JP7303101B2 - Model generation device, estimation device, model generation method and estimation method - Google Patents

Description

TECHNICAL FIELD The present invention relates to a model generating device and the like for generating a model for estimating the injection rate of a coagulant to be injected into influent water flowing into a water purification plant.

A known conventional technology is a technology for inputting input data such as the quality of influent water flowing into a water purification plant into a regression model and estimating the injection rate (chemical injection rate) of a flocculant that flocculates suspended solids in the influent water. (See Patent Documents 1 to 3 below).

JP-A-2002-119956 JP 2005-329359 A Japanese Patent Application Laid-Open No. 2007-061800

However, in the above-described conventional technology, the cost of water purification treatment for purifying influent water, especially the flocculation treatment for flocculating suspended solids in the influent water with a flocculant, the treatment of sludge generated in the flocculation treatment, and the There is a problem that the cost of the filtering process after the flocculation process is not taken into consideration. Therefore, if the coagulant is injected at the estimated chemical injection rate, there is a risk that the cost of these treatments will increase.

An object of one aspect of the present invention is to realize a model generation device and the like that can realize water purification treatment at reduced cost.

In order to solve the above problems, a model generation device according to one aspect of the present invention is a model generation device that generates a regression model for estimating the injection rate of a coagulant to be injected into influent water flowing into a water purification plant. There is, the turbidity of the influent before injecting the coagulant, the injection rate of the coagulant injected into the influent, the consumption cost of the coagulant injected into the influent, and the coagulant into the influent A set of data sets including the sludge treatment cost for treating the sludge generated by injecting and the filtration cost for backwashing when filtering the influent after removing the sludge, a classifying unit that classifies into a plurality of groups based on the turbidity; and a specifying unit that specifies, for each group, the injection rate that minimizes the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost. , from a set of data sets containing the representative value of turbidity and the specified injection rate for each group, the representative value is an explanatory variable, and the specified injection rate is a regression model with the objective variable and a model generator that generates

According to the above configuration, the regression model generated by the model generation device includes the turbidity of the influent before the injection of the flocculant, the consumption cost of the flocculant injected into the influent of the turbidity, and the flocculant into the influent. Correspondence with the coagulant injection rate that minimizes the total value of the sludge treatment cost for treating the sludge generated by injecting the becomes a regression model that shows As a result, by estimating the injection rate of the coagulant into the influent using the regression model, the injection rate of the coagulant that minimizes the above total value can be estimated according to the turbidity of the influent. can. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration costs.

Further, in the model generation device according to the aspect of the present invention, the identification unit includes an intermediate model generation unit that generates an intermediate model, which is a regression model with the injection rate and the total value as variables, for each of the groups. The injection rate that minimizes the total value may be estimated for each group using the generated intermediate model.

According to the above configuration, the regression model generated by the model generation device is generated in consideration of the correspondence relationship between the injection rate of the coagulant and the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost. become a thing. As a result, it is possible to improve the accuracy of estimating the chemical injection rate that minimizes the total cost value.

Further, in the model generation device according to one aspect of the present invention, the model generation unit generates the regression model from the set including only data sets in which the turbidity at the outlet of the sedimentation basin of the influent is within a predetermined range. may

According to the above configuration, a regression model generated only from a data set in which the turbidity of the water at the sedimentation basin outlet is equal to or less than a predetermined threshold can be used for estimating the injection rate. From this, it is possible to estimate the flocculant injection rate at which the turbidity of the influent at the sedimentation basin outlet will be below a predetermined threshold. By injecting the coagulant at the injection rate, the possibility of highly turbid water flowing out from the sedimentation basin and further from the water purification plant is reduced. As a result, the turbidity of the filtered water in the rapid filtration basin in the latter stage has been reduced, the filtration cost has been further reduced by reducing the frequency of backwashing, and the risk of infection with cryptosporidium, etc. due to drinking the discharged water has been reduced. can do.

Further, in order to solve the above problems, an estimating device according to one aspect of the present invention is an estimating device for estimating the injection rate of a coagulant to be injected into influent water flowing into a water purification plant, wherein the coagulant is injected turbidity of the influent before the injection, the injection rate of the coagulant injected into the influent, the consumption cost of the coagulant injected into the influent, and the treatment of the sludge generated by injecting the coagulant into the influent In each of a plurality of groups classified based on the turbidity, a data set including the sludge treatment cost and the filtration treatment cost for backwashing when filtering the influent after removing the sludge A regression model that uses the representative value of the turbidity as an explanatory variable and the injection rate that minimizes the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost as an objective variable, and a current inflow into the water purification plant and an estimating unit for estimating the injection rate of the coagulant to be injected into the current influent based on the turbidity of the current influent, which is the influent.

