてはＮＯｘ触媒１７の温度が最低限６５０℃以上のときにイオウ成分が放出され得るものとして、この６５０℃をイオウ成分の放出可能温度の最低温度としているのである。もちろん、例えばＮＯｘ触媒１７に流入する排ガスの空燃比などのその他の条件により、このイオウ放出可能温度はいろいろな値に設定され得るものである。
【０１１１】
次いで、ステップＳ８４以下において、上記イオウ放出処理時間Ｔｒと、排気温センサ２７によって検出された排ガス温度Ｔｍｐとに基づいて、イオウ放出量を推定し、該イオウ放出量からイオウ残存量Ｑｚｓを推定する。
【０１１２】
まず、このステップＳ８４以下で行うイオウ残存量Ｑｚｓの推定動作の概略を説明する。ステップＳ１３のイオウ付着量Ｑｓｓの推定動作で述べたように、排ガスは図２２に示すようにＮＯｘ触媒１７の上流部分から先に流れ込む。したがって、上記ステップＳ２０で実行するイオウ放出処理においても、イオウ放出可能温度にまで昇温された排ガスはＮＯｘ触媒１７の上流部分から先に流れ込み、該上流部分が先に昇温されて、イオウの放出は該上流部分において優先して始まる。つまり、ＮＯｘ触媒１７は一様にはイオウ成分が放出除去されず、イオウ成分は排ガスの通過経路に沿って偏って除去され、偏って残存する。
【０１１３】
それゆえ、イオウ放出量ないし残存量Ｑｚｓを推定するにあたり、ＮＯｘ触媒１７を、図２２に示すように、排ガスの通過経路に沿って一般にｎ個のブロック（図例では１０個）に分割して考え、各ブロック毎にイオウ残存量Ｑｚｓ［ｊ］（ｊ＝１〜分割数ｎ）を推定して、その総和（Ｑｚｓ［１］＋…＋Ｑｚｓ［ｎ］）を触媒１７全体のイオウ残存量Ｑｚｓとするのである。これにより、精度のよいイオウ放出量ないしイオウ残存量の推定を図ることが可能となる。明らかに、図２２において左側の上流側ブロックのイオウ残存量Ｑｚｓ［ｊ］は、下流側ブロックのそれに比べて少なく推定される。
【０１１４】
なお、各ブロック毎に推定したイオウ残存量Ｑｚｓ［ｊ］の総和（Ｑｚｓ［１］＋…＋Ｑｚｓ［ｎ］）に代えて、各ブロック毎に推定したイオウ残存量Ｑｚｓ［ｊ］の平均値（（Ｑｚｓ［１］＋…＋Ｑｚｓ［ｎ］）／ｎ）を触媒１７のイオウ残存量Ｑｚｓを代表する値として取り扱うようにしてもよい。
【０１１５】
この実施の形態においては、一例として、図２２に示すように、ＮＯｘ触媒１７を１０個のブロックに分割して考えている（ｊ＝１〜１０）。ステップＳ８４では、イオウ放出処理時間Ｔｒと排ガス温度Ｔｍｐとに基づいて、全ＮＯｘ触媒１７の容積のうち、イオウ成分が完全に放出除去された領域（完全除去領域）ＳＡの割合（％）、換言すればイオウ成分が完全に放出除去されたブロックの数を、図２３に示すような特性のマップから設定する。図２３に示すように、完全除去領域ＳＡは、放出処理時間Ｔｒが長くなるほど、また、排ガス温度Ｔｍｐが高くなるほど大きくなる。つまり、完全除去領域ＳＡに属するブロックの数が多くなる。
【０１１６】
図２３のマップは、例えば、排ガス温度Ｔｍｐが７００℃のときは、イオウ放出処理時間Ｔｒが時刻ｔ１１に到達したときに、また、排ガス温度Ｔｍｐが６５０℃のときは、それより遅い時刻ｔ１２に到達したときに、１０個のブロック全部からイオウが放出され、完全除去領域ＳＡが１００％となることを示している。
【０１１７】
次のステップＳ８５では、同じくイオウ放出処理時間Ｔｒと排ガス温度Ｔｍｐとに基づいて、ＮＯｘ触媒１７のうち、イオウ成分が部分的に放出除去された領域（部分除去領域）ＳＢの割合（％）（ブロック数）を、図２４に示すような特性のマップから設定する。図２４に示すように、部分除去領域ＳＢは、少なくとも放出処理時間Ｔｒが長くなるに従って小さくなり、上記の時刻ｔ１１又はｔ１２で、完全除去領域ＳＡが１００％となったときに消滅する。
【０１１８】
なお、イオウ成分が全く除去されていない未除去領域ＳＣの割合（％）（ブロック数）は、上記完全除去領域ＳＡの割合と部分除去領域ＳＢの割合とを加えた値を、全ＮＯｘ触媒１７の容積から減算することにより求められる。
【０１１９】
次のステップＳ８６では、同じくイオウ放出処理時間Ｔｒと排ガス温度Ｔｍｐとに基づいて、部分除去領域ＳＢにおけるイオウ除去率α（％）を、図２５に示すような特性のマップから設定する。図２５に示すように、部分除去領域ＳＢにおけるイオウ除去率α（％）は、この実施の形態においては、放出処理時間Ｔｒとは無関係に、排ガス温度Ｔｍｐが高くなるほど大きくなるように設定されている。
【０１２０】
なお、部分除去領域ＳＢにおけるイオウ残存率β（％）は、（１００−α）％であり、また、完全除去領域ＳＡにおけるイオウ除去率は１００％（イオウ残存率は０％）、未除去領域ＳＣにおけるイオウ除去率は０％（イオウ残存率は１００％）である。
【０１２１】
次のステップＳ８７では、以上得られた各データから最終的にＮＯｘ触媒１７全体のイオウ残存量Ｑｚｓを算出する。そのためには、例えば、各領域ＳＡ，ＳＢ，ＳＣ毎に放出されたイオウ成分の量を推定し、その総和を求め、そして、その値をステップＳ１３で推定したイオウ付着量Ｑｓｓから減算する。あるいは、各領域ＳＡ，ＳＢ，ＳＣ毎に放出されたイオウ成分の量を推定し、その値を各領域ＳＡ，ＳＢ，ＳＣ毎に付着したイオウ成分の量（例えば「Ｑｓｓ／ｊ（ブロックの数）」とする）から減算し、そして、その総和を求めてもよい。
【０１２２】
ここで、完全除去領域ＳＡに属するブロック（上流側ブロック）の各イオウ残存量Ｑｚｓ［ｊ］はゼロであり、未除去領域ＳＣに属するブロック（下流側ブロック）の各イオウ残存量Ｑｚｓ［ｊ］はステップＳ１３で推定された各イオウ付着量（Ｑｓｓ／ｊ）のままであり、そして、部分除去領域ＳＢに属するブロック（中央部のブロック）の各イオウ残存量Ｑｚｓ［ｊ］は、（（Ｑｓｓ／ｊ）×（１００−α））であるから、ＮＯｘ触媒１７全体の総イオウ残存量Ｑｚｓは、例えば数６に従って求められる。
【０１２３】
【数６】

