JP7733294B2 - Water quality simulation method, water quality simulation device, and method for determining the number of carriers for treating pollutants by a fluidized carrier method - Google Patents

Description

The present invention relates to a water quality simulation method, a water quality simulation device, and a method for determining the number of carriers required to treat pollutants using a flowing carrier method.

Conventionally, a biological treatment method using an aerobic fluidized bed for water treatment is known, which can selectively and efficiently remove thiacyanate ions (see, for example, Patent Document 1).

In addition, a water quality simulation method is known in which pollutants (COD) in ammonia water are fractionated into phenol, thiocyanate, thiosulfate, and persistent COD, and the decomposition rate of each component is indirectly measured by an oxygen consumption experiment (see, for example, Patent Document 2 ).

Japanese Patent Application Laid-Open No. 2016-112556 Patent No. 5435658

For example, coke oven wastewater (ammonia water) undergoes physicochemical treatment, followed by the activated sludge process, a biological water treatment method, to treat COD (Chemical Oxygen Demand) components that meet wastewater standards before being discharged. Inadequate treatment of the COD component, thiocyanide (SCN), in the ammonia water, or the production of nitrite from ammonia can result in an increase in COD in the treated water. For this reason, stable and efficient ammonia water treatment is required in production activities.

However, the technology described in Patent Document 1 biologically treats the water to be treated using actual equipment that uses an aerobic fluidized bed, and does not anticipate simulating water quality.

Furthermore, the technology described in Patent Document 2 does not take into account the interspecies competition that occurs between the microbial groups treating each pollutant, and the impact that this interspecies competition has on post-treatment water quality. For this reason, it is difficult to accurately simulate post-treatment water quality using the method described in Patent Document 2.

The present invention aims to accurately simulate post-treatment water quality by taking into account the effects of interspecies competition among multiple microbial groups that decompose each of the multiple different pollutants contained in the water to be treated, and to accurately determine the number of carriers required to treat pollutants using the fluidized carrier method.

The gist of this disclosure is as follows:

(1) A water quality simulation method executed by a processor in a treatment process for biologically treating water to be treated, comprising:
A water quality simulation method that simulates the water quality after treatment of the treated water by multiple microbial groups based on a model formula having a term that represents interspecies competition between multiple different microbial groups that decompose each of multiple different pollutants contained in the treated water, and a term that represents that the higher the density of any one of the multiple microbial groups, the more the other microbial groups decrease .

(2) The water quality simulation method described in (1) above, wherein the model formula further includes, for each of the plurality of microbial groups, a term representing self-proliferation, in which the microbial groups grow in response to the concentration of the pollutants to be decomposed , and a term representing extinction of the microbial groups due to overcrowding, and a term representing self-decomposition, in which the higher the density of the microbial groups, the more the microbial groups decrease .

(3) The water quality simulation method according to (1) or (2), wherein the model formula further includes a term indicating that , for each of the plurality of pollutants, the greater the concentration of the pollutant to be decomposed, the more the growth of the microbial population is suppressed due to toxicity.

(4) A water quality simulation method according to any one of (1) to (3) above, which calculates the decomposition rate of each of the plurality of pollutants using the model formula, and fits parameters representing the model formula based on the actual measured values of the decomposition rates of the plurality of pollutants and the calculation results of the decomposition rates using the model formula.

(5) A water quality simulation method according to any one of (1) to (3) above, which calculates the concentration of each of the plurality of pollutants in the treated water using the model formula, and fits parameters representing the model formula based on the actual measured values of the concentrations of each of the plurality of pollutants in the treated water when they reach a steady state and the calculation results of the concentrations of each of the pollutants using the model formula.

(6) A water quality simulation method described in any one of (1) to (3) above, which calculates the concentration of each of the multiple pollutants in the treated water using the model formula, and fits parameters representing the model formula based on the actual measured values of the concentrations of each of the multiple pollutants in the treated water at each measurement point and the calculation results of the concentrations of each of the pollutants using the model formula corresponding to each measurement point.

(7) A water quality simulation method as described in (6) above, which predicts the concentration of each of the plurality of pollutants in the treated water at a later arbitrary point in time from the value at an arbitrary point in time.

(8) A water quality simulation method according to any one of (4) to (7) above, in which the water quality of the treated water after treatment is simulated based on a plurality of model formulas obtained by the fitting, the model formulas corresponding to parameters for which the value of the evaluation function used in the fitting is equal to or less than a predetermined value.

(9) A water quality simulation method according to any one of (1) to (8) above, characterized in that the treatment process is a fluidized carrier method.

(10) A water quality simulation method according to any one of (1) to (9) above, wherein the pollutants include at least one of thiocyanate, phenol, and thiosulfate.

(11) A water quality simulation device for a treatment process that biologically treats water to be treated, the water quality simulation device comprising a processor,
The processor:
a parameter fitting unit that fits parameters of the model formula based on actual measurement values of decomposition rates of a plurality of pollutants obtained by introducing a plurality of microbial groups that decompose each of the plurality of different pollutants into the water to be treated, and calculated values of decomposition rates of a plurality of the pollutants calculated from a model formula having a term that represents interspecific competition between the plurality of microbial groups in the water to be treated, the term representing that the higher the density of any one of the plurality of microbial groups, the greater the decrease in other microbial groups;
a water quality simulation unit that simulates the water quality of the treated water after treatment using the model formula to which the parameters have been fitted;
A water quality simulation device comprising:

(12) A method for determining the number of carriers required to treat pollutants using the flowing carrier method, using the water quality simulation method described in (9) above.

The present invention has the advantage of enabling accurate simulation of post-treatment water quality by taking into account the effects of interspecies competition between multiple microbial groups that decompose each of the multiple different pollutants contained in the water to be treated, and accurately determining the number of carriers required to treat pollutants using the flowing carrier method.

FIG. 1 is a schematic diagram showing a model according to the present embodiment. FIG. 1 is a schematic diagram showing a model according to the present embodiment. 3 is a flowchart showing a processing procedure of a wastewater quality simulation method according to the present embodiment. 1 is a block diagram showing an example of the configuration of a wastewater quality simulation device according to an embodiment of the present invention; FIG. 1 is a schematic diagram showing a biological treatment device. FIG. 1 is a characteristic diagram showing the removal rate (decomposition rate) of thiocyanate per day of operation. FIG. 1 is a characteristic diagram showing the removal rate (decomposition rate) of phenol per day of operation. FIG. 1 is a characteristic diagram showing the removal rate (decomposition rate) of thiosulfate per day of operation. FIG. 1 is a characteristic diagram showing the thiocyanate removal rate per day of operation. FIG. 1 is a characteristic diagram showing the phenol removal rate per day of operation. FIG. 1 is a characteristic diagram showing the thiosulfate removal rate per day of operation. FIG. 1 is a schematic diagram showing a method for collecting a sponge carrier from a water treatment reactor, placing it in a beaker, and conducting a culture test to identify the autolysis rate (d _i ). FIG. 13 is a characteristic diagram showing the difference in thiocyanate concentration in the culture test shown in FIG. 12 carried out on each collection day. FIG. 10 is a characteristic diagram showing the results of comparing calculated values (Calculation) with measured values (Experiment). FIG. 10 is a characteristic diagram showing the results of comparing calculated values (Calculation) with measured values (Experiment). FIG. 10 is a characteristic diagram showing the results of comparing calculated values (Calculation) with measured values (Experiment). FIG. 10 is a characteristic diagram showing the results of comparing a toxic model with a non-toxic model. FIG. 1 is a characteristic diagram showing the thiocyanate concentration obtained from experimental results in operation of a water treatment reactor using thiocyanate as a pollutant. FIG. 10 is a characteristic diagram showing the actual measured value of thiocyanate concentration, the simulation results of thiocyanate concentration obtained by a non-steady state model taking into account time changes, and the simulation results of microorganism concentration. FIG. 10 is a characteristic diagram showing the concentration of thiocyanate ions in treated water obtained from the results of monitoring thiocyanate ions. FIG. 10 is a characteristic diagram showing the concentration of thiocyanate ions in treated water obtained from the results of monitoring thiocyanate ions. FIG. 10 is a characteristic diagram showing both simulation results and actual measurement values when the number of sponge carriers 21 is changed. 19 is a characteristic diagram showing the simulation results based on the experimental results shown in FIG. 18, and FIG. 19 is a characteristic diagram showing the simulation results when the parameters are changed. FIG. 20 is a characteristic diagram showing an example of a simulation of the results of another water treatment reactor using the same parameters as in FIG. 19 based on the experimental results shown in FIG. 18 . FIG. 24 is a characteristic diagram showing an example of a simulation of the results of another water treatment reactor using the same parameters as in FIG. 23 based on the experimental results shown in FIG. 18 . FIG. 19 is a characteristic diagram showing the results of shading the range between the maximum and minimum values in each simulation result calculated using a set of 81 parameters obtained from the measurement data of FIG. 18. FIG. 26 is a characteristic diagram showing the results of simulating the results of another water treatment reactor using all of the 81 parameter sets obtained from the measurement data of FIGS. 24 and 25.

