JP2007229550A - Sewage treatment plant operation support arrangement - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable a significant reduction in operation efficiency, operation cost or the like, and the water quality of inflow water to fall into a prescribed range by automatically adjusting a model parameter even when an operator having no special knowledge operates a sewage treatment plant. <P>SOLUTION: The simulation of a sewage treatment process 2 is executed by automatically adjusting the model parameter to a step simulator 8a from the model parameter set on a model parameter setting part 4, a plant condition set on a plant condition setting part 5, an operation condition set on an operation condition set part 6, data for simulation output from a database apparatus 3, data for parameter identification or the like while storing measured data output from the sewage treatment process 2, analyzed data obtained by the analysis action of a water quality analyst and others, or the like. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention is a sewage treatment plant operation support apparatus that significantly reduces the operating cost while automatically adjusting model parameters and performing on-line simulation regardless of whether there is experience or knowledge of models and processes, and improving the quality of discharged water. About.

  In conventional sewage treatment plants, organic substances have been removed mainly by a process called activated sludge process. However, in recent years, eutrophication has progressed in closed waters such as lakes and bays. Therefore, A2O (Anaerobic − The introduction of advanced sewage treatment processes such as Anoxic and Oxic is progressing.

  The introduction of advanced treatment processes will improve the quality of effluent water, while increasing operating costs. For this reason, in the sewage treatment plant, efficient operation is required so as to reduce the operation cost as much as possible while maintaining the quality of discharged water. As a tool to support this, a process simulator incorporating a biological reaction model such as an activated sludge model (ASM: Activated S1udge Model, a mathematical model published by IWA (International Water Association)) has been developed. However, the activated sludge model has a large number of model parameters of 60 or more, and the parameter values are different for each sewage treatment plant and each process and are difficult to adjust. In many cases, the required simulation accuracy cannot be obtained without parameter adjustment. At present, an engineer with specialized knowledge conducts parameter identification experiments and makes manual adjustments based on experience and knowledge. However, this operation is very complicated and requires time.

  In addition, some of the water quality components that are input to the activated sludge model are continuously measured by online sensors, but continuous measurement by online sensors is not performed for all items. There are also items that are spotted and analyzed, and items that are not measured at all. For this reason, regarding these input water qualities, there is a problem that driving support by online simulation cannot be performed unless some estimation is performed.

  As described above, a process simulator incorporating a biological reaction model such as an activated sludge model has a large number of model parameters, and this adjustment requires a very high level of expertise and time. It was.

  In the conventional example described in Patent Document 1, as shown in FIG. 10, while accumulating process data output from the sewage treatment process 101 in the data accumulating device 102, a technique such as GA (genetic algorithm) is used. The parameter setting device 103 processes the process data and optimizes the model parameters, thereby reducing the burden on the operator. However, some of the 60 or more model parameters do not change the quality of the output water even if the parameters are changed, and there is a problem that a large amount of calculation time is required for searching parameters including these.

  Further, in the conventional example described in Patent Document 2, as shown in FIGS. 11 and 12, a parameter selection processing unit 111 or the like is used, sensitivity analysis is performed, and only a highly sensitive parameter is selected as a change parameter. Yes.

  By combining the inventions of Patent Documents 1 and 2, it is possible to reduce the calculation time. However, each model parameter is a parameter that is highly sensitive to multiple water quality items, a parameter that is highly sensitive to one water quality item, anaerobic conditions, anoxic conditions, aerobic conditions, etc. Etc. It has various features. In the inventions of Patent Documents 1 and 2, such parameter characteristics are not considered, and it is considered that more efficient and realistic parameter search is possible by considering such characteristics.

  Also, since the output for certain input data is not uniquely determined, even if the output result for certain data can be predicted with high accuracy by parameter adjustment, the output result for other input data cannot always be predicted. For this reason, in the adjustment that does not consider the parameter characteristics, there is a possibility that an unrealistic solution that suits only certain input data may be derived, and there is a high possibility that this can be avoided by considering the parameter characteristics.

Moreover, when using an activated sludge model (ASM) as a water quality reaction model, it is necessary to input items that are not normally measured in a treatment plant. When performing an online simulation, it is necessary to generate time-series input data of items that are not normally measured. For water quality prediction, it is necessary to generate future time-series data for simulation. For items that are not measured, a method of inputting a constant value based on a single sample or a method of repeatedly using daily fluctuation data for a day is often used. In this method, the influence of dilution of the influent water quality concentration especially in rainy weather cannot be taken into consideration, and the simulation accuracy may be greatly affected.
JP 2000-167585 A JP 2002-102878 A

  As described above, the conventional advanced processing process has the following problems.

(1) Since parameter adjustment that greatly affects simulation accuracy needs to be performed by a person with specialized knowledge, labor costs are required.

(2) Several automatic adjustment techniques have been proposed, but take a long time.

(3) Input items for which continuous measurement is not performed may greatly affect simulation accuracy.

(4) In order to predict future water quality, it is necessary to generate future time-series data.

  In view of the above circumstances, the present invention automatically adjusts the parameters, and even when an operator who does not have specialized knowledge operates the sewage treatment plant, the operation efficiency, the operation cost, etc. can be significantly reduced, An object of the present invention is to provide a sewage treatment plant operation support device that can keep the quality of influent water within a specified range.

  In order to achieve the above object, the present invention is a sewage treatment plant operation support apparatus that performs operation support by simulating a sewage treatment process, and is measured by input operation information and sensors in the sewage treatment process. A database that accumulates water quality information and water quality analysis result information analyzed offline in time series, and data necessary for automatic adjustment of model parameters of the water quality model based on the information stored in this database, and influent water quality A data generation unit for generating data necessary for evaluation, a setting unit for setting each condition relating to the installation status and operation status of the sewage treatment process, each condition set in these setting units, and the database Based on each stored information and each data generated by the data generation unit, model parameters Quality of the sewage treatment process while executing the parameter switching in consideration of symptoms, and by calculating the operating costs, is characterized by comprising a process simulator to optimize each model parameter.

  According to the present invention, even when an operator who does not have expert knowledge operates a sewage treatment plant by automatically adjusting parameters, it is possible to greatly reduce the operation efficiency, operation cost, etc. The water quality can be kept within the specified range.

<< First Embodiment >>
FIG. 1 is a block diagram showing a first embodiment of a sewage treatment plant operation support apparatus according to the present invention.

<Configuration and operation>
The sewage treatment plant operation support apparatus 1 shown in this figure accumulates measurement data output from the sewage treatment process 2, analysis data input by a water quality analyst, etc. in a time series, while simulating data and parameter identification data. And the like. In addition, the model parameter setting unit 4 in which the model parameters of the sewage treatment process 2 are set, the plant condition setting unit 5 in which the plant conditions of the sewage treatment process 2 are set, and the operating conditions when operating the sewage treatment process 2 are An operating condition setting unit 6 to be set, and an automatic adjustment command setting unit 7 for outputting an automatic adjustment command for model parameters according to the operation contents of the operator are provided. Furthermore, when a model parameter automatic adjustment command is output from the automatic adjustment command setting unit 7, the model parameter set in the model parameter setting unit 4 of the sewage treatment process 2 is set in the plant condition setting unit 5. Based on the plant conditions, the operating conditions set in the operating condition setting unit 6, the simulation data output from the database device 3, the parameter identification data, etc., the model parameters are automatically adjusted, and the operation content of the sewage treatment process 2 A process simulator 8 for simulating the above is provided.