According to the above configuration, the turbidity of the influent before the injection of the flocculant, the consumption cost of the flocculant injected into the influent of the turbidity, and the treatment of the sludge generated by injecting the flocculant into the inflow A regression model showing the correspondence relationship between the coagulant injection rate that minimizes the total value of the sludge treatment cost and the filtration treatment cost of backwashing in the filtration treatment of influent water after sludge removal, and the current influent water Turbidity is used to estimate the flocculant injection rate into the current influent. Thereby, the injection rate of the flocculant that minimizes the total value can be estimated according to the turbidity of the influent water. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration costs.

Further, in order to solve the above problems, a model generation device according to an aspect of the present invention includes an injection rate of a coagulant injected into influent water flowing into a water purification plant, a consumption cost of the coagulant, the influent water The total value of the sludge treatment cost for treating the sludge generated by injecting the flocculant into the sludge and the filtration treatment cost for the backwash treatment when filtering the influent after the sludge is removed. A model generation device for generating a learning model for estimating a relationship, comprising: water quality data including turbidity of said influent before injecting flocculant, said injection rate, said consumption cost, and said sludge treatment A model generation unit is provided for generating a learning model for outputting the correspondence relationship with the water quality data as input by performing machine learning using the cost and the filtration cost as teacher data.

According to the above configuration, the learning model generated by the model generation device is input with the water quality data of the influent, and the injection rate of the coagulant to be injected into the influent and the coagulant injected at the injection rate. consumption cost, sludge treatment cost for treating sludge generated by injecting flocculant into influent water, and backwash treatment cost for filtration treatment of influent water after sludge removal. It becomes the learning model to be output. Accordingly, by using the correspondence output from the learning model, it is possible to specify the flocculant injection rate that minimizes the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration costs.

Further, in order to solve the above problems, an estimation device according to an aspect of the present invention includes an injection rate of a coagulant injected into influent water flowing into a water purification plant, a consumption cost of the coagulant, and Correspondence relationship between the total value of the sludge treatment cost for treating the sludge generated by injecting the coagulant and the filtration treatment cost for the backwashing treatment when the influent water after the sludge is removed is filtered An estimating device that estimates the water quality data including the turbidity of the influent before injecting the coagulant, the injection rate of the coagulant injected into the influent, and the amount of the coagulant injected into the influent Consumption cost, sludge treatment cost for treating sludge generated by injecting coagulant into the influent, and filtration treatment for backwashing when filtering the influent after removing the sludge Using a learning model generated by performing machine learning with cost and as teacher data, the turbidity of the current influent, which is the influent currently flowing into the water purification plant before injecting the coagulant, is An estimating unit is provided for estimating the correspondence relationship using water quality data including water quality data as an input.

According to the above configuration, by inputting the water quality data of the influent, the injection rate of the coagulant to be injected into the influent, the consumption cost of the coagulant injected at the injection rate, and the injection of the coagulant. A learning model that outputs the correspondence relationship between the sludge treatment cost for treating the sludge generated by the process and the total value of the backwash filtration treatment cost for the filtration treatment of influent water after sludge removal, and the water quality data of the current influent water and to estimate the correspondence. Accordingly, using the estimated correspondence relationship, it is possible to specify the flocculant injection rate that minimizes the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration costs.

In order to solve the above problems, a model generation method according to one aspect of the present invention provides a model generation method for generating a regression model for estimating an injection rate of a coagulant to be injected into influent water flowing into a water purification plant. A model generation method performed by an apparatus, comprising: the turbidity of the influent before adding coagulant, the injection rate of the coagulant injected into the influent, and the consumption cost of the coagulant injected into the influent and the sludge treatment cost for treating the sludge generated by injecting the coagulant into the influent, and the filtration treatment cost for backwashing when filtering the influent after removing the sludge. A classification step of classifying a set of data sets containing the From a set of data sets including a specifying step of specifying an injection rate, the representative value of the turbidity for each group, and the specified injection rate, the representative value is used as an explanatory variable, and the specified injection a model generation step of generating a regression model with rate as the objective variable.

According to the above configuration, the regression model generated by the model generation method includes the turbidity of the influent before the injection of the flocculant, the consumption cost of the flocculant injected into the influent of the turbidity, and flocculation into the influent. Correspondence with the coagulant injection rate that minimizes the total value of the sludge treatment cost for treating the sludge generated by injecting the agent and the filtration treatment cost for backwashing when filtering the influent after sludge removal It becomes a regression model that shows the relationship. As a result, by estimating the injection rate of the coagulant into the influent using the regression model, the injection rate of the coagulant that minimizes the above total value can be estimated according to the turbidity of the influent. can. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration costs.

