JP2932280B2 - Vehicle failure diagnosis device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a failure diagnosis apparatus for a vehicle in which a failure cause, that is, a failure site can be known based on a failure symptom of the vehicle.

(Prior Art) Modern vehicles tend to be more and more complex in structure. For this reason, when the vehicle is malfunctioning, that is, when a trouble symptom occurs in the vehicle, even a considerably experienced mechanic tends to have difficulty in identifying a part that causes the trouble symptom. . This imposes a heavy burden on quick repairs.

For this reason, recently, as disclosed in Japanese Patent Application Laid-Open No. 62-6856, an apparatus called an expert system has been proposed which can identify a part causing a failure symptom from a failure symptom. . This is because the causal relationship between the failure symptom and the component failure that causes the failure is stored as a failure tree, and the failure symptom is sequentially input to finally find the cause of the failure, that is, the component that has failed, and find this. Is displayed. By storing the probability of the causal relationship between the failure symptom and the component failure that causes the failure symptom, the failed component can be more accurately found.

By the way, even if the cause of a failure according to the failure information is inferred by this kind of device, the user, that is, the person who performs the maintenance, may not be able to fully understand the inference result.

In many cases, a number of suspected failure components having the same certainty may be inferred.

One major reason for the above causes is that
The input information, that is, the failure information itself is not appropriate, and in extreme cases, the failure symptom may be wrong. However, the user does not readily realize that the input information is inappropriate, which is a major problem in effectively using this type of system.

On the other hand, as shown in FIG. 13, a first storage means storing a causal relationship between a failure symptom and a cause of failure as first failure information, and a causal relationship between a part and a failure symptom caused by the failure of the part. A second storage unit that stores the relationship as second failure information; a third failure unit that stores a causal relationship between a part and a failure symptom that does not occur when the part is normal as third failure information; Input means for inputting a failure symptom, based on the failure symptom input by the input means,
First inference means for retrieving the first failure information and inferring a failure cause; and a failure symptom input by the input means for a component corresponding to the failure cause inferred by the first inference means. A second inference means for searching the second failure information to infer a component corresponding to a failure cause, and a third inference means for searching the third failure information to infer a component that does not cause a failure symptom; And a notifying means for notifying the inference result inferred by each inference means.

By adopting such a configuration, if the failure symptom is inappropriate, the component suspected of having failed is simultaneously notified that it is non-failure, that is, normal. When such a contradiction occurs, the user can easily determine that the failure symptom inputted by the user is inappropriate.

Further, in order to further improve the user, as described in the above-mentioned publication, it is also possible to store the certainty factor (probability) of the causal relationship together in each storage means, and to notify this certainty factor. Good.

(Problems to be Solved by the Invention) However, in spite of re-input of the failure symptom, there may be a case where the parts which have been informed that they have failed again are simultaneously informed that they are normal. The cause may be that the input failure symptom is still inappropriate, or that there is a problem with the failure tree, etc. In the case where the parts reported to have failed are inspected in order from the one with the highest degree of certainty, there are the following problems.

In other words, if two parts that are reported to be faulty have the same certainty, and one of them is also reported to be normal, the other part that is reported only to be failed is replaced by the other part. It is preferred that inspection be given priority. However, since the two parts have the same certainty, when the parts are simply inspected starting from the part having the high degree of certainty, the one part that has been reported as normal may be preferentially inspected. Efficiency may be reduced. this is,
It is considered that the cause is that the priority of the parts to be inspected is not correctly represented by the confidence.

Accordingly, an object of the present invention is to provide a method of accurately representing the priority of parts to be inspected with certainty even when there is an error such as an inappropriate input of a failure symptom, thereby preventing deterioration of work efficiency. Is to do.