【０１２４】
なお、式中に記した「ＳＢ」、「ＳＣ」は、それぞれＮＯｘ触媒１７のうちの部分除去領域、未除去領域の割合（％）を示すものとする。そして、同じく式中に記した「（ｊ×ＳＢ／１００）」，「（ｊ×ＳＣ／１００）」は、それぞれ部分除去領域、未除去領域に属するブロックの数を示すものとする。
【０１２５】
ここで求められたＮＯｘ触媒１７全体の総イオウ放出量または総イオウ残存量Ｑｚｓは、一回一回のイオウ放出処理について放出されたイオウ放出量またはイオウ残存量であって、各イオウ放出処理の結果放出されたイオウ放出量またはイオウ残存量の累積量ではない。そして、該放出量あるいは残存量を求めるためのパラメータであるイオウ放出処理時間Ｔｒは、ステップＳ８１，Ｓ８２で、排ガス温度Ｔｍｐがイオウ放出可能温度（６５０℃）以上とならなくなったときにリセットされるから、イオウ放出が連続して実行された時間を示し、途切れ途切れにイオウ放出が行なわれた時間の累積ではない。
【０１２６】
例えば、１分間のイオウ放出処理を５回行っても、採用されるイオウ放出処理時間Ｔｒは、各イオウ放出処理すべてにおいて１分間であるから、図２３〜図２５に示すマップからは、排気ガス温度Ｔｍｐが同じであれば、常に同じ完全除去領域ＳＡの割合、同じ部分除去領域ＳＢの割合、および同じ部分除去領域ＳＢにおけるイオウ除去率αが読み出される。したがって、結果的に、同じイオウ放出量、同じイオウ残存量が算定されることになり、イオウ放出量またはイオウ残存量が更新されることはない。つまり、各１分間のイオウ放出処理の結果算定された各イオウ放出量またはイオウ残存量をすべて累積総和するものではないのである。
【０１２７】
したがって、例えば、同じ５分間のイオウ放出処理であっても、このように１分間のイオウ放出処理を５回行ったときに比べて、５分間のイオウ放出処理を１回行ったときの方が、イオウ放出量が多くなり、イオウ残存量が少なくなる。
【０１２８】
そして、例えば、１分間のイオウ放出処理が行なわれたのちに、次に２分間のイオウ放出処理が行なわれたときに、イオウ放出量あるいは残存量の値が、その２分間のイオウ放出の結果算定された、より大きいイオウ放出量、あるいは、より小さい残存量の値に更新されることになる。
【０１２９】
つまり、イオウ放出処理は、前述したように、昇温された排ガスが先に流れ込むＮＯｘ触媒１７の上流部分から優先して始まり、加熱が遅れるＮＯｘ触媒１７の下流部分は時間がある程度経過しないとなかなかイオウが除去されない。そして、一回のイオウ放出処理が終了して、次に再びイオウ放出処理が行われるまでの間に、ＮＯｘ触媒１７はまた温度が低下するから、この二回目のイオウ放出処理においても、再度、ＮＯｘ触媒１７は上流部分から加熱され始める。したがって、このときに、また前回と同じ、またはそれ以下の時間で、この二回目のイオウ放出処理が終了すると、この二回目のイオウ放出処理では、全く何も新しくイオウが除去されることのないまま終わることになるのである。したがって、イオウ放出連続時間Ｔｒが短くなればなるほど、全体のイオウ除去率が低下し、イオウ放出連続時間Ｔｒが長くなればなるほど、全体のイオウ除去率が向上するということができる。
【０１３０】
ところで、一般に、リッチ運転時及び理論空燃比運転時は、リーン運転時に比べて、排ガス温度Ｔｍｐが高くなり、自然にイオウ放出可能温度（例えば６５０℃）またはそれ以上にまで上昇することがある。つまり、ステップＳ２０のイオウ放出処理を行わなくても、リッチ運転時又は理論空燃比運転時は、イオウ放出処理をしているのと同じ効果が得られる場合があるのである（図５における符号カ、キ）。
【０１３１】
したがって、例えば、リッチ運転時又は理論空燃比運転時（λ≦１）であって、且つ、排ガス温度がイオウ放出可能温度以上のとき（Ｔｍｐ≧６５０℃）は、イオウ放出処理時における上記ステップＳ２１のイオウ放出量ないしイオウ残存量Ｑｚｓの推定手法と同じ手法を用いて、放出除去されたイオウ成分の量を推定し、それをステップＳ１３で推定したイオウ付着量Ｑｓｓから減じる等して該推定イオウ付着量Ｑｓｓを修正するとよい。
【０１３２】
これにより、イオウ付着量Ｑｓｓの推定精度がより向上し、無駄に早い時期にイオウ放出処理が開始されたり、あるいは無駄に長くイオウ放出処理が続行されるという、実際より多い量のイオウ付着量Ｑｓｓを推定した場合に生じる不具合が回避されることになる。
【０１３３】
メインフローに戻り、次に、ステップＳ２２で、上記のステップＳ２１で求めた総イオウ残存量Ｑｚｓがゼロであるか否かを判定する。つまり、付着したイオウ成分がすべて除去され、ＮＯｘ触媒１７のＮＯｘ吸蔵能力ないし浄化能力が完全に回復したか否かを判定するのである。
【０１３４】
そして、イオウ残存量Ｑｚｓがゼロとなるまでは、ステップＳ２３で、車速Ｖが第二のイオウ放出実行許可条件としての第二の判定用所定車速Ｖ２以上であることを確認しつつ、ステップＳ２０のイオウ放出処理およびステップＳ２１のイオウ残存量Ｑｚｓの推定を繰り返す。ここで、図５に示すように、上記第二所定車速Ｖ２は、第一所定車速Ｖ１よりも低い車速に設定されている。これにより、ステップＳ１９で肯定的判定がなされて、いったんイオウ放出処理が開始したのちは、車速が第一所定車速Ｖ１程度にまで低下しても、該イオウ放出処理が解除され難くなり、排ガス温度の低下が最小限度にくいとめられる。
【０１３５】
そして、ステップＳ２２でイオウ残存量Ｑｚｓがゼロであると判定されたときに、ステップＳ２７に進んで、イオウ除去要求フラグＦを０にリセットし、また、イオウ付着経過時間を示すイオウ被毒時間Ｔｓ、および後述する運転特性カウンタをそれぞれリセットする。そののち、ステップＳ１１にリターンして、イオウ被毒時間Ｔｓを再び最初から計測していくことになる。
【０１３６】
ここで、ステップＳ２７に到達するにはいろいろなパターンがある。そのうちの最善のパターンは、第一、第二の両イオウ放出実行許可条件（車速Ｖが判定用所定車速Ｖ１，Ｖ２以上であること）が、ステップＳ１６でフラグＦが１にセットされた当初から継続して満足され続け、ステップＳ１９およびステップＳ２３で一度も否定的判定がなされず、イオウ放出処理が当初から円満に遂行され、そして終了した場合である（図５における符号ク）。このパターンにおいては、ＮＯｘ触媒１７の機能低下が最も低く抑制される。この第一のパターンは、ステップＳ１９およびステップＳ２３で常に肯定的判定がなされることから明らかなように、走行中、より多くの時間、より高い車速を維持する傾向の多い運転者の場合に該当する確率が大きい。
【０１３７】
残るステップＳ２４〜Ｓ２６、およびステップＳ２８〜Ｓ３０は、そのような上記第一パターン以外の経路を経るときの処置である。
【０１３８】
すなわち、上記ステップＳ２３で、車速Ｖが第二所定車速Ｖ２以上でないと判定されたとき、つまり、イオウ放出処理実行中にかなり減速状態となり、分割噴射やリタード制御を実行しても、排ガス温度Ｔｍｐが目標温度Ｔｍｐｓまで上昇せず、効率のよいイオウ放出処理が続けられなくなったとき、あるいはエンジン出力が不安定になる可能性が生じたときは（図５における符号ケ）、ステップＳ２４，Ｓ２５を実行して、排ガス昇温のための燃料の分割噴射および点火時期のリタード制御を解除し、また、図１７のイオウ放出処理時の空燃比マップから図２の通常時の空燃比マップに切り換えて、強制的な理論空燃比運転を解除し、燃料カット（Ｆ／Ｃ）を復活させる。
【０１３９】
そのうえで、ステップＳ２６で、イオウ残存量Ｑｚｓが所定のイオウ放出処理終了許可判定量Ｑｓ２以下であるか否かを判定する。ここで、この終了許可判定量Ｑｓ２は、ゼロではない比較的小さい値に設定されている。その結果、イオウ残存量Ｑｚｓが上記終了許可判定量Ｑｓ２以下である場合（図５における破線コ）は、ステップＳ２７に進んで、イオウ放出処理が当初から円満に遂行され、そして終了した場合と同じに扱う。すなわち、イオウ除去要求フラグＦ、イオウ被毒時間Ｔｓ、および後述する運転特性カウンタをそれぞれリセットし、ステップＳ１１にリターンしたときには、イオウ被毒時間Ｔｓを再び最初から計測していくのである。
【０１４０】
これに対し、ステップＳ２６で、イオウ残存量Ｑｚｓがイオウ放出処理終了許可判定量Ｑｓ２以下でないとき（図５における実線サ）、つまり、まだ多くのイオウ成分が除去されずに残っているときは、前述のステップＳ１９でイオウ放出実行許可条件が満足されていないと判定されたときと同様に、ステップＳ２８以下に進む。
【０１４１】
ステップＳ２８では、運転特性カウンタをカウントアップする。すなわち、ステップＳ１９で、イオウ放出実行許可条件が運転者の運転操作により満足されず、その結果、ＮＯｘ触媒１７のイオウ付着量Ｑｓｓが所定量Ｑｓ１以上となってイオウを除去する必要が生じていても、イオウ放出処理を実質的に開始できない状況や、あるいは、ステップＳ２３で、第二のイオウ放出実行許可条件が運転者の運転操作により満足されず、その結果、いったん開始したイオウ放出処理を、未だ多くのイオウ成分が残存しているけれども（Ｑｚｓ＞Ｑｓ２）、途中で中断しなければならない状況が発生した回数を計測するのである。
【０１４２】
したがって、この運転特性カウンタのカウント数が多いほど、運転者の運転操作は、より多くの時間をより低い車速で走行する傾向のものであるといえる。換言すれば、フラグＦが１にセットされたのちイオウ放出処理が速やかに円満に終了し難い運転特性のものであるといえる。
【０１４３】
次いで、ステップＳ２９で、上記運転特性カウンタのカウント数に基づいて、イオウ放出処理実行時の排ガス温度Ｔｍｐの補強的な目標値である第２目標温度Ｔｍｐ２を設定する。この第２目標温度Ｔｍｐ２は、上記カウント数が多いほど、つまり、イオウ放出処理が速やかに円満に終了し難いときほど、例えば７００〜７５０℃などに高く設定される。
【０１４４】
そして、上記ステップＳ１８において、この補強的な第２目標温度Ｔｍｐ２と、ステップＳ１７で設定される標準的な第１目標温度Ｔｍｐ１とのうち、高いほうの温度が最終的な目標排ガス温度Ｔｍｐｓに選択されるから（図５の符号シ参照）、これにより、実行され難いイオウ放出処理がいったん実行されたときには、排ガスないしＮＯｘ触媒１７がより高い温度に昇温されて、イオウ成分が効率よく速やかに放出除去され、イオウ放出処理が短時間で確実に円満に終了することが図られることになる。
【０１４５】
そして、さらに、ステップＳ３０で、ＮＯｘ触媒１７のイオウ被毒の抑制、およびＮＯｘ吸蔵の抑制を図る。つまり、ＮＯｘ触媒１７にはかなりな量のイオウ成分が付着していて（Ｑｓｓ≧Ｑｓ１またはＱｚｓ＞Ｑｓ２）、ＮＯｘ吸蔵能力および触媒機能が低下しているのであるから、このような状態でリーン運転を行うと、排ガス中に多量に含まれるＮＯｘが触媒１７に吸蔵されずに大気中に放出されることになる。
【０１４６】
したがって、例えば、図２に示す目標空燃比マップにおけるリーン運転領域Ａを縮小して、リーン運転自体を規制又は禁止することにより、排気ガス中にＮＯｘが多量に含まれないようにし、ＮＯｘ触媒１７の触媒機能が低下していることに起因するＮＯｘの大気中への放出を抑制するのである。
【０１４７】
また、リーン運転を規制、禁止することにより、同時に、イオウの付着速度も鈍化され、さらにイオウがＮＯｘ触媒１７に付着するのを抑制することもできる（イオウ被毒の抑制）。図５において、実線サの傾き（推定イオウ付着量Ｑｓｓの傾き）が小さいのはこのことを示している。
【０１４８】
その場合に、そのリーン領域Ａを縮小する程度は、運転特性カウンタのカウント数に応じて決定するとより好ましい。該カウント数の大小がＮＯｘ触媒１７の触媒機能の低下の程度を表わしているからである。明らかに、該カウント数が大きいほど、触媒機能がより低下しているから、リーン運転領域Ａをより大きく縮小する。
【０１４９】
そして、運転特性カウンタがステップＳ２７でリセットされないうちは、イオウ除去要求フラグＦも継続して１にセットされているから、リターンしたときに、ステップＳ１５，Ｓ１６をスキップし、ステップＳ１９で第１のイオウ放出実行許可条件が満足されたときに、イオウ放出処理が実行されることになる。しかし、例えば、イオウ除去要求フラグＦが１にセットされている状態で、運転者の低車速走行傾向によってイオウ放出が一向に最後まで（Ｑｚｓ＝０まで、あるいはＱｚｓ≦Ｑｓ２まで）遂行されないときは、イグニッションスイッチがＯＦＦされたのち、またはＯＮされたのちに、強制的にイオウ放出処理を実行するようにしてもよい。これにより、走行する必要のない場合に、運転者の走行の支障とならずに、イオウを放出処理することができる。特に、エンジン始動時は、空燃比をリッチにして暖機運転するのが通例であるから、そのことを利用して効率よくイオウ放出が行える。
【０１５０】
ただし、この場合、例えばブザーや警告灯などの警告手段によって、運転者にイオウ放出処理を実行していることを報知することが好ましい。特に、イグニッションスイッチをＯＦＦとしたのにエンジンが停止しないことの違和感や、故障であるとの誤認を防ぐことができる。
［推定イオウ付着量の補正制御−イオウ含有量予測値Ｋｆｓｓの算出］
図２６は推定イオウ付着量Ｑｓｓの補正制御のために行う空燃比制御の具体的動作の一例を示すタイムチャート、及び図２７は同フローチャートである。イオウ被毒解消制御のステップＳ１３（イオウ付着量Ｑｓｓの推定動作）で述べたように、一般に、燃料中に含まれるイオウの量は一定ではなくバラツキがあるから、例えば給油をすると、エンジン１に供給される燃料のイオウ含有量がそれまでのものとは変化する。
【０１５１】
したがって、同じ燃料供給量（噴射量）であっても、例えば、図２６に例示したように、給油によってイオウ含有量が多くなると、ＮＯｘ触媒１７へのイオウ付着量が多くなり、鎖線で示したように、被毒の状態が推定よりも早く進んで、それまで推定していたイオウ付着量Ｑｓｓ（実線）より多い量のイオウが付着する。その結果、すでにイオウが判定基準量Ｑｓ１以上付着しているのに、イオウ放出処理が開始されず、ＮＯｘ触媒１７の浄化能力低下の状態が無駄に続いたり、また、イオウ放出処理が開始された後においても、まだイオウがゼロ又は終了許可判定量Ｑｓ２以上残存しているのに、イオウ放出処理が終了され、ＮＯｘ触媒１７の浄化能力が元通りには回復しない等といった不具合が生じる。
【０１５２】
逆に、給油によってイオウ含有量が少なくなると、ＮＯｘ触媒１７へのイオウ付着量が少なくなり、被毒の状態が推定よりも遅く進んで、それまで推定していたイオウ付着量Ｑｓｓより少ない量のイオウが付着する。その結果、まだイオウが判定基準量Ｑｓ１以上付着していないのに、イオウ放出処理が開始されて、無駄に燃料が費やされたり、また、イオウ放出処理が開始された後においても、すでにイオウがゼロ又は終了許可判定量Ｑｓ２以下に除去されているのに、イオウ放出処理が続行され、やはり無駄に燃料が費やされる等といった不具合が生じる。
【０１５３】
そこで、イオウ放出処理を過不足なく行うために、前述のステップＳ１３、より詳しくはステップＳ５４で、燃料供給量Ｑｆのみに基づいて推定した基本イオウ増加量Ｑｓａを燃料のイオウ含有量Ｋｆｓｓで補正し、推定イオウ付着量と実際のイオウ付着量との間のずれを是正して、イオウ付着量の推定精度を向上させるようにしたのである。そして、そのための処理、すなわち推定付着量補正のための空燃比のリッチ化が行なわれる。
【０１５４】
その結果、このエンジン１においては、ＮＯｘ浄化制御における吸蔵能力回復のためのリッチ化、ＮＯｘ触媒１７のイオウ被毒解消制御におけるイオウ放出促進のためのリッチ化、及びこの推定イオウ付着量補正のためのリッチ化というそれぞれ別異の目的のための複数種類の空燃比のリッチ化が行なわれる。なお、図２６において、符号ナで示したリッチ化は、イオウ被毒解消制御で行なわれる空燃比のリッチ化、符号ニで示したリッチ化は、吸蔵能力回復のための空燃比のリッチ化、及び符号ヌで示したリッチ化は、この推定付着量補正のためのリッチ化である。
【０１５５】
このリッチ化は図２７のフローチャートに従って行なわれる。このプログラムは前述のイオウ被毒解消制御と独立して実行され、まず、ステップＳ９１で給油直後であること、ステップＳ９２でイオウ除去要求フラグＦがゼロにリセットされていること、ステップＳ９３でリーン運転中であること、及び、ステップＳ９４で推定イオウ付着量Ｑｓｓが予め設定された補正制御を行うための判定基準量Ｑｓ３以上であることが確認される。ここで、この判定基準量Ｑｓ３は、前述の判定基準量Ｑｓ１よりも小さい値である。
【０１５６】
そして、上記条件が全て満足されているときに、ステップＳ９５に進んで、前述のＮＯｘ放出処理における図４のステップＳ２で算出される修正ＮＯｘ吸蔵量Ｑｎｏｘｓが読み込まれる。この修正ＮＯｘ吸蔵量Ｑｎｏｘｓは、前述したように、基本ＮＯｘ吸蔵量Ｑｎｏｘａが推定イオウ付着量Ｑｓｓで修正されたものである。
【０１５７】
次いで、ステップＳ９６で、上記修正ＮＯｘ吸蔵量Ｑｎｏｘｓに基いて、この推定付着量補正のためのリッチ化の条件を決定する。その場合に、図２８に示すように、リッチ化条件（リッチ化の時間又はリッチ度合の少なくともいずれか）は、修正ＮＯｘ吸蔵量Ｑｎｏｘｓが多くなるに従い、リッチ化時間が長くなり、あるいはリッチ度合が高くなるように設定される。
【０１５８】
そして、ステップＳ９７で、そのようにして求められたリッチ化条件が実現するように空燃比のリッチ化を実行すると共に、ステップＳ９８で、そのときのＯ₂センサでなる第２空燃比センサ２８の出力値を検出する。特に、ここでは、第２空燃比センサ２８の出力最大値Ｅを検出する。ここで、第２空燃比センサ２８は、酸素濃度が多いほど、すなわちリーン側で出力値が小さく、酸素濃度が少ないほど、すなわちリッチ側で出力値が大きくなる特性を有する。
【０１５９】
なお、図２６では、このときの第２空燃比センサ２８の出力部分のみを拡大して示した。
【０１６０】
次いで、ステップＳ９９で、上記出力最大値Ｅに基いて、ＮＯｘ吸蔵量の補正係数Ｋｎｏｘを設定する。その場合に、図２９に示すように、ＮＯｘ吸蔵量補正係数Ｋｎｏｘは、出力最大値Ｅが大きくなるに従い、小さくなるように設定される。つまり、ＮＯｘ触媒１７下流の空燃比が、酸素濃度が低くリッチ側にシフトするに従い、ＮＯｘ吸蔵量が小さな値に補正されることになる。換言すれば、イオウ付着量が大きな値に補正されることになる。
【０１６１】
なお、この場合、補正係数Ｋｎｏｘが１に近づくに従って、イオウ付着量の推定値Ｑｓｓが実際の値にほぼ近似していたことになる。
【０１６２】
次いで、ステップＳ１００で、数７に従い、上記補正係数Ｋｎｏｘに基いて、修正ＮＯｘ吸蔵量Ｑｎｏｘｓを補正することにより、ＮＯｘ吸蔵量の検出値（検出ＮＯｘ吸蔵量）Ｑｎｏｘｋを算出する。つまり、この推定付着量補正のためのリッチ化の実行の時点で、ＮＯｘ触媒１７に吸蔵されていたＮＯｘ吸蔵量を、このリッチ化で得られた酸素濃度の検出結果に基いて補正することになる。
【０１６３】
【数７】

【０１６４】
さらに、ステップＳ１０１で、数８に従い、上記検出ＮＯｘ吸蔵量Ｑｎｏｘｋと、修正ＮＯｘ吸蔵量Ｑｎｏｘｓとに基いて、イオウ付着量の検出値（検出イオウ付着量）Ｑｓｋを算出する。つまり、この推定付着量補正のためのリッチ化の実行の時点で、ＮＯｘ触媒１７に付着していたと推定していたイオウ付着量Ｑｓｓを、このリッチ化で得られた酸素濃度の検出結果に基いて補正することになる。
【０１６５】
【数８】

【０１６６】
次いで、ステップＳ１０２で、数９に従い、上記検出イオウ付着量Ｑｓｋと、推定イオウ付着量Ｑｓｓとに基いて、イオウ含有量の予測値Ｋｆｓｓを更新する。この更新値Ｋｆｓｓは、次に行なわれるイオウ被毒解消制御のステップＳ５４で用いられることになる。
【０１６７】
【数９】

