JP2002062927A - Waste water treatment process simulator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute the optimization of the model constants of a waste water treatment process according to conditions such as calculation time, precision, stability and the like. SOLUTION: This waste water treatment process simulator comprises a quality function storage part 108 for modeling and storing the waste water treatment process as a quality function containing a quality factor of waste water and the model constant; a model constant optimization means 105 provided with a plurality of optimization arithmetic means for optimizing the value of the model constant by use of the measured values of the quality factor of waste water, an optimization means priority determining means 101 for determining the priority of the optimization arithmetic means on the basis of a given selecting condition, an optimization means control device 102 for making the model constant optimization means 105 perform the optimizing operation of the model constant by use of the optimization arithmetic means on the basis of the determined priority, and a calculation result judgment means 103 for judging the propriety of the model constant obtained by the optimizing operation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device or a driving support device to which a simulator for a wastewater treatment process such as sewage treatment is applied.

[0002]

2. Description of the Related Art In general, current water purification processes, sewage treatment processes, waste water treatment processes and the like are operated and managed by operators who have experience and knowledge by monitoring and judging the situation. In such a plant, the role played by the operator is large, and there are problems such as (1) difficulty in labor saving, (2) human error is likely to occur, and (3) there is no guarantee that the operating conditions are optimal. As one of solutions to these problems, automatic driving control or driving support using a simulator based on a mathematical model has been proposed. The mathematical model referred to here is a biological / chemical reaction model that simulates the state in which the substance contained in each treated water is changed by microorganisms and oxidizing agents, and a material balance by inflow and outflow, and in some cases, physical Refers to those that calculate the model at the same time. Among them, IAWQ (International Conference on Water Quality,
The IAWQ model proposed by International Association on Water Quality is well known. As a method of controlling a process using the IAWQ model, for example, there is a method described in JP-A-10-235333. This technology simulates and controls the sewage treatment plant using an IAWQ model or a mathematical model equivalent to it, so that the connection status of each tank, input / output items, control method, etc. can be easily set on the screen. It was done.

In an actual wastewater treatment process, the quality and conditions of inflow water to be treated fluctuate depending on the season and water temperature.
Even in a biological reaction model simulating a wastewater treatment process, the values of various constants (model constants) included in the biological reaction model vary depending on each process, season, and water temperature. For this reason, when applying the simulator based on the mathematical model described above, it is necessary to optimize the value of the model constant according to the season and the water temperature. Conventionally, it was necessary to carry out a detailed experiment in order to optimize each model constant. Therefore, experimental know-how was required, and the analysis of experimental data was complicated. For this reason, it has been difficult to operate a process simulator provided in a computer system at a site online.

As a solution to this, without experimentation,
The value of the model constant is calculated for each inflow water quality (season, water temperature)
Various techniques for optimizing appropriately (optimization calculation technique) have been proposed. For example, P.475 of the 36th Sewage Works Presentation Meeting (1999), describes the results of optimizing model constants using the Monte Carlo method. Also, JP-A-5-253590 and JP-A-10-253
235333 proposes optimizing model constants using a simplex method.

[0005]

The method of optimizing the model constants on a brute force basis, such as the Monte Carlo method, has an advantage in that it does not fall to a local minimum. However, since the amount of calculation for optimization is large, it takes time until a solution is obtained, and when the search range or the number of model constants is further increased, the amount of calculation dramatically increases. For example, assuming that the number of model constants to be optimized is 10, doubling the search range increases the amount of calculation by a factor of 2 times, that is, 1024 times. To shorten the calculation time, it is conceivable to narrow the search range. However, if the search range is made narrower than necessary, the possibility that the optimum point exists outside the range increases.

[0006] On the other hand, the simplex method of the nonlinear programming is one of the methods for directly finding an optimal solution, and the amount of calculation is smaller than that of the brute force method, and the time until the solution is obtained is shorter. However, if there is a local equilibrium point, it may converge toward that point and may not converge to an optimal solution.

[0007] As a typical method that directly converges toward an optimal solution other than the simplex method, there are an optimal regulator method, a steepest descent method, and a Newton method. In either case, the amount of calculation is smaller than that of the simplex method, and the time required to obtain an optimal solution is shorter, but the stability (convergence of calculation) is slightly inferior in JP-A-5-253590.