  The sewage treatment plant operation support apparatus 1 accumulates measurement data output from the sewage treatment process 2, analysis data obtained by an analysis operation of a water quality analyst, and the like in the database device 3, so that simulation data, parameters Generate identification data. Further, when the adjustment start condition set by the operator is satisfied and the automatic adjustment command is output from the automatic adjustment command setting unit 7, the model parameter set in the model parameter setting unit 4 and the plant condition setting unit 5 are set. Based on the plant conditions being set, the operating conditions set in the operating condition setting unit 6, the simulation data output from the database device 3, the parameter identification data, etc., the process simulator 8 automatically adjusts the model parameters, The simulation of the sewage treatment process 2 is performed, and the water quality prediction result and the operation cost result are presented to the operator.

  The database device 3 includes a database unit 3A, a simulation data generation unit 9, and a parameter identification data generation unit. In the database section 3A, offline data and online data collected by a service company that partially manages each sewage treatment plant, and sensor information obtained from sensors installed in the sewage treatment process 2 (inflow rate, return flow) Operational data such as flow rate, sludge extraction amount, aeration air flow rate, circulation flow rate, etc., water temperature, MLSS, DO, pH, UV, ammonia nitrogen, phosphoric acid phosphorus, etc.) and water quality analysts at sewage treatment plants Information on the results of water quality analysis (COD, BOD, TOC, ammonia nitrogen, nitrate nitrogen, phosphoric acid phosphorus, etc.) is taken in and stored in time series.

  Then, based on the data stored in the database unit 3A, the simulation data generation unit 9 generates simulation input data and supplies it to the process simulator 8. In addition, the parameter identification data generation unit 10 generates input data for automatic parameter adjustment, and evaluation target water quality data in the reaction tank or inflow water that is an evaluation target of the automatic parameter adjustment result, and these are generated as a process simulator 8. To supply.

  The model parameter setting unit 4 is a part set by the operator of the sewage treatment plant, and sets model parameters of the water quality model stored in the process simulator 8. There is ASM2d (Activated S1udge Model No.2d) published by lWA (International Hydrological Society) as the most commonly used water quality model in Japan. This model has 64 model parameters. If an identification experiment is performed in advance and the value is known, the operator of the sewage treatment plant inputs the value to the model parameter setting unit 4. The model parameters are input to the model parameter storage unit 12 in the process simulator 8 and are excluded from the model parameters to be automatically adjusted.

  The plant condition setting unit 5 is a part set by the operator of the sewage treatment plant. The process flow of the sewage treatment process 2 (A2O process, AO process, circulation nitrification denitrification process, etc.), each reaction tank, and the volume of the sedimentation site Civil conditions such as, and equipment conditions of equipment such as pumps and blowers in the sewage treatment process 2 are set. By changing the conditions input to the plant condition setting unit 5, it is possible to simulate water quality and cost when the facility conditions of the sewage treatment plant 2 are changed. When the equipment condition of the sewage treatment plant 2 is not changed, it is only set first, and it is not necessary to set after that.

  The operation condition setting unit 6 is a part set by an operator of the sewage treatment plant, and sets operation conditions such as a return amount, a circulation amount, a sludge extraction amount, an aeration air amount, and a reaction tank DO (dissolved oxygen) concentration. By setting various operating conditions and calculating the water quality value and the cost value in the process simulator 8, it is possible to determine what kind of operation can be performed at low cost and good water quality. .

  The automatic adjustment command setting unit 7 is a part set by the operator of the sewage treatment plant. When the start condition of the automatic adjustment of the model parameter is set by the operator and the start condition is satisfied, the automatic adjustment command is output. Then, automatic adjustment of model parameters is started in the parameter automatic adjustment unit 11 of the process simulator 8.

  The process simulator 8 includes an automatic parameter adjustment unit 11 and a model parameter storage unit 12. When an automatic adjustment command is output from the automatic adjustment command setting unit 7, the parameter automatic adjustment unit 11 is operated to change model parameters. While making adjustments and updating the model parameters stored in the model parameter storage unit 12, a process simulation is performed using the model parameters stored in the model parameter storage unit 12, and the water quality prediction result, operation cost is given to the operator. Present prediction results.

  In this case, the parameter automatic adjustment unit 11 includes a parameter value set in the model parameter setting unit 4, a process flow, civil engineering conditions, and equipment conditions set in the plant condition setting unit 5, and a parameter identification generation unit 10. The generated parameter automatic adjustment input data and parameter adjustment result evaluation data are taken in, and under the plant conditions set in the plant condition setting section 5, the operating conditions, the inflow water condition of the parameter automatic adjustment input data, etc. Perform a simulation with input conditions. Then, an error evaluation between the water quality value of the simulation result and the water quality value of the evaluation data generated by the parameter identification data generation unit 10 is performed. If the error is large, the model parameter is automatically adjusted, and the water quality of the simulation result is Change the value so that it approaches the water quality value of the evaluation data. Thereafter, when the error between the water quality value of the simulation result and the water quality value of the evaluation data becomes equal to or less than a preset error determination value, the parameter group obtained by the parameter automatic adjustment unit 11 is stored in the model parameter storage unit 12. Supply.

  The model parameter storage unit 12 stores the parameter value input to the model parameter setting unit 4. When the water quality model in the process simulator 8 is ASM2d, there are 64 model parameters, and default values are stored as initial values for data other than those set by the model parameter setting unit 4. Each time a parameter group is output from the model parameter automatic adjustment unit 11, it is captured and stored.

<effect>
(1) Since the model parameters are automatically adjusted, the operator does not have to spend effort on parameter adjustments, and the water quality simulation accuracy of the process simulator is improved, so that more practical operation support for the treatment plant is possible. .

(2) Since data for simulation input and data for parameter identification are generated using data stored on the database device 3, the operator only sets the plant conditions and operating conditions without performing new work. It is possible to carry out simulation.

(3) Optimize model parameters by generating parameter identification data corresponding to major items such as sludge concentration, ammonia nitrogen concentration, nitrate nitrogen concentration, and phosphoric acid phosphorus concentration, and evaluating the contents of the simulation Is possible.

(4) It is possible to promote effective utilization of data by accumulating offline data and online data collected by a service company in the database device 3 and generating simulation input data and parameter identification data. .

<Modification of First Embodiment>
(1) The process model incorporated in the process simulator 8 is not limited to ASM2d, and may be any model as long as it is a model describing a biological reaction.

(2) The process simulator 8 does not calculate the water quality and the operation cost, but may calculate only the water quality.

(3) The automatic adjustment command setting unit 7 may be set so that an automatic adjustment command is periodically issued. For example, it may be set such that automatic adjustment is performed at 10:00 on the 10th of every month.

<< Second Embodiment >>
Next, a second embodiment of the sewage treatment plant operation support apparatus according to the present invention will be described. Since the basic configuration is the same as that of the first embodiment, description will be made with reference to FIG.

<Configuration and operation>
The parameter automatic adjustment unit 11 of the sewage treatment plant operation support apparatus 1 according to the second embodiment is configured to automatically adjust model parameters in consideration of parameter characteristics such as sensitivity of model parameters. In the second embodiment, parameters are adjusted in a manner almost the same as the manual adjustment of model parameters based on knowledge of models and knowledge of processes, and parameters for adjusting all water qualities at once. A more realistic and practical parameter value is obtained than the adjustment method.