【０１６８】
さらに、ステップＳ１０３で、上記検出イオウ付着量Ｑｓｋの値をそのまま推定イオウ付着量Ｑｓｓの値であるとして取り扱うようにする。
【０１６９】
以上により、第２空燃比センサ２８の検出結果から、推定イオウ付着量Ｑｓｓが補正され、ＮＯｘ吸蔵量Ｑｎｏｘｓも修正され、燃料のイオウ含有量Ｋｆｓｓも予測されることになる。
【０１７０】
このように、燃焼室４から排出される排ガスの空燃比を強制的にリッチ化し、そのときＮＯｘ触媒１７の下流の空燃比センサ２８で検出される空燃比に基いて、該触媒１７への推定イオウ付着量Ｑｓｓが補正される。このリッチ雰囲気での酸素濃度は、ＮＯｘ触媒１７が現にイオウ被毒している程度によって変化し、また、そのイオウ被毒の程度を反映するから、燃料供給量Ｑｆに基いて推定したイオウ付着量Ｑｓｓを、上記酸素濃度に基いて、つまりＮＯｘ触媒１７下流の空燃比に基いて補正することにより、燃料のイオウ含有量のバラツキに起因する推定精度への影響が是正される。
【０１７１】
その結果、燃料のイオウ含有量をいちいち測定したりコントロールユニット２０に入力操作したりすることなく、空燃比のリッチ化という普通に行なわれる手法を用いて、燃料のイオウ含有量のバラツキによってイオウ成分付着量Ｑｓｓの推定精度が低下するという不具合に、合理的、且つ効率的に対処することが可能となり、ひいては、過不足なく良好にイオウ被毒解消制御を実行することが可能となる。
【０１７２】
なお、ステップＳ９８では、第２空燃比センサ２８の検出結果として、空燃比センサ２８の出力最大値Ｅを検出するようにしたが、これに代えて、あるいはこれと共に、図２６に示したように、リーン状態からリッチ状態に切り換えたときの該センサ２８の出力値の小から大への立上りの遅れ時間Ｄや、該センサ２８の出力値の波形で囲まれる面積Ｍを検出してもよい。
【０１７３】
また、ＮＯｘ触媒１７のＮＯｘ吸蔵能力の回復を図るために本来的に必ず行なわれる図４に示すリッチ化を利用して、触媒１７下流の空燃比を検出し、推定イオウ付着量Ｑｓｓを補正してもよい。空燃比をリッチ化する回数が徒に増えることがなく、リーン運転がむやみに阻害されて燃費性能が損なわれることが抑制される。
【０１７４】
一方、以上説明したように、図４に示すリッチ化と推定付着量補正のためのリッチ化とを独立して実行し、特に、少なくとも、給油がされた後、イオウの放出処理が最初に完了した（Ｆ＝０）後において、推定イオウ付着量Ｑｓｓが所定量Ｑｓ３以上増加したときに、排ガスの空燃比をリッチ化することにより、推定イオウ付着量の補正にとって適切な時期にリッチ化及び空燃比の検出を行うことができて、補正の精度が一層向上する。その場合でも、吸蔵能力回復のためのリッチ化に比べてはるかに少ない頻度でリッチ化を行えば済むから、空燃比をリッチ化する回数が徒に増えることもなく、リーン運転がむやみに阻害されて燃費性能が損なわれることも抑制される。
【０１７５】
さらに、イオウの放出処理が完了した（Ｆ＝０）後であるから、推定イオウ付着量Ｑｓｓの起算値がゼロとなり、該付着量Ｑｓｓが所定量Ｑｓ３以上増加したかどうかの判定精度が向上する。併せて、給油後に最初にイオウ放出処理が行われた場合において上記のタイミングでリッチ化が行なわれるから、給油によって変化する燃料のイオウ含有量のバラツキによる影響が早期に是正されることになる。
【０１７６】
そして、ステップＳ９６において、リッチ化度合やリッチ化時間等のリッチ化条件が、結局、その時点での推定イオウ付着量Ｑｓｓに基いて決定されるから、ＮＯｘの放出という観点からは、吸蔵ＮＯｘが、過不足のないリッチ化度合で、あるいは過不足のない時間だけ、適正に放出処理され、その結果、未処理の吸蔵ＮＯｘが触媒１７に残存したり、逆に、過度にリッチ化されて燃費性能が低下するというような不具合が抑制される一方で、推定イオウ付着量の補正という観点からは、吸蔵ＮＯｘが全て酸素に還元分解され、充分量の酸素が発生するから、ノイズの少ない明確な空燃比検出値（上記の遅れ時間、面積、出力最大値等）が得られ、補正の精度が向上することになる。
【０１７７】
【発明の効果】
以上説明したように、本発明によれば、燃焼室から排出される排ガスの空燃比を強制的にリッチ化し、そのときＮＯｘ触媒の下流で検出される酸素濃度に基いて、該触媒への推定イオウ付着量が補正される。そして、このリッチ雰囲気での酸素濃度は、ＮＯｘ触媒が現にイオウ被毒している程度によって変化し、また、そのイオウ被毒の程度を反映するから、燃料供給量に関連する値に基いて推定したイオウ付着量を上記酸素濃度に基いて補正することにより、燃料のイオウ含有量のバラツキに起因する推定精度への影響が是正される。その結果、燃料のイオウ含有量をいちいち測定したり入力操作したりすることなく、空燃比のリッチ化という普通に行なわれる手法を用いて、燃料のイオウ含有量のバラツキによってイオウ成分付着量の推定精度が低下するという不具合に、合理的、且つ効率的に対処することが可能となり、ひいては、過不足なく良好にイオウ放出処理を実行することが可能となる。
【０１７８】
また、本発明によれば、吸蔵能力回復のためのリッチ化を利用して、触媒下流の酸素濃度を検出し、推定イオウ付着量を補正することができるから、空燃比をリッチ化する回数が徒に増えることがなく、リーン運転がむやみに阻害されて燃費性能が損なわれることが抑制される。
【０１７９】
一方、本発明によれば、吸蔵能力回復のためのリッチ化とはまた別に、推定付着量補正のためのリッチ化が行なわれる。したがって、推定イオウ付着量の補正にとって適切な酸素濃度検出値が得られ、該補正の精度が向上し、ひいては、過不足のない良好なイオウ放出処理の実行が担保される。しかも、その場合に、吸蔵能力回復のためのリッチ化に比べてはるかに少ない頻度でリッチ化を行えば済むから、空燃比をリッチ化する回数が徒に増えることもなく、リーン運転がむやみに阻害されて燃費性能が損なわれることも抑制される。
【０１８０】
また、本発明によれば、推定付着量補正のためのリッチ化を行う時期の一例として、イオウ放出処理が行われていったんゼロになった推定イオウ付着量が再び所定量まで増加した時期にリッチ化が行なわれるから、イオウ被毒がある程度の状態にまで進んだ時期を選択してリッチ化が行なわれることになり、前述したように、推定イオウ付着量の補正にとって適切な酸素濃度検出値が得られ、該補正の精度が向上する。また、推定イオウ付着量の起算値がゼロであることから、該付着量が所定量以上増加したかどうかの判定精度が増す。併せて、給油後に最初にイオウ放出処理が行われた場合において上記のタイミングでリッチ化が行なわれるから、給油によって変化する燃料のイオウ含有量のバラツキによる影響が早期に是正される。
【０１８１】
さらに、本発明によれば、リッチ化度合やリッチ化時間等のリッチ化条件が、推定イオウ付着量に基いて定められるから、ＮＯｘの放出という観点からは、吸蔵ＮＯｘが、過不足のないリッチ化度合で、あるいは過不足のない時間だけ、適正に放出処理される。その結果、未処理の吸蔵ＮＯｘが触媒に残存したり、逆に、過度にリッチ化されて燃費性能が低下するというような不具合が抑制される。一方、推定イオウ付着量の補正という観点からは、吸蔵ＮＯｘが全て酸素に還元分解され、充分量の酸素が発生するから、明確な酸素濃度検出値が得られ、上記補正の精度が向上する。
【０１８２】
本発明は、ＮＯｘ吸蔵還元型触媒を備えるエンジン一般に広く好ましく適用可能である。
【図面の簡単な説明】
【図１】 本発明の実施の形態に係るエンジンの制御システム構成図である。
【図２】 同エンジンの空燃比マップである。
【図３】 同エンジンのＮＯｘ浄化制御の一例を示すタイムチャートである。
【図４】 同フローチャートである。
【図５】 同エンジンのイオウ被毒解消制御の一例を示すタイムチャートである。
【図６】 同メインフローチャートである。
【図７】 同フローチャートのサブルーティンを表わすフローチャートである。
【図８】 同じくサブフローチャートである。
【図９】 同制御で用いられる特性図である。
【図１０】 同じく特性図である。
【図１１】 同じく特性図である。
【図１２】 同じく特性図である。
【図１３】 同制御の原理図である。
【図１４】 同じく原理図である。
【図１５】 同じく特性図である。
【図１６】 同じくサブフローチャートである。
【図１７】 同じく特性図である。
【図１８】 同じく特性図である。
【図１９】 同じく特性図である。
【図２０】 同じくサブフローチャートである。
【図２１】 同じくサブフローチャートである。
【図２２】 同じく原理図である。
【図２３】 同じく特性図である。
【図２４】 同じく特性図である。
【図２５】 同じく特性図である。
【図２６】 同エンジンの推定付着量補正制御の一例を示すタイムチャートである。
【図２７】 同じくフローチャートである。
【図２８】 同制御で用いられる特性図である。
【図２９】 同じく特性図である。
【図３０】 本発明の原理図である。
【図３１】 同じく原理図である。
【図３２】 同じく原理図である。
【符号の説明】
１ エンジン
４ 燃焼室
５ 点火プラグ
６ インジェクタ
１６ 三元触媒
１７ ＮＯｘ触媒（ＮＯｘ吸蔵還元型触媒）
２０ コントロールユニット（推定手段、放出手段、リッチ化手段、補正手段）
２８ 第２空燃比センサ（検出手段）[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an engine exhaust gas purification device, and particularly to the technical field of an engine exhaust gas purification device that includes a NOx occlusion reduction type catalyst that purifies NOx in exhaust gas generated during lean operation.
[0002]
[Prior art]
A three-way catalyst provided in an engine for vehicles or the like efficiently purifies harmful components such as CO (carbon monoxide), HC (hydrocarbon), and NOx (nitrogen oxide) contained in exhaust gas. Since the window is limited to a narrow range in the vicinity of the theoretical air-fuel ratio (λ = 1), the NOx purification rate decreases in a so-called lean burn engine or the like that employs a direct injection stratified combustion method for the purpose of improving fuel efficiency in recent years. There is a problem. Therefore, NOx occlusion reduction type catalyst that stores NOx in exhaust gas when the air-fuel ratio is lean (oxygen excess state) and reduces and releases the stored NOx when the air-fuel ratio becomes rich (oxygen deficient state). (Lean NOx catalyst) is provided in the exhaust passage.
[0003]
In this case, if the lean operation continues for a long time, the NOx catalyst becomes saturated with NOx. Therefore, in order to recover the NOx occlusion capacity of the catalyst, regularly or when the occluded NOx amount exceeds a predetermined occlusion amount. Further, NOx is released from the catalyst by enriching the air-fuel ratio of the exhaust gas. Of course, NOx is also released when the rich operation or the stoichiometric air-fuel ratio operation is performed according to the driving operation of the driver, and the NOx occlusion capacity of the catalyst is recovered.
[0004]
On the other hand, since NOx catalyst uses barium (Ba), sulfur components such as SOx (sulfur oxide) contained in the fuel are likely to adhere, and NOx occlusion ability, that is, NOx purification ability, decreases due to the increase in the amount of adhesion. There is a problem of sulfur poisoning. In order to cope with this, if the amount of the sulfur component adhering to the NOx catalyst becomes equal to or larger than the predetermined adhering amount, for example, the ignition timing is retarded or the fuel is dividedly injected to raise the exhaust gas temperature. It is known to raise the temperature and release the sulfur component from the NOx catalyst.
[0005]
In that case, since the sulfur adhesion amount, which is a criterion for determining whether to start the sulfur release process, cannot be directly measured, it is generally obtained by estimation. Therefore, if the estimation accuracy is lowered and there is an error in the estimated amount, various problems occur.
[0006]
For example, when a larger amount of sulfur is estimated than the actual amount of sulfur, the sulfur release process is started immediately even though not so much sulfur is deposited, or it is wasteful even if sulfur has already been completely removed. In addition, the sulfur release treatment is continued, which is not preferable from the viewpoint of fuel consumption. On the other hand, when it is estimated that the amount of sulfur attached is less than the actual amount, the sulfur release treatment has not started yet even though a considerable amount of sulfur has already been attached, or the sulfur has not been completely removed. Sulfur release treatment has already been completed, both of which are undesirable from the standpoint of eliminating sulfur poisoning.
[0007]
Therefore, the amount of sulfur component that passes through the NOx catalyst without being occluded is excluded from the estimated value, or the estimated sulfur adhesion amount is corrected based on the exhaust gas temperature or the sulfur adhesion amount already attached to the catalyst. There are a number of proposals for improving the estimation of the amount of adhesion.
[0008]
[Problems to be solved by the invention]
By the way, as disclosed in Japanese Patent Laid-Open No. 10-1000063, the estimated sulfur adhesion amount is first determined based on the supply amount of fuel which is a sulfur component generation source. However, since the amount of sulfur contained in the fuel is usually not constant and varies, as long as the basic value of the amount of sulfur attached is simply estimated based on the amount of fuel supplied, there are various Even if this correction is performed, the estimation accuracy cannot be drastically improved. However, it is not possible to measure the sulfur content of the fuel one by one, and even if the sulfur content is known, the value must be input to the in-vehicle computer every time the fuel changes.
[0009]
Therefore, an object of the present invention is to rationally and efficiently cope with the problem that the estimation accuracy of the amount of sulfur adhesion decreases due to the variation in the sulfur content of the fuel. Hereinafter, the present invention will be described in detail including other problems.
[0010]
[Means for Solving the Problems]
That is, the present inventors paid attention to oxygen generated by reduction of NOx stored in the NOx catalyst when the exhaust gas air-fuel ratio is switched from lean to rich. The oxygen concentration is proportional to the amount of occluded NOx, and the amount of occluded NOx is influenced by the amount of sulfur adhering to the NOx catalyst, so that the oxygen concentration when enriched eventually becomes the sulfur of the NOx catalyst. The present invention has been completed by utilizing the fact that it reflects the poisoning state.
[0011]
That is, in order to solve the above-mentioned problem, the invention according to claim 1 in the scope of claims of the present application,
A NOx occlusion reduction type catalyst that occludes NOx components in exhaust gas in an excess oxygen atmosphere and reduces and releases NOx components occluded due to a decrease in oxygen concentration is provided in the exhaust passage, and sulfur components adhering to the catalyst are reduced. An estimation means for estimating the amount based on a value related to the fuel supply amount, and a release for releasing the sulfur component from the catalyst when the sulfur component adhesion amount estimated by the estimation means exceeds a predetermined adhesion amount An exhaust emission control device for an engine, comprising: means for downstream of the catalyst Oxygen concentration Detecting means for detecting the exhaust gas and forcibly enriching the air-fuel ratio of the exhaust gas discharged from the combustion chamber Sulfur component estimation Enrichment means, and Sulfur component estimation The detection means when the air-fuel ratio of exhaust gas is enriched by the enrichment means Oxygen concentration detected by And a correcting means for correcting the sulfur component adhesion amount estimated by the estimating means.
[0012]
According to the present invention, the air-fuel ratio of the exhaust gas discharged from the combustion chamber is forcibly enriched and then detected downstream of the NOx catalyst. Oxygen concentration Based on this, the estimated sulfur adhesion amount to the catalyst is corrected. More specifically, first, as shown in FIG. 30, NOx in the exhaust gas is occluded in the barium of the NOx catalyst in a lean atmosphere. At this time, the amount of occluded NOx is affected by the sulfur poisoning of the catalyst, and the amount of occluded NOx decreases as the sulfur poisoning amount increases. Next, as shown in FIG. 31, in a rich atmosphere, the stored NOx is promoted by platinum (Pt) as an oxidation-reduction reaction catalyst, and nitrogen (N ₂ ) And oxygen (O ₂ ) And released. At this time, since the released oxygen is derived from the occluded NOx, the oxygen concentration in the rich atmosphere is proportional to the occluded NOx amount.
[0013]
Therefore, the oxygen concentration in the rich atmosphere varies depending on the degree to which the NOx catalyst is actually poisoned with sulfur and reflects the degree of sulfur poisoning. Therefore, the amount of sulfur deposition estimated based on the value related to the fuel supply amount Oxygen concentration above By correcting based on the above, the influence on the estimation accuracy due to the variation in the sulfur content of the fuel is corrected.
[0014]
As a result, it is possible to estimate the amount of sulfur component adhering by the variation in the sulfur content of the fuel, using the usual method of enriching the air-fuel ratio without measuring and inputting the sulfur content of the fuel one by one. It becomes possible to deal with the problem of reduced accuracy reasonably and efficiently, and consequently, it is possible to execute the sulfur release process satisfactorily without excess or deficiency.
[0015]
The inventors of the present invention downstream of the NOx catalyst Oxygen concentration As a means for detecting, for example, the output value is smaller as the oxygen concentration is higher, that is, the output value is smaller on the lean side, and the output value is larger as the oxygen concentration is lower, that is, the rich side. ₂ As a result of intensive studies and examinations using sensors, the knowledge shown in FIG. 32 was obtained. In the figure, the solid line a is the sensor output when the sulfur poisoning amount is large and therefore the occluded NOx amount is small, and the broken line b is the sensor output when the sulfur poisoning amount is small and therefore the occluded NOx amount is large. It is. That is, in the latter case b, when the lean state is switched to the rich state, the rise x from the small to the large sensor output is delayed, and the area surrounded by the waveform of the sensor output when the rich time is constant c. It has been found that y becomes smaller and the maximum value z of the sensor output becomes smaller.
[0016]
Based on this knowledge, in one preferable aspect of the present invention, at least one of the delay time x, the area (integrated value) y, or the output maximum value z represents the amount of occluded NOx or sulfur poisoning. Adopted as an index, the adhesion amount of sulfur component is corrected according to the value.
[0017]
Note that “forcedly” here means that lean driving should be performed in order to improve the original fuel efficiency considering the driving conditions, but the lean driving is interrupted to achieve different purposes. Meaning that is intentional.
[0018]
Next, the invention according to claim 2 is the invention according to claim 1, wherein the catalyst is removed from the catalyst periodically or when the amount of NOx component occluded in the catalyst becomes equal to or greater than a predetermined occlusion amount. NOx component releasing means for enriching the air-fuel ratio of the exhaust gas to release the NOx component is provided, Estimated sulfur component The enrichment means is the NOx component release means.
[0019]
According to the present invention, the air-fuel ratio downstream of the catalyst is detected by utilizing the enrichment (enrichment for restoring the storage capacity) that is necessarily performed in order to restore the NOx storage capacity of the NOx catalyst. Since the estimated sulfur adhesion amount can be corrected, the number of times the air-fuel ratio is enriched does not increase easily, and it is suppressed that lean operation is obstructed unnecessarily and fuel efficiency is impaired.
[0020]
Next, the invention according to claim 3 is the invention according to claim 1, wherein the catalyst is removed from the catalyst periodically or when the amount of NOx component occluded in the catalyst becomes equal to or greater than a predetermined occlusion amount. NOx component releasing means for enriching the air-fuel ratio of the exhaust gas to release the NOx component is provided, Estimated sulfur component The enrichment means enriches the air-fuel ratio of the exhaust gas at a frequency less than the enrichment frequency performed by the NOx component release means.
[0021]
In contrast to the invention described in claim 2, according to the present invention, first, separately from the enrichment for restoring the storage capacity, the enrichment for correcting the estimated adhesion amount according to the present invention, that is, the catalyst Downstream Oxygen concentration Is detected, and enrichment is performed to correct the estimated sulfur adhesion amount. Therefore, it is appropriate for correcting the estimated sulfur adhesion amount. Oxygen concentration A detection value is obtained, and the accuracy of the correction is improved.
[0022]
In other words, the enrichment for correcting the estimated adhesion amount according to the present invention ultimately releases the NOx occluded by the catalyst, so as described above, the enrichment for restoring the occlusion capacity is used. May be. However, since the enrichment according to the present invention is concerned with the sulfur content of the fuel and looks at the sulfur poisoning state of the catalyst, it is meaningful to perform enrichment when sulfur poisoning is not progressing much. Na Oxygen concentration detection value Will not be obtained.
[0023]
However, the inherent enrichment for recovery of the storage capacity is performed periodically or when the storage NOx amount exceeds a predetermined value regardless of the sulfur poisoning state of the catalyst. It may be performed when sulfur poisoning is not progressing, and using the release of NOx at this time Oxygen concentration Even if detected, an appropriate detection value may not be obtained from the viewpoint of correcting the estimated sulfur adhesion amount. Moreover, the NOx occlusion rate is generally much larger than the sulfur deposition rate, and therefore the enrichment for restoring the occlusion capacity is frequently performed many times while the sulfur poisoning proceeds to a certain level. Therefore, it is probable that the enrichment for restoring the storage capacity is performed when the sulfur poisoning is not progressing so much.
[0024]
Therefore, when the enrichment for correcting the estimated adhesion amount according to the present invention is combined with the enrichment for restoring the storage capacity, it is not useful for maintaining and improving the correction accuracy. Oxygen concentration There is a high possibility that only the detected value can be obtained, which is wasteful, and when the estimated sulfur adhesion amount is corrected based on the detection result, the credibility of the obtained sulfur adhesion amount is also lowered.
[0025]
On the other hand, in the invention according to the third aspect, the enrichment for correcting the estimated adhesion amount is performed independently of the enrichment for restoring the storage capacity, so that, for example, sulfur poisoning is in a certain state. Can be selected and the enrichment according to the present invention can be performed. As a result, it is suitable for the correction of the estimated sulfur adhesion amount. Oxygen concentration A detection value is obtained, and the accuracy of the correction is improved. As a result, it is possible to secure a good sulfur release process without excess or deficiency.
[0026]
Moreover, in this case, since enrichment is performed by selecting the time when sulfur poisoning has progressed to a certain level, enrichment can be performed at a much lower frequency than enrichment for restoring the storage capacity. . Therefore, the number of times that the air-fuel ratio is enriched does not increase, and the lean operation is obstructed and the fuel consumption performance is prevented from being impaired.
[0027]
Next, the invention described in claim 4 is the invention described in claim 3, wherein Sulfur component estimation The enrichment means is an exhaust gas exhaust when the amount of adhesion of the sulfur component estimated by the estimation means increases by a predetermined amount or more after the release means first releases the sulfur component after refueling. It is characterized by enriching the fuel ratio.
[0028]
According to the present invention, an example of the timing for performing enrichment for correcting the estimated adhesion amount is shown. First, in this case, since the enrichment is performed when the estimated sulfur adhesion amount that once became zero after the sulfur release process has been performed again increases to a predetermined amount, the time when the sulfur poisoning has progressed to a certain level is determined. And the enrichment according to the present invention is performed, and as described above, it is appropriate for the correction of the estimated sulfur adhesion amount. Oxygen concentration A detection value is obtained, and the accuracy of the correction is improved. In this case, since the estimated value of the estimated sulfur adhesion amount is zero, the determination accuracy of whether or not the adhesion amount has increased by a predetermined amount or more is increased.
[0029]
In addition, since the enrichment is performed at the above timing when the sulfur release process is first performed after refueling, the influence of the variation in the sulfur content of the fuel that changes due to refueling is corrected at an early stage.
[0030]
Next, the invention according to claim 5 is the invention according to any one of claims 1 to 4, Estimated sulfur component The enrichment means determines at least one of the degree of enrichment or the enrichment time based on the adhesion amount of the sulfur component estimated by the estimation means.
[0031]
According to this invention, the enrichment conditions such as the enrichment degree and enrichment time are determined based on the estimated sulfur adhesion amount. More specifically, for example, the stored NOx amount is corrected based on the estimated sulfur adhesion amount, and the enrichment condition for releasing the stored NOx is determined based on the corrected stored NOx amount.
[0032]
Therefore, from the viewpoint of NOx release, the occluded NOx is appropriately released at an enrichment degree without excess or deficiency, or for a time without excess or deficiency. As a result, problems such as untreated occluded NOx remaining in the catalyst or conversely being excessively enriched to reduce fuel consumption performance are suppressed.
[0033]
On the other hand, from the viewpoint of correcting the estimated sulfur adhesion amount, all of the stored NOx is reduced and decomposed into oxygen, and a sufficient amount of oxygen is generated. Oxygen concentration A detection value is obtained, and the accuracy of the correction is improved. Hereinafter, the present invention will be described in more detail through embodiments of the invention.
[0034]
DETAILED DESCRIPTION OF THE INVENTION
[System configuration]
FIG. 1 is a control system configuration diagram of a direct injection stratified combustion engine 1 according to the present embodiment. The main body 2 of the engine 1 is provided with a plurality of combustion chambers 4 (only one of which is shown) defined by pistons 3. An ignition plug 5 faces the center of the upper portion of the combustion chamber 4, and an injector 6 that directly injects fuel into the combustion chamber 4 faces the side portion.
[0035]
An intake passage 9 and an exhaust passage 10 are connected to the combustion chamber 4 via an intake valve 7 and an exhaust valve 8. An air cleaner 11, an air flow sensor 12, a throttle valve 13, and a surge tank 14 are disposed in the intake passage 9 from the upstream side. The downstream side of the surge tank 14 is an independent intake passage 9a branched for each cylinder, and the downstream end of the independent intake passage 9a facing the combustion chamber 4 is divided into a first passage 9b and a second passage 9c. . A swirl generation valve 15 is provided in the second passage 9c, and when the valve 15 is closed, a swirl is generated in the combustion chamber 4 by intake air introduced from the first passage 9b.
[0036]
The exhaust passage 10 has a three-way catalyst 16 that simultaneously removes CO, HC, and NOx in the exhaust gas near the stoichiometric air-fuel ratio (A / F = 14.7), and NOx storage that stores and reduces NOx in the exhaust gas. A reduction catalyst (hereinafter simply referred to as “NOx catalyst”) 17 is arranged in series. The NOx catalyst 17 stores NOx that flows in without being purified by the three-way catalyst 16 when the air-fuel ratio is leaner than the stoichiometric air-fuel ratio (λ> 1), for example, and suppresses discharge to the outside. For example, when the air-fuel ratio is in the vicinity of the stoichiometric air-fuel ratio or in a richer state (λ ≦ 1), the NOx stored is oxidized and reduced with CO and HC in the exhaust gas to be decomposed into oxygen and nitrogen.
[0037]
The NOx catalyst 17 is composed of a NOx absorbent containing barium as a main component and supporting an alkali metal such as potassium, magnesium, strontium, lanthanum, or rare earth, and a noble metal having a chemical reaction catalytic action such as platinum. Decorate.
[0038]
An exhaust gas recirculation passage 18 is provided between the upstream side of the three-way catalyst 16 in the exhaust passage 10 and the upstream side of the surge tank 14 in the intake passage 9 to return a part of the exhaust gas flowing in the exhaust passage 10 to the intake passage 9. It has been. The exhaust gas recirculation passage 18 is provided with an exhaust gas recirculation amount adjustment valve 19 that adjusts the recirculation amount of the exhaust gas in the vicinity of the junction with the intake passage 9.
[0039]
The control unit (ECU) 20 of the engine 1 includes a signal from the air flow sensor 12 that detects the intake air amount, a signal from the throttle opening sensor 21 that detects the opening of the throttle valve 13, and an exhaust recirculation amount adjustment valve 19. A signal from the recirculation amount sensor 22 for detecting the opening, a signal from the boost sensor 23 for detecting the intake negative pressure in the surge tank 14, a signal from the fuel pressure sensor 24 for detecting the pressure of the fuel supplied to the injector 6, A signal from a water temperature sensor 25 for detecting the temperature of the cooling water in the engine body 2 is provided on the upstream side of the three-way catalyst 16 and supplied to the combustion chamber 4 from the residual oxygen concentration in the exhaust gas discharged from the combustion chamber 4. Detects whether the air-fuel ratio of the air-fuel mixture is richer or leaner than the stoichiometric air-fuel ratio ₂ A signal from a first air-fuel ratio sensor 26 comprising a sensor, a signal from an exhaust temperature sensor 27 provided between the three-way catalyst 16 and the NOx catalyst 17 and detecting the exhaust gas temperature immediately before flowing into the NOx catalyst 17, NOx An O is provided downstream of the catalyst 17 and detects the residual oxygen concentration in the exhaust gas that has passed through the NOx catalyst 17. ₂ A signal from the second air-fuel ratio sensor 28, which is a sensor, a signal from the engine rotation sensor 29 that detects the number of revolutions of the engine 1, and a signal from the accelerator opening sensor 30 that detects the amount of depression of an accelerator pedal (not shown). A signal from the intake air temperature sensor 31 for detecting the temperature of the intake air, a signal from the atmospheric pressure sensor 32 for detecting the atmospheric pressure, and the like are input, and the throttle valve 13 according to the operating state of the engine 1 indicated by these signals. The throttle opening control by outputting control signals to the actuator 33 for driving the exhaust gas, the exhaust gas recirculation amount adjusting valve 19, the injector 6, the actuator 34 for driving the swirl generation valve 15, the ignition circuit 35 for igniting the spark plug 5, etc. Exhaust gas recirculation control, fuel injection amount control, fuel injection timing control, swirl generation control, ignition timing control, NOx catalyst NOx purification control by 7, comprehensively perform sulfur NOx catalyst 17 poisoning elimination control and the like.
[Air-fuel ratio map]
FIG. 2 is a target air-fuel ratio map of the engine 1. The engine operating region with the engine speed and the engine load as parameters is a first region A set in a low rotation to medium rotation and low load to medium load region, and a first region A set in a high rotation and high load region. It is divided into two regions B, a third region C set between these regions A and B, and a fourth region D set as a low load region equal to or higher than a predetermined engine speed.
[0040]
The first region A having the highest operation frequency is a lean operation region in which the air-fuel ratio is made larger than the theoretical air-fuel ratio (λ> 1). During the lean operation in the region A, fuel is injected during the compression stroke (late injection), and the fuel is unevenly distributed in the vicinity of the spark plug 5 to cause stratified combustion. During lean operation, NOx in the exhaust gas is occluded by the NOx catalyst 17, and both fuel efficiency and exhaust performance are improved.
[0041]
The second region B which is an operation region during high speed operation or acceleration is a rich operation region in which the air-fuel ratio is made smaller than the theoretical air-fuel ratio (λ <1). During the rich operation in this region B, fuel is injected during the intake stroke (pre-injection), and the fuel is sufficiently vaporized and atomized in the combustion chamber 4. During the rich operation, NOx stored in the NOx catalyst 17 and CO, HC react with each other by oxidation-reduction reaction, and a good torque is obtained and the exhaust performance is improved.
[0042]
The third region C is a stoichiometric air-fuel ratio operation region where the air-fuel ratio is the stoichiometric air-fuel ratio (λ = 1). During the stoichiometric air-fuel ratio operation in this region C, fuel is injected during the intake stroke (said injection) as in the rich operation, and the fuel is sufficiently vaporized and atomized in the combustion chamber 4. During the theoretical air-fuel ratio operation, CO, HC and NOx in the exhaust gas are simultaneously purified by the three-way catalyst 16.
[0043]
The fourth region D is a fuel cut (F / C) region in which fuel injection into the combustion chamber 4 is stopped.
[NOx purification control by NOx catalyst-NOx release processing]
FIG. 3 is a time chart showing an example of a specific operation of air-fuel ratio control performed for NOx purification control. As the lean operation continues in the lean operation region A, the amount of NOx stored in the NOx catalyst 17 increases. If this is left as it is, it will become saturated and the purification capacity of the NOx catalyst 17 will decrease, so that the stored NOx is decomposed and released into oxygen and nitrogen to recover the storage capacity of the catalyst 17, that is, the storage capacity recovery. The air-fuel ratio is enriched for the purpose.
[0044]
For example, as shown by a broken line a in the figure, the enrichment shifts the air-fuel ratio of the exhaust gas to at least a richer side than the air-fuel ratio during lean operation at a predetermined period T1 for a predetermined time t1 (for example, This is achieved by setting the stoichiometric air-fuel ratio as shown in the figure or making it richer than that. In this case, the processing cycle T1 is experimentally set in advance in a shorter cycle than when the NOx catalyst 17 is in the NOx saturation state. Further, the processing time t1 is experimentally set in advance as a time during which all of the stored NOx is reduced and released in relation to the degree of enrichment.
[0045]
Instead of such regular enrichment, the enrichment may be performed according to the flowchart shown in FIG. 4, for example, in order to perform the reduction and release of the occluded NOx better and more appropriately. The time variation of the NOx occlusion amount in this case is shown by a solid line a in FIG.
[0046]
That is, in step S1, the basic value (basic NOx occlusion amount) Qnoxa of the NOx occlusion amount is estimated based on the operating state of the engine 1 such as the engine speed and the engine load. Next, in step S2, the corrected NOx occlusion amount Qnoxs is calculated by correcting the basic NOx occlusion amount Qnoxa based on the estimated sulfur adhesion amount Qss used in the sulfur poisoning elimination control described later, according to Equation 1.
[0047]
[Expression 1]