The method of optimizing model constants using an AI method based on a knowledge database or a neural network is an effective method for obtaining an optimal solution once the inference system is completed. The computation time is considered to be faster than methods such as the simplex method, and the stability is considered to be good because it is based on past data. However, the inference system is inadequate at the time of newly starting the apparatus or until data is sufficiently accumulated, and a certain learning period is required.

[0009] Each of the above-mentioned techniques used for optimizing the model constant has different characteristics. Typical features are the length of calculation time, high and low accuracy, and stability. That is, when only one of the various optimization methods described above is provided in the wastewater treatment process simulator, the required conditions, that is, the calculation time, accuracy, and stability may not be satisfied. In some cases, a solution may not be obtained by that method.

The present invention has been made in view of the above background, and has been made to optimize a model constant of a wastewater treatment process in accordance with required conditions (calculation time, accuracy, and stability). Aim.

[0011]

According to the present invention, in order to achieve the above object, first, a plurality of types of optimization calculation means are provided so as to be selectable. Next, the features of the plurality of optimization calculation means were numerically expressed. Then, there is provided means for determining a priority order based on the numerical values of the plurality of optimization calculation means for given selection conditions.

That is, a first means of the present invention for achieving the above object is to provide a water quality function storage unit for modeling a waste water treatment process and storing it as a water quality function including a water quality factor and a model constant of the waste water, And a model constant optimizing means having an optimization calculating means for optimizing the value of the model constant using actual measured values of the wastewater treatment process, wherein a wastewater treatment process is simulated using the water quality function. In the simulator, the model constant optimizing means includes a plurality of types of optimizing operation means, and an optimizing means priority determining means for deciding a priority order of the plurality of optimizing operation means based on a given selection condition. And a ranking determining means.

One of the plurality of types of optimization calculation means is
For example, it is desirable to use a brute force method such as a Monte Carlo method.

Further, there are provided a calculation result judging means for judging the validity of the model constant value obtained by a plurality of types of optimization calculating means, and an optimization result displaying means for simultaneously displaying a valid result and an invalid result. You may do so.

The selection condition for selecting the optimization operation means defines, for example, the length of time required for the optimization operation.

At the time of the optimal calculation, the measured values of the water quality factors are used. If the number of the obtained measured values is small, the state variable estimating means for estimating the value of the water quality factor of the wastewater required for the model constant optimization calculation is provided. It is desirable to provide an estimated water quality factor display means for displaying the value of the estimated water quality factor of the wastewater. The estimated water quality factor may be used for controlling the actual wastewater treatment process.

Further, a database storing values of general model constants and model constants optimized in the past is provided, and values of model constants determined using the model constant optimizing means are stored in the database. Is compared with the values of the model constants stored in, and it is judged by logical inference whether or not the values are within the theoretically valid range. A device abnormality diagnosis means for diagnosing that the measuring device is abnormal may be provided.

[0018]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows a main part of the configuration of a waste water treatment process simulator according to a first embodiment of the present invention. The illustrated wastewater treatment process simulator 100 includes an optimization means priority determination means 101, an optimization means control device 10
2, calculation result determination means 103, calculation result display means 10
4. Model constant optimizing means 105, measured value input unit 10
7. It is configured to include the water quality function storage unit 108.

In this embodiment, a plurality of operation methods for optimizing the model constant are prepared, and the operation method most suitable for the condition required for the optimization operation is automatically selected for optimization. An operation is performed.

Hereinafter, each component of the present embodiment will be described.

The model constant optimizing means 10 shown in FIG.
Reference numeral 5 includes a plurality of optimization calculation means (hereinafter referred to as an optimization calculation method) for performing a calculation for optimizing a model constant. In the present embodiment, a brute force method, a simplex method, and a Newton method are provided. The brute force method takes a lot of calculation time but has good stability. On the other hand, the calculation time of the Newton method is short, but if the calculation target is nonlinear, the stability may be poor. The simplex method has features intermediate between the two.

In addition, the model constant optimizing means 10
Examples of the optimization calculation method that can be used for the 5 include a genetic algorithm method, an optimal regulator method, a steepest descent method,
A Monte Carlo method or the like can be adopted, and the combination may be made in consideration of the characteristics of each method according to the assumed condition of model constant optimization. In contrast to the brute force method represented by the Monte Carlo method, the simplex method, Newton method, and the like are called gradient methods. The plurality of optimization calculation methods provided in the model constant optimization means 105 are of different types, and one of them is preferably a brute force method.