Several embodiments of the present invention will be described below with reference to the drawings. However, these descriptions are intended to merely exemplify preferred embodiments of the present invention and are not intended to limit the present invention to such specific embodiments.

In this embodiment, a model is constructed to simulate the decomposition of multiple pollutants contained in the water to be treated when the water is biologically treated using an aerobic fluidized bed. Because the multiple pollutants contained in the water to be treated are decomposed using an aerobic fluidized bed, multiple microbial groups that decompose these multiple pollutants are attached to the fluidized carriers.

These multiple microbial groups multiply by themselves while breaking down pollutants. Furthermore, these multiple microbial groups die due to self-decomposition. Therefore, when building a model, the self-proliferation and self-decomposition of the microbial groups are taken into account.

Furthermore, the inventors focused on the interspecific competition between these multiple microbial groups and discovered that the accuracy of the model could be improved by building it while taking into account the impact of interspecific competition on the decomposition of each pollutant.

Furthermore, when the concentration of pollutants is high, the pollutants may act as toxic components and inhibit the growth of microbial communities or even kill them. The inventors have found that the accuracy of the model can be further improved by building it after taking into account the impact of these toxic components on the decomposition of each pollutant.

FIG. 1 is a schematic diagram showing a model for simulating the decomposition of multiple pollutants according to this embodiment. As shown in FIG. 1 , this embodiment uses coke wastewater as an example of the water to be treated. Thiocyanate contained in the coke wastewater is decomposed by a thiocyanate-decomposing microbial group S, phenol is decomposed by a phenol-decomposing microbial group P, and thiosulfate is decomposed by a thiosulfate-decomposing microbial group _S2 . The wastewater from which the pollutants have been decomposed is discharged as treated water. Note that, since there are multiple types of microorganisms that decompose thiocyanate, these microorganisms will be collectively referred to as a microbial group. Similarly, since there are multiple types of microorganisms that decompose phenol, these microorganisms will be collectively referred to as a microbial group, and since there are multiple types of microorganisms that decompose thiosulfate, these microorganisms will be collectively referred to as a microbial group.

When treating water biologically using an aerobic fluidized bed, each microbial group attaches to the fluidized carrier. The inventors discovered that as each microbial group grows while decomposing their respective pollutants, they compete with each other on the fluidized carrier, and this competition results in a decrease in the microbial groups. When the amount of each microbial group attached to the fluidized carrier is relatively small, the deaths due to competition among the microbial groups are relatively low. However, when the amount of each microbial group attached to the fluidized carrier becomes saturated due to self-renewal and they can no longer attach, interspecies competition intensifies, and the microbial groups decrease.

For this reason, the inventors constructed a model shown in Figure 1 in which microbial groups were separated into groups to treat each of the three pollutant components (phenol, thiocyanate, and thiosulfate) in coke oven wastewater, and also constructed the competitive system model shown in Figure 2, taking into account the competitive relationships between these microbial groups.

The constructed model includes multiple parameters that determine the characteristics of the model. The inventors performed parameter fitting so that the decomposition rates of each pollutant obtained from the model matched the decomposition rates of each pollutant obtained from experiments, and completed the model by applying the parameters obtained through parameter fitting to the model.

By using the model completed in this way, the post-treatment water quality of the coke oven wastewater, which is the water to be treated, can be calculated. Therefore, the concentration of each pollutant in the treated water can be predicted. Furthermore, because the model parameters are related to water quality, it is possible to control the post-treatment water quality by controlling the various physical quantities related to the parameters.

Furthermore, for example, there are no clear regulations regarding the amount of fluid carrier required for processing; the standard value is 33% (v/v), and if processing capacity is insufficient, the amount is increased to 50%, 66%, etc. However, excessive increases in volume can result in unnecessary increases in costs. In light of this situation, a model that can accurately predict processing performance is needed, but such a model did not exist.

The model of this embodiment makes it possible to predict changes in water quality relative to the total surface area of the carrier. Therefore, it is also possible to calculate the exact amount of carrier to be added to maintain post-treatment water quality at a specified standard.

Figure 3 is a flowchart showing the processing steps of the wastewater quality simulation method according to this embodiment. First, in step S10, a model formula is constructed to calculate the decomposition rates of multiple pollutants in the water to be treated. Next, in step S12, actual measured values of the decomposition rates of the multiple pollutants are obtained. Next, in step S14, parameters of the model formula are fitted based on the calculated values of the decomposition rates using the model formula and the actually measured decomposition rates. Next, in step S16, the model formula with the fitted parameters is used to calculate the wastewater quality. Processing ends after step S16.

Figure 4 is a block diagram showing an example configuration of a wastewater quality simulation device according to this embodiment. More specifically, Figure 4 is a schematic diagram showing the functional blocks of a processor 100 included in the water quality simulation device. The processor 100 of the water quality simulation device includes a parameter fitting unit 110 that fits model equation parameters based on actual measurements of the decomposition rates of multiple pollutants in the wastewater and calculated decomposition rates using a model equation, and a water quality simulation unit 120 that simulates post-treatment water quality using the model equation to which the parameters have been fitted. Each of these units included in the processor 100 is a functional module implemented, for example, by a computer program running on the processor 100. In other words, the functional blocks of the processor 100 are composed of the processor 100 and a program (software) for operating the processor 100. The program may be stored in memory included in the water quality simulation device or on an externally connected recording medium. Alternatively, each of these units included in the processor 100 may be a dedicated arithmetic circuit provided in the processor 100.

The present embodiment will be described in detail below based on examples in the order of [1] to [13].
[1] Operation of a water treatment reactor using thiocyanate as a pollutant. [2] Operation of a water treatment reactor using phenol as a pollutant. [3] Operation of a water treatment reactor using thiosulfate as a pollutant. [4] Operation of a water treatment reactor using thiocyanate, phenol, and thiosulfate as pollutants. [5] Construction of a competitive system model (non-toxic).
[7] Identification of autolysis rate (d _i ) [8] Simplification by partial non-dimensionalization of the one-type model (with toxicity)
[9] Simplification of the three-type model by partial non-dimensionalization (with toxicity)
[10] Parameter fitting of the model formula [11] Simulation results of three types of models [12] Comparison of the toxic model and the non-toxic model (type 1)
[13] Use of model formulas [14] Parameter fitting considering unsteady states [15] Method for determining the number of carriers using the flow carrier method [16] Simulation method using multiple parameter sets

First, in [1] to [4], we will explain the extraction of water treatment data for model analysis using treated water simulating coke oven wastewater (ammonia water). In [1] to [3], a water treatment reactor is operated and actual measured values of water treatment data are extracted for each of the cases where thiocyanate is the pollutant, where phenol is the pollutant, and where thiosulfate is the pollutant. In [4], a water treatment reactor is operated and actual measured values of water treatment data are extracted for the cases where thiocyanate, phenol, and thiosulfate are the pollutants. Note that the actual measurement of water treatment data can be performed in a manner similar to that described in Patent Document 1 mentioned above.

[1] Operation of a water treatment reactor using thiocyanate as a pollutant Industrial water and natural seawater were mixed in a volume ratio of 2:3 to obtain a solvent, and the solutes shown in Table 1 were dissolved in the concentrations shown in Table 1 to prepare artificial wastewater (water to be treated).

As shown in Figure 5, an integrated biological treatment device 20 was prepared, in which a biological treatment zone 20a and a settling zone 20b are separated from each other by a partition 23 within a single tank, and the zones are interconnected below the partition 23. A 10mm x 10mm x 10mm sponge carrier 21 (fluid carrier (AQ-1 manufactured by Kanto Inoac)) and high-concentration activated sludge as a microbial inoculum were placed in a plastic bottle, kneaded thoroughly by hand, and left to soak overnight with a lid on, allowing the microorganisms to adhere to the sponge carrier 21. The activated sludge added here is preferably an inoculum rich in thiocyanate-decomposing microorganisms. However, since the activated sludge used to treat coke oven wastewater containing all of thiocyanate, thiosulfate, and phenol is used as the microbial inoculum, the activated sludge may be the same as the activated sludge added in steps [2] to [4] described below.

The sponge carriers 21 (500 pieces) and activated sludge prepared in this way were placed into the biological treatment area 20a of the biological treatment device 20, and the biological treatment device 20 was prepared. The amount placed was determined to be approximately 20% of the apparent volume. Since the volume of the biological treatment area 20a was 3.4 L, the carrier placement rate was 14.7% (v/v), but the apparent volume was approximately 20%.