[0048]
Here, the estimated sulfur adhesion amount Qss is 1 from zero when the adhesion amount when NOx is substantially not occluded in the NOx catalyst 17 is 1 and the adhesion amount when no NOx is adhering is zero. Is substituted into the above formula as a value (ratio).
[0049]
If the corrected NOx storage amount Qnoxs is equal to or greater than the predetermined storage amount Qnox1 in step S3, the air-fuel ratio of the exhaust gas is shifted to the rich side at least as compared with the air-fuel ratio during lean operation as described above in step S4. (For example, the stoichiometric air-fuel ratio is set as illustrated in FIG. 3 or richer than that).
[0050]
In step S5, when the enrichment time has passed the predetermined time t2, in step S6, the enrichment is terminated and the air-fuel ratio is returned to the original lean state. In this case, the processing time t2 is experimentally set in advance as a time during which the predetermined NOx occlusion amount Qnox1 is all reduced and released in relation to the degree of enrichment.
[0051]
The NOx occluded in the NOx catalyst 17 is discharged periodically or irregularly as described above, and, for example, as shown by chain lines C and D in FIG. Even when the operating region is switched from the lean operating region A to the stoichiometric air-fuel ratio operating region C or when the operating region is switched to the rich operating region B, the NOx storage capacity of the catalyst 17 is restored.
[0052]
Further, the occluded NOx is released from the NOx catalyst 17 in the same manner as the air-fuel ratio is enriched in sulfur poisoning elimination control and correction control of the estimated sulfur adhesion amount Qss used in the control.
[NOx catalyst sulfur poisoning elimination control]
FIG. 5 is a time chart showing an example of a specific operation of the sulfur poisoning elimination control of the NOx catalyst 17. Since the NOx catalyst 17 contains barium, there is a problem of sulfur poisoning in which sulfur components in the fuel adhere to the catalyst 17. If this is left as it is, the NOx purification capacity is lowered, and therefore, the treatment for releasing the adhering sulfur and restoring the purification capacity of the catalyst 17, that is, the temperature rise of the NOx catalyst 17 for eliminating sulfur poisoning is performed.
[0053]
The temperature rise of the NOx catalyst 17 is achieved by raising the exhaust gas temperature by split injection of fuel or retarding the ignition timing. Sulfur poisoning progresses more easily when the air-fuel ratio is lean, and becomes slower as it becomes richer. For example, as shown by the chain line in the figure, when the air-fuel ratio is the stoichiometric air-fuel ratio, sulfur poisoning is almost stagnant, and as shown by the chain line in the figure, the air-fuel ratio becomes richer than the stoichiometric air-fuel ratio. The release of sulfur begins. Therefore, in this sulfur poisoning elimination control, the NOx catalyst 17 is heated up after enriching the air-fuel ratio to promote the release of sulfur.
[0054]
In addition, the sulfur deposition rate in the lean state is much slower than the NOx occlusion rate. Therefore, this sulfur poisoning elimination control is executed less frequently than the above-mentioned NOx purification control. For example, the NOx purification control is repeatedly performed many times while the sulfur poisoning elimination control is performed once. In the time charts of FIG. 5 and FIG. 3, the time intervals are not necessarily represented the same.
[0055]
FIG. 6 is a main flowchart showing an example of a specific operation of sulfur poisoning elimination control. This program is repeatedly executed regardless of, for example, fuel injection amount control, fuel injection timing control, ignition timing control, etc. during the period when the ignition switch is on.
[0056]
First, in step S11, the sulfur poisoning time Ts is measured. The sulfur poisoning time Ts is an elapsed time from the time when the sulfur poisoning of the NOx catalyst 17 is almost completely eliminated. That is, when the process proceeds to step S11 for the first time after the sulfur poisoning has been almost completely eliminated, the sulfur poisoning time Ts is newly calculated from zero, and otherwise, the previous measurement is continued.
[0057]
Next, in step S12, the continuous duration Tln of the lean operation is measured. That is, as shown in FIG. 7, it is determined whether or not the lean operation is performed in step S41. When the lean operation is performed, the lean continuation time Tln is measured in step S42, while when the lean operation is not performed, the lean continuation time Tln is determined in step S43. To reset. However, when the lean continuation time Tln is reset, the last value of the lean continuation time Tln measured in step S42 is stored in the memory. Thus, when the lean operation is interrupted by the NOx purification control, interrupted according to the driving operation of the driver, or interrupted by the correction control of the estimated sulfur adhesion amount Qss described later, the lean that has been performed until then is performed. The continuous duration of operation Tln remains in the memory.
[0058]
Returning to the main flow, next, in step S13, the sulfur adhesion amount Qss to the NOx catalyst 17, that is, the sulfur poisoning state of the NOx catalyst 17, is estimated. The estimation of the sulfur adhesion amount Qss is performed according to the flowchart of FIG. 8. After reading various data in step S51, the previous sulfur adhesion amount Qss is estimated in step S52, and the current sulfur adhesion amount Qss is calculated. The amount of fuel (fuel supply amount) Qf injected from the injector 6 until the estimation is calculated.
[0059]
Next, in step S53, based on the fuel supply amount Qf, the sulfur adhesion amount increased from the previous sulfur adhesion amount Qss to the current sulfur adhesion amount Qss (ie, per unit time). A basic value (basic sulfur increase amount) Qsa of an instantaneous value of the amount of sulfur adhering to the NOx catalyst 17 is estimated. Here, as shown in FIG. 9, the basic sulfur increase amount Qsa is estimated so as to be substantially proportional to the supply amount Qf of the fuel that is the source of sulfur.
[0060]
Next, in step S54, the corrected basic sulfur is corrected by correcting the basic sulfur increase amount Qsa based on the predicted sulfur content Kfss of the fuel calculated by the correction control of the estimated sulfur adhesion amount Qss described later according to Equation 2. An increase amount Qsb is calculated.
[0061]
[Expression 2]