  As shown in FIG. 2, the automatic parameter adjustment unit 11 includes a simulation calculation processing unit 13 that performs simulation calculation using plant conditions, simulation input data, model parameters, and the like when an automatic adjustment command is input. A simulation result storage unit 14 that stores the results, an individual simulation error determination unit 15 that compares the actual data and the simulation results to determine the error, and a check unit 16 that checks whether the model parameters have been optimized, Based on the error information, a change parameter group selection unit 17 that selects a change parameter, an adjustment target water quality, and the like, a parameter optimization unit 18 that optimizes a specified model parameter, and a parameter setting change unit that changes the setting of the model parameter 19 and

  When the automatic adjustment command is output from the automatic adjustment command setting unit 7, the plant condition setting unit 5 is switched while changing the model parameter to be changed based on the degree of influence (sensitivity) and the like within a range satisfying the constraint conditions. Process simulation is performed using the plant conditions set in the above, and the model parameters stored in the model parameter storage unit 12 are optimized.

  The simulation calculation unit 13 takes in the simulation input data (inflow water quality data and operation data) generated by the simulation data generation unit 9 under the plant conditions set in the plant condition setting unit 5 as input conditions, and model parameters. A simulation is performed using the model parameters stored in the storage unit 12. The simulation calculation is performed until each water quality value converges, and the convergence value is stored in the simulation result storage unit 14.

The simulation error determination unit 15 calculates an error between the water quality value stored in the simulation result storage unit 14 and the evaluation target water quality data generated by the parameter identification data creation unit 10. When the water quality model is ASM2d, 19 water quality components are calculated. However, it can be said that not all of these water quality components are measured. In particular, organic components are classified into many categories such as readily degradable, slow degradable, non-degradable soluble, non-degradable floating, but there is no sewage treatment plant that classifies and analyzes components in this way. Therefore, only the component measured offline or online or the component that can be estimated from the measured value is generated by the parameter identification data creation unit 10 and becomes the evaluation target data. A simulation error E _{k, l} between the simulation result of the water quality component to be evaluated and the water quality data to be evaluated is calculated by the equation (1).

[Equation 1]
_{Ek, l} = | _{Yk, l-} Xk _{, l} | (k = 1,..., M, l = 1,..., N) (1)
However, E _{k, l} : Simulation error at point 1 of water quality component k Y _{k, l} : Simulation value at point 1 of water quality component k X _{k, l} : Water quality value to be evaluated at point 1 of water quality component k m: Evaluation target Number of water qualities n: Number of measurement points The check unit 16 determines that a good model parameter has been obtained when all evaluation target water quality components are within a preset error determination value, and the model parameter group. Is stored in the model parameter storage unit 12 in the process simulator 8, and when the simulation error of any one or more water quality components is larger than a preset error judgment value, the error information of each water quality component is changed by the parameter selection unit. 17

  The change parameter selection unit 17 is set with matrix information in which model parameters and respective water quality components are arranged in a matrix, and parameter information describing constraint conditions such as upper and lower limit values of each model parameter.

An example of matrix information is shown in Table 1.

For the sake of simplicity, this matrix is described with respect to six model parameters out of 64 model parameters. However, this matrix is actually a matrix having a higher order. The process flow is an A2O (anaerobic / anoxic / aerobic) process, and the water quality to be evaluated is phosphoric acid concentration S _{po4, oxic} , ammonia concentration S _{NH4, oxic} , nitric acid concentration S in the aerobic tank which is the last stage of the biological reaction tank. _{N03, oxic} . When the model parameter is moved to the 10th side, the arrow on the matrix indicates that the water quality to be evaluated moves to the 10th side, and the arrow that moves to the 1st side indicates that the water quality to be evaluated moves to the 10th side. The thickness of the arrow is described according to the amount of change in the water quality to be evaluated when the model parameter is changed, and the thicker the amount of change is thicker. In addition, the case where the amount of change is almost negligible is shown as blank. Model parameters that have little effect on all water quality are all blank on the matrix and can be removed from the matrix. Moreover, the colored part on the matrix is the water quality to be adjusted for each model parameter. Each model parameter has its own meaning. For example, μ _AUT is a model parameter related to the growth rate of nitrifying bacteria that converts ammonia into nitric acid, and b _AUT is a model parameter related to the killing rate of nitrifying bacteria.

That is, these μ _AUT and b _AUT are model parameters suitable for adjusting ammonia and nitric acid, but as is apparent from Table 1, by adjusting these, the phosphorus concentration is changed. Since there is a model parameter such as η _N03 shown in Table 1 for adjusting the denitrification reaction related to nitrogen gas from nitric acid, these μ _AUT and b _AUT are generally used to adjust the ammonia concentration. It is used. For this reason, the adjustment target water quality of these model parameters is ammonia concentration _{SNH4, oxic} . By describing such know-how information about model parameters in the matrix, more realistic and practical parameter adjustment can be performed.

  The degree of influence on the adjustment target water quality written on the rightmost side of the matrix shown in Table 1 indicates the degree of influence of each model parameter on the adjustment target water quality. For example, the change parameter is changed by 10%. It is expressed by the rate of change of the water quality subject to adjustment.

Then, the change parameter group selection unit 17 selects, as the change parameter, the parameter having the greatest influence among all the model parameters having the water quality having a large simulation error in the equation (1) as the adjustment target water quality. In the example of Table 1, it is assumed that the simulation error is larger than the evaluation judgment value for all of the evaluation target water quality phosphoric acid concentration S _{po4, oxic} , ammonia concentration S _{NH4, oxic} , and nitric acid concentration S _{N03, oxic} . In this case, μ _AUT having the greatest influence is selected as the adjustment parameter, and ammonia S _{NH4, oxic} is selected as the adjustment target water quality. Then, the selected adjustment parameter, the information on the adjustment target water quality, and the constraint information on the upper and lower limit values of the model parameter set in the parameter information are sent to the parameter optimization unit 18.

When it is determined by the check unit 16 that the parameter optimization process has not been completed, the parameter optimization unit 18 performs sequential quadratic programming, GA (genetic algorithm), tabu search, SA (simulated). The adjustment parameter selected by the change parameter group selection unit 17, the information on the quality of the adjustment target water, and the constraint information on the upper and lower limit values of the model parameter are captured by a metaheuristic numerical optimization method such as annealing), and the parameter setting change unit 19 is The simulation calculation unit 13, the simulation result storage unit 14, the individual simulation error determination unit 15, and the check unit 16 are operated while starting and adjusting the parameter values, and within the constraint conditions, the water quality to be adjusted, for example, S _{NH4 Oxic} simulation error is minimized The model parameters are optimized by repeating the simulation calculation, storing the simulation result of the convergence value, and evaluating the simulation error.

When the simulation error of _{SNH4, oxic} , which is the adjustment target water quality, is minimized, the error evaluation of each evaluation target water quality is performed again.

If at this _{point, S N03, oxic, S po4} , oxic is in error decision value _{out, S NH4,} long simulation error of _Oxic the following evaluation criterion value, _{S N03,} oxic be evaluated _{quality, S PO4,} Y _po4 having the greatest influence on _oxic is selected as the adjustment parameter. On the other hand, if the simulation error of _{SNH4 and oxic} is also larger than the evaluation judgment value, b _AUT becomes the target parameter, but the parameter information includes a constraint that the growth rate and death rate do not move simultaneously as a constraint condition of the model parameter. In this case, b _AUT is not an adjustment parameter. Therefore, in this case, _Ypo4 having the greatest influence on the evaluation target water quality other than μ _AUT and b _AUT is selected as the adjustment parameter, and _{SPO4 , oxic,} which is the adjustment target water quality of _Ypo4 , is selected as the adjustment target water quality.