The optimizing means priority determining means 101 includes:
The selection condition 121 relating to the model constant optimizing means 105 is manually or automatically input from outside. The selection condition 121 in this case is, for example, an optimization calculation method used preferentially, a calculation time, stability, and the like. The calculation time is the calculation time required from the start of the optimization operation to the convergence of the operation result and the time required for processing related to input and output. The calculation time is desirably given quantitatively, for example, within one hour, but in the present embodiment,
It is qualitatively given as "short" or "long". Stability is qualitatively given as "good", "normal", "poor", and so on. The optimizing means priority determining means 101 fetches the above-mentioned selection conditions 121, determines the priorities 109 for the optimization calculation methods provided in the plurality of model constant optimizing means 105, and optimizes the priorities 109. Output to the means control device 102.

Here, a method of determining the priority order 109 will be specifically described with a simple example. First, it is assumed that the calculation time and the stability are input as the selection conditions 121. These selection conditions 121 are the optimization means priority order determination means 101
Are first quantified into a calculation time index t and a stability index s. Each index is dimensionless and takes a value from 1 to 10. The closer the calculation time index t is to 1, the shorter the calculation time, and the closer it is to 10, the longer the calculation time. The closer the stability index s is to 1, the lower the stability is.
It shows that stability is so good that it is close to. For example, if the selection condition is “the calculation time is as short as possible and the stability is about normal”, the value of the index t is 1 and the value of the index s is 5
Is applied.

Based on the values of the calculation time index t and the stability index s, priorities are assigned to the brute force method, the simplex method, and the Newton method using a polygonal function as shown in FIG. In FIG. 2, ft is a time evaluation function,
fs is a stability evaluation function. The number in parentheses indicates the optimization calculation method of the model constant optimizing means 105.
Represents the Newton method, 2 represents the simplex method, and 3 represents the brute force method. The sum of the value of ft and the value of fs is calculated as the total evaluation function J
The higher the value of J, the higher the priority.

For example, it is assumed that the input selection conditions are digitized and (calculation time index t, stability index s) = (1, 1). From the function shown in FIG. 2, ft (1) = 1, ft (2) =
0.5, ft (3) = 0, fs (1) = 1, fs (2) = 0.5, fs (3) =
Since it becomes 0, the total evaluation function J for each method is obtained from each sum. ft (1) + fs (1) = J (1) = 2 ft (2) + fs (2) = J (2) = 1 ft (3) + fs (3) = J (3) = 0 The result is Newton The method has the highest priority, the simplex method has the second priority, and the brute force method has the lowest priority. The calculation time index t
The fact that the stability index s is 1 is equivalent to the case where the stability is low, and therefore the calculation should be completed immediately. Using the Newton's method with priority is based on the input selection conditions. And appropriate.

If the numerical selection condition is (calculation time index t, stability index s) = (10, 10), ft (1) =
0, ft (2) = 0.5, ft (3) = 1, fs (1) = 0, fs (2) = 0.
5, fs (3) = 1. The total evaluation function J is as follows. ft (1) + fs (1) = J (1) = 0 ft (2) + fs (2) = J (2) = 1 ft (3) + fs (3) = J (3) = 2 This indicates that the win method has the highest priority, followed by the simplex method, and the Newton method has the lowest priority. The fact that the calculation time index t is 10 and the stability index s is 10 corresponds to the case where “the calculation time may be long, so that a stable solution is desired”, and the brute force method is used with priority. That is appropriate.

Assuming that the selection condition further quantified is ((calculation time index t, stability index s) = (10, 1), ft (1) = 0 and ft (2) = 0 in the same manner as described above. .5, ft (3) = 1, fs
(1) = 1, fs (2) = 0.5, fs (3) = 0, and the total evaluation function
J is J (1) = 1, J (2) = 1, and J (3) = 1. this is,
Since this corresponds to a condition with a weak restriction such as "the calculation time may be long and the stability may be poor", there is no problem using any model constant optimizing means, and it is difficult to determine the superiority or inferiority. When the values of the total evaluation functions J are equal, it is desirable to request the operator to correct the selection condition 121.

(Calculation time index t, stability index s) = (1, 1)
0), ft (1) = 1, ft (2) = 0.5, ft (3) = 0, f
Since s (1) = 0, fs (2) = 0.5, and fs (3) = 1, J (1)
= 1, J (2) = 1, J (3) = 1. This corresponds to a strict restriction such as "select a model constant optimizing means having a short calculation time and good stability". . Also in this case, it is desirable to request the operator to correct the selection condition 121.