The water to be treated 24 was then introduced into the biological treatment device 20 prepared in this manner, and activated sludge was added as a microbial inoculant source. During the first stage of treatment, the water to be treated 24 was introduced so that its hydraulic retention time was 24 hours. The water to be treated 24 in each biological treatment device 20 was aerated 22 to form an aerobic fluidized bed, allowing the microorganisms to acclimate. Treatment was also carried out while adjusting the pH to around 7.5 using a 5 wt% aqueous sodium hydroxide solution. The treated water 26 was then discharged from the biological treatment device 20.

The thiocyanate ion concentration was measured in the treated water in the biological treatment area 20a of each biological treatment device 20 to monitor thiocyanate ions. Monitoring was performed approximately twice a week.

After this first-stage treatment had stabilized, water to be treated 24 was introduced so that the amount of thiocyanate ions introduced per day was as shown in Table 2. Each stage was operated for at least two weeks, and once it was confirmed that the fluctuations in the treated water concentration at each stage had stabilized, the system moved on to the next stage. Monitoring was carried out approximately twice a week.

Figure 6 is a characteristic diagram showing the thiocyanate removal rate (decomposition rate) per day of operation, obtained from the results of thiocyanate ion monitoring. The thiocyanate removal rate is obtained from the amount of thiocyanate ions flowing in per day and the thiocyanate ion concentration of the treated water in the biological treatment area 20a, obtained through monitoring. In other words, the thiocyanate removal rate is obtained from the difference between the amount of thiocyanate ions flowing into the biological treatment area 20a and the amount of thiocyanate ions that were not completely treated in the biological treatment area 20a (residual thiocyanate ions in the treated water).

[2] Operation of water treatment reactor using phenol as a pollutant Industrial water and natural seawater were mixed in a volume ratio of 2:3 to obtain a solvent, and the solutes shown in Table 3 were dissolved in the concentrations shown in Table 3 to prepare artificial wastewater (water to be treated).

Furthermore, a 10mm x 10mm x 10mm sponge carrier 21 [fluid carrier (AQ-1 manufactured by Kanto Inoac)] and high-concentration activated sludge as a microbial inoculum were placed in a plastic bottle, kneaded thoroughly by hand, and left to soak overnight with the lid on, allowing the microorganisms to adhere to the sponge carrier 21. It is preferable that the activated sludge placed here be an inoculum rich in phenol-degrading microorganisms.

The sponge carriers 21 (500 pieces) and activated sludge prepared in this manner were placed into the biological treatment area 20a of the biological treatment device 20.

The water to be treated 24 was introduced into the biological treatment device 20 prepared in this manner, and activated sludge was added as a microbial inoculant source. During the first stage of treatment, the water to be treated 24 was introduced so that the hydraulic retention time of the water to be treated 24 was 24 hours. Furthermore, the water to be treated 24 in each biological treatment device 20 was aerated 22 to form an aerobic fluidized bed, allowing the microorganisms to acclimate. Furthermore, treatment was carried out while adjusting the pH to around 7.5 using a 5 wt% sodium hydroxide aqueous solution. The treated water 26 treated in the biological treatment device 20 was then discharged from the biological treatment device 20.

The phenol concentration was measured in the treated water in the biological treatment area 20a of each biological treatment device 20 to monitor the phenol concentration. Monitoring was performed approximately twice a week.

After this first stage treatment had stabilized, treated water 24 was introduced so that the amount of phenol introduced per day was as shown in Table 4. Each stage was operated for at least two weeks, and once it was confirmed that the fluctuations in the treated water concentration at each stage had stabilized, the system moved on to the next stage. Monitoring was carried out approximately twice a week.

Figure 7 is a characteristic diagram showing the phenol removal rate (decomposition rate) per day of operation obtained from the results of phenol monitoring. The phenol removal rate is obtained from the amount of phenol flowing in per day and the phenol concentration in the treated water in the biological treatment area 20a obtained by monitoring. In other words, the phenol removal rate is obtained from the difference between the amount of phenol flowing into the biological treatment area 20a and the amount of phenol that was not completely treated in the biological treatment area 20a (residual phenol in the treated water).

[3] Operation of a water treatment reactor using thiosulfate as a pollutant Industrial water and natural seawater were mixed in a volume ratio of 2:3 to obtain a solvent, and the solutes shown in Table 5 were dissolved in the concentrations shown in Table 5 to prepare artificial wastewater (water to be treated).

Furthermore, a 10mm x 10mm x 10mm sponge carrier 21 [fluid carrier (AQ-1 manufactured by Kanto Inoac)] and high-concentration activated sludge as a microbial inoculum were placed in a plastic bottle, kneaded thoroughly by hand, and left to soak overnight with the lid on, allowing the microorganisms to adhere to the sponge carrier 21. It is preferable that the activated sludge placed here be an inoculum containing a high concentration of thiosulfate-decomposing microorganisms.

The sponge carriers 21 (500 pieces) and activated sludge prepared in this manner were placed into the biological treatment area 20a of the biological treatment device 20.

The water to be treated 24 was introduced into the biological treatment device 20 prepared in this manner, and activated sludge was added as a microbial inoculant source. During the first stage of treatment, the water to be treated 24 was introduced so that the hydraulic retention time of the water to be treated 24 was 24 hours. Furthermore, the water to be treated 24 in each biological treatment device 20 was aerated 22 to form an aerobic fluidized bed, allowing the microorganisms to acclimate. Furthermore, treatment was carried out while adjusting the pH to around 7.5 using a 5 wt% sodium hydroxide aqueous solution. The treated water 26 treated in the biological treatment device 20 was then discharged from the biological treatment device 20.

The thiosulfate ion concentration was measured for the treated water in the biological treatment area 20a of each biological treatment device 20, and the thiosulfate ion concentration was monitored. Monitoring was performed approximately twice a week.

After this first-stage treatment had stabilized, water to be treated 24 was introduced so that the amount of thiosulfate ions introduced per day was as shown in Table 6. Each stage was operated for at least two weeks, and once it was confirmed that the fluctuations in the treated water concentration at each stage had stabilized, the system moved on to the next stage. Monitoring was carried out approximately twice a week.

Figure 8 is a characteristic diagram showing the thiosulfate removal rate (decomposition rate) per day of operation obtained from the results of thiosulfate monitoring. The thiosulfate removal rate is obtained from the amount of thiosulfate ions flowing in per day and the thiosulfate ion concentration of the treated water in the biological treatment area 20a obtained by monitoring. In other words, the thiosulfate removal rate is obtained from the difference between the amount of thiosulfate ions flowing into the biological treatment area 20a and the amount of thiosulfate ions (thiosulfate ions in the treated water) that were not completely treated in the biological treatment area 20a.

[4] Operation of a water treatment reactor using thiocyanate, phenol, and thiosulfate as pollutants In the above [1] to [3], one of thiocyanate, phenol, and thiosulfate was used as the pollutant, and a water treatment reactor was operated for each pollutant separately to extract water treatment data. Next, a water treatment reactor was operated using the three pollutants thiocyanate, phenol, and thiosulfate to extract water treatment data.

Artificial wastewater (water to be treated) was prepared by dissolving the solutes shown in Table 7 at the concentrations shown in Table 7 in a solvent obtained by mixing industrial water and natural seawater in a volume ratio of 2:3.

Furthermore, a 10mm x 10mm x 10mm sponge carrier 21 [fluid carrier (AQ-1 manufactured by Kanto Inoac)] and high-concentration activated sludge as a microbial inoculum were placed in a plastic bottle, kneaded thoroughly by hand, and left to soak overnight with the lid on, thereby allowing the microorganisms to adhere to the sponge carrier 21. It is preferable that the activated sludge placed here be an inoculum rich in thiocyanide-decomposing microorganisms, phenol-decomposing microorganisms, and thiosulfate-decomposing microorganisms.

The sponge carriers 21 (500 pieces) and activated sludge prepared in this manner were placed into the biological treatment area 20a of the biological treatment device 20, and the biological treatment device 20 was prepared.

The water to be treated 24 was introduced into the biological treatment device 20 prepared in this manner, and activated sludge was added as a microbial inoculant source. During the first stage of treatment, the water to be treated 24 was introduced so that the hydraulic retention time of the water to be treated 24 was 24 hours. Furthermore, the water to be treated 24 in each biological treatment device 20 was aerated 22 to form an aerobic fluidized bed, allowing the microorganisms to acclimate. Furthermore, treatment was carried out while adjusting the pH to around 7.5 using a 5 wt% sodium hydroxide aqueous solution. The treated water 26 treated in the biological treatment device 20 was then discharged from the biological treatment device 20.