  In addition, after the first optimization is completed, the simulation error is determined, the degree of influence is compared with the evaluation target water quality whose error is larger than the determination value, and the one with the highest degree of influence is selected as the change parameter. The water quality to be adjusted for the model parameter is selected, and the model parameter is optimized under the constraint conditions set in the parameter information.

  Even when a simulation error falls within an error determination value for a certain evaluation target water quality, moving the model parameter for another evaluation target water quality may deviate from the error determination value. In such a case, the model parameter once moved is again a candidate for the adjustment parameter.

  Through such simulation processing, error determination processing, change parameter group selection processing, parameter optimization processing, etc., when the error falls below the determination value for all evaluation target water qualities, or affects the evaluation target water quality When there are no more changed parameters, or when the parameter selection / optimization loop exceeds a preset number of times, the calculation is terminated, and the adjusted parameter value is input to the model parameter storage unit 12 of the process simulator 8.

<effect>
(1) By setting the water quality to be adjusted for each model parameter and making automatic adjustments based on that information, it is almost the same as an expert manually adjusting model parameters based on model knowledge and process knowledge. The model parameters can be adjusted with the accuracy of This makes it possible to derive more realistic and practical parameter values than the parameter adjustment method that collectively adjusts all the water qualities of the conventional invention.

(2) By adjusting the model parameters in descending order of influence, automatic adjustment can be performed more efficiently.

(3) Since the water quality to be adjusted is one, if the simulation error is a positive error or a negative error, it will be decided whether to move the model parameter in the plus or minus direction from the current value. It can be squeezed efficiently.

<Modification of Second Embodiment>
(1) The number of model parameters selected by the change parameter selection unit 17 is not one, and a plurality of model parameters may be used as long as they are adjusted for the same adjustment target water quality. For example, although in the above example is Y _PO4 and q _pHA as a model parameter for the adjusted water _{S po4, oxic,} which select simultaneously as change parameters, even those optimized by the optimization technique such as GA good.

(2) The information stored as the matrix information is not limited to the format shown in Table 1, but represents the change direction of each water quality when each model parameter is changed and the magnitude of the change of each model parameter with respect to the evaluation target water quality. Anything is acceptable.

(3) The water quality to be evaluated is not limited to that shown in Table 1, but any water quality component calculated by a water quality model that can be measured or estimated online or offline can be obtained. The water quality may be the target water quality.

(4) The determination equation for the simulation error is not limited to the equation (1), and may be determined based on, for example, a value obtained by squaring the simulation error.

(5) The adjustment target water quality for each model parameter is not necessarily one. For example, in the matrix of Table 1, the adjustment target water quality of μ _AUT is S _{NH4, oxic} , but it also affects S _{N03, oxic} , so S _{NH4, oxic} and S _{N03, oxic} are the adjustment target water quality, It may be one that minimizes the simulation error expressed by equation (2).

[Equation 2]
E _μAUT = w1 · E _{NH4, oxic} + w2 · E _{N03, oxic} (2)
E _μAUT : Simulation error of water quality to be adjusted with respect to parameter μAUT E _{NH4, oxic} : Simulation error of ammonia concentration in aerobic tank E _{N03, oxic} : Simulation error of nitric acid concentration in aerobic tank w1, w2: Weight coefficient (6 ) The method of evaluating the degree of influence does not indicate the rate of change of the water quality subject to adjustment when the model parameter is moved by 10%, but evaluates the rate of change by varying what percentage the model parameter can take. It may be. Moreover, even if it is not expressed by the rate of change, the degree of influence may be indicated by the actual fluctuation range of water quality.

(7) The matrix information describing the relationship between the parameters of the activated sludge model and the water quality is created from the sensitivity analysis results that have been implemented in advance. Since the sensitivity of the parameters of the activated sludge model does not depend so much on the treatment process, the matrix information was obtained from the sensitivity analysis results at other treatment sites, even though it was not necessarily created from the sensitivity analysis results at that treatment site. The matrix information may be used as it is.

<< Third Embodiment >>
Next, a third embodiment of the sewage treatment plant operation support apparatus according to the present invention will be described. Note that the basic configuration of the third embodiment is the same as that of the first embodiment, and will be described with reference to FIG.

<Configuration and operation>
The parameter automatic adjustment unit 11 according to the third embodiment automatically adjusts the model parameter in consideration of not only the sensitivity of the model parameter but also the degree of interference. In this case, as matrix information set in the changed parameter group selection unit 17 shown in FIG. 2, as shown in Table 2, not only the aerobic tank but also the anaerobic tank, an anaerobic tank, the evaluation target water quality, and the evaluation target The one describing the degree of interference with water quality is used. And, it is more realistic than the parameter adjustment method that adjusts all the water qualities by adjusting the parameters almost equivalent to manual adjustment of model parameters based on knowledge about the model and knowledge about the process. This is to obtain a practical parameter value.

Looking at the matrix in Table 2, mu _AUT, and that a plurality of water quality than adjusted water thus varied as b _AUT, such as changing only one adjustment target water quality in the as q _pHA characteristics I understand that there is. In such a case, it is more efficient to adjust what affects more water quality components first and to adjust what affects only one water quality later. The degree of interference in Table 2 indicates how many water quality other than the adjustment target water quality have an effect. The model parameter having the highest degree of interference is selected as the adjustment parameter from among all the model parameters having the water quality with the simulation error of the evaluation target water quality larger than the error determination value in the equation (1) as the adjustment target water quality.

When all the simulation errors are larger than the error determination value for the water quality to be evaluated, the change parameter candidates are μ _AUT and b _AUT having the highest degree of interference. Among these, μ _AUT having the highest influence on the adjustment target water quality is the first change parameter. After the μ _AUT is optimized, the constraint that the growth rate and the death rate do not move simultaneously is set in the parameter information, and b _AUT cannot be used as a change parameter. Therefore, after μ _AUT and b _AUT , K _NH4 having a high degree of interference is selected as the second change parameter.

  Thus, in the third embodiment, a model parameter having a higher degree of interference than a model parameter having a higher degree of influence on the evaluation target water quality is selected.

<effect>
(1) Since the possibility of adjusting the model parameter once adjusted is reduced by sequentially adjusting from the model parameter having a high interference degree, the calculation is performed more efficiently than the adjustment method considering only the influence degree. it can. For this reason, calculation time can be shortened.

(2) By using data for anaerobic conditions, anaerobic conditions, and aerobic conditions, practical parameter automatic adjustment closer to reality can be performed, and simulation accuracy can be further improved.

<Modification of Third Embodiment>
(1) The number of model parameters selected by the change parameter selection unit 17 is not one, and a plurality of model parameters may be used as long as they are adjusted for the same adjustment target water quality.

(2) The information written in the matrix information is not limited to the format shown in Table 2, but may represent the direction of change of each water quality when each model parameter is changed and the magnitude of change of each model parameter with respect to the evaluation target water quality. Anything is acceptable.

(3) The water quality to be evaluated is not limited to the one shown in Table 2, but the water quality components calculated in the water quality model can be actually measured online or offline, or can be estimated. Any water quality may be used as an evaluation target.