(Calculation time index t, stability index s) = (5,
5), ft (1) = 0.5, ft (2) = 1, ft (3) = 0, f
Since s (1) = 0, fs (2) = 1 and fs (3) = 0.5, J (1) =
0.5, J (2) = 2, and J (3) = 0.5. This indicates that the simplex method has the highest priority, followed by the simplex method and the brute force method. The fact that the calculation time index t is 5 and the stability index s is 5 corresponds to the case where “the calculation time is short to some extent and a stable solution is desired”, and it is appropriate to preferentially use the simplex method. is there.

As described in the above example, the optimizing means priority determining means 101 determines the priorities 109 for a plurality of optimizing calculation methods, and outputs the determination result.

The optimizing means control device 102 is connected to the output side of the optimizing means priority determining means 101,
Model constant optimizing means 105 and calculation result determining means 103
It is connected to the. The optimization means control device 102 is a model constant optimization means 1 having a plurality of optimization calculation methods.
05 is controlled in accordance with the priority order 109 input from the optimization means priority order determination means 101, the optimization operation of the model constants is performed, and the optimization result set 110 which is the sum of the optimization results 112 is determined as the calculation result. Output to means 103.

The measured value input section 107 has an input side connected to the wastewater treatment process 119 and an output side connected to the model constant optimizing means 105. The actually measured value input unit 107 captures at least the actually measured value 120 of the soluble organic matter concentration among the water quality factors of the wastewater treatment process 119, and in some cases, performs processing / correction for a measurement error, a measurement error, and the like to obtain a corrected actual value 116. Generated and given to the model constant optimizing means 105.

The model constant optimizing means 105 is connected to the optimizing means control device 102, the measured value input unit 107, and the water quality function storage unit 108. Model constant optimization means 10
5 is controlled by the optimization means control device 102, and the water quality factor calculation value 118 input from the water quality function storage unit 108 is converted into the actually measured value input unit 1 by using the specified optimization calculation method.
The model constant 117 is adjusted so as to coincide with the corrected actual measurement value 116 input from 07 and output to the water quality function storage unit 108.

The water quality function storage unit 108 is connected to the model constant optimizing unit 105 and includes a mathematical model (a mathematical model of the wastewater treatment process) including model constants to be optimized by the model constant optimizing unit 105. I have. The mathematical model provided in the water quality function storage unit 108 is IA
ASM No. 2 proposed by WQ is typical, but other mathematical models can also be used. The water quality function storage unit 108 first calculates the water quality factor calculation value 118 using the provided mathematical model, and
And then fetches the model constant 117 given from the model constant optimizing means 105, and outputs the model constant 1
The water quality factor calculation value 118 calculated using the calculation result 17 is output to the model constant optimization means 105.

The calculation result judging means 103 has an input side connected to the optimizing means control device 102 and an output side connected to the optimizing result display means 104. Calculation result determination means 10
3 takes in the optimization result set 110 output from the optimization means control device 102, determines whether the result is a valid value, and outputs the optimization determination result 115 to the optimization result display means 104. .

The determination of validity is performed by the following method, for example. (1) The value of the optimized model constant is compared with a preset upper and lower limit range, and if the value is within the range, it is determined that the value is appropriate. For example, the value of the saturation constant of dissolved oxygen is theoretically 0.
It cannot be less than or equal to a certain value.
If the value of the optimized model constant is within this range, it is determined to be appropriate. (2) Among the optimized model constant values, some model constants are compared based on a preset association table. If the relationship between the model constants matches the rules of the association table, it is determined that the obtained model constant is valid. For example, the growth amount of an aerobic microorganism is determined by the difference between the growth rate and the death rate. In a tank with a large amount of dissolved oxygen, the growth rate is theoretically higher than the death rate. If the value of the optimized model constant satisfies this tendency, it is determined that the model constant is appropriate.

As a result of the determination on the validity, if the set of optimization results 110 output from the optimizing means control device 102 is not valid, the calculation result determining means 103 notifies the optimizing means control device 102 of that fact. Upon receiving this notification, the optimizing unit control device 102 causes the optimizing operation to be performed in the above-described procedure by the next priority optimizing operation method, and outputs the obtained result to the calculation result determining unit 103 again. The calculation result determination unit 103 determines the validity of the input optimization result set 110 in the above-described procedure. This procedure is repeated until a reasonable result is obtained.