The concentrations of thiocyanate ions, phenol, and thiosulfate ions were measured for the treated water in the biological treatment area 20a of each biological treatment device 20, and the concentrations of thiocyanate ions, phenol, and thiosulfate ions were monitored. Monitoring was performed approximately twice a week.

After this first-stage treatment had stabilized, water to be treated 24 was introduced so that the amounts of thiocyanate ions, phenol, and thiosulfate ions introduced per day were as shown in Table 8. Each stage was operated for at least two weeks, and once it was confirmed that the fluctuations in the treated water concentrations at each stage had stabilized, the system moved on to the next stage. Monitoring was carried out approximately twice a week.

Figure 9 is a characteristic diagram showing the thiocyanate ion removal rate per day of operation, obtained from the results of monitoring thiocyanate ions in the treated water in the biological treatment area 20a. Figure 10 is a characteristic diagram showing the phenol removal rate per day of operation, obtained from the results of monitoring phenol in the treated water in the biological treatment area 20a. Figure 11 is a characteristic diagram showing the thiosulfate removal rate per day of operation, obtained from the results of monitoring thiosulfate in the treated water in the biological treatment area 20a.

[5] Construction of a competitive model (non-toxic)
The inventors considered three pollutant components (phenol = P, thiocyanate = S, thiosulfate = _S2 ) in coke oven wastewater as shown in Figure 1, and separated the microorganisms attached to the fluidized carrier into three microbial groups that mainly treat each of the pollutant components. Taking into account the competitive relationships between these microbial groups, a competitive system model was constructed as shown in Figure 2. Because the treatment process is a fluidized carrier method, the constructed mathematical model describes the amount of microbial groups as the amount per unit surface area of the carrier.

As shown in Figure 1, thiocyanide contained in the coking wastewater is decomposed by the thiocyanide-decomposing microbial group S, phenol contained in the coking wastewater is decomposed by the phenol-decomposing microbial group P, and thiosulfate contained in the coking wastewater is decomposed by the thiosulfate-decomposing microbial group _S2 . Each microbial group grows by treating thiocyanide, phenol, and thiosulfate. Furthermore, each microbial group decreases due to interspecific competition as shown in Figure 2. Furthermore, each microbial group decreases due to autolysis. For this reason, the changes in the thiocyanide-decomposing microbial group, phenol-decomposing microbial group, and thiosulfate microbial group are expressed by the following formulas (1), (2), and (3), respectively.

In each of equations (1), (2), and (3), the first term on the right-hand side is a term related to self-replication (self-replication term). The second term on the right-hand side is a term related to interspecific competition (interspecific competition term). The third term on the right-hand side is a term related to self-decomposition (self-decomposition term).

The self-replication term has a positive (+) sign because the density b _i (i = S, P, S ₂ ) of the microbial community increases due to self-replication. On the other hand, the interspecific competition term has a negative (-) sign because the higher the density b _i of the microbial community, the more intense the interspecific competition becomes and the more microbial communities are reduced. Also, the autolysis term has a negative (-) sign because the higher the density b _i of the microbial community, the more microbial communities are reduced.

In the self-growth term, f _i (c _i ) (i = S, P, S ₂ ) is defined by the following equation (4). _{α i} (i = S, P, S ₂ ) represents the growth rate. _ci represents the concentration of the pollutant in the wastewater, c _S is the thiocyanide concentration, c _P is the phenol concentration, and c _S2 is the thiosulfate concentration. _ki is the death rate per unit time due to overcrowding, _Ki is the standard microbial population density per unit area, and β _i is the pollutant concentration that gives a growth rate that is half the maximum growth rate.

In the interspecific competition term, A _ij (i, j = S, P, S ₂ ) represents the interspecific competition coefficient. According to the interspecific competition term, the higher the density of the other two competing microbial species, the smaller the change in density b _i on the left side.

In the autolysis term, d _i (i = S, P, S ₂ ) represents the autolysis rate. According to the autolysis term, the higher the density b _i of the microbial population, the smaller the change in density b _i on the left side. The autolysis term represents the natural death of a certain proportion of the microbial population due to autolysis.

The inventors also constructed models expressed by the following formulas (5), (6), and (7) as decomposition models for three pollutant components. In each of formulas (5), (6), and (7), c _i (i = S, P, S ₂ ) represents the concentration of each pollutant component, c _S is the concentration of thiathioic acid in the wastewater, c _S is the concentration of phenol in the wastewater, and c _S2 is the concentration of thiosulfate in the wastewater.

In each of equations (5), (6), and (7), the left side is the value obtained by multiplying the change in concentration _ci of each pollutant component by the volume V of the biological treatment device 20. Also, in each of equations (5), (6), and (7), L is the surface area of the sponge carrier 21.

In each of equations (5), (6), and (7), the first term on the right side is the pollutant decomposition reaction term, and h _i (b _i ) (i = S, P, S ₂ ) is expressed by the following equation (8): Equation (8) means that h _i (b _i ) changes depending on the concentration b _i of the microbial group.

Furthermore, in each of equations (5), (6), and (7), the second term on the right-hand side is the inflow term of the water to be treated, which is a term related to the concentration of pollutants due to the inflow of the water to be treated 24 into the biological treatment device 30. _Pi (i = S, P, _S2 ) is the inflow rate, which is defined as the reciprocal of the hydraulic retention time. Therefore, the product of the inflow rate and the hydraulic convection time is the amount of water to be treated 24 flowing into the biological treatment area 20a of the biological treatment device 20 per hour. _Ciin ⁽ i = S, ^P , _S2 ⁾ is the inflow concentration of each pollutant (type i pollutant). Specifically, _cSin , _cPin , ^and _cS2in are values calculated from the proportion of each pollutant in the inflowing water to be treated 24.

[6] Simplification of the first-class model by partial non-dimensionalization (variable transformation) (non-toxic)
Of the constructed models, the model for decomposing one type of pollutant, excluding the interspecies competition term, was partially non-dimensionalized, and equations (1) to (3) and (5) to (7) were simplified to the following equations (9) and (10).

In equations (9) and (10), K _i , k _i , the total surface area L of the sponge carrier, and the volume V of the biological treatment area are non-dimensional. The details of the simplification are as follows: First, replace S, P, and S ₂ in equations (1) to (3) with i, and then replace B i with B _i = (k _i /α _i * K _i ) * b _i . Furthermore, in equations (1) to (3), the interspecies competition term (second term) on the right-hand side is removed. Furthermore, by setting _{R i} = α _i * r _i (L/V) and Γ i = (k _i /α _i * K _i ) γ _i and rearranging the equations, equations (9) and (10) are obtained. This non-dimensionalization is intended to non-dimensionalize the density b _i of the microbial population. The pollutant concentration c _i is not non-dimensional because the inflow concentration c _i ⁱⁿ is included in equation (10) as an experimental parameter. Time is also not non-dimensionalized.

[7] Identification of autolysis rate (d _i ) When determining the parameters of the model [5] or [6] through the experiments [1] to [4], it is difficult to distinguish between the growth and decomposition of microorganisms in a normal treatment system where both occur simultaneously. Therefore, the autolysis rate d _i (i = S, P, S ₂ ) (autolysis rate) was first determined based on experiments in different treatment systems.

In operation of the water treatment reactor [1] using thiocyanate as a pollutant, artificial wastewater was prepared from the solutes shown in Table 1, excluding sodium thiocyanate, from the 197th day onward to create starvation conditions for the thiocyanate-decomposing microorganisms. As shown in Figure 12, 30 sponge carriers 21 were collected from the water treatment reactor on days 190, 197, 203, 206, 209, 210, and 215. They were then placed in a beaker 30 containing 300 mL of artificial wastewater prepared with the composition shown in Table 1, and a 3-hour incubation test was conducted. The difference in thiocyanate concentration before and after the 3-hour incubation was calculated. Figure 13 is a characteristic diagram showing the difference in thiocyanate concentration in the incubation test conducted on each sampling date.

As a result, as shown in Figure 13, the concentration difference decreased from day 197 onwards and remained constant from day 203 onwards. It is believed that from day 203 onwards, the artificial wastewater was diluted by the solvent contained in the sponge carrier 21, resulting in an apparent concentration difference. In other words, on days 190 and 197, the concentration difference is thought to have increased due to dilution caused by the carrier being added to beaker 30 and the decomposition of thiocyanate. On the other hand, from day 203 onwards, as mentioned above, the concentration difference is thought to be due solely to dilution. Here, day 197 was the day sodium thiocyanate was removed from the artificial wastewater, and there is a possibility that thiocyanate still remained in the reactor. Therefore, it is believed that the ability to decompose thiocyanate was lost over the five days from day 198 to day 203, i.e., the thiocyanate-decomposing microorganisms were completely wiped out.