(4) The determination equation for the simulation error is not limited to the equation (1), and may be determined based on, for example, a value obtained by squaring the simulation error.

(5) The adjustment target water quality for each model parameter is not necessarily one. For example, in the matrix of Table 2, the adjustment target water quality of _Ypo4 is set to _{Sp04 , anaerobic} , but it also affects _{Sp04, oxic} , so _{Sp04, anaerobic} and _{Sp04, oxic} are _set as the adjustment target water quality, ( It may be one that minimizes the simulation error represented by equation (3).

[Equation 3]
E _Yp04 = w1 · E _{p04, anaerobic} + w2 · E _{p04, oxic} ... (3)
However, E _Yp04 : Simulation error of water quality to be adjusted with respect to parameter Y _p04 E _{p04, anaerobic} : Simulation error of phosphate concentration in anaerobic tank E _{p04, oxic} : Simulation error of phosphate concentration in aerobic tank w1, w2: Weighting factor (6) The degree of interference is not necessarily the number of waters that interfere, but may be a numerical value in some form in consideration of the magnitude of interference.

<< 4th Embodiment >>
Next, a fourth embodiment of the sewage treatment plant operation support apparatus according to the present invention will be described. Note that the basic configuration of the fourth embodiment is the same as that of the first embodiment, and will be described with reference to FIG.

<Configuration and operation>
The parameter automatic adjustment unit 11 of the sewage treatment plant operation support apparatus 1 according to the fourth embodiment classifies model parameters for each water quality to be evaluated in consideration of parameter characteristics such as sensitivity of model parameters. The model parameters are automatically adjusted sequentially. In the fourth embodiment, as in the second embodiment, parameter adjustment is performed in the same manner as an expert manually adjusts model parameters based on model knowledge and process knowledge, and all water quality is collected at once. Thus, a more realistic and practical parameter value is obtained than the parameter adjustment method to be adjusted. The model parameter group information stored in the changed parameter group selection unit 17 in FIG. 3 is classified in advance according to the sensitivity and parameter characteristics of the model parameters for each evaluation target water quality. For example, for simplicity, only the six parameters shown in Tables 1 and 2 are described. Μ _AUT , b _AUT , K _NH4 is the first group, η _NO3 is the second group, Y _PO4 , q _PHA is the third group. Grouped as a group. The first group is a parameter group for adjusting ammonia (S _NH4 ), the second group is for adjusting nitric acid (S _NO3 ), and the third group is a parameter group for adjusting phosphoric acid (S _PO4 ). If the simulation error expressed by the above equation (1) is greater than or equal to the evaluation determination value for ammonia, the second group is greater than or equal to the evaluation determination value for nitric acid. Three groups of parameters are selected and their parameters are optimized. If the simulation error is greater than or equal to the evaluation judgment value for all target water qualities, first optimize the parameters of the first group that affect many components, and then optimize the parameters of the second and third groups. Therefore, efficient adjustment can be performed.

<Effect>
(1) By setting the water quality to be adjusted for each model parameter, grouping them in advance, and automatically adjusting based on that information, the expert can manually adjust the model parameters based on the knowledge about the model and the knowledge about the process. The model parameters can be adjusted with almost the same accuracy as the adjustment. This makes it possible to derive more realistic and practical parameter values than the parameter adjustment method that collectively adjusts all the water qualities of the conventional invention.

(2) Efficient adjustment can be performed by adjusting a parameter group that affects more parameters first.

<Modification of Fourth Embodiment>
(1) The parameter classification method is not limited to 3 groups. For example, an adjustment parameter group is registered in advance for each of anaerobic tank, anoxic tank, and aerobic tank for each of ammonia, nitric acid, and phosphoric acid, and 3 × 3 9 It may be composed of a parameter group of groups, and may be anything as long as it is composed of two or more groups.

<< 5th Embodiment >>
FIG. 4 is a block diagram showing a fifth embodiment of a sewage treatment plant operation support apparatus according to the present invention. In this figure, parts corresponding to those in FIG. 1 are denoted by the same reference numerals.

<Configuration and operation>
The sewage treatment plant operation support device 51 shown in this figure is different from the sewage treatment plant operation support device 1 shown in FIG. 1 in that a simulation accuracy determination unit 20 is provided, and a simulation error of the process simulator 8 is set in advance. When this is the case, this is detected by the simulation accuracy determination unit 20 so that the process simulator 8 automatically adjusts the model parameters.

The simulation accuracy determination unit 20 receives the actual value of the evaluation target water quality from the database device 3. In addition, a simulation value of the relevant water quality is input from the process simulator 8. An error between the actual value of the evaluation target water quality and the simulation value is evaluated by, for example, the equation (4).

ΔE: Weighted mean square error between actual value and simulation value X _kl : Actual value of evaluation target water quality k at point l at time t Y _kl : Simulation of evaluation target water quality k at point l at time t Value W _kl : Weighting factor of evaluation target water quality k at point l m: Number of evaluation target water qualities n: Number of measurement points And, if the value of equation (4) is greater than or equal to a preset error value, simulation accuracy The determination unit 20 transmits an automatic adjustment command to the parameter automatic adjustment unit 11 in the process simulator 8. The parameter automatic adjustment unit 11 is an algorithm used in any of the first to fourth embodiments described above, recalculates model parameters, and adds new model parameters to the model parameter storage unit 12 in the process simulator 8. Set.

<effect>
(1) With such a configuration, the simulation accuracy can always be kept good.

<Modification of Fifth Embodiment>
(1) The simulation accuracy determination unit 20 may display a message on the monitoring screen for determining whether or not to issue an automatic adjustment command to the operator, instead of directly issuing the automatic adjustment command.

<< 6th Embodiment >>
Next, a sixth embodiment of the sewage treatment plant operation support apparatus according to the present invention will be described. Since the basic configuration is the same as that of the first embodiment, description will be made with reference to FIG.

<Configuration and operation>
The simulation data generation unit 9 of the sewage treatment plant operation support apparatus 1 in the sixth embodiment uses the actual measurement data input online to generate simulation input data of water quality items that are not measured online. I have to. Thereby, the process simulator 8 is caused to perform a simulation with higher accuracy.

The simulation data generation unit 9 measures the measurement data output from each sensor provided in the sewage treatment process 2, for example, the inflow water output from the total nitrogen meter 22 provided in the settling tank 21 as shown in FIG. The total nitrogen measurement value is taken in, and in the procedure described below, the simulation input data of the water quality item that is not actually measured, for example, the time series data of the readily decomposable organic substance concentration _SF (t) which is one of the model input components create.

First, in order to perform a simulation with the process simulator 8 incorporating ASM2d, the organic matter concentration of the influent water is required, but there is almost no case where the organic matter concentration of the influent water is measured online. The organic matter concentration of the influent water is analyzed once a day by the water quality analyst at the treatment plant for COD _Mn (chemical oxygen demand by permanganic acid) and BOD (biochemical oxygen demand), which are organic matter indicators. Degree. In addition, a day-to-day test is carried out once a month to several times a year, and this day-to-day test is conducted every 1 hour to several hours to conduct an analytical test.

In such a situation, the COD _Mn value of the influent water analyzed by the day-to-day test is stored on the database device 3 in the format of Table 3, for example. When the inflow water of a sewage treatment plant is in fine weather, it depends on the activities and business activities of the people living in the area, so it can be said that it shows almost the same daily fluctuation pattern although there are seasonal fluctuations. Data between non-continuous data in Table 3 is obtained by linear interpolation processing, and a value obtained by dividing by the average value of COD _Mn is defined as a daily variation pattern function FCOD _Mn (t) of CODMn. This variation pattern function shows the daily variation with respect to the average value. That is, the function describes the coefficient of variation.