It should be noted that the optimizing means control device 102 does not wait for the validity judgment by the calculation result judging means 103,
The optimization calculation of the model constant may be performed by sequentially using the optimization calculation methods having the highest priority.

The optimization result display means 104 is connected on the input side to the calculation result determination means 103, and the calculation result determination means 10
The optimization determination result 115 input from Step 3 and the model constant optimizing means (method) used in the optimization calculation are displayed on an instrument panel or a monitor so as to be understood by the operator. The contents displayed here are the name of the optimization operation method used for the optimization operation and the valid model constants obtained as a result of the optimization operation. However, in some cases, the performance of each optimization operation method is compared. Alternatively, an invalid result may be displayed at the same time.

According to the present embodiment, the optimization calculation method that matches the specified selection condition is selected, and the optimization calculation of the model constant is performed by the selected optimization calculation method. It is possible to optimize the model constant under conditions such as accuracy, stability, and the like. In addition, by displaying the result of the optimization operation including an invalid result, an inexperienced operator can learn and judge whether or not the optimization result is appropriate.

A second embodiment of the present invention will be described with reference to FIG. In this embodiment, the optimal regulator method 30 in modern control theory is used as one of the optimization calculation methods of the model constant optimizing means 105 in the first embodiment. FIG. 3 shows a flow of the optimization calculation by the optimum regulator method 30. Briefly describing this method, first, a differential matrix A relating to a plurality of water quality factors of a water quality function (mathematical model) and a differential matrix B relating to a plurality of model constants are numerically generated. An appropriate state feedback matrix F is obtained from the values of the differential matrices A and B, and the values of the model constants are sequentially adjusted for optimization. The state feedback matrix F is
It is obtained by solving the Riccati matrix equation given the gains Q and R with moderate stability and convergence speed.

The optimization calculation was actually performed using the waste water treatment process simulator of the present embodiment. The target of calculation is two anaerobic tanks R1 and R2 and two anaerobic tanks R3 and R
This is a small experimental apparatus for treating actual sewage, comprising four and four aerobic tanks R5 to R8. Fig. 4 shows the flow of this device.
Shown in The capacity of each tank is 27L. Aerobic tanks R5 to R8
Is configured so that bubbles of air are blown out from the bottom thereof and rise in the treated water, and a settling pond is connected to the outlet side of the most downstream aerobic tank R8. Part of the settling sludge in the settling basin is circulated to the anaerobic tank R1 as returned sludge. Further, a part of the water to be treated is circulated from the aerobic tank R8 to the anoxic tank R3 as a circulating liquid.

In the illustrated apparatus, the water to be treated is first introduced into the anaerobic tank R1, and passes through each tank in order from the anaerobic tank R1 to the aerobic tank R8. The water to be treated is purified while passing through each tank, and is guided from the aerobic tank R8 to the sedimentation basin. After the sludge is separated in the settling basin, the water to be treated is discharged.

With respect to the mathematical model of the apparatus having the above-described configuration, optimization of model constants was separately performed for each tank. After optimizing the model constants, the results regarding ammonia nitrogen in the water quality obtained by the simulation are compared with the values obtained by the simulation before the optimization of the model constants and the target values (actually measured values) in FIG. Show. The horizontal axis is the tank number. The calculated values before optimization of model constants were obtained using model constants optimized in other sewage treatment plants. As shown in this figure, the calculated value before optimization largely deviates from the target value. However, as a result of optimizing the model constants by the optimal regulator method 30, the calculated water quality value obtained by simulation is almost equal to the target value. I do.

The optimal regulator method of the present embodiment is a method for directly finding an optimal solution based on a differential gradient.
The time required to optimize the model constants is short. The optimal regulator method can obtain better results than the Newton method in terms of stability by changing the gain given by the Riccati matrix equation. For reference, FIG. 2 shows a time evaluation function ft and a stability evaluation function fs of the optimal regulator method 30.
Is indicated by a broken line.

Next, a third embodiment of the present invention will be described with reference to FIG. The present embodiment is different from the embodiment shown in FIG. 1 in that a state variable estimating unit 41 is provided connected to the actual measurement value input unit 107.
In that an estimated water quality factor display means 42 is provided in connection with the output side. The other configuration is the same as that of the first embodiment, and the same configuration is denoted by the same reference numeral and the description is omitted. The state variable estimating means 41 calculates the corrected actual measurement value 116.
And has a function of estimating the value of an item that can be estimated among water quality factors that have not been actually measured, and the estimated result is displayed on an instrument panel or a monitor by the estimated water quality factor display means 42.