[0062]
Here, the predicted value Kfss is substituted into the above equation as a coefficient (ratio) when the sulfur content of the fuel considered so far is 1. That is, since there is a variation in the sulfur content of the fuel, the basic sulfur increase amount Qsa simply estimated based on the fuel supply amount Qf has a deviation from the actual sulfur adhesion amount. This corrects and improves the estimation accuracy of the amount of sulfur adhesion.
[0063]
For example, if the same fuel is used and the sulfur content does not change, the fuel sulfur content prediction value (coefficient) Kfss is approximately 1. The calculation of the predicted value Kfss will be described in detail in the correction control of the estimated sulfur adhesion amount.
[0064]
Next, in steps S55 to S57, correction coefficients K1, K2, and K3 for the corrected basic sulfur increase amount Qsb are set. That is, in step S55, the first correction coefficient K1 is set based on the exhaust gas temperature Tmp detected by the exhaust temperature sensor 27, and in step S56, the second correction coefficient K2 is set based on the lean operation duration Tln measured in step S42. In step S57, the third correction coefficient K3 is set based on the previous value (previously sulfur deposition amount) Qss [i-1] of the estimated sulfur deposition amount.
[0065]
Here, as shown in FIG. 10, the first correction coefficient K1 is set to a small value regardless of whether the predetermined exhaust gas temperature Tmpa is higher or lower than the peak. Further, as shown in FIG. 11, the second correction coefficient K2 is set to a small value regardless of whether the predetermined lean operation duration time Tlna is longer or shorter than the peak. Further, as shown in FIG. 12, the third correction coefficient K3 is set to a small value when it exceeds a predetermined existing sulfur adhesion amount Qssa [i-1].
[0066]
In particular, the reason why the second correction coefficient K2 has the above characteristics is as follows. That is, when the lean operation duration time Tln is long, as shown in FIG. 13, a large amount of NOx and sulfur are already attached to barium used in the NOx catalyst 17. Therefore, it becomes difficult for sulfur to newly adhere to barium, and the amount of sulfur deposited per unit time decreases.
[0067]
Further, when the lean continuation time Tln is short, as shown in FIG. 14, the contact time between sulfur and barium becomes short, and it becomes difficult to form a strong bond between them. Therefore, when switching to the stoichiometric air-fuel ratio operation or rich operation next time, sulfur is likely to be released in the same manner as NOx, and the amount of sulfur adhered per unit time is also reduced.
[0068]
In other words, even if the cumulative sulfur adhesion amount Qss increases with the passage of time, for example, even in the same 5-minute lean operation, the 5-minute lean operation is performed as compared with the case where the 1-minute lean operation is performed five times. When the lean operation (corresponding to the predetermined lean operation duration time Tlna shown in FIG. 11) is performed once, the cumulative total amount of the sulfur adhesion amount increases. Further, for example, even in the case of the same 20-minute lean operation, a 5-minute lean operation (corresponding to the predetermined lean operation duration time Tlna shown in FIG. 11 as compared with the case where the 20-minute lean operation is performed once) ) Is performed four times, the cumulative total amount of sulfur deposition is also increased.
[0069]
Therefore, it is advantageous from the viewpoint of making it difficult for sulfur to adhere to the NOx catalyst 17 to make the time during which the lean operation is continuously performed shorter or longer than the predetermined time Tlna. For example, when the air-fuel ratio enrichment for recovery of the storage capacity is periodically performed in the NOx purification control described above (reference number a in FIG. 3), the cycle T1 may be set shorter than the predetermined time Tlna.
[0070]
Returning to the sub-flow of FIG. 8, next, in step S58, the corrected basic sulfur increase amount Qsc is calculated by multiplying the corrected basic sulfur increase amount Qsb by the first to third correction coefficients K1, K2, and K3. To do.
[0071]
[Equation 3]

[0072]
In step S59, the current estimated sulfur deposition amount Qss is calculated by adding the corrected sulfur increase amount Qsc to the existing sulfur deposition amount Qss [i-1] according to Equation 4.
[0073]
[Expression 4]

[0074]
The sulfur poisoning starts preferentially from the upstream portion of the NOx catalyst 17 into which the exhaust gas first flows. That is, the sulfur component does not uniformly adhere to the NOx catalyst 17, and the sulfur component adheres unevenly along the exhaust gas passage. Therefore, in estimating the sulfur adhesion amount Qss, the NOx catalyst 17 is generally divided into n blocks along the exhaust gas passage, and the sulfur adhesion amount Qss [j] (j = 1 to 1) is determined for each block. The division number n) is estimated, and the sum (Qss [1] +... + Qss [n]) can be used as the sulfur adhesion amount Qss of the entire catalyst 17. In this case, the sulfur adhesion amount Qss [j] of the upstream block is estimated to be larger than that of the downstream block. Alternatively, the average value ((Qss [1] +... + Qss [n]) / n) of the sulfur adhesion amount Qss [j] estimated for each block may be handled as a value representing the sulfur adhesion amount Qss of the catalyst 17. it can. Such a concept will be described in more detail later in the estimation of the sulfur residual amount Qzs in step S21.
[0075]
Returning to the main flow, next, in step S14, it is determined whether or not the flag F is set to 1. In short, this flag F is a flag that serves as an indicator of whether the NOx catalyst 17 is in a state where sulfur poisoning should be eliminated, and is set to 1 when it is to be eliminated, and reset to zero when it is not. Is done.
[0076]
In step S14, when the flag F is not set to 1, in step S15, it is determined whether or not the sulfur adhesion amount Qss estimated in step S13 is equal to or larger than a predetermined determination reference amount Qs1. If it is greater than or equal to the reference amount Qs1, the flag F is set to 1 in step S16, and the process proceeds to step S17 and subsequent steps. On the other hand, when the estimated sulfur adhesion amount Qss is less than the determination reference amount Qs1 in step S15, the process returns to step S12. If the flag F is already set to 1 in step S14, steps S15 and S16 are skipped and the process immediately proceeds to step S17 and subsequent steps.
[0077]
Here, the determination reference amount Qs1 is set to, for example, a sulfur adhesion amount when the NOx purification capacity of the NOx catalyst 17 is reduced to 80%. That is, the flag F is a sulfur removal request flag that indicates whether or not an amount of sulfur component that decreases so that the NOx purification capacity of the NOx catalyst 17 affects the NOx emission is attached to the catalyst 17. .
[0078]
Next, in step S17, a first target temperature Tmp1, which is a standard target value of the exhaust gas temperature Tmp when the sulfur release process is performed on the NOx catalyst 17, is set. The first target temperature T1 is set based on the sulfur poisoning time Ts measured in step S11. In this case, as shown in FIG. 15, the first target temperature Tmp1 is set to a temperature between, for example, 650 ° C. and 700 to 750 ° C., and when the sulfur poisoning time Ts is long, compared to when it is short. Set to a high temperature. Further, the first target temperature Tmp1 may be further corrected based on the sulfur adhesion amount Qss estimated in step S13. Also in this case, according to FIG. 15, the first target temperature Tmp1 is corrected to a higher temperature when the estimated sulfur adhesion amount Qss is large than when it is small.
[0079]
Next, in step S18, a final target value Tmps of the exhaust gas temperature Tmp when the sulfur release process is executed is set. That is, the first target temperature Tmp1 set in step S17 is compared with the second target temperature Tmp2 set in step S29 described later, and the higher one is finally selected as the target exhaust gas temperature Tmps.
[0080]
Note that the second target temperature Tmp2 is generally set to a higher value than the first target temperature Tmp1. In addition, the first target temperature Tmp1 is always set, whereas the second target temperature Tmp2 is not always set. That is, when the second target temperature Tmp2 is not set, the first target temperature Tmp1 is set as the final target temperature Tmps, and when the second target temperature Tmp2 is set, the second target temperature Tmp2 is set as the final target temperature Tmps. Tend to.
[0081]
In the sulfur release process in step S20, fuel split injection and ignition timing retard are performed so that the final target exhaust gas temperature Tmps is realized, and as a result, the sulfur component adhering to the NOx catalyst 17 is removed. The In that case, the final target temperature may be a temperature that is a predetermined temperature higher than the final target exhaust gas temperature Tmps, which is normally set, for a certain period at the start of the sulfur release process in step S20. Thereby, the rising of the exhaust gas temperature Tmp is promoted, and the sulfur component starts to be removed immediately after the start of the release process. For example, when the first target temperature Tmp1 is selected as the final target exhaust gas temperature Tmps, the chain line shown in FIG. 15 exemplifies a high target temperature that is employed only for a certain period at the start of the sulfur release process. .
[0082]
Next, in step S19, it is determined whether or not the sulfur release execution permission condition is satisfied. Here, it is determined whether or not the vehicle speed V is equal to or higher than a predetermined vehicle speed V1 that does not cause a problem even if the sulfur release process is executed.
[0083]
The sulfur release execution permission condition is set by the vehicle speed because the exhaust gas temperature Tmp is a target temperature Tmps (for example, this temperature) at which the exhaust gas temperature Tmp is a sulfur releaseable temperature even if the partial injection or the ignition timing retard is performed for the sulfur release process at the low vehicle speed. In the embodiment, it does not rise to Tmps ≧ 650 ° C. from FIG. 15, and an efficient sulfur release process is not realized, and when the divided injection and the ignition timing retard are performed for the sulfur release process at a low vehicle speed. The reason is that the engine output becomes excessively unstable.
[0084]
However, the conditions for permitting the execution of sulfur release are not limited to this. In general, other conditions may be used for the purpose of eliminating a situation in which some trouble is accompanied by the execution of the sulfur release process or the accompanying trouble becomes relatively large. You may set using a parameter.
[0085]
And when it determines with the vehicle speed V being the said predetermined vehicle speed V1 or more, it progresses to step S20 and performs a sulfur discharge | release process. On the other hand, if it is determined that the vehicle speed V is not equal to or higher than the predetermined vehicle speed V1, the process proceeds to step S28 without executing the sulfur release process.
[0086]
The sulfur release process in step S20 is performed according to the flowchart shown in FIG. First, in step S61, the normal target air-fuel ratio map shown in FIG. 2 that is currently used is switched to a target air-fuel ratio map for sulfur release processing as shown in FIG. 17, for example. Here, in the target air-fuel ratio map at the time of the sulfur release process, the entire operation region is the stoichiometric air-fuel ratio region C. That is, the lean operation currently performed is prohibited to forcibly set the stoichiometric air-fuel ratio operation, and fuel cut (F / C) is prohibited. By performing the theoretical air-fuel ratio operation, the CO and HC concentrations in the exhaust gas are increased, and an environment in which sulfur components attached to the NOx catalyst 17 are easily released is generated. Further, by prohibiting the fuel cut, the fuel is always supplied, and the sulfur release process using the fuel injection control is stably performed.
[0087]
Next, in step S62, it is determined whether or not the current operation state is in the divided injection region. That is, it is determined whether or not the current operating state of the engine 1 is within an operating region in which the temperature of the exhaust gas for sulfur release is increased by split fuel injection. Here, the divided injection region is set to a middle-rotating load region (all portions shaded) in the map at the time of the sulfur release process of FIG.
[0088]
If it is the split injection region, the fuel late injection amount Qfks and the late injection timing θfks are set in steps S63 and S64. Here, the latter-stage injection means that the fuel is injected during the compression stroke, as in the normal time map of FIG. 2, and the vaporization of the fuel does not progress as compared with the first-stage injection in which the fuel is injected during the intake stroke. , Unburned components increase.
[0089]
The late injection amount Qfks in step S63 is set by multiplying the total fuel injection amount Qfts separately determined from the intake air amount and the like by the late injection amount coefficients K4 and K5 (both values less than 1) according to Equation 5. It is done by.
[0090]
[Equation 5]