(Means for Solving the Problem, Function) In order to achieve the above object, according to the present invention, a first storage means for storing a causal relationship between a failure symptom and a failure cause as first failure information; A second storage means for storing as a second failure information a causal relationship between the component and a failure symptom caused by the failure of the component; and a third causal relationship between the component and a symptom occurring when the component is normal as the third failure information. Based on the stored third failure means, an input means for inputting a failure symptom of the vehicle, and a failure symptom input by the input means,
First inference means for retrieving the first failure information and inferring a failure cause; and a failure symptom input by the input means for a component corresponding to the failure cause inferred by the first inference means. Second inference means for retrieving the second failure information and inferring a component corresponding to the cause of failure together with the probability thereof, and third inference means for retrieving the third failure information and inferring the component together with a probability of normality When the same part is inferred by the second inference means and the third inference means, it is determined whether or not the input information by the input means is appropriate, and the inadequacy is notified, and the probability of inference by each inference means is determined. A probability correction unit that corrects the probability of the larger value in accordance with the probability of the smaller value, and deletes the inference result of the inference unit that infers the smaller value of the probability. (2) a part corresponding to the cause of failure based on the inference result inferred by the third inference means and the correction result corrected by the probability correction means, and a finally determined probability and a part in a normal state; With such a configuration, a part suspected of failure and a part considered to be normal are inferred together, and it is inferred that the part is faulty. If there is a part that is inferred to be normal, the probability correction means notifies the invincibility of the input information and corrects the probability of the larger value in accordance with the probability of the smaller value. And
The inference result with the smaller probability is deleted, and the reduced-corrected probability is reported by the reporting means as the probability of the part.

(Example) Hereinafter, an example of the present invention will be described with reference to the attached drawings.

FIG. 1 is a block diagram showing an overall system of the present invention, and this system is configured using a computer. In FIG. 1, each part is roughly divided into 5
Are represented by two blocks 1, 2, 3, 4, 5. That is, it comprises a failure symptom input unit 1, a knowledge base 2, an inference unit 3, a probability correction unit 4, and a notification unit 5.

The failure symptom input section is for a user of the present system to input a failure symptom.

The knowledge base 2 includes first to third storage units. The first storage unit stores a component causing a failure symptom as first failure information together with a probability of a causal relationship thereof, and stores the component in the form of a failure tree as described in the above publication. It is a memory of. In addition, the second storage unit stores, as the second failure information, the reliability of the causal relationship between the component and the failure information generated when the component fails. Further, the third storage unit stores, as third failure information, a causal relationship between a component and a symptom that occurs when the component is normal, together with its certainty factor.

The inference unit 3 performs first to third inferences. That is, first, the first storage unit in the storage base is searched to infer the parts which are likely to have failed and the probability thereof.

Second, for the part inferred as described above,
The failure information is searched to infer that this component has failed and its probability according to the failure symptom.

Third, the third failure information is searched to infer that the component is normal and its probability according to the failure symptom.

The results inferred as the second and third are sent to the probability correction unit 4 that mainly corrects the probability. Here, the magnitude of the value of a certain part is compared with the probability of being inferred to be faulty and the probability of being inferred to be normal. Then, the probability with the larger value is reduced and corrected in accordance with the magnitude of the ratio between the probability and the probability with the smaller value, while the component with the smaller probability is deleted from the inference result. As will be described later in detail, when it is necessary to reduce and correct the probability, the probability correction unit also determines that the input failure information is inappropriate.

The results determined by the inference unit 3 and the probability correction unit 4 are reported to a reporting unit 5. This notification can be simply a result of whether or not the input failure symptom is appropriate, or a notification of a component suspected to have failed and a notification of a normal component.

The above-mentioned failure symptom input unit 1 is constituted by, for example, a keyboard. An external storage device, such as a floppy disk or a hard disk, is used for each storage unit in the knowledge base 2 from the viewpoint that a large storage capacity is required. As the inference notifying unit 5, for example, a CRT is used, and in addition to this, a printer can also be used. Of course, the inference unit 3 and the probability correction unit 4 are configured by a CPU.

The storage contents of the first storage unit in the knowledge base 2 are schematically represented in a fault tree format as shown in FIG. This is a causal relationship, for example, in which no cranking occurs as a failure symptom, and the cause becomes more concrete as it goes downward, so that it eventually reaches a part that seems to be defective. And the degree of causality of the tree's hierarchical relationship is indicated by its probability. FIG. 3 schematically shows the structure shown in FIG. Such a fault tree is specifically stored as a number of rules as shown in FIG. That is, X is a failure symptom (possibly a cause at a lower level of the tree), Y is a cause one level below X (a component at the lower level of the tree), and N is a probability (causality). Relationship confidence). As shown in FIG. 5, for example, LOW, MIDDLE, HIGH
Thus, roughly three-stage classification may be used together, or only this rough classification may be used.