In formula (9), if the concentration c _i of the pollutant component (i) such as thiocyanate in the culture test is set to 0, formula (9) becomes the following formula (11).

According to the model of equation (11), the amount of microorganisms B _i (t) at time t is expressed by the following equation (12), where B _i (0) is the amount of microorganisms at the initial time.

When equation (12) is solved for d _i , the following equation (13) is obtained.

The unit time in the equation system is 1 hour, and according to the experimental results described above, it is believed that the microbial community was wiped out 5 days from the initial state, i.e., 120 hours later. In the model equation, the microbial community will never be wiped out, i.e., B _i (t1) = 0 at a certain time t1. Therefore, if we consider the experimental results to indicate that the microbial mass decreased to about 1% of the initial amount in 120 hours, below the observation limit, the autolysis rate d _i can be calculated as follows:

From the above, the autolysis rate _ds of the thiocyanate-decomposing microorganisms was estimated to be approximately 0.0384. The above values are also used as the autolysis rates _di for phenol and thiosulfate, but the autolysis rates of phenol and thiosulfate may be determined individually using a similar method. The calculation of the autolysis rate described above is only an example, and autolysis coefficients proposed in known activated sludge models may be used instead.

[8] Simplification of the Type 1 model by partial non-dimensionalization (toxicity)
When the concentration of pollutants is high, the pollutants may act as toxic components and inhibit or kill the growth of the corresponding microorganisms. In particular, the remaining pollutants that cannot be completely treated by the microorganisms may be toxic. For this reason, a term related to toxicity (the third term on the right side) was added to equation (9), which non-dimensionally represents k _i , K _i , the total surface area L of the sponge carrier, and the volume V of the biological treatment device 20, as shown in equation (14) below. The newly added term related to toxicity is close to zero when the concentration c _i of the pollutants is sufficiently small, but becomes more toxic as the concentration c _i of the pollutants becomes sufficiently large, suppressing the growth of the microorganisms and even reducing them.

[9] Simplification of the three-type model by partial non-dimensionalization (with toxicity)
A model for wastewater containing all three pollutants (S: thiocyanate, P: phenol, S2: thiosulfate) was made dimensionless as shown in the following equations (15), (16), and (17). Equations (15) to (17) can be obtained by adding an interspecies competition term to equation (9), making it dimensionless in the same way as equation (14), fractionating it into three pollutant types (i = S, P, _S2 ), and adding a term related to toxicity (the third term on the right-hand side). When B _i =(k _i /α _i *K _i )*b _i in the non-dimensionalization, in the interspecific competition term, ((α _S *K _S )/k _S )*((α _P *K _P )/k _P )*A ₁₂ =A' ₁₂ , ((α _S *K _S )/k _S )*((α _S2 *K _S2 )/k _S2 )*A ₁₃ =A' ₁₃ , ((α _P *K _P )/k _P )*((α _S *K _S )/k _S )*A ₂₁ =A' ₂₁ , ((α _P *K _P )/k _P )*((α _S2 *K _S2 )/k _S2 )*A ₂₃ =A' ₂₃ , ((α _S2 *K _S2 )/k _S2 )*((α _S *K _S )/k _S )*A ₃₁ =A' ₃₁ , ((α _S2 *K _S2 )/k _S2 )*((α _P *K _P )/k _P )*A ₃₂ =A' ₃₂ . Furthermore, equations (18) to (20) can be obtained by dividing equation (15) into three types of pollutants (i=S, P, S ₂ ).

[10] Parameter Fitting of the Model Equations The parameters in equations (15) to (20) are determined by fitting the constructed model to the results of a water treatment experiment using wastewater. That is, the model equations (15) to (20) are numerically calculated according to the conditions of the water treatment experiment described in [1] to [4] to determine the amount (density) of the microbial community and the pollutant concentration at each time point. The decomposition rate of the pollutants is then obtained and compared with the results of the water treatment experiment. For the numerical calculation, a known method such as the Runge-Kutta method is used. When comparing the numerical calculation with the experimental results, it is preferable to use a criterion based on relative error, since the decomposition rate varies greatly at each time point. Note that at the initial time, the amount of the microbial community was set to _bi (0) = 1, and the pollutant concentration was set to _c (0) = 0. This is done in the hope that the state obtained from the model equations will quickly converge to a state in which the pollutant components are sufficiently decomposed. The actual measured values of the decomposition rate of each pollutant have been obtained from the experiments described in [1] to [4], so there is no need to fit these parameters.

[10.1] Details of Parameter Fitting Consider the case where a steady state can be achieved under each experimental condition. In this case, a subscript n is assigned to each experimental condition. That is, the sponge surface area _Ln , _inflow rate _Pn , and inflow pollutant concentration ^cinn are determined for each experimental condition n (1≦n≦N), and numerical calculations are performed assuming the sponge surface area is _Ln , the inflow rate is _Pn , and the inflow pollutant concentration is _cnin ^. Furthermore, the measured value is _cn . When parameter fitting is performed under a steady state, the measured value _cn is the average value of the measured values under each experimental condition n. Note that the measured value _cn may also be the value obtained when a steady state is reached under each experimental condition n. In this case, a numerical experiment is performed as follows, and the numerical experimental value c^ _n (λ) corresponding to the parameter λ is calculated. Note that experimental condition n is a serial number specifying each experimental condition, and N is the total number of experimental conditions. The experimental conditions here are defined by the hydraulic retention time, the concentration of the inflowing material, and the total surface area of the sponge, but there may be some experimental conditions that provide the same conditions.

1) Appropriately set the starting conditions ^b ₀ (λ) and c^ ₀ (λ). Typically, these are set as ^b ₀ (λ) = 1 and c^ ₀ (λ) = 0, in the hope that the decomposition of pollutant components is sufficiently performed.
2) The numerical solutions of the model equations with ^b _n-1 (λ) and c^ _n-1 (λ) as initial values are defined as ^-b _n (t; λ) and c ^- _n (t; λ), where ^b _n (λ) and c^ _n (λ) are defined as in the following equations (21) and (22).

The mathematical expression is as above, but in practice, the numerical calculation is performed until it is deemed to have converged sufficiently numerically, and then it is stopped at that point and used as a substitute for the limit value.
The cost function E(λ) is calculated using the above numerical experimental value c^ _n (λ). The cost function E(λ) is set appropriately, but in this example, the definition of the following equation (23) is used, utilizing the relative error of the decomposition rate.

In this example, sampling was performed using the Metropolis-Hastings method with E(λ) as the likelihood function, and parameters that minimize E(λ) were searched for.

Among the parameters, _αi , _βi , _ωi , _δi , _Ri , _Γi , _ni , _mi , _qi (i = S, P, _S2 ) are all parameters that are not related to the interspecies competition term. Therefore, these parameters are determined by fitting the parameters so that the decomposition rate of each pollutant obtained from the results of water treatment experiments using wastewater containing one type of pollutant as described in [1] to [3] above matches the decomposition rate of each pollutant determined from the model for one type of pollutant excluding the interspecies competition term (Equation (14), Equation (10)). Alternatively, these parameters can be determined by fitting parameters so that the concentrations of each pollutant obtained from the results of water treatment experiments using wastewater containing one type of pollutant as described in [1] to [3] above match the concentrations of each pollutant determined from a model (Equation (14), Equation (10)) targeting one type of pollutant excluding the interspecies competition term. Whether fitting parameters based on the decomposition rate of each pollutant or fitting parameters based on the concentration of each pollutant is performed can be selected appropriately depending on the water treatment process being designed. For example, fitting based on the decomposition rate is performed when calculating the treatment capacity of the water treatment process, particularly the maximum capacity. Furthermore, fitting based on the concentration is performed when calculating the remaining (untreated) concentration of each pollutant and determining whether it satisfies the effluent standards. Note that fitting to determine parameters unrelated to the interspecies competition term is referred to as the first-stage fitting. Here, since R _i = α _i * r _i (L/V), once R _i is determined, r _i is determined. Furthermore, since Γ _i =(α _i *k _i /K _i )γ _i , once Γ _i is found, γ _i is found.

For example, when one type of pollutant is thiathione and the parameters α _S , β _S , ω S , δ S , R S , Γ _S , n _S , m _S , and _{q S} that are not related to the interspecies competition term are fitted from the decomposition rate of thiocyanate obtained from the water _treatment experiment results in [ ₁ ] above, parameter fitting is performed in steps 1 to 8 shown in Table 2 so that the decomposition rate of thiocyanate calculated from equations (14) and (10) matches the decomposition rate of thiocyanate obtained from the experiment shown in Figure 6.