[0091]
Here, as shown in FIG. 18, the first late injection amount coefficient K4, which is the fourth correction coefficient, is the target exhaust gas temperature Tmps (the target exhaust gas temperature Tmps is increased by a predetermined temperature for a certain period at the start of the sulfur release process). The higher the target temperature Tmps, the larger the value is set. Further, as shown in FIG. 19, the second late injection amount coefficient K5, which is the fifth correction coefficient, is determined based on the actual exhaust gas temperature Tmp, and is set to a larger value as the actual exhaust gas temperature Tmp is lower. Therefore, the higher the target exhaust gas temperature Tmps and the lower the actual exhaust gas temperature Tmp, the greater the late injection amount Qfks. As a result, the unburned components of the fuel further increase and the temperature of the NOx catalyst 17 is raised more quickly.
[0092]
On the other hand, in step S64, the late injection timing θfks is set to be slower as the engine load is lower. Therefore, as the load is low, the late injection timing θfks is delayed, the vaporization of fuel does not progress further, the unburned components increase, and the NOx catalyst 17 is heated more quickly.
[0093]
The late injection amount Qfks and the late injection timing θfks set in this way are output as control signals to the injector 6 and are used to control the fuel injection amount and fuel injection timing for the engine 1.
[0094]
Next, in step S65, it is determined whether or not the retard (retarding) control region. That is, it is determined whether or not the current operation region of the engine 1 is within the operation region in which the temperature of the exhaust gas for sulfur release is increased by the ignition timing retard. Here, the retard control region is set so as to overlap the low load side region (the portion hatched densely below line D) in the divided injection region in the map at the time of the sulfur release process of FIG. Yes. That is, in this retard control region, both fuel split injection and ignition timing retard are performed.
[0095]
If it is the retard control area, the retard amount is set in step S66. In particular, in step S66, the retard amount is feedback controlled so that the actual exhaust gas temperature Tmp converges to the target temperature Tmps.
[0096]
The setting of the retard amount and the feedback control thereof in step S66 are performed according to the flowchart shown in FIG. First, in step S71, the actual exhaust gas temperature Tmp detected by the exhaust gas temperature sensor 27 is read, and in step S72, it is determined whether or not the execution of retard control is started.
[0097]
Then, the process proceeds to step 73 only once at the start of the retard control, and it is determined whether or not the actual exhaust gas temperature Tmp is equal to or higher than the target temperature Tmps. As a result, when the actual exhaust gas temperature Tmp is equal to or higher than the target temperature Tmps, a preset high-temperature retard amount is set as the basic retard amount in step S74. On the other hand, when the actual exhaust gas temperature Tmp is not equal to or higher than the target temperature Tmps, a preset low-temperature retard amount is set as the basic retard amount in step S75.
[0098]
Here, the amount of retard for low temperature is set to be larger than the amount of retard for high temperature. As a result, when the exhaust gas temperature Tmp is low, a higher temperature increase is achieved than when the exhaust gas temperature Tmp is high.
[0099]
On the other hand, when the retard control is not started, that is, when the retard control has already started, the process proceeds directly to step S76 to determine whether or not the actual exhaust gas temperature Tmp is equal to or higher than the target temperature Tmps. As a result, when the actual exhaust gas temperature Tmp is equal to or higher than the target temperature Tmps, it is further determined in step S77 whether or not the actual exhaust gas temperature Tmp is equal to or higher than a temperature (Tmps + ΔTmp) higher than the target temperature Tmps by the hysteresis increment ΔTmp. If so, the retard amount is decreased by a predetermined amount in step S78.
[0100]
On the other hand, when the actual exhaust gas temperature Tmp is equal to or higher than the target temperature Tmps and lower than the temperature (Tmps + ΔTmp) higher by the hysteresis increment ΔTmp in step S77, the retard amount is maintained as it is.
[0101]
Furthermore, when it is determined in step S76 that the actual exhaust gas temperature Tmp is not equal to or higher than the target temperature Tmps, the retard amount is increased by a predetermined amount in step S79.
[0102]
The retard amount of the ignition timing thus set is output as a control signal to the ignition circuit 35 of the spark plug 5 and used for controlling the ignition timing for the engine 1.
[0103]
As described above, as apparent from the flowchart of the sulfur release process of FIG. 16 and the air-fuel ratio map at the time of the sulfur release process of FIG. 17, the operating state of the engine 1 is on the high load side in the middle rotation load region ( In the case of an affirmative determination in step S62 and a negative determination in step S65), only split injection is performed to raise the temperature of the exhaust gas. Further, when the operating state of the engine 1 is on the low load side in the middle-rotation load region (in the case of an affirmative determination in step S62 and an affirmative determination in step S65), the divided injection and the retard of the ignition timing are detected. It is done together.
[0104]
However, when the operating state of the engine 1 is not originally in the middle rotation load region (in the case of negative determination in step S62), neither split injection nor ignition timing retard is performed. That is, the temperature of the exhaust gas is not substantially increased during low rotation and low load and during high rotation and high load.
[0105]
This is because the exhaust gas temperature Tmp is originally low in the low rotation and low load region, and even if the exhaust gas is heated, it may not rise to the target temperature Tmps at which sulfur release processing can be performed. Since the output state of the engine 1 may become unstable, the purpose is to avoid such waste.
[0106]
On the other hand, in the high rotation and high load region, the exhaust gas temperature Tmp is originally high, and there is a possibility that sulfur emission may be performed without raising the exhaust gas temperature. When the exhaust gas temperature is raised, the NOx absorbent and therefore the NOx catalyst. Again, it is the purpose of avoiding such waste because 17 can be damaged by becoming too hot.
[0107]
Further, by taking such measures, problems such as deterioration in fuel consumption and torque reduction that accompany the temperature rise of exhaust gas using split injection and retard can be suppressed to the minimum necessary.
[0108]
Furthermore, even when the operating state of the engine 1 is in the middle-rotation load region and the exhaust gas is heated, basically, the exhaust gas is heated by split injection, and the exhaust gas temperature is relatively low. Since retard is also performed only at low loads where it is necessary to raise the temperature to a large extent, the problem of torque reduction due to retard is still suppressed to the minimum necessary.
[0109]
Returning to the main flow, next, in step S21, the sulfur remaining amount Qzs remaining in the NOx catalyst 17 is estimated. The sulfur remaining amount Qzs is estimated according to the flowchart of FIG. 21. First, in step S81, whether the exhaust gas temperature Tmp is equal to or higher than the lowest sulfur dischargeable temperature (for example, 650 ° C. in this embodiment). If the exhaust gas temperature Tmp is less than 650 ° C., the sulfur release processing time Tr is reset in step S82, and if it is 650 ° C. or higher, the sulfur release processing time Tr is measured in step S83. To do.
[0110]
That is, the sulfur component can be released when the exhaust gas temperature Tmp and thus the temperature of the NOx catalyst 17 is at least 650 ° C. or more in the same manner as when the first target temperature Tmp1 is set in FIG. This 650 ° C. is set as the lowest temperature at which the sulfur component can be released. Of course, the sulfur releaseable temperature can be set to various values depending on other conditions such as the air-fuel ratio of the exhaust gas flowing into the NOx catalyst 17, for example.
[0111]
Next, in step S84 and the subsequent steps, the sulfur release amount is estimated based on the sulfur release processing time Tr and the exhaust gas temperature Tmp detected by the exhaust temperature sensor 27, and the sulfur residual amount Qzs is estimated from the sulfur release amount. .
[0112]
First, an outline of the estimation operation of the sulfur remaining amount Qzs performed in step S84 and subsequent steps will be described. As described in the operation of estimating the sulfur adhesion amount Qss in step S13, the exhaust gas flows first from the upstream portion of the NOx catalyst 17, as shown in FIG. Therefore, also in the sulfur release process executed in step S20, the exhaust gas that has been heated to the sulfur releaseable temperature flows first from the upstream portion of the NOx catalyst 17, and the upstream portion is heated first, so that the sulfur Release begins preferentially in the upstream part. In other words, the NOx catalyst 17 does not release and remove the sulfur component uniformly, and the sulfur component is removed unevenly along the exhaust gas passage and remains unevenly.
[0113]
Therefore, in estimating the sulfur emission amount or the remaining amount Qzs, the NOx catalyst 17 is generally divided into n blocks (10 in the illustrated example) along the exhaust gas passage route as shown in FIG. In view of this, the residual sulfur amount Qzs [j] (j = 1 to the number of divisions n) is estimated for each block, and the sum (Qzs [1] +... + Qzs [n]) is calculated as the residual sulfur amount Qzs of the entire catalyst 17. It is. As a result, it is possible to accurately estimate the amount of released sulfur or the remaining amount of sulfur. Obviously, the sulfur residual amount Qzs [j] in the left upstream block in FIG. 22 is estimated to be smaller than that in the downstream block.
[0114]
In addition, instead of the sum of the remaining sulfur amount Qzs [j] estimated for each block (Qzs [1] +... + Qzs [n]), an average value of the remaining sulfur amount Qzs [j] estimated for each block ( (Qzs [1] +... + Qzs [n]) / n) may be treated as a value representative of the sulfur residual amount Qzs of the catalyst 17.
[0115]
In this embodiment, as an example, as shown in FIG. 22, the NOx catalyst 17 is divided into 10 blocks (j = 1 to 10). In step S84, based on the sulfur release processing time Tr and the exhaust gas temperature Tmp, the ratio (%) of the area (completely removed area) SA in which the sulfur component is completely released and removed out of the volume of the total NOx catalyst 17, in other words Then, the number of blocks from which the sulfur component has been completely released is set from a characteristic map as shown in FIG. As shown in FIG. 23, the complete removal area SA becomes larger as the discharge processing time Tr becomes longer and the exhaust gas temperature Tmp becomes higher. That is, the number of blocks belonging to the complete removal area SA increases.
[0116]
The map in FIG. 23 shows, for example, when the exhaust gas temperature Tmp is 700 ° C., when the sulfur release processing time Tr reaches the time t11, and when the exhaust gas temperature Tmp is 650 ° C., at a later time t12. When reached, sulfur is released from all 10 blocks, indicating that the complete removal area SA is 100%.
[0117]
In the next step S85, based on the sulfur release processing time Tr and the exhaust gas temperature Tmp, the ratio (%) of the area (partial removal area) SB in which the sulfur component is partially released and removed in the NOx catalyst 17 ( The number of blocks) is set from a characteristic map as shown in FIG. As shown in FIG. 24, the partial removal region SB becomes smaller as at least the release processing time Tr becomes longer, and disappears when the complete removal region SA becomes 100% at the time t11 or t12.
[0118]
The ratio (%) (number of blocks) of the unremoved area SC from which no sulfur component has been removed is the value obtained by adding the ratio of the completely removed area SA and the ratio of the partially removed area SB to the total NOx catalyst 17. It is obtained by subtracting from the volume of.
[0119]
In the next step S86, the sulfur removal rate α (%) in the partial removal region SB is set from a characteristic map as shown in FIG. 25 based on the sulfur release processing time Tr and the exhaust gas temperature Tmp. As shown in FIG. 25, in this embodiment, the sulfur removal rate α (%) in the partial removal region SB is set to increase as the exhaust gas temperature Tmp increases regardless of the release processing time Tr. Yes.
[0120]
In addition, the sulfur residual rate β (%) in the partial removal region SB is (100−α)%, the sulfur removal rate in the complete removal region SA is 100% (the sulfur residual rate is 0%), and the unremoved region The sulfur removal rate in SC is 0% (the sulfur residual rate is 100%).
[0121]
In the next step S87, the sulfur remaining amount Qzs of the entire NOx catalyst 17 is finally calculated from each data obtained above. For this purpose, for example, the amount of sulfur components released for each region SA, SB, SC is estimated, the sum is obtained, and the value is subtracted from the sulfur adhesion amount Qss estimated in step S13. Alternatively, the amount of the sulfur component released for each region SA, SB, SC is estimated, and the value is calculated based on the amount of the sulfur component attached to each region SA, SB, SC (for example, “Qss / j (number of blocks ) "), And the sum may be obtained.
[0122]
Here, each sulfur remaining amount Qzs [j] of the block (upstream block) belonging to the completely removed area SA is zero, and each sulfur remaining amount Qzs [j] of the block (downstream block) belonging to the unremoved area SC. Remains the sulfur adhesion amount (Qss / j) estimated in step S13, and the sulfur residual amount Qzs [j] of the block (center block) belonging to the partial removal region SB is ((Qss). / J) × (100−α)), the total sulfur residual amount Qzs of the entire NOx catalyst 17 is obtained, for example, according to Equation 6.
[0123]
[Formula 6]