FIG. 10 is a flowchart showing a procedure for searching for information stored in the first storage unit, that is, failure information, and inferring a failure cause (part) corresponding to the failure information together with its probability. Hereinafter, FIG. 10 will be described. First, in P (step) 1, a leading rule (a rule in which A in FIG. 3 is the highest order) is fetched. Then
In P2, look at X in the rule and search for any other rules with this X. Then, among the rules having the X, the rule having the highest certainty factor N is selected.

After P2, in P3, it is determined whether or not Y in the selected rule matches the failure symptom. If they match, the current relationship between X, Y, and N is temporarily held at P4. Then, after Y is set to X in P5 (Y is updated as X of the rule), it is determined whether there is a rule caused by X in P6. By the judgment of this P6
If it is NO, it will be completed as it is. If the determination in P6 is YES, the processes in and after P2 are repeated.

If the determination in P3 is NO, in P7, a rule with a lower confidence level N is selected, and then in P8, it is determined whether a similar rule still exists. Then, if there is another rule, the process returns to P3, and if not, at P9, a lower level (the rule at the bottom of the fault tree) is temporarily stored. It is to be noted that the arrival at P9 is a time when the site that eventually causes the failure symptom cannot reach a sufficiently lower level of the failure tree.

The second storage unit in the knowledge base 2 is stored in the form of an inference tree as shown in FIG. Starting from a component (a battery is shown in FIG. 6), a failure symptom that occurs when this component fails along with its confidence is stored in the second storage unit. "Value. That is, as the inference tree goes from the upper level to the lower level, the fault symptom or the cause of the fault is shown in more detail.

The third storage unit in the knowledge base 2 is in the form of an inference tree as shown in FIG. The third storage unit stores, as a tip, a failure symptom that does not occur when the part is normal, and in a reverse view, a symptom when the part is normal. Also in this third storage unit,
The probability is given, but to distinguish it from the case of Fig. 6,
As the probability, a value of “−” is used. Each such inference tree is specifically stored as a rule, similarly to the case of a fault tree.

FIG. 7 shows a schematic diagram of the sixth inference tree, and FIG. 9 shows a schematic diagram of the inference tree of FIG.

Now, while using the inference tree shown in FIGS. 6 and 8,
In particular, the point of determining whether input information is appropriate will be described in detail.

First, a failure symptom is input in S1 of FIG. 11, and then, in S2, the inference tree of FIG. 6 and FIG. 8 is used for a part which seems to have failed as described in FIG. That inference is made. FIG. 12 shows the details of S2. That is, at T1, a component suspected to have failed is read out using the inference tree in FIG. Then, by the processing of T2 and T3, the parts corresponding to the failure symptoms and their probabilities (+ values) are inferred using the inference tree shown in FIG. Parts considered to be normal and their probabilities (values of-) are inferred.

In S3 of FIG. 11, each component inferred in S2, that is, the probability of the component group is output, and in S4, the output component group has a probability of “+”;
Those having a probability of “−” are sorted (rearranging in order of the probability).

In S5, it is determined whether or not there is a component having a certain level of certainty (probability) of both “+” and “−”. If the determination in S5 is YES, in S6, two types of rules that give the certainty of "+" and "-" are extracted. Thereafter, it is determined whether or not there is a relative relationship between the check items for these two rules. Specifically, "Do you turn on the headlamps?"
The input of the failure symptom that answered the question format can be regarded as having a relative relationship. If the determination in S7 is YES, it is determined that there is inconsistency in the input information, and it is determined in S8 whether the input information has been re-input once. If it has not been re-entered, the user is asked whether or not to re-enter in S9, and if NO in S10, it is determined that re-entry has not been rejected, and then in S11, detailed inspection procedures are displayed. I do. Then, based on the result of the inspection performed according to the displayed inspection procedure, new input information (failure symptom) is input in S12.

When YES is determined in S8, or YE is determined in S10.
In the case of S, the process proceeds to S13. In this S13,
The ratio of the certainty factor of “+” and “−” is obtained, and a coefficient for resetting the certainty factor having a larger value is obtained. As a specific example, the positive confidence that the fuel injection is not airtight is 0.8, and the negative confidence that the fuel injection is normal is
If it is 0.6, the ratio of the certainty factors is set to 0.8: 0.6 = 100: 75. Further, the coefficient for resetting the certainty is set inconsistently as follows according to the ratio of the certainty.