Furthermore, the competitive coefficients _A'12 , _A'13 , _A'21 , _A'23 , _A'31 , and _A'32 can be determined by a second-stage fitting, in which the results of a water treatment experiment using wastewater containing the three pollutants described above in [4] are fitted to the calculated values of the model of the above-mentioned equations (15) to (20). In this case, since parameters other than _A'12 , _A'13 , _A'21 , _A'23 , _A'31 , and _A'32 have already been determined in the first-stage fitting, the second-stage fitting is performed with these parameters incorporated into equations (15) to (20).

In this case, parameter fitting is performed in stages 1 to 9 shown in Table 8 so that the decomposition rate of each pollutant calculated from equations (15) to (20) matches the decomposition rate of each pollutant obtained in the experiments shown in Figures 9 to 11. Alternatively, parameter fitting is performed in stages 1 to 9 shown in Table 8 so that the concentration of each pollutant calculated from equations (15) to (20) matches the concentration of each pollutant obtained in the experiments shown in Figures 9 to 11.

[11] Simulation results for three types of models The results of determining parameters using the above method are shown below. First, based on the actual measurement data obtained from water treatment reactors operated in [1], [2], and [3] using one type of pollutant, thiocyanate, phenol, or thiosulfate, parameter fitting was performed for the model (Equation (14), Equation (10)) constructed in [8] targeting one type of pollutant. The value determined in [7] was used for the autolysis rate (d _i ). As a result, the parameters were estimated as shown in Table 9 below. Note that the parameters n _i , m _i , and q _i were fitted within the range of natural numbers.

Next, based on the data from the water treatment reactor operated in [4] with all three pollutants - thiocyanate, phenol, and thiosulfate - parameter fitting was performed using the parameters determined in the first stage of fitting to determine the coefficients of the interspecies competition terms in the model (equations (15) to (20)) for the three pollutants constructed in [9]. As a result, the following parameters were estimated.

Using the model [9] with the estimated parameters, the decomposition rates of each pollutant, thiocyanate (SCN), phenol (P), and thiosulfate (S2), were calculated and compared with the measured decomposition rates obtained in the experiment [4]. Figures 14 to 16 are characteristic diagrams showing the results of comparing the calculated and measured values. Figure 14 shows the comparison results for thiocyanate, Figure 15 shows the comparison results for phenol, and Figure 16 shows the comparison results for thiosulfate. In Figures 14 to 16, the vertical axis of the upper graph represents the decomposition rate, and the vertical axis of the lower graph represents the ratio ((calculated value - measured value)/measured value) obtained by subtracting the measured value from the calculated value and dividing it by the measured value. The results of Figures 14 to 16 demonstrate that the wastewater quality simulation method of the present invention was able to precisely simulate the measured values.

[12] Comparison of toxic and non-toxic models (1 type)
Next, for the model for one pollutant constructed in [6] and [8], parameter fitting was performed based on the measured data of the water treatment reactor operated in [1] using thiocyanate as the pollutant, and the simulation results were compared with the measured values. The autolysis rate (d _i ) was the value calculated in [7].

In the toxic model, stages 1 to 9 shown in Table 2 in [1] above were designated as Conditions 1 to 9, and parameter fitting was performed so that the water treatment rate calculated from Equation (14) and Equation (10) for each condition matched the decomposition rate obtained in the experiment shown in Figure 6.

In addition, in the non-toxic model, stages 1 to 9 shown in Table 2 in [1] above were set as conditions 1 to 9, and parameter fitting was performed so that the water treatment rate calculated from equations (9) and (10) for each condition matched the decomposition rate obtained in the experiment shown in Figure 6.

Figure 17 is a characteristic diagram showing the results of comparing the toxic model and the non-toxic model. The top of Figure 17 shows the results of comparing the calculated values and the measured values for the non-toxic model. The bottom of Figure 17 shows the results of comparing the calculated values (Calculation) and the measured values (Experiment) for the toxic model. Again, the vertical axis of the left-hand graph shows the decomposition rate, and the vertical axis of the right-hand graph shows the ratio ((Calculation value - Measured value) / Measured value) obtained by subtracting the measured value from the calculated value and dividing it by the measured value. According to the results of Figure 17, the measured values (Experiment) and simulation results (Calculation) differed significantly for the non-toxic model, but by using a model that takes toxicity into account, simulation results that were nearly identical to the measured values were obtained.

[13] Utilization of the model formulas Once the models of formulas (15) to (20) are completed as a result of parameter fitting, it becomes possible to simulate the decomposition rate of each pollutant, as shown in Figures 14 and 15. It also becomes possible to simulate the concentration of each microbial group.

Furthermore, according to equations (15) to (20), the concentrations c _S , c _P , ^and c S2 of each pollutant in the treated water 26 after decomposition treatment can be determined based on the inflow concentrations c _S ⁱⁿ , c _P ⁱⁿ , and c _S2 in of each pollutant. Therefore, it is possible to control the concentration of each pollutant in the _treated water 26 to be below a desired value.

For example, in equations (15) to (20), as described above, R _i =α _i *r _i (L/V). Therefore, by changing the value of L for the value of R _i obtained by fitting, the concentrations c _S , c _P , and c _S2 of each pollutant when these values are changed can be determined. This allows the required surface area L of the sponge carrier 21 to be determined when controlling the concentration of each pollutant in the treated water 26 to a desired value, making it possible to add just the right amount of sponge carrier 21 according to the target concentration of each pollutant. This prevents costs from being increased by adding too much sponge carrier 21. Furthermore, by adding just the right amount of sponge carrier 21, the sponge carrier 21 is occupied by the necessary microbial population, making it possible to eliminate unnecessary microbial populations.

[14] Parameter fitting taking into account unsteady states In the above explanation, it was assumed that a steady state could be achieved under each experimental condition, and the average value of the measurement data at each stage was used as a representative value, and this value was used as the removal rate per day. On the other hand, in the case of a model that solves a steady state problem, it is possible to evaluate the treatment capacity in a stable state after reaching a steady state under certain operating conditions, but because the actual increase and decrease of microorganisms includes a time element, it is preferable to use a model that can evaluate the treatment capacity in an unsteady state taking time changes into account.

Figure 18 is a characteristic diagram showing the thiocyanate concentration of treated water in the biological treatment area 20a obtained from experimental results during operation of the water treatment reactor described above in [1] using thiocyanate as a pollutant. Stages 1 to 8 in Figure 18 correspond to stages 1 to 8 in Table 2, respectively. As shown in Figure 18, even under the same experimental conditions (stages), the pollutant concentration changes from moment to moment, and the water quality data changes over time. For this reason, we will build a model that can simulate water quality in an unsteady state, taking time changes into account. A model that can simulate water quality in an unsteady state makes it possible to determine the pollutant concentration, which changes from moment to moment. This makes it possible to obtain information such as, for example, how many days it will take for the concentration of pollutants remaining in the treated water to recover to a level that meets specifications, or how many days it will take for the concentration of pollutants to deteriorate to a level that exceeds specifications.

When treating it as a steady state, the subscript n is assigned to each experimental condition, but when treating it as an unsteady state, the subscript n is assigned to each measurement. That is, the total sponge surface area _Ln ^, inflow rate _Pn , and inflow pollutant component concentration _cnin are determined corresponding to measurement n (1≦n≦N). Also, the measured value is _cn . At this time, a numerical experiment is performed as follows, and the numerical experimental value c^ _n (λ) corresponding to the parameter λ is calculated.

1) Appropriately set the starting conditions ^b(0;λ) and c^(0;λ). Typically, these are given as ^b(0;λ) = 1 and c^(0;λ) = 0, with the expectation that pollutant components are being sufficiently decomposed. Unlike when treating the system as a steady state, when treating the system as a non-steady state, the starting conditions may affect the value of the cost function. Since estimating appropriate starting conditions increases the number of parameters that need to be estimated, in practice it is considered desirable to operate the system under conditions that allow sufficient decomposition of pollutant components only in the early stages until a steady state is reached.

2) At time tn _-1 , set b(tn _-1 ) = ^b(tn _-1 ; λ) and c(tn _-1 ) = c^(tn _-1 ; λ) and numerically integrate the model equation up to time _tn . At this time, the numerical calculation is performed under the operating conditions of sponge surface area _Ln , inflow rate _Pn , and inflow pollutant concentration _cn ⁱⁿ .

The cost function E(λ) is calculated using the above numerical experimental value c^ _n (λ). The cost function E(λ) is set appropriately, but here the cost function E(λ) of the following equation (24) is used.

In equation (24), ε is a parameter associated with the measurement error, and in this example, ε = 1.0. Furthermore, sampling is performed using the Metropolis method with E(λ) as the likelihood function to explore the parameters that minimize E(λ).