[0124]
It should be noted that “SB” and “SC” described in the formula indicate the ratio (%) of the partially removed region and the unremoved region in the NOx catalyst 17, respectively. Similarly, “(j × SB / 100)” and “(j × SC / 100)” shown in the equation indicate the number of blocks belonging to the partially removed area and the unremoved area, respectively.
[0125]
The total amount of released sulfur or the total amount of remaining sulfur Qzs of the NOx catalyst 17 obtained here is the amount of released sulfur or the amount of remaining sulfur for each sulfur release treatment, and the amount of each sulfur release treatment. It is not the cumulative amount of sulfur released or the amount of residual sulfur released as a result. The sulfur release processing time Tr, which is a parameter for obtaining the released amount or the remaining amount, is reset when the exhaust gas temperature Tmp does not become equal to or higher than the sulfur releaseable temperature (650 ° C.) in steps S81 and S82. Indicates the time at which sulfur release was continuously performed, and is not an accumulation of the time at which sulfur release was performed intermittently.
[0126]
For example, even if the sulfur release process for 1 minute is performed five times, the sulfur release process time Tr employed is 1 minute in all the sulfur release processes. Therefore, from the maps shown in FIGS. If the temperature Tmp is the same, the ratio of the same complete removal area SA, the ratio of the same partial removal area SB, and the sulfur removal rate α in the same partial removal area SB are always read out. Accordingly, as a result, the same sulfur release amount and the same sulfur residual amount are calculated, and the sulfur release amount or the sulfur residual amount is not updated. That is, it is not a cumulative sum of all sulfur release amounts or residual sulfur amounts calculated as a result of the sulfur release treatment for 1 minute each.
[0127]
Therefore, for example, even in the same 5-minute sulfur release treatment, when the 5-minute sulfur release treatment is performed once compared to when the 1-minute sulfur release treatment is performed five times as described above, , The amount of sulfur released increases and the amount of sulfur remaining decreases.
[0128]
Then, for example, when the sulfur release process is performed for 1 minute and then the sulfur release process is performed for 2 minutes, the value of the sulfur release amount or the remaining amount is the result of the sulfur release for 2 minutes. It will be updated to the calculated larger sulfur release or smaller residual value.
[0129]
That is, as described above, the sulfur release process starts preferentially from the upstream portion of the NOx catalyst 17 into which the heated exhaust gas flows first, and the downstream portion of the NOx catalyst 17 where the heating is delayed is likely to have a certain amount of time. Sulfur is not removed. Then, since the temperature of the NOx catalyst 17 decreases again after one sulfur release process is completed and the next sulfur release process is performed again, in the second sulfur release process again, The NOx catalyst 17 starts to be heated from the upstream portion. Therefore, at this time and when the second sulfur release process is completed at the same time or less than the previous time, no new sulfur is removed in the second sulfur release process. It will end. Therefore, it can be said that as the sulfur release continuous time Tr becomes shorter, the overall sulfur removal rate decreases, and as the sulfur release continuous time Tr becomes longer, the overall sulfur removal rate improves.
[0130]
By the way, generally, at the time of rich operation and stoichiometric air-fuel ratio operation, the exhaust gas temperature Tmp becomes higher than that at the time of lean operation, and may naturally rise to a sulfur releaseable temperature (for example, 650 ° C.) or higher. That is, even if the sulfur release process of step S20 is not performed, the same effect as that of the sulfur release process may be obtained during the rich operation or the stoichiometric air-fuel ratio operation (reference symbol in FIG. 5). , Ki).
[0131]
Therefore, for example, when the rich operation or the theoretical air-fuel ratio operation (λ ≦ 1) and the exhaust gas temperature is equal to or higher than the sulfur releaseable temperature (Tmp ≧ 650 ° C.), the above step S21 in the sulfur release process is performed. The amount of sulfur component released and removed is estimated using the same method as the method for estimating the amount of released sulfur or the remaining amount of sulfur Qzs, and the estimated sulfur is subtracted from the sulfur adhesion amount Qss estimated in step S13. The adhesion amount Qss may be corrected.
[0132]
As a result, the estimation accuracy of the sulfur adhesion amount Qss is further improved, and the sulfur release process is started at a wasteful time, or the sulfur release process is continued for a long time. The trouble which arises when estimating is avoided.
[0133]
Returning to the main flow, next, in step S22, it is determined whether or not the total sulfur residual amount Qzs obtained in step S21 is zero. That is, it is determined whether or not all attached sulfur components are removed and the NOx storage capacity or purification capacity of the NOx catalyst 17 is completely recovered.
[0134]
Until the remaining sulfur amount Qzs becomes zero, in step S23, while confirming that the vehicle speed V is equal to or higher than the second predetermined vehicle speed V2 for determination as the second sulfur release execution permission condition, The sulfur release process and the estimation of the remaining sulfur amount Qzs in step S21 are repeated. Here, as shown in FIG. 5, the second predetermined vehicle speed V2 is set to a vehicle speed lower than the first predetermined vehicle speed V1. Thus, after a positive determination is made in step S19 and the sulfur release process is started once, even if the vehicle speed decreases to about the first predetermined vehicle speed V1, the sulfur release process becomes difficult to be canceled, and the exhaust gas temperature It is said that the decrease in the level is difficult to minimize.
[0135]
When it is determined in step S22 that the residual sulfur amount Qzs is zero, the process proceeds to step S27, the sulfur removal request flag F is reset to 0, and the sulfur poisoning time Ts indicating the sulfur deposition elapsed time. , And an operation characteristic counter to be described later is reset. After that, the process returns to step S11, and the sulfur poisoning time Ts is measured again from the beginning.
[0136]
Here, there are various patterns for reaching step S27. The best pattern among them is that the first and second sulfur release execution permission conditions (the vehicle speed V must be equal to or higher than the predetermined vehicle speeds V1 and V2 for determination) are from the beginning when the flag F is set to 1 in step S16. This is a case where the user continues to be satisfied and the negative determination is never made in step S19 and step S23, and the sulfur release process is completed from the beginning and completed (reference sign in FIG. 5). In this pattern, the functional degradation of the NOx catalyst 17 is suppressed to the lowest level. This first pattern corresponds to the case of a driver who tends to maintain a higher vehicle speed for a longer period of time, as is apparent from the positive determinations made at steps S19 and S23. The probability of doing is great.
[0137]
The remaining steps S24 to S26 and steps S28 to S30 are treatments when passing through routes other than the first pattern.
[0138]
That is, when it is determined in step S23 that the vehicle speed V is not equal to or higher than the second predetermined vehicle speed V2, that is, the vehicle is considerably decelerated during execution of the sulfur release process, and the exhaust gas temperature Tmp is obtained even when the divided injection and the retard control are executed. Is not raised to the target temperature Tmps and efficient sulfur release processing cannot be continued, or when there is a possibility that the engine output becomes unstable (reference numeral in FIG. 5), steps S24 and S25 are performed. When executed, the fuel split injection and the ignition timing retard control for raising the exhaust gas temperature are canceled, and the air-fuel ratio map during the sulfur release process of FIG. 17 is switched to the normal air-fuel ratio map of FIG. The forced theoretical air-fuel ratio operation is canceled, and the fuel cut (F / C) is restored.
[0139]
After that, in step S26, it is determined whether or not the remaining sulfur amount Qzs is equal to or smaller than a predetermined sulfur release processing end permission determination amount Qs2. Here, the end permission determination amount Qs2 is set to a relatively small value that is not zero. As a result, when the sulfur remaining amount Qzs is equal to or smaller than the end permission determination amount Qs2 (dashed line in FIG. 5), the process proceeds to step S27, and the sulfur release process is performed from the beginning to the full and is the same as when the process ends. To deal with. That is, when the sulfur removal request flag F, the sulfur poisoning time Ts, and an operation characteristic counter to be described later are reset and the process returns to step S11, the sulfur poisoning time Ts is measured again from the beginning.
[0140]
On the other hand, when the sulfur remaining amount Qzs is not less than or equal to the sulfur release processing end permission determination amount Qs2 in step S26 (solid line in FIG. 5), that is, when many sulfur components still remain without being removed, As in the case where it is determined in step S19 that the sulfur release execution permission condition is not satisfied, the process proceeds to step S28 and subsequent steps.
[0141]
In step S28, the operation characteristic counter is counted up. That is, in step S19, the sulfur release execution permission condition is not satisfied by the driving operation of the driver, and as a result, the sulfur adhesion amount Qss of the NOx catalyst 17 becomes equal to or greater than the predetermined amount Qs1, and it is necessary to remove sulfur. However, the situation in which the sulfur release process cannot be substantially started, or in step S23, the second sulfur release execution permission condition is not satisfied by the driving operation of the driver. Although many sulfur components still remain (Qzs> Qs2), the number of occurrences of a situation that must be interrupted is measured.
[0142]
Therefore, it can be said that as the count of the driving characteristic counter increases, the driving operation of the driver tends to travel more time at a lower vehicle speed. In other words, after the flag F is set to 1, it can be said that the sulfur release process has a driving characteristic that is difficult to complete quickly.
[0143]
Next, in step S29, a second target temperature Tmp2, which is a reinforcing target value of the exhaust gas temperature Tmp when the sulfur release process is executed, is set based on the count value of the operation characteristic counter. The second target temperature Tmp2 is set higher, for example, at 700 to 750 ° C., for example, as the number of counts is larger, that is, when the sulfur release process is more difficult to complete quickly.
[0144]
In step S18, the higher one of the reinforcing second target temperature Tmp2 and the standard first target temperature Tmp1 set in step S17 is selected as the final target exhaust gas temperature Tmps. Therefore, when the sulfur release process that is difficult to be executed is once executed, the exhaust gas or the NOx catalyst 17 is heated to a higher temperature, and the sulfur component is efficiently and quickly supplied. As a result, the sulfur release process is reliably completed in a short time.
[0145]
Further, in step S30, the sulfur poisoning of the NOx catalyst 17 is suppressed and the NOx occlusion is suppressed. That is, a considerable amount of sulfur component is attached to the NOx catalyst 17 (Qss ≧ Qs1 or Qzs> Qs2), and the NOx occlusion ability and the catalyst function are lowered. When NO is performed, NOx contained in a large amount in the exhaust gas is not stored in the catalyst 17 but released into the atmosphere.
[0146]
Therefore, for example, by reducing the lean operation region A in the target air-fuel ratio map shown in FIG. 2 and restricting or prohibiting the lean operation itself, the exhaust gas does not contain a large amount of NOx, and the NOx catalyst 17 This suppresses the release of NOx into the atmosphere due to the decrease in the catalytic function.
[0147]
In addition, by restricting and prohibiting lean operation, at the same time, the deposition rate of sulfur is slowed, and further, sulfur can be prevented from adhering to the NOx catalyst 17 (suppression of sulfur poisoning). In FIG. 5, the fact that the slope of the solid line (the slope of the estimated sulfur adhesion amount Qss) is small indicates this.
[0148]
In that case, it is more preferable that the extent of reducing the lean region A is determined according to the count number of the operation characteristic counter. This is because the count number represents the degree of decrease in the catalytic function of the NOx catalyst 17. Obviously, the larger the count number is, the more the catalytic function is lowered, so the lean operation area A is further reduced.
[0149]
As long as the operation characteristic counter is not reset in step S27, the sulfur removal request flag F is also continuously set to 1, so when returning, the steps S15 and S16 are skipped, and the first in step S19. When the sulfur release execution permission condition is satisfied, the sulfur release process is executed. However, for example, in a state where the sulfur removal request flag F is set to 1, when the release of sulfur is not performed all the way to the end (until Qzs = 0 or until Qzs ≦ Qs2) due to the low vehicle speed traveling tendency, The sulfur release process may be forcibly executed after the ignition switch is turned off or turned on. Thereby, when it is not necessary to travel, the sulfur can be released without hindering the traveling of the driver. In particular, when the engine is started, it is usual to warm up the engine with a rich air-fuel ratio, so that it can be used to efficiently release sulfur.
[0150]
In this case, however, it is preferable to notify the driver that the sulfur release process is being executed by warning means such as a buzzer or a warning light. In particular, it is possible to prevent a sense of incongruity that the engine does not stop even when the ignition switch is turned off, or a misconception that it is a failure.
[Correction control of estimated sulfur adhesion amount-calculation of sulfur content prediction value Kfss]
FIG. 26 is a time chart showing an example of a specific operation of air-fuel ratio control performed for correction control of the estimated sulfur adhesion amount Qss, and FIG. 27 is the same flowchart. As described in step S13 of sulfur poisoning elimination control (operation for estimating the amount of adhered sulfur Qss), generally, the amount of sulfur contained in the fuel is not constant and varies. The sulfur content of the supplied fuel changes from the previous one.
[0151]
Therefore, even if the fuel supply amount (injection amount) is the same, for example, as illustrated in FIG. 26, when the sulfur content increases due to refueling, the amount of sulfur adhering to the NOx catalyst 17 increases, which is indicated by a chain line. In this way, the poisoning state proceeds faster than estimated, and a larger amount of sulfur adheres than the estimated sulfur adhesion amount Qss (solid line). As a result, the sulfur release process is not started even when sulfur is already attached to the determination reference amount Qs1 or more, the state of the NOx catalyst 17 being depleted in purification capacity is wasted, or the sulfur release process is started. Even after that, although sulfur is still zero or the remaining allowance determination amount Qs2 or more remains, the sulfur release process is ended, and the purification ability of the NOx catalyst 17 is not restored to the original state.
[0152]
Conversely, when the sulfur content is reduced by refueling, the amount of sulfur attached to the NOx catalyst 17 is reduced and the poisoning state progresses later than estimated, and the amount of sulfur attached is less than the estimated amount of sulfur attached Qss. Sulfur adheres. As a result, even though sulfur has not yet adhered to the criterion amount Qs1 or more, the sulfur release process is started and fuel is wasted, or even after the sulfur release process is started, Has been removed to zero or less than the end permission determination amount Qs2, but the sulfur release process is continued and fuel is wasted.
[0153]
Therefore, in order to perform the sulfur release process without excess or deficiency, the basic sulfur increase amount Qsa estimated based only on the fuel supply amount Qf is corrected by the fuel sulfur content Kfss in the above-described step S13, more specifically, step S54. The deviation between the estimated sulfur adhesion amount and the actual sulfur adhesion amount was corrected to improve the estimation accuracy of the sulfur adhesion amount. Then, processing for that purpose, that is, enrichment of the air-fuel ratio for correcting the estimated adhesion amount is performed.
[0154]
As a result, in the engine 1, the enrichment for the storage capacity recovery in the NOx purification control, the enrichment for the sulfur release promotion in the sulfur poisoning elimination control of the NOx catalyst 17, and the correction of the estimated sulfur adhesion amount are performed. A plurality of types of air-fuel ratio enrichment are performed for different purposes, i.e., enrichment. In FIG. 26, the enrichment indicated by the symbol “n” is the enrichment of the air-fuel ratio performed by the sulfur poisoning elimination control, and the enrichment indicated by the symbol “d” is the enrichment of the air-fuel ratio for recovering the storage capacity, The enrichment indicated by the symbol N is a enrichment for correcting the estimated adhesion amount.
[0155]
This enrichment is performed according to the flowchart of FIG. This program is executed independently of the above-described sulfur poisoning elimination control. First, in step S91, it is immediately after refueling, in step S92, the sulfur removal request flag F is reset to zero, and in step S93, the lean operation is performed. In step S94, it is confirmed that the estimated sulfur adhesion amount Qss is equal to or larger than the determination reference amount Qs3 for performing the preset correction control. Here, the determination reference amount Qs3 is a value smaller than the above-described determination reference amount Qs1.
[0156]
When all the above conditions are satisfied, the process proceeds to step S95, and the corrected NOx occlusion amount Qnoxs calculated in step S2 of FIG. 4 in the NOx releasing process described above is read. The corrected NOx occlusion amount Qnoxs is obtained by correcting the basic NOx occlusion amount Qnoxa with the estimated sulfur adhesion amount Qss, as described above.
[0157]
Next, in step S96, the enrichment condition for correcting the estimated adhesion amount is determined based on the corrected NOx occlusion amount Qnoxs. In this case, as shown in FIG. 28, the enrichment condition (at least one of the enrichment time and the richness) is that the enrichment time becomes longer or the richness is increased as the corrected NOx storage amount Qnoxs increases. Set to be higher.
[0158]
In step S97, the air-fuel ratio is enriched so that the enrichment condition thus obtained is realized, and in step S98, the current O ₂ The output value of the second air-fuel ratio sensor 28 which is a sensor is detected. In particular, here, the maximum output value E of the second air-fuel ratio sensor 28 is detected. Here, the second air-fuel ratio sensor 28 has a characteristic that the higher the oxygen concentration, that is, the smaller the output value on the lean side, and the smaller the oxygen concentration, that is, the larger the output value on the rich side.
[0159]
In FIG. 26, only the output portion of the second air-fuel ratio sensor 28 at this time is shown enlarged.
[0160]
Next, in step S99, based on the output maximum value E, a correction coefficient Knox for the NOx occlusion amount is set. In this case, as shown in FIG. 29, the NOx occlusion amount correction coefficient Knox is set so as to decrease as the output maximum value E increases. That is, as the air-fuel ratio downstream of the NOx catalyst 17 shifts to the rich side where the oxygen concentration is low, the NOx occlusion amount is corrected to a smaller value. In other words, the sulfur adhesion amount is corrected to a large value.
[0161]
In this case, as the correction coefficient Knox approaches 1, the estimated value Qss of the sulfur adhesion amount approximates to the actual value.
[0162]
Next, in step S100, the detected NOx occlusion amount (detected NOx occlusion amount) Qnoxk is calculated by correcting the corrected NOx occlusion amount Qnoxs based on the correction coefficient Knox according to Equation 7. That is, the NOx occlusion amount stored in the NOx catalyst 17 at the time of execution of the enrichment for correcting the estimated adhesion amount is corrected based on the detection result of the oxygen concentration obtained by the enrichment. Become.
[0163]
[Expression 7]