In an actual wastewater treatment process, IAWQ ASM N
o.2 Sewage treatment microbial reaction model (mathematical model)
It is rare that all of the water quality factors included in the water quality are measured at the sewage treatment plant, and in many cases, only actual measured values of some water quality factors are given. As a result, there is a possibility that a problem will occur when the model constants are optimized using the optimal regulator method 30 or other optimization calculation methods. State variable estimating means 41
Takes the corrected actual measurement value 116 and estimates the value of an estimable item among the unmeasured water quality factors. The estimated result (water quality factor estimation value 43) is displayed on the instrument panel or monitor by the estimated water quality factor display means 42. Both the estimated value and the corrected actual measurement value set 40 are taken into the model constant optimizing means 105 and used for the optimization calculation.

The operator confirms the estimation result of the state variable estimating means 41 via the estimated water quality factor displaying means 42, and if the estimated value of the water quality factor 43 is not appropriate, it can be set so as not to be used for the optimization calculation. . Whether or not the water quality factor estimation value 43 is appropriate can be determined based on the upper and lower limits of the water quality factor and the relationship between the water quality factors, as in the case of the validity determination of the model constant described above. Alternatively, the determination may be made by verifying whether the estimated value is appropriate for the combination of the actually measured values of the model constants. When it is determined that the estimated value is not appropriate, it is desirable to use the previously used estimated value without using the estimated value.

According to the present embodiment, in addition to the effects of the first embodiment, even if some of the water quality factors required for the optimization of the model constants are not measured, the optimization of the model constants is performed. In this case, it is possible to increase apparently the water quality conditions to be given, which has the effect of improving the stability and reliability of the optimization calculation. Further, since the value of the estimated water quality factor is displayed, the operator can confirm the value of the estimated water quality factor. Therefore, when the value of the estimated water quality factor is out of the range assumed by the operator, the value of the corresponding water quality factor is
Since it can be instructed not to use for the model constant optimization calculation, the use of an incorrect estimation value of the water quality factor prevents the calculation from proceeding in an undesirable direction.

A fourth embodiment of the present invention will be described with reference to FIG. This embodiment is different from the third embodiment in that the estimated value of the water quality factor output from the measured value input unit 107 is corrected on the input side of the process control unit 50 for controlling the wastewater treatment process 119. The set of measured values 40 is
The point is that it is configured to be input. The other configuration is the same as that of the third embodiment, and the same reference numerals are given and the description is omitted.

As described in the third embodiment, there are few cases where all the water quality factors included in the sewage treatment microbial reaction model are measured, and in many cases only actual values of some water quality factors are given. . As a result, in order to use the wastewater treatment process simulator 100 for control, it may be necessary to estimate a deficient measured value. Therefore,
The state variable estimating means 41 is incorporated in the wastewater treatment process simulator, and the state variable estimating means 41 takes in the corrected actual measurement value 116 to estimate the value of an estimable item among the water quality factors not measured. The estimated water quality factor estimation value 43 is displayed on the instrument panel or monitor by the estimated water quality factor display means 42, and the estimated water quality factor value and the corrected actual measurement set 40 are stored in the model constant optimization means 10
5 and used for optimization calculations. In the present embodiment, the water quality factor estimation value and the corrected actual measurement value set 4
0 is simultaneously output to the process control means 50 for controlling the wastewater treatment process 119. The operator confirms the result through the estimated water quality factor display means 42, and obtains the estimated water quality factor 4
If 3 is not appropriate, it is desirable that it can be set so that it is not used for optimization calculation or control of model constants.

According to the present embodiment, in addition to the effects of the third embodiment, even in a facility with insufficient measuring equipment, the apparent measured value can be increased, and more appropriate control can be performed. There is an effect that it becomes possible.

A fifth embodiment of the present invention will be described with reference to FIG. This embodiment is different from the first embodiment in that a calculation result determination unit 103A is provided in place of the calculation result determination unit 103, and a device abnormality diagnosis function is added to the calculation result determination unit 103. The difference is that a database 60 and a device abnormality diagnosis result display means 62 are provided in connection with 103A. The other configuration is the same as that of the first embodiment, so the same reference numerals are given and the description is omitted.