Here, based on the experimental results shown in Figure 18, i.e., the experimental results of [1] above, a model capable of simulating water quality in an unsteady state using thiocyanate as the pollutant was constructed. The results of the first stage of parameter fitting were α _S ≒ 0.0491, β _S ≒ 0.0072, ω _S ≒ 0.0115, R _S ≒ 1.6453, γ _S ≒ 0.292, δ _S ≒ 0.3017, n _S = 1, ms = 1, and qs = 3. Note that by applying a similar method to the experimental results of [2] to [4] above, it is possible to construct models capable of simulating water quality in an unsteady state for other pollutants, including the three types of pollutants. The constructed model makes it possible _to predict the concentration of each of multiple pollutants in the treated water from the value at any point in time to the value at any later point in time, depending on the total surface area of the sponge L _n , the inflow rate _{P n} , and the inflow pollutant component concentration c n ⁱⁿ .

Figure 19 is a characteristic diagram showing the actual measured values of thiocyanate concentration based on the experimental results shown in Figure 18, and the simulation results of thiocyanate concentration obtained using a non-steady-state model that takes time changes into account (characteristics shown by the solid line). The vertical axis represents thiocyanate concentration, and the horizontal axis represents the number of days. As shown in Figure 19, the model capable of simulating water quality under non-steady-state conditions shows that the thiocyanate concentration in the treated water rises sharply at treatment stage 7. This makes it possible to predict when treatment will deteriorate. Furthermore, as shown in Figure 19, the model capable of simulating water quality under non-steady-state conditions shows that immediately after a change in treatment stage (such as immediately after switching from stage 3 to 4, or from stage 5 to 6), the growth of treating microorganisms cannot keep up, causing a temporary rise in the thiocyanate concentration in the treated water, resulting in a temporary deterioration in treatment.

[15] Method for determining the number of carriers using the flowing carrier method Next, a method for determining the required number of carriers by determining model parameters from experiments in which the number of carriers is changed will be described. Since the required number of carriers is obtained from simulation, the number of carriers can be optimized and excessive increase in the amount of carriers can be prevented.

To build the model, a water treatment reactor using thiocyanate as the pollutant was operated in the same manner as in [1] above, and then the number of carriers was changed while the water treatment reactor was operated, and parameter fitting was performed from the measurement results. Here, we will explain the construction of a model that can simulate water quality in an unsteady state when thiocyanate is used as the pollutant, based on the operation of this water treatment reactor. By applying this model construction method to the experimental results in [1] to [4] above, it is possible to build a model that can simulate water quality in an unsteady state for three types of pollutants.

First, as in [1] above, industrial water and natural seawater were mixed in a volume ratio of 2:3 to obtain a solvent, and the solutes shown in Table 11 below were dissolved in the concentrations shown in the table below to prepare artificial wastewater (water to be treated).

As shown in Figure 5, an integrated biological treatment device 20 was prepared, in which a biological treatment area 20a and a sedimentation area 20b are separated from each other by a partition 23 within a single tank, and the areas are interconnected below this partition 23. A 10mm x 10mm x 10mm sponge carrier 21 [fluid carrier (AQ-1 manufactured by Kanto Inoac)] and high-concentration activated sludge as a microbial inoculum source were placed in a plastic bottle, kneaded thoroughly by hand, and left to soak overnight with the lid on, allowing the microorganisms to adhere to the sponge carrier 21.

The sponge carriers 21 (500 pieces) and activated sludge prepared in this manner were placed into the biological treatment area 20a of the biological treatment device 20, and the biological treatment device 20 was prepared. Since the volume of the biological treatment area 20a was 3.4 L, the carrier loading rate was 14.7% (v/v).

The water to be treated 24 was then introduced into the biological treatment device 20 prepared in this manner, and activated sludge was added as a microbial inoculant source. During the first stage of treatment, the water to be treated 24 was introduced so that the hydraulic retention time was 48 hours. The water to be treated 24 in each biological treatment device 20 was aerated 22 to form an aerobic fluidized bed, allowing the microorganisms to acclimate. Treatment was also carried out while adjusting the pH to around 7.5 using a 5 wt% aqueous sodium hydroxide solution. The treated water 26 was then discharged from the biological treatment device 20.

The thiocyanate ion concentration was measured in the treated water in the biological treatment area 20a of each biological treatment device 20 to monitor thiocyanate ions. Monitoring was performed approximately twice a week.

After this first-stage treatment had stabilized, water to be treated 24 was introduced so that the amount of thiocyanate ions introduced per day was as shown in Table 12 below. Each stage was operated for at least two weeks, and once it was confirmed that the fluctuations in the treated water concentration at each stage had stabilized, the system moved on to the next stage. Monitoring was carried out approximately twice a week.

Figure 20 is a characteristic diagram showing the thiocyanate ion concentration in the treated water obtained from the results of thiocyanate ion monitoring.

After operating the water treatment reactor according to Table 12, 300 sponge carriers 21 were collected and placed in the biological treatment area 20a of a new biological treatment device 20 to prepare the biological treatment device 20. Note that in order to conduct a more precise experiment, assuming that microorganisms that have peeled off from the sponge carriers 21 have accumulated at the bottom or that microorganisms are attached to the walls of the biological treatment device 20, 300 sponge carriers 21 were collected and placed in the biological treatment area 20a of the new biological treatment device 20. However, for simplicity, it is also possible to collect all of the sponge carriers 21 from the biological treatment device 20 and return the 300 sponge carriers 21 to the biological treatment area 20a of the same biological treatment device 20.

The water to be treated 24 was introduced into the biological treatment device 20 prepared in this manner so that the hydraulic residence time was 24 hours, and the amount of thiocyanate ions introduced per day was 460 mg SCN/L. Furthermore, the water to be treated 24 in each biological treatment device 20 was aerated 22, and treatment was carried out while adjusting the pH to around 7.5 using a 5 wt% aqueous sodium hydroxide solution. The treated water 26 was then discharged from the biological treatment device 20.

After this first stage treatment had stabilized, sponge carriers 21 were removed so that the number of sponge carriers 21 was as shown in Table 13 below. Each stage was operated for at least two weeks to confirm that the fluctuations in treated water concentration at each stage had stabilized before moving on to the next stage. Monitoring was carried out approximately twice a week.

Figure 21 is a characteristic diagram showing the thiocyanate ion concentration in the treated water obtained from the results of thiocyanate ion monitoring. When determining parameter values, parameter values that minimize the cost function are selected, but if there are multiple parameter values with similar cost function values, selection can be difficult.

Here, the problem was treated as an unsteady problem (with time changes) as explained in [14], parameter values were determined, and the number of sponges required was simulated. Based on actual measurement data obtained from an operating water treatment reactor using thiocyanate as a pollutant, parameter fitting was performed on the model (Equation (14), Equation (10)) for one type of pollutant constructed in [8]. The value calculated in [7] was used for the autolysis rate (d _i ). As a result, the parameters were estimated as shown in Table 14 below. Note that the parameters n _i , m _i , and q _i were fitted within the range of natural numbers.

Figure 22 is a characteristic diagram showing both simulation results and actual measurements when the number of sponge carriers 21 is changed. Figure 22 shows the simulation results and actual measurements for stages 1 to 3 listed in Table 13. The constructed model makes it possible to predict the concentration of each of multiple pollutants in the treated water from a value at any point in time to a value at any subsequent point in time, depending on the total sponge surface area _Ln , the inflow rate _Pn , and the inflow pollutant concentration _cn ⁱⁿ . In particular, since the parameters are determined taking into account changes in the total sponge surface area _Ln , the thiocyanate concentration corresponding to changes in the total sponge surface area _Ln can be determined more accurately. The number of sponge carriers 21 is calculated from the total sponge surface area _Ln based on the surface area per sponge. As shown in Figure 22, the simulation results clearly show that the thiocyanate concentration in the treated water increases as the number of sponge carriers 21 decreases, which is consistent with the trend of the actual measurements. In particular, at the third stage (200 sponge carriers), the simulation results show that a relatively high concentration of thiocyanate remains in the treated water, making it impossible to operate the water treatment reactor under such conditions. On the other hand, at the second stage (250 sponge carriers), the simulation results show that the thiocyanate concentration is suppressed. Therefore, it can be seen that the minimum number of sponge carriers in the second stage (=250) is required to suppress the thiocyanate concentration. In this way, the required number of sponge carriers can be determined from the simulation results. Note that the microbial concentration shown in Figure 22 is shown as a relative amount, with the maximum amount adhering to the surface of the sponge carrier 21 set to 1.