[0164]
Further, in step S101, a detected value (detected sulfur adhesion amount) Qsk of the sulfur adhesion amount is calculated based on the detected NOx occlusion amount Qnoxk and the corrected NOx occlusion amount Qnoxs according to Equation 8. That is, the sulfur adhesion amount Qss estimated to have adhered to the NOx catalyst 17 at the time of execution of the enrichment for the estimated adhesion amount correction is based on the detection result of the oxygen concentration obtained by the enrichment. Will be corrected.
[0165]
[Equation 8]

[0166]
Next, in step S102, the predicted value Kfss of the sulfur content is updated based on the detected sulfur adhesion amount Qsk and the estimated sulfur adhesion amount Qss according to Equation 9. This updated value Kfss is used in step S54 of the sulfur poisoning elimination control to be performed next.
[0167]
[Equation 9]

[0168]
Furthermore, in step S103, the value of the detected sulfur adhesion amount Qsk is handled as it is as the value of the estimated sulfur adhesion amount Qss.
[0169]
As described above, from the detection result of the second air-fuel ratio sensor 28, the estimated sulfur adhesion amount Qss is corrected, the NOx occlusion amount Qnoxs is also corrected, and the sulfur content Kfss of the fuel is also predicted.
[0170]
In this way, the air-fuel ratio of the exhaust gas discharged from the combustion chamber 4 is forcibly enriched, and the estimation to the catalyst 17 is performed based on the air-fuel ratio detected by the air-fuel ratio sensor 28 downstream of the NOx catalyst 17 at that time. The sulfur adhesion amount Qss is corrected. The oxygen concentration in the rich atmosphere changes depending on the degree of NOx catalyst 17 actually poisoned with sulfur, and reflects the degree of sulfur poisoning. Therefore, the amount of sulfur adhesion estimated based on the fuel supply amount Qf By correcting Qss based on the oxygen concentration, that is, based on the air-fuel ratio downstream of the NOx catalyst 17, the influence on the estimation accuracy due to the variation in the sulfur content of the fuel is corrected.
[0171]
As a result, it is possible to measure the sulfur content of the fuel according to the variation in the sulfur content of the fuel by using the usual method of enriching the air-fuel ratio without measuring the sulfur content of the fuel or inputting it to the control unit 20. The problem that the estimation accuracy of the adhesion amount Qss is lowered can be dealt with rationally and efficiently, and as a result, the sulfur poisoning elimination control can be executed well without excess or deficiency.
[0172]
In step S98, as the detection result of the second air-fuel ratio sensor 28, the output maximum value E of the air-fuel ratio sensor 28 is detected, but instead of or together with this, as shown in FIG. Alternatively, the delay time D of the rise of the output value of the sensor 28 from small to large when the lean state is switched to the rich state, or the area M surrounded by the waveform of the output value of the sensor 28 may be detected.
[0173]
Further, by utilizing the enrichment shown in FIG. 4 that is necessarily performed in order to restore the NOx storage capacity of the NOx catalyst 17, the air-fuel ratio downstream of the catalyst 17 is detected, and the estimated sulfur adhesion amount Qss is corrected. May be. The number of times the air-fuel ratio is enriched does not increase suddenly, and it is suppressed that the lean operation is obstructed and the fuel efficiency is impaired.
[0174]
On the other hand, as described above, the enrichment shown in FIG. 4 and the enrichment for correcting the estimated adhesion amount are performed independently, and at least after the fuel is supplied, the sulfur release process is completed first. (F = 0), when the estimated sulfur deposition amount Qss increases by a predetermined amount Qs3 or more, the air-fuel ratio of the exhaust gas is enriched, thereby enriching and emptying at an appropriate time for correcting the estimated sulfur deposition amount. The detection of the fuel ratio can be performed, and the correction accuracy is further improved. Even in that case, since the enrichment needs to be performed much less frequently than the enrichment for restoring the storage capacity, the number of times of enriching the air-fuel ratio does not increase and the lean operation is obstructed unnecessarily. This also prevents the fuel efficiency from being impaired.
[0175]
Furthermore, since the sulfur release process is completed (F = 0), the estimated value of the estimated sulfur adhesion amount Qss becomes zero, and the determination accuracy of whether the adhesion amount Qss has increased by a predetermined amount Qs3 or more is improved. . In addition, since the enrichment is performed at the above timing when the sulfur release process is first performed after refueling, the influence of the variation in the sulfur content of the fuel that changes due to refueling is corrected early.
[0176]
In step S96, the enrichment conditions such as the enrichment degree and enrichment time are finally determined based on the estimated sulfur adhesion amount Qss at that time, so from the viewpoint of NOx release, the stored NOx is In addition, it is properly released at a richness level without excess or deficiency, or for a time without excess or deficiency. As a result, untreated occluded NOx remains in the catalyst 17, or conversely, it is excessively enriched and fuel consumption From the viewpoint of correcting the estimated sulfur adhesion amount, all the occluded NOx is reduced and decomposed into oxygen, and a sufficient amount of oxygen is generated. An air-fuel ratio detection value (the above-mentioned delay time, area, output maximum value, etc.) is obtained, and the correction accuracy is improved.
[0177]
【The invention's effect】
As described above, according to the present invention, the air-fuel ratio of the exhaust gas discharged from the combustion chamber is forcibly enriched and detected at that time downstream of the NOx catalyst. Oxygen concentration Based on this, the estimated sulfur adhesion amount to the catalyst is corrected. The oxygen concentration in the rich atmosphere changes depending on the degree of NOx catalyst poisoning, and reflects the degree of sulfur poisoning. Therefore, the oxygen concentration is estimated based on the fuel supply amount. The amount of sulfur deposited Oxygen concentration above By correcting based on the above, the influence on the estimation accuracy due to the variation in the sulfur content of the fuel is corrected. As a result, it is possible to estimate the amount of sulfur component adhering by the variation in the sulfur content of the fuel, using the usual method of enriching the air-fuel ratio without measuring and inputting the sulfur content of the fuel one by one. It becomes possible to deal with the problem of reduced accuracy reasonably and efficiently, and consequently, it is possible to execute the sulfur release process satisfactorily without excess or deficiency.
[0178]
Moreover, according to the present invention, the downstream of the catalyst is utilized by utilizing the enrichment for restoring the storage capacity. Oxygen concentration Therefore, the number of times of enriching the air-fuel ratio does not increase, and the lean operation is obstructed and the fuel consumption performance is prevented from being impaired.
[0179]
On the other hand, according to the present invention, the enrichment for correcting the estimated adhesion amount is performed separately from the enrichment for restoring the storage capacity. Therefore, it is appropriate for correcting the estimated sulfur adhesion amount. Oxygen concentration A detection value is obtained, and the accuracy of the correction is improved. As a result, it is possible to secure a good sulfur release process without excess or deficiency. Moreover, in that case, since the enrichment needs to be performed with a much lower frequency than the enrichment for restoring the storage capacity, the number of times of enriching the air-fuel ratio does not increase and the lean operation is unavoidable. It is suppressed that the fuel efficiency is impaired due to the inhibition.
[0180]
In addition, according to the present invention, as an example of the timing for performing the enrichment for the estimated adhesion amount correction, the estimated sulfur adhesion amount once reduced to zero after the sulfur release process is performed is increased to the predetermined amount again. Therefore, the enrichment is performed by selecting the time when sulfur poisoning has progressed to a certain level, and as described above, it is appropriate for the correction of the estimated sulfur adhesion amount. Oxygen concentration A detection value is obtained, and the accuracy of the correction is improved. Further, since the estimated value of the estimated sulfur adhesion amount is zero, the determination accuracy of whether or not the adhesion amount has increased by a predetermined amount or more is increased. In addition, since the enrichment is performed at the above timing when the sulfur release process is performed for the first time after refueling, the influence of the variation in the sulfur content of the fuel that changes due to refueling is corrected early.
[0181]
Furthermore, according to the present invention, the enrichment conditions such as the enrichment degree and enrichment time are determined based on the estimated sulfur adhesion amount. Therefore, from the viewpoint of NOx release, the stored NOx is rich without excess or deficiency. The release treatment is performed properly at a chemical degree or for a time that is not excessive or insufficient. As a result, problems such as untreated occluded NOx remaining in the catalyst or conversely being excessively enriched to reduce fuel consumption performance are suppressed. On the other hand, from the viewpoint of correcting the estimated sulfur adhesion amount, all of the stored NOx is reduced and decomposed into oxygen, and a sufficient amount of oxygen is generated. Oxygen concentration A detection value is obtained, and the accuracy of the correction is improved.
[0182]
The present invention can be preferably applied to an engine having a NOx storage reduction catalyst in general.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of an engine control system according to an embodiment of the present invention.
FIG. 2 is an air-fuel ratio map of the engine.
FIG. 3 is a time chart showing an example of NOx purification control of the engine.
FIG. 4 is a flowchart of the same.
FIG. 5 is a time chart showing an example of sulfur poisoning elimination control of the engine.
FIG. 6 is a main flowchart of the same.
FIG. 7 is a flowchart showing a subroutine of the flowchart.
FIG. 8 is also a sub-flowchart.
FIG. 9 is a characteristic diagram used in the control.
FIG. 10 is also a characteristic diagram.
FIG. 11 is also a characteristic diagram.
FIG. 12 is also a characteristic diagram.
FIG. 13 is a principle diagram of the control.
FIG. 14 is also a principle view.
FIG. 15 is also a characteristic diagram.
FIG. 16 is also a sub-flowchart.
FIG. 17 is also a characteristic diagram.
FIG. 18 is also a characteristic diagram.
FIG. 19 is also a characteristic diagram.
FIG. 20 is also a sub-flowchart.
FIG. 21 is also a sub-flowchart.
FIG. 22 is the same principle diagram.
FIG. 23 is also a characteristic diagram.
FIG. 24 is also a characteristic diagram.
FIG. 25 is also a characteristic diagram.
FIG. 26 is a time chart showing an example of estimated adhesion amount correction control of the engine.
FIG. 27 is also a flowchart.
FIG. 28 is a characteristic diagram used in the control.
FIG. 29 is also a characteristic diagram.
FIG. 30 is a principle diagram of the present invention.
FIG. 31 is the same principle diagram.
FIG. 32 is the same principle diagram.
[Explanation of symbols]
1 engine
4 Combustion chamber
5 Spark plug
6 Injector
16 Three-way catalyst
17 NOx catalyst (NOx storage reduction catalyst)
20 Control unit (estimating means, releasing means, enriching means, correcting means)
28 Second air-fuel ratio sensor (detection means)

Claims (5)

A NOx occlusion reduction type catalyst that occludes NOx components in exhaust gas in an excess oxygen atmosphere and reduces and releases NOx components occluded due to a decrease in oxygen concentration is provided in the exhaust passage, and sulfur components adhering to the catalyst are reduced. An estimation means for estimating the amount based on a value related to the fuel supply amount, and a release for releasing the sulfur component from the catalyst when the sulfur component adhesion amount estimated by the estimation means exceeds a predetermined adhesion amount An exhaust purification device for an engine comprising: a detecting means for detecting an oxygen concentration downstream of the catalyst; and a sulfur component estimation enriching means for forcibly enriching an air-fuel ratio of exhaust gas discharged from the combustion chamber If, sulfur air-fuel ratio of the exhaust gas by the sulfur component estimation enrichment means is estimated by the estimating means based on the oxygen concentration detected by said detecting means when it is rich Exhaust purification apparatus for an engine and having a correction means for correcting the minute adhesion amount. There is provided NOx component releasing means for enriching the air-fuel ratio of the exhaust gas in order to release the NOx component from the catalyst periodically or when the amount of NOx component stored in the catalyst exceeds a predetermined storage amount. 2. The engine exhaust gas purification apparatus according to claim 1, wherein the sulfur component estimation enrichment means is the NOx component release means. There is provided NOx component releasing means for enriching the air-fuel ratio of the exhaust gas in order to release NOx component from the catalyst periodically or when the amount of NOx component stored in the catalyst exceeds a predetermined storage amount. 2. The engine exhaust gas purification apparatus according to claim 1, wherein the sulfur component estimation enrichment means enriches the air-fuel ratio of the exhaust gas at a frequency less than the enrichment frequency performed by the NOx component release means. . The sulfur component estimation enrichment means is at least when the amount of adhesion of the sulfur component estimated by the estimation means increases by a predetermined amount or more after the release means first releases the sulfur component after refueling. 4. The exhaust emission control device for an engine according to claim 3, wherein the air-fuel ratio of the exhaust gas is enriched. The sulfur component estimation enrichment means determines at least one of the degree of enrichment or the enrichment time based on the amount of adhesion of the sulfur component estimated by the estimation means. The engine exhaust gas purification apparatus according to any one of the above.

Images