  On the other hand, in the sewage treatment process 2 shown in FIG. 5, the total nitrogen concentration is measured online. In a sewage treatment plant connected to a combined sewer that treats both rainwater and sewage, the total nitrogen concentration decreases due to rainwater in rainy weather. Even in sewage treatment plants connected to a diversion sewer system that treats only sewage, the total nitrogen concentration of the inflow water is thin in most sewage treatment plants (it is considered to be caused by rainwater entering). When the total nitrogen concentration is thus reduced, it is considered that not only the total nitrogen concentration but also the organic matter concentration not measured is diluted in the same manner.

Assuming that the total nitrogen concentration and COD _Mn are diluted with the same tendency, as shown in the equation (5), the total nitrogen concentration meter 22 is added to the coefficient of variation described in the above-mentioned daily fluctuation pattern function FCOD _Mn (t). This dilution effect can be taken into account by taking the form multiplied by the measured value of.

[Equation 5]
COD _Mn (t) = COD _average · FCOD _Mn (t) · TN (t) (5)
However, COD _Mn (t): COD _Mn estimated value data at time t COD _average : COD _Mn average data during day-to-day test FCOD _Mn (t): COD _Mn variation pattern function TN (t): at time t Total nitrogen meter measurement value Moreover, since the readily decomposable organic matter input to ASM2d incorporated in the process simulator 8 is indicated by CODcr (oxygen demand by potassium dichromate), COD _Mn as shown in the equation (6) Is obtained by multiplying the CODcr conversion coefficient b _COD by the readily decomposable organic matter ratio a _SF in the organic matter. These conversion coefficients are best if data can be obtained at this sewage treatment plant, but if the data cannot be obtained, values can be determined from existing literature values.

[Equation 6]
S _F (t) = a _SF · b _COD · COD _Mn (t) (6)
_However, S F (t): easily decomposable organic matter at time t Concentration a _SF: easily decomposable organic matter ratio of organic matter b _COD: FCOD _Mn → transform coefficients _{b COD} be converted to CODcr
COD _Mn (t): COD _Mn estimated value data at time t <Effect>
(1) A simulation with higher accuracy can be performed by creating simulation input data of unmeasured water quality items in consideration of dilution effects from other online data.

(2) As shown in the equation (6), the conversion factor b _COD for converting COD _Mn to CODcr is multiplied by the readily decomposable organic matter ratio a _SF in the organic matter, and the easily decomposable organic matter required for ASM2d is CODcr can be obtained, and the time and load required for calculation can be reduced.

<Modification of Sixth Embodiment>
(1) The generated data is not limited to S _F (t), and any input data may be used.

(2) The on-line sensor to be used is not limited to the total nitrogen meter 22 but may be any device that measures the water quality concentration, such as a total phosphorus meter, a turbidity meter, and an SS meter.

(3) The estimation formula is not limited to the formula (5), and any formula may be used as long as it describes the relationship between the target data and the online data.

<< 7th Embodiment >>
Next, a seventh embodiment of the sewage treatment plant operation support apparatus according to the present invention will be described. Since the basic configuration is the same as that of the first embodiment, description will be made with reference to FIG.

<Configuration and operation>
The simulation data generation unit 9 of the sewage treatment plant operation support apparatus 1 according to the seventh embodiment includes a total nitrogen measurement value of inflow water and a water quality test result that is performed about once a day in a daily test, for example, a water quality test of COD _Mn . By using the results and creating time-series data of the readily decomposable organic substance concentration S _F (t), which is one of the input components of the model, a more accurate simulation can be performed.

The simulation data generation unit 9 measures the measurement data output from each sensor provided in the sewage treatment process 2, for example, inflow water output from the total nitrogen meter 22 provided in the settling tank 21 as shown in FIG. Incorporating the total nitrogen measurement value and COD _Mn water quality test results analyzed in daily water quality tests, using equations (8) to (10) to calculate COD _Mn (t) and actually measuring Non-simulated input data, for example, time-series data of the readily decomposable organic substance concentration S _F (t) that is one of the input components of the model is created.

[Equation 7]
COD _Mn (t) = COD _Mnaverage · FCOD _Mn (t) · TN (t) · k (T) (8)
COD _Mn (T) = COD _Mnaverage · FCOD _Mn (T) · TN (T) · k (T−ΔT) (9)
k (T) = COD _Mnday (T) / COD _Mn (T) (10)
However, COD _Mn (t): COD _Mn estimated value data at time t COD _Mnverage : COD _Mn average data during day-to-day test COD _Mnday (T): COD _Mn data of daily water quality test at time T COD _Mn (T): COD _Mn estimated value data at time T COD _Mnverage : COD _Mn average data during day-to-day test FCOD _Mn (t): Daily variation pattern function of COD _Mn TN (t): Total nitrogen meter measured value at time t k ( T): Correction coefficient k (T−ΔT): Correction coefficient before one step At this time, the latest COD _Mn water quality test result is input to the COD _Mnday (T) by a water quality analyst. However, since the water quality test has several hours from the sampling to the result being _obtained , both COD _Mndday (T) and COD _Mn (T) are past values.

Also, before COD _Mndday (T) is input to the database, COD _Mn (t) is calculated using k (T) of the previous step, and when COD ^Mndday (T) is input to the database, ^kD (T) is calculated by equations (9) and (10), updated, and input to k (T) in equation (8).

Then, a (8) _COD is calculated by the formula Mn (t), (6) formula and is used, which is one of the simulation input data _{S F} (t) is computed.

<effect>
(1) Create simulation input data for unmeasured water quality items from other online data, and then correct the data based on the data measured on that day to reflect the current water quality status from the dilution information of other water quality items At the same time, the water quality of the target water quality of the day can be reflected, so simulation with higher accuracy becomes possible.

<Modification of the seventh embodiment>
(1) The estimation formula is not limited to the formulas (8) to (10), and any formula can be used as long as it describes the relationship between the target data and the online data and uses the water quality test result of the target data for the day. Such a formula may be used.

<< Eighth Embodiment >>
Next, an eighth embodiment of the sewage treatment plant operation support apparatus according to the present invention will be described. Since the basic configuration is the same as that of the first embodiment, description will be made with reference to FIG.

<Configuration and operation>
As shown in FIG. 6, the simulation data generation unit 9 of the sewage treatment plant operation support apparatus 1 in the eighth embodiment includes an inflow amount prediction unit 23, a similar inflow pattern selection unit 24, and a future simulation data generation unit 25. Based on current weather forecast information, current pump well water level information, past weather forecast information, past pump well water level information, etc., current inflow water quality information will be used as future simulation data. This means that data was created so that future simulations could be performed.

  The inflow prediction unit 23 predicts a future sewage inflow based on past weather prediction information, past inflow, past pump well water level information, current weather prediction information, and current pump well water level information. Thus, the future inflow data and the inflow prediction data are calculated and sent to the similar inflow pattern selection unit 24.