The database 60 stores general model constant values and model constant values obtained by past optimization calculations. The device abnormality diagnosis function of the calculation result determination means 103A compares the past data 61 stored in the database 60 with the value of the model constant obtained by the optimization operation, and obtains the value obtained by the optimization operation by logical inference. It is evaluated whether the model constant (the set of optimization results 110) is within a predicted range, or whether the combination of values of the model constant is a possible combination during normal operation. As a result, if the value of the optimized model constant is outside the predicted range,
Alternatively, if the combination of the values of the model constants cannot occur during normal operation, it is determined that a device such as a measuring device for measuring water quality or a blower that blows air into the water to be treated is abnormal. The determination result of the device abnormality diagnosis function of the calculation result determination unit 103A is incorporated into the device abnormality diagnosis result display unit 62 as the abnormality diagnosis result 63. If a device is determined to be abnormal, the name of the device is changed to the instrument panel. Alternatively, it is displayed on the monitor as an abnormal device.

According to the present embodiment, even if each device does not have a detector for detecting an abnormality, it is possible to diagnose the abnormality of the device. In addition to the upper and lower limit values that can be taken by each model constant, the ratio and combination of a plurality of correlated model constant values can also be set as the allowable range of the model constant, so that the device abnormality diagnosis can be performed in a detailed and appropriate manner. This can be performed based on conditions.

A sixth embodiment of the present invention will be described with reference to FIG. This embodiment is different from the first embodiment in that a standard input generation unit 80 whose output side is connected to a water quality function storage unit 108 and an actual measurement value input unit 107, and an input side is connected to the water quality function storage unit 108. Transfer function estimating means 81 connected, and PI connecting input side to transfer function estimating means 81
D gain determining means 82 is provided, and the output side of the PID gain determining means 82 is connected to the process control means 50 for controlling the wastewater treatment process 119. The other configuration is the same as that of the first embodiment, and thus the same reference numerals are given and the description is omitted.

The standard input 83 such as a pulse input, a step input, a ramp input, or a combination of a plurality of them generated by the standard input generating means 80 is input to the water quality function storage unit 108. The standard input is control data for controlling the quality data of the water to be treated and the wastewater treatment process 119. The water quality function storage unit 108 has a response output (water quality factor calculation value) 8 corresponding to the standard input 83.
4 is output to the transfer function estimating means 81. The transfer function estimating means 81 includes a standard input 83 and a response output 84 corresponding thereto.
, The transfer function of the wastewater treatment process 119 is estimated.
The estimated process transfer function 85 is input to the PID gain determining means 82. PID gain determining means 82
Is based on the input estimated process transfer function 85,
An appropriate stable and responsive PID gain 86 is calculated and output to the process control means 50. The process control means 50 sets the gain of the control device based on the PID gain 86, and gives the control amount 87 to the wastewater treatment process 119. Here, PID control is assumed,
If the control system includes any of P, I, and D control elements, an appropriate gain can be similarly provided.

According to the present embodiment, the control gain of the control device for performing the PID control can be set to an appropriate value in a short time.

Needless to say, each of the above embodiments achieves the object of the present invention even when some of them are combined with each other.

[0062]

As described above, according to the present invention,
In optimizing the value of the model constant of the mathematical model of the wastewater treatment process, the value can be optimized by an optimizing means appropriately selected according to each selection condition. Therefore,
Conditions required in actual wastewater treatment work, that is,
Even if the accuracy is sacrificed to some extent, it is possible to perform the wastewater treatment corresponding to conditions such as starting the treatment early or controlling the water quality as close to the target as possible because there is enough time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a functional configuration of a wastewater treatment process simulator according to a first embodiment of the present invention.

FIG. 2 is a graph showing an example of a function for calculating a priority according to the first embodiment of the present invention.

FIG. 3 is a flowchart showing a procedure of a model constant optimization operation according to a third embodiment of the present invention.

FIG. 4 is a system diagram illustrating a schematic configuration of a model constant optimization calculation target experimental device according to a third embodiment of the present invention.

FIG. 5 is a graph showing water quality in a case where a model constant optimization calculation according to a third embodiment of the present invention is performed, in comparison with a target value and a water quality calculated by a mathematical model before the model constant optimization.

FIG. 6 is a block diagram showing a functional configuration of a wastewater treatment process simulator according to a fourth embodiment of the present invention.

FIG. 7 is a block diagram showing a functional configuration of a wastewater treatment process simulator according to a fifth embodiment of the present invention.

FIG. 8 is a block diagram showing a functional configuration of a wastewater treatment process simulator according to a sixth embodiment of the present invention.