[16] Simulation Method Using Multiple Parameter Sets It is generally preferable to determine optimal parameter values for each target process as the model parameter values. This is because optimal parameter values often differ for each process. Figure 23 shows simulation results based on the experimental results shown in Figure 18, and is a characteristic diagram showing simulation results when parameters are changed, unlike Figure 19. As described above, the parameter values in Figure 19 are α _S ≒ 0.0491, β _S ≒ 0.0072, ω _S ≒ 0.0115, R _S ≒ 1.6453, γ _S ≒ 0.292, δ _S ≒ 0.3017, n _S = 1, ms = 1, and qs = 3. Figure 23 shows an example of a simulation of the experimental results shown in Figure 18 using parameters different from those in Figure 19, with parameter values being α _S ≈ 0.0472, β _S ≈ 0.0036, ω _S ≈ 0.0101, R _S ≈ 2.4694, γ _S ≈ 0.0485, δ _S ≈ 0.3694, n _S = 1, ms = 1, and qs = 3. Although Figures 19 and 23 use different parameter values, both are similar to the experimental values and produce good simulation results.

On the other hand, Figure 24 is a characteristics diagram showing an example of a simulation of the results of another water treatment reactor using the same parameters as Figure 19, based on the experimental results shown in Figure 18. Similarly, Figure 25 is a characteristics diagram showing an example of a simulation of the results of another water treatment reactor using the same parameters as Figure 23, based on the experimental results shown in Figure 18.

As shown in Figures 24 and 25, when parameter values determined from the results of one water treatment reactor are used to simulate the results of another water treatment reactor, completely different simulation results may result. In particular, in Figure 24, the simulation results show that almost no pollutant concentrations are detected in the treated water up to the fourth stage, meaning that stages 1 to 3 are good treatment, but in reality, the pollutant concentrations in the treated water increase in stages 2 and 3, and treatment deteriorates.

As mentioned above, limiting the model formula parameters to one parameter can lead to over-learning, which can lead to results that are completely different from the actual measurement results when predicting pollutant components in other water treatment reactors. For this reason, we propose a method that simultaneously applies simulation results based on multiple parameters, making predictions possible even when the target process changes.

First, a set of parameters to be searched is defined as Λ, an appropriate value of ε is determined, and _Λε is defined by the following equation (25).

Next, in order to give the lower and upper limits of b and c such that the cost function (evaluation function) E(λ) of equation (25) is equal to or smaller than ε, b _min (t), b _max (t), c _min (t), and c _max (t) are defined for Λ _ε as follows:

Next, the estimated value m of b(t) and c(t) is output as the following interval:

Here, ε = 2.0 was set, and parameters were sampled using the Metropolis method for the measurement data in Figure 18. Of the approximately 7,000 parameters sampled, there were 81 parameter sets (combinations) that satisfied E(λ) < ε = 2.0. Figure 26 is a characteristic diagram showing the results of simulations calculated using these 81 parameter sets, with the range between the maximum and minimum values filled in.

Figure 27 shows the results of simulating the results of another water treatment reactor using all 81 parameter sets determined by sampling parameters using the Metropolis method on the measurement data of Figures 24 and 25 in a manner similar to that of Figure 26. Figure 27 also shows the results of shading the range between the maximum and minimum values for each simulation result calculated using a set of 81 parameters. As shown in Figure 27, the results show that treatment deteriorates from the second stage onwards. This demonstrates that by using multiple parameter values determined in a water treatment reactor, it is possible to predict deterioration in water treatment in another reactor.

As described above, by simulating the post-treatment water quality of the water to be treated based on multiple model equations obtained by fitting, which correspond to parameters for which the value of the evaluation function used in fitting is below a predetermined value, it is possible to predict deterioration in water treatment in a reactor other than the water treatment reactor from which actual measurements were obtained.

20 Biological treatment device 20a Biological treatment area 20b Sedimentation area 21 Sponge carrier 22 Aeration 23 Partition wall 24 Water to be treated 26 Treated water treated by the biological treatment device

Claims (11)

A water quality simulation method executed by a processor in a treatment process for biologically treating water to be treated, comprising:
a model formula including a term representing interspecific competition among a plurality of different microbial groups that decompose each of a plurality of different pollutants contained in the water to be treated, the term representing that the higher the density of any one of the plurality of microbial groups, the greater the reduction of other microbial groups; and a simulation of the water quality after treatment by the plurality of microbial groups .
The plurality of microbial groups are a microbial group capable of decomposing a first pollutant, a microbial group capable of decomposing a second pollutant, and a microbial group capable of decomposing a third pollutant,
The water quality simulation method , wherein the model formula expresses the change per unit time of the decomposition microorganism group of the first pollutant by the following formula (1) :
In equations (1) and (4), i = S, P, S2 _, where S represents the first pollutant, P represents the second pollutant, and S2 _represents the third pollutant, and fi ₍ ci ₎ is defined by the following equation (4): In equations (1) and (4), bi represents _the density of the decomposing microorganisms for each pollutant, αi _represents the growth rate, and ci _represents the concentration of the pollutant in the wastewater. Ki _is the death rate per unit time due to overcrowding, Ki is _the standard microbial population density per unit area, βi is _the pollutant component concentration that gives a growth rate of half the maximum growth rate, and ni _is a power coefficient. In equation (1), A12 _and A13 _are interspecies competition coefficients, and di _is the autolysis rate.
2. The water quality simulation method of claim 1, wherein the model formula further includes a term that indicates that, for each of the plurality of pollutants, the higher the concentration of the pollutant being decomposed, the smaller the change per unit time in the density of the microbial population that decomposes the pollutant due to toxicity. 3. The water quality simulation method according to claim 1, further comprising: calculating the decomposition rate of each of the plurality of pollutants using the model formula; and fitting parameters representing the model formula based on the actual measured values of the decomposition rates of the plurality of pollutants and the calculation results of the decomposition rates using the model formula. 3. The water quality simulation method according to claim 1 or 2, wherein the concentration of each of the plurality of pollutants in the treated water is calculated using the model formula, and parameters representing the model formula are fitted based on the actual measured values of the concentrations of each of the plurality of pollutants in the treated water when they reach a steady state and the calculation results of the concentrations of each of the pollutants using the model formula. 3. The water quality simulation method according to claim 1 or 2, wherein the concentration of each of the plurality of pollutants in the treated water is calculated using the model formula, and parameters representing the model formula are fitted based on the actual measured values of the concentrations of each of the plurality of pollutants in the treated water at each measurement point and the calculation results of the concentrations of each of the pollutants using the model formula corresponding to each measurement point. The water quality simulation method according to claim 5 , wherein the concentration of each of the plurality of pollutants in the treated water is predicted from a value at an arbitrary time point to a value at a later arbitrary time point. The water quality simulation method according to any one of claims 3 to 6, wherein the water quality after treatment of the treated water is simulated based on a plurality of model formulas obtained by the fitting, the model formulas corresponding to parameters for which the value of the evaluation function used in the fitting is equal to or less than a predetermined value. The water quality simulation method according to any one of claims 1 to 7 , wherein the treatment process is a fluidized carrier method. The water quality simulation method according to any one of claims 1 to 8 , wherein the pollutants include at least one of thiocyanate, phenol, and thiosulfate. A water quality simulation device for a treatment process that biologically treats water to be treated, the water quality simulation device comprising:
The processor:
a parameter fitting unit that fits parameters of the model formula based on actual measurement values of decomposition rates of a plurality of pollutants obtained by introducing a plurality of microbial groups that decompose each of the plurality of different pollutants into the water to be treated, and calculated values of decomposition rates of a plurality of the pollutants calculated from a model formula having a term that represents interspecific competition between the plurality of microbial groups in the water to be treated, the term representing that the higher the density of any one of the plurality of microbial groups, the greater the decrease in other microbial groups;
a water quality simulation unit that simulates the water quality of the treated water after treatment using the model formula to which the parameters have been fitted;
Equipped with
The plurality of microbial groups are a microbial group capable of decomposing a first pollutant, a microbial group capable of decomposing a second pollutant, and a microbial group capable of decomposing a third pollutant,
The model formula expresses the change per unit time of the decomposition microorganism group of the first pollutant by the following formula (1) :
In equations (1) and (4), i = S, P, S2 _, where S represents the first pollutant, P represents the second pollutant, and S2 _represents the third pollutant, and fi ₍ ci ₎ is defined by the following equation (4): In equations (1) and (4), bi represents _the density of the decomposing microorganisms for each pollutant, αi _represents the growth rate, and ci _represents the concentration of the pollutant component in the wastewater. Ki _is the death rate per unit time due to overcrowding, Ki _is the standard microbial population density per unit area, βi is _the pollutant component concentration that gives a growth rate of half the maximum growth rate, and ni _is a power coefficient. In equation (1), A12 _and A13 _are interspecies competition coefficients, and di _is the autolysis rate.
A method for determining the number of carriers required to treat pollutants by a flowing carrier method, using the water quality simulation method according to claim 8 .