  The similar inflow pattern selection unit 24 searches the past inflow data stored in the database device 3 using the future inflow data and the inflow prediction data, and is most similar to the future inflow data and the inflow prediction data. In addition to selecting the past inflow data, the inflow water quality information during the period corresponding to the selected inflow data is extracted. Then, the inflow water quality time series data included in the inflow water quality information is divided by the average inflow water quality during the prediction period to create an inflow water quality variation pattern function describing the variation coefficient of the inflow water quality, and the future simulation data generation unit 25 Send to.

  The future simulation data generation unit 25 takes in the future inflow data output from the inflow prediction unit 23 and sends it to the process simulator 8a, and applies it to the inflow water quality variation pattern function output from the similar inflow pattern selection unit 24. By multiplying the current measured value or the current estimated value of the water quality, future influent water quality data is created and sent to the process simulator 8a.

<effect>
(1) Since future simulation data is automatically generated, future simulation becomes possible, and driving methods can be examined.

<< Ninth Embodiment >>
Next, a ninth embodiment of the sewage treatment plant operation support apparatus according to the present invention will be described. Since the basic configuration is the same as that of the first embodiment, description will be made with reference to FIG.

<Configuration and operation>
The simulation data generation unit 9 of the sewage treatment plant operation support apparatus 1 according to the ninth embodiment includes a pattern function generation unit 26 and a future simulation data generation unit 27, as shown in FIG. Based on the current inflow water quality information, etc., inflow water quality time-series data, etc., which will be the data for future simulation, are created so that future simulations can be performed.

  The pattern function generation unit 26 includes a rainy day pattern function generation unit 28 and a fine weather pattern function generation unit 29. Whether the rainy weather pattern function is used by taking in weather information such as rainfall information or the rainy day pattern Select whether to use the function, and generate a fluctuation pattern function of the daily water quality and quantity.

  The fine weather pattern function generation unit 28 calculates the daily water quality / water quantity variation pattern function in fine weather based on the clear water quality / water quantity information (daily water quality / water quantity) stored in the database device 3. Generate. This is generated, for example, by averaging the daily fluctuation patterns of water volume and water quality on a clear day for 10 days before that day.

  In addition, the rainy day pattern function generation unit 29 stores the actual rainfall information, which is stored in the database device 3 as the rainfall information, the rainfall information, the rainfall intensity, the rainfall duration, and the like. Are compared to select similar rainfall, and a variation pattern function describing the variation coefficient is generated from the inflow water amount / inflow water quality pattern in the past similar rainfall.

  The data generation unit 27 multiplies the variation coefficient included in the variation pattern function output from the pattern function generation unit 26 by the current measured value or the current estimated value of the inflow amount data or the inflow water quality data, thereby Create water volume and quality data.

<effect>
(1) Future simulation data is automatically generated from past clear weather data and rainy weather data, enabling future simulations and examining driving methods.

<< 10th Embodiment >>
FIG. 8 is a block diagram showing a tenth embodiment of a sewage treatment plant operation support apparatus according to the present invention. In this figure, parts corresponding to those in FIG. 1 are denoted by the same reference numerals.

<Configuration and operation>
The sewage treatment plant operation support device 61 shown in this figure is different from the sewage treatment plant operation support device 1 shown in FIG. 1 in that an operation condition setting unit 6 that outputs an operation condition manually set by an operator is used instead of an operation condition setting unit 6. The operation optimization unit 30 for optimizing the operation content is provided by evaluating the cost and the quality of the influent water in an integrated manner, and the operation cost is reduced while ensuring a constant inflow water quality.

  The operation optimization unit 30 includes a cost evaluation function setting unit 31 that converts the quality of the influent water into an actual operation cost using a cost evaluation function and evaluates the operation cost and the quality of the influent water in an integrated manner. The inflow water quality taken in from the process simulator 8 is multiplied by a conversion coefficient to obtain the actual operation cost, and an evaluation function corresponding to the operation cost and the inflow water quality is calculated and registered in advance. An operating condition that minimizes the cost evaluation value of the evaluation function is obtained using an optimization algorithm such as that disclosed in Japanese Patent Application Laid-Open No. 2004-171531. Then, this is supplied to the sewage treatment process 2 to control the operation content and to the process simulator 8 to automatically adjust the model parameters.

<effect>
(1) By optimizing the operation with the process model in which the model parameters are automatically adjusted, the actual condition can be simulated more, so that it is possible to search for highly practical operation conditions.

(2) Since the operation condition that optimizes the quality and cost of the influent water is automatically obtained, operation management is more efficient.

<< 11th Embodiment >>
FIG. 9 is a block diagram showing an eleventh embodiment of a sewage treatment plant operation support apparatus according to the present invention. In this figure, parts corresponding to those in FIG. 1 are denoted by the same reference numerals.

<Configuration and operation>
The sewage treatment plant operation support device 71 shown in this figure is different from the sewage treatment plant operation support device 1 shown in FIG. 1 in that a water quality support center connected to each treatment site 32 via a communication line 33 such as the Internet. 34, the process simulator 8 is arranged, and the parameter optimization processing of each processing site 32 is performed by the process simulator 8.

  In each processing site 32, a model parameter setting unit 4, a plant condition setting unit 5, an operation condition setting unit 6, an automatic adjustment command unit 7, and a database device 3 are installed. Unit 5, operating condition setting unit 6, set value information set in automatic adjustment command unit 7, simulation input data generated by database device 3, and parameter identification data are communicated via communication line 33 This is transmitted to the process simulator 8 of the center 34.

  Each time information is supplied from each processing site 32, the process simulator 8 of the water quality support center 34 performs automatic adjustment of model parameters and simulation based on such information, and determines the quality of the influent water and the operating cost. The calculation result is transmitted to the processing site 34 of the information transmission source, and the operation content is optimized.

<effect>
(1) Since only one process simulator 8 for simulating a plurality of processing sites 32 is required, the cost for hardware can be reduced.

(2) Since the person who performs maintenance of the process simulator 8 may be concentrated in the water quality support center 34, the cost required for maintenance of the process simulator 8 can be reduced.

<Modification of Eleventh Embodiment>
(1) An optimization operation device may be provided in the water quality support center 34, and the optimum operation condition obtained by this optimization operation device may be distributed to the processing site 32.

(2) A database is provided in the water quality support center 34, and process parameter values and water quality analysis values at each treatment site 32 are transmitted to the water quality support center 34 and stored in the database, and model parameters are optimized and simulated. In another embodiment, model parameters, simulation results, etc. may be distributed to each processing site 32.

The block diagram which shows embodiment of the sewage treatment plant operation assistance apparatus by this invention. The block diagram which shows the detailed circuit structural example of a parameter automatic adjustment part. The block diagram which shows the other example of a detailed circuit structure of a parameter automatic adjustment part. The block diagram which shows other embodiment of the sewage treatment plant operation assistance apparatus by this invention. The block diagram which shows an example of a sewage treatment process. The block diagram which shows the detailed circuit structural example of a database. The block diagram which shows the detailed circuit structural example of the data generation part for simulation. The block diagram which shows other embodiment of the sewage treatment plant operation assistance apparatus by this invention. The block diagram which shows other embodiment of the sewage treatment plant operation assistance apparatus by this invention. The block diagram which shows the outline | summary of the parameter setting apparatus which is a prior art example. The block diagram explaining the parameter setting technique which is a prior art example. The block diagram which shows the detailed circuit structural example of the parameter selection process part in the prior art example shown in FIG.