Ratio coefficient 100: 0 to 100: 30 0.9 100: 31 to 100: 70 0.7 100: 71 to 100: 95 0.5 100: 96 to 100: 100 − Therefore, in this case, 0,5 is adopted as the coefficient. . In S14, a process of obtaining a new certainty factor by multiplying the certainty factor having the larger value by the coefficient obtained in S13 is performed. Specifically, in the case of the above example, since 0.8 * 0.5 = 0.4, 0.4 is the final certainty that the fuel injection is poorly airtight. The part with the lower certainty factor is finally deleted from the inference result in S14. Therefore, S17 described later
In, fuel injection is normal,
There is no notification that the certainty factor is 0.6.

When the process of S14 is completed or when the determination of S7 is NO, the process proceeds to S15. In this S15, for all the parts having both the “+” and “−” values,
It is determined whether or not the above-described processing has been performed. If the determination is NO, the process returns to S5. If the determination is YES, in S16, all the component groups have a probability of “+” and a value of “−”. Are sorted (probability sorting), and in S17, the estimation result is displayed.

As is clear from the above description, according to the present embodiment,
For example, as in the specific example described above, when the positive certainty that the fuel injection is poorly airtight is 0.8 and the negative certainty that the fuel injection is normal is 0.6, the positive certainty is finally displayed as 0.4. Will be done. Therefore, the priority order of the parts to be inspected is correctly represented by the confidence, and the work efficiency of the parts inspection is improved.

Further, in the above embodiment, with respect to parts having positive and negative certainty factors, only the part with the highest certainty factor is finally displayed. This is to prevent the user from feeling uncomfortable by simultaneously performing the notification that the component is normal and the notification that the component has failed.

(Effect) As is clear from the above description, the present invention can notify the user of an inappropriate failure symptom when the failure symptom is inappropriate. In this case, it is possible to correctly recognize the priority of components to be inspected by appropriately correcting the probability of components that are suspected to be faulty, and to report the suspected components and normal components. Can be minimized,
The work efficiency of parts inspection can be further improved.

[Brief description of the drawings]

FIG. 1 is an overall system diagram showing an embodiment of the present invention. FIG. 2 is a diagram showing an example of a fault tree. FIG. 3 is a diagram schematically showing FIG. 2; FIG. 4 is a diagram showing an example in which the contents of FIGS. 2 and 3 are stored. FIG. 5 is a diagram showing a setting example of a certainty level and a certainty factor. FIG. 6 is a diagram showing an example of an inference tree about a part and a failure symptom that occurs when the part fails. FIG. 7 is a diagram schematically showing FIG. 6; FIG. 8 is a diagram showing an example of an inference tree for a part and a failure symptom that does not occur when the part is normal. FIG. 9 is a diagram schematically showing FIG. 10 to 12 are flowcharts showing control examples using the present invention. FIG. 13 is a block diagram showing an example of a failure diagnosis device.

────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-62-6858 (JP, A) JP-A-63-79031 (JP, A) JP-A-62-6855 (JP, A) (58) Field (Int.Cl. ⁶ , DB name) G01M 17/00 B60S 5/00 G06F 9/44

Claims (1)

(57) [Claims] A first storage means for storing a causal relationship between a failure symptom and a cause of failure as first failure information, and a causal relationship between a component and a failure symptom caused by a failure of the component as second failure information. Second storage means for storing; a third storage means for storing a causal relationship between a part and a symptom occurring when the part is normal as third failure information; an input means for inputting a failure symptom of the vehicle; First inference means for retrieving the first failure information based on the failure symptom input by the means and inferring a failure cause; and inputting a part corresponding to the failure cause inferred by the first inference means. A second inference means for retrieving the second failure information based on the failure symptom inputted by the means and inferring a component corresponding to the failure cause together with the probability thereof; Third inference means for inferring the part together with the probability that the same part is inferred by the second inference means and the third inference means. , Of the probabilities inferred by the respective inference means, the probability of the larger value is reduced and corrected according to the probability of the smaller value, and the inference result of the inference means inferring the smaller value probability And a component corresponding to the cause of failure based on the inference result inferred by the second inference unit and the third inference unit and the correction result corrected by the probability inference unit, and finally determining the component. And a notifying means for notifying the determined probability and the parts in a normal state.