FIG. 9 is a block diagram showing a functional configuration of a wastewater treatment process simulator according to a seventh embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 40 Estimated value and corrected actual measurement value set 41 State variable estimating means 42 Estimated water quality factor display means 43 Water quality factor estimated value 50 Process control means 60 Database 61 Past data 62 Equipment abnormality diagnosis result display means 63 Abnormal diagnosis result 80 Standard input generation Means 81 Transfer function estimating means 82 PID gain determining means 83 Standard input 84 Response output 85 Estimated process transfer function 86 PID gain 87 Control amount 100 Wastewater treatment process simulator 101 Optimizing means priority determining means 102 Optimizing means control device 103, 103A Calculation result determination means 104 Optimization result display means 105 Model constant optimization means (plural) 107 Actual measurement value input unit 108 Water quality function storage unit 109 Priority order 110 Optimization result set 111 Execution signal 112 Optimization result 115 Optimization determination result 116 corrected Found 117 model constants 118 Water factor computation value 119 wastewater treatment process 120 Found 121 selected condition

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G05B 13/04 G05B 13/04 G06F 19/00 110 G06F 19/00 110 F term (Reference) 5B049 AA06 BB07 DD01 EE03 EE31 EE41 5H004 GA15 GA18 GA27 GA28 GB08 HA04 KC12 KC28 KC35 KC45 KC48 5H223 AA01 BB01 CC01 FF05

Claims (8)

[Claims] 1. A water quality function storage unit for modeling a waste water treatment process and storing it as a water quality function including a water quality factor and a model constant of the waste water, and optimizing a value of the model constant using an actually measured value of the water quality factor of the waste water. A wastewater treatment process simulator for simulating a wastewater treatment process using the water quality function, wherein the model constant optimization means comprises a plurality of types of optimization operation. Wastewater characterized by comprising optimization operation means, and optimizing means priority determination means for determining priorities of the plurality of optimization operation means based on given selection conditions. Processing process simulator. 2. The wastewater treatment process simulator according to claim 1, wherein one of said plurality of optimization calculation means is based on a brute force method. 3. The wastewater treatment process simulator according to claim 1 or 2, wherein an optimization means control device for optimizing and calculating a value of a model constant by using the model constant optimization means having a higher priority first. It is provided with a calculation result determining means for judging the validity of the result of the optimization calculation and classifying the invalid result and the valid result, and an optimization result displaying means for simultaneously displaying the valid result and the invalid result. A wastewater treatment process simulator characterized by the following. 4. The wastewater treatment process simulator according to claim 1, wherein the selection condition is a qualitative or quantitative item that defines the length of time required for the optimization operation. A wastewater treatment process simulator characterized by including: 5. The wastewater treatment process simulator according to claim 1, wherein the value of the water quality factor of the wastewater required for the model constant optimization calculation is determined based on the actually measured value of the water quality factor. A wastewater treatment process simulator comprising: a state variable estimating means for estimating; and an estimated water quality factor displaying means for displaying an estimated value of a water quality factor of wastewater. 6. The wastewater treatment process simulator according to claim 5, wherein the state variable estimating means is based on an actually measured value of the water quality factor, and the water quality factor of the wastewater required for the model constant optimization calculation and the control required for the control. A wastewater treatment process simulator for estimating a value of a wastewater quality factor, and comprising a process control means for controlling a wastewater treatment process based on the measured value and the estimated value of the wastewater quality factor. 7. The wastewater treatment process simulator according to claim 1, wherein a database storing values of general model constants and values of model constants optimized in the past. Comparing the value of the model constant determined using the model constant optimization means with the value of the model constant stored in the database,
Device abnormality diagnosis that determines whether or not the value of the model constant that is determined is theoretically reasonable by logical inference, and that the device for measuring water quality factors is abnormal if the determined model constant value is not theoretically valid A wastewater treatment process simulator, comprising: means for displaying a diagnosis result of the device abnormality diagnosis means. 8. The wastewater treatment process simulator according to claim 1, wherein a standard input generating means for generating a standard input comprising a step input, an impulse input, a ramp input, or a combination thereof. Transfer function estimating means for estimating a transfer function of a wastewater treatment process from a response output calculated by inputting the standard input to the water quality function storage unit, and a PID control device using the estimated transfer function of the wastewater treatment process A wastewater treatment process simulator comprising: PID gain determination means for determining a gain of the wastewater treatment process.