Explanation of symbols

1, 51, 61, 71: Sewage treatment plant operation support device 2: Sewage treatment process 3: Database 4: Model parameter setting unit (setting unit)
5: Plant condition setting part (setting part)
6: Operating condition setting part (setting part)
7: Automatic adjustment command setting part (setting part)
8: Process simulator 9: Simulation data generation unit (data generation unit)
10: Data generation unit for parameter identification (data generation unit)
12: Model parameter storage unit 11: Parameter automatic adjustment unit 13: Simulation calculation unit 14: Simulation result storage unit 15: Individual simulation error determination unit 16: Check unit 17: Change parameter selection unit 18: Parameter optimization unit 19: Parameter setting Change unit 20: Simulation accuracy determination unit 21: Precipitation tank 22: Total nitrogen meter 23: Inflow amount prediction unit (future inflow water quality pattern function generation unit)
24: Similar inflow pattern selection unit (future inflow water quality pattern function generation unit)
25: Data generation unit for future simulation 26: Pattern function generation unit 27: Data generation unit 28: Pattern function generation unit during fine weather 29: Pattern function generation unit during rainy weather 30: Operation optimization unit 31: Cost evaluation function setting unit 32: Treatment site 33: Communication line 34: Water quality support center

Claims (18)

A sewage treatment plant operation support device that performs operation support by simulating a sewage treatment process,
A database that accumulates input operation information, water quality information measured by sensors in the sewage treatment process, and water quality analysis result information analyzed offline, in time series;
Based on the information stored in this database, a data generator for generating data necessary for automatic adjustment of model parameters of the water quality model and water quality data for influent water required for process simulation,
A setting unit in which plant conditions and operation of the sewage treatment process are set, and
Based on each condition set in these setting units, each information stored in the database, and each data generated by the data generation unit (switching of model parameters to be adjusted in consideration of characteristics of model parameters) Calculating the water quality of the sewage treatment process and having an automatic parameter adjustment unit that optimizes each model parameter so that the calculated value approaches the actual water quality value collected in the database. A process simulator to calculate,
A sewage treatment plant operation support apparatus, comprising: In the sewage treatment plant operation support device according to claim 1,
The process simulator includes parameter information describing the upper and lower limit values of each model parameter, the constraint conditions of each model parameter, the direction and magnitude of change of each evaluation target water quality when each model parameter is changed, and each model. The adjustment target water quality with respect to the parameter and the matrix information in which the degree of influence on the adjustment target water quality is described in a matrix, and based on the model parameter constraint conditions and the matrix information, the sewage treatment is performed while switching the model parameter to be adjusted. A parameter automatic adjustment unit that calculates the water quality of the process and optimizes each model parameter so that the calculated value approaches the actual water quality value collected in the database,
A sewage treatment plant operation support apparatus, comprising: In the sewage treatment plant operation support device according to claim 2,
In the matrix information, the water quality to be evaluated is described in a matrix form according to anaerobic conditions, anoxic conditions, and aerobic conditions.
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to claim 1,
The process simulator includes parameter information in which upper and lower limits of each model parameter and constraint conditions for each model parameter are described, parameter group information in which parameters are grouped into two or more groups in advance according to the characteristics of each model parameter, While switching the model parameter group to be adjusted, calculate the water quality of the sewage treatment process, and optimize the model parameters of each group sequentially so that the calculated value approaches the actual water quality value collected in the database Automatic parameter adjustment section,
A sewage treatment plant operation support apparatus, comprising: In the sewage treatment plant operation support device according to claim 2 or 4,
The parameter automatic adjustment unit preferentially selects and adjusts model parameters that affect the water quality of a plurality of items,
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 5,
A simulation accuracy judgment unit that evaluates whether the error between the water quality data measured online or offline and the water quality data obtained by simulation is within the judgment value,
The simulation accuracy determination unit generates a parameter automatic adjustment command and outputs it to the process simulator when the error exceeds a determination value.
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 6,
The data generation unit is based on offline time-series data of target data for a certain day accumulated in the database, and on-line data correlated with the offline time-series data at a certain time of data not measured online. Currently generating data,
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 6,
The data generation unit estimates data that is not measured online based on offline time-series data of target data for a certain day accumulated in the database and online data correlated with the offline time-series data, Using the representative data measured at a point in time on the day, the estimated data is corrected, and current data at a certain point in time on the day is generated.
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 6,
The data generation unit predicts the amount of sewage inflow one day, several hours ahead, or several minutes ahead based on at least one of weather forecast information, pump well level information, and past statistical information. Compare the inflow pattern one day, several hours, or several minutes in the day with the past inflow pattern accumulated in the database, and from the water quality information corresponding to the most similar past inflow pattern A future influent water quality pattern function generation unit that creates a future influent water quality pattern function that is one day, several hours ahead, or several minutes ahead of the day,
A future simulation data generation unit that generates future simulation data based on the future inflow water quality pattern function generated by the future inflow water quality pattern function generation unit and the current data that is measured online or offline;
A sewage treatment plant operation support apparatus, comprising: In the sewage treatment plant operation support device according to any one of claims 1 to 6,
The data generation unit, based on the information accumulated in the database, a fine weather pattern function generation unit for generating an inflow water amount / water quality pattern in fine weather,
Based on the information accumulated in the database, a rainy day pattern function generation unit that generates an inflow water amount / water quality pattern in rainy weather,
Depending on the weather conditions at that time, select either the fine weather pattern function generation unit or the rainy day pattern function generation unit to generate the inflow water amount / water quality pattern. Or generate future simulation data based on the current data measured offline.
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 6,
The data generation unit multiplies data that is measured online or data generated by the method described in any one of claims 7 to 10 by a certain ratio, and is either online or offline. Finding unmeasured data,
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 6,
The data generator is a model parameter evaluation data necessary for evaluating the quality of the sewage treatment process and discharged water, and includes any one of sludge concentration, ammonia nitrogen concentration, nitrate nitrogen concentration, and phosphorous acid phosphorus concentration. Generate corresponding data,
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 6,
The data generation unit is a daily average of data measured online, an average of daily offline data, or a daily average of data generated by the method according to any one of claims 7 to 10. Use the values to generate data necessary for automatic parameter adjustment, data required for water quality assessment within the sewage treatment process and effluent,
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 6,
The data generation unit uses measurement data output online or off-line from each sensor arranged along the flow direction of the sewage treatment process, and the data necessary for automatic parameter adjustment, in the sewage treatment process and Generate data necessary for evaluating the quality of discharged water,
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 6,
Operating conditions necessary to minimize the cost value of this cost evaluation function by converting the quality of the discharged water into operating cost and using a cost evaluation function that evaluates the operating cost and the quality of the discharged water in a unified manner Operation optimization unit,
A sewage treatment plant operation support apparatus, comprising: In the sewage treatment plant operation support device according to any one of claims 1 to 15,
One or more sewage treatment sites having the database, the data generation unit, and the setting unit;
Each condition set in the setting unit of the sewage treatment site, each information stored in the database, each data generated by the data generation unit is taken in via the communication line, and each of these conditions and each information Using each data, calculate the water quality and operating cost of the sewage treatment process, optimize each model parameter, and supply the optimized model parameter to the sewage treatment site via the communication line. A support center,
A sewage treatment plant operation support apparatus, comprising: In the sewage treatment plant operation support device according to claim 16,
An internet line is used as the communication line.
A sewage treatment plant operation support device. In the sewage treatment plant operation support device according to any one of claims 1 to 15 or 17,
The database captures and accumulates offline data and online data collected by a service company,
A sewage treatment plant operation support device.

Images