Further, in order to solve the above problems, an estimation method according to one aspect of the present invention is an estimation method executed by an estimation device for estimating the injection rate of a coagulant injected into influent water flowing into a water purification plant, , the turbidity of the influent before injecting the flocculant, the injection rate of the flocculant injected into the influent, the consumption cost of the flocculant injected into the influent, by injecting the flocculant into the influent A plurality of data sets classified based on the turbidity, including a sludge treatment cost for treating the generated sludge and a filtration treatment cost for backwashing when filtering the influent after removing the sludge A regression model with the representative value of the turbidity in each of the groups as an explanatory variable and the injection rate at which the total value of the consumption cost, the sludge treatment cost and the filtration treatment cost is the minimum as an objective variable; an estimating step of estimating the injection rate of the coagulant to be injected into the current influent based on the turbidity of the current influent, which is the influent currently flowing into the water treatment plant.

According to the above configuration, the turbidity of the influent before the injection of the flocculant, the consumption cost of the flocculant injected into the influent of the turbidity, and the treatment of the sludge generated by injecting the flocculant into the inflow A regression model showing the correspondence relationship between the coagulant injection rate that minimizes the total value of the sludge treatment cost and the filtration treatment cost of backwashing in the filtration treatment of influent water after sludge removal, and the current influent water Turbidity is used to estimate the flocculant injection rate into the current influent. Thereby, the injection rate of the flocculant that minimizes the total value can be estimated according to the turbidity of the influent water. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration costs.

Further, in order to solve the above problems, a model generation method according to an aspect of the present invention includes an injection rate of a coagulant injected into influent water flowing into a water purification plant, a consumption cost of the coagulant, the influent water The total value of the sludge treatment cost for treating the sludge generated by injecting the flocculant into the sludge and the filtration treatment cost for the backwash treatment when filtering the influent after the sludge is removed. A model generation method executed by a model generation device that generates a learning model for estimating a relationship, wherein the water quality data including the turbidity of the influent before injecting the flocculant, the injection rate, and the consumption A model generation step of generating a learning model that outputs the correspondence relationship with the water quality data as input by performing machine learning using the cost, the sludge treatment cost, and the filtration treatment cost as teacher data.

According to the above configuration, the learning model generated by the model generation method is input with the water quality data of the inflow, and the injection rate of the coagulant to be injected into the inflow and the injection rate. Correspondence with the total value of the consumption cost of the coagulant, the sludge treatment cost required to treat the sludge generated by injecting the coagulant into the influent, and the backwash treatment cost when the influent is filtered after the sludge is removed It becomes a learning model that outputs relationships. Accordingly, by using the correspondence output from the learning model, it is possible to specify the flocculant injection rate that minimizes the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration costs.

Further, in order to solve the above problems, an estimation method according to one aspect of the present invention includes an injection rate of a coagulant injected into influent water flowing into a water purification plant, a consumption cost of the coagulant, and Correspondence relationship between the total value of the sludge treatment cost for treating the sludge generated by injecting the coagulant and the filtration treatment cost for the backwashing treatment when the influent water after the sludge is removed is filtered An estimation method executed by an estimation device that estimates the water quality data including the turbidity of the influent before the coagulant is injected, the injection rate of the coagulant injected into the inflow, and the inflow Consumption cost of injected coagulant, sludge treatment cost for treatment of sludge generated by injecting coagulant into the influent, and backwashing when filtering the influent after removing the sludge Using a learning model generated by performing machine learning with the filtration cost for treatment and the training data, the current inflow that is the influent that is currently flowing into the water purification plant before injecting the coagulant An estimating step of estimating the correspondence relationship using water quality data including water turbidity as an input.

According to the above configuration, by inputting the water quality data of the influent, the injection rate of the coagulant to be injected into the influent, the consumption cost of the coagulant injected at the injection rate, and the injection of the coagulant. A learning model that outputs the correspondence relationship between the sludge treatment cost for treating the sludge generated by the process and the total value of the backwash filtration treatment cost for the filtration treatment of influent water after sludge removal, and the water quality data of the current influent water and to estimate the correspondence. Accordingly, using the estimated correspondence relationship, it is possible to specify the flocculant injection rate that minimizes the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration costs.

The model generating device and the estimating device according to each aspect of the present invention may be implemented by a computer. A control program for the model generating device and a control program for the estimating device that cause the model generating device and the estimating device to be realized by a computer, and a computer-readable recording medium recording them are also included in the scope of the present invention.

According to one aspect of the present invention, the cost of flocculation treatment for flocculating suspended solids in influent water with a flocculant, treatment of sludge generated in the flocculation treatment, and filtration treatment in the latter stage of the flocculation treatment is reduced. Water purification treatment can be realized.

BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows an example of the outline|summary of the chemical-feeding-rate estimation system which concerns on Embodiment 1 of this invention. It is a block diagram which shows an example of a principal part structure of the chemical-feeding-rate estimation system shown in FIG. 3 is a diagram showing an example of a regression model generated by the model generation device shown in FIG. 2; FIG. 3 is a flowchart showing an example of the flow of regression model generation processing executed by the model generation device shown in FIG. 2; 5 is a diagram showing an example of the relationship between pre-injection turbidity and sedimentation basin outlet turbidity included in the data set used in the regression model generation process shown in FIG. 4. FIG. 3 is a flowchart showing an example of the flow of estimation processing executed by the estimation device shown in FIG. 2; It is a figure which shows an example of the outline|summary of the chemical-feeding-rate estimation system which concerns on Embodiment 2 of this invention. FIG. 8 is a block diagram showing an example of the main configuration of the chemical-feeding rate estimation system shown in FIG. 7; FIG. FIG. 9 is a flowchart showing an example of the flow of learning model generation processing executed by the model generation device shown in FIG. 8; FIG. 9 is a flowchart showing an example of the flow of estimation processing executed by the estimation device shown in FIG. 8; In the modification of this invention, it is the schematic which shows an example of the river which a water purification plant water-intakes. It is a figure which shows the calculation example of the total precipitation contained in the data set used in the modification of this invention.

[Embodiment 1]

<Outline of chemical dosing rate estimation system>
Drawing 1 is a figure showing an example of an outline of chemical injection rate presumption system 100 concerning this embodiment. The chemical injection rate estimation system 100 calculates the chemical injection rate, which is the injection rate of the coagulant injected into the water flowing into the water purification plant (influent water, hereinafter referred to as “raw water”), from the raw water before the coagulant is injected. It is estimated according to the turbidity (hereinafter referred to as "pre-injection turbidity"). Specifically, the chemical dosing rate is the sum of the consumption cost of the coagulant, the cost of treating the sludge generated by the injection of the coagulant, and the cost of filtration including backwashing (hereinafter referred to as "cost is the dosing rate that minimizes the total value. Here, the flocculant is a chemical for flocculating and precipitating suspended substances contained in raw water. Also, the dosing rate is the ratio of the coagulant injected into the raw water to the raw water. Conventionally, the dosing rate is determined using the result of a jar test, past knowledge, a predetermined calculation formula, and the like. The injection of the coagulant according to the present disclosure may be the injection of the coagulant in the coagulation-sedimentation rapid filtration method, or the injection of the coagulant as a pretreatment in the membrane filtration method.

The consumption cost of the flocculant is the cost of the flocculant consumed by injecting it into the raw water. As an example, the amount of flocculant injected into the raw water and the purchase price (unit price) per unit amount of the flocculant. determined by the product of The cost of sludge treatment is the sum of the cost of thickening and dehydrating the sludge generated by injecting the coagulant and the cost of transporting the sludge outside the water purification plant and disposing of it. , can be calculated from the injection amount of the flocculant. The cost required for the filtration process is the cost required for the backwashing process when the raw water after the coagulant is injected is filtered, and for example, the cost of electricity required for the backwashing process. Hereinafter, the cost of sludge treatment will be referred to as "sludge treatment cost", and the cost of filtration treatment will be referred to as "filtration treatment cost".

(Generation of regression model)
The chemical injection rate estimation system 100 uses the pre-injection turbidity as an explanatory variable in order to realize the estimation of the chemical injection rate that minimizes the total cost value according to the pre-injection turbidity, and the total cost value becomes the minimum. Generate a regression model with chemical injection rate as the objective variable. Chemical dosing rate estimation system 100 accepts input of a set of data sets D1 including pre-injection turbidity, chemical dosing rate, consumption cost, sludge treatment cost and filtration cost, that is, a plurality of data sets D1 (in FIG. 1 a1). The data set D1 is the turbidity before injection of raw water that has flowed into the water purification plant in the past, the chemical dosing rate of the coagulant actually injected into the raw water, the consumption cost of the coagulant injected at the chemical dosing rate, This includes the sludge treatment cost for the sludge generated by the injection of the coagulant and the filtration treatment cost for backwashing when the raw water is filtered after the sludge is removed.

Subsequently, the chemical injection rate estimation system 100 classifies the set into a plurality of groups based on the pre-injection turbidity, and for each group, an intermediate regression model that is a regression model in which the chemical injection rate and the total cost value are variables. (b1 in FIG. 1).

Furthermore, the chemical dosing rate estimation system 100 uses the pre-injection turbidity of each intermediate regression model as an explanatory variable, and generates a regression model with the chemical dosing rate that minimizes the total cost value in each intermediate regression model as the objective variable. (b2 in FIG. 1).

(Estimation of chemical injection rate)
The chemical dosing rate estimation system 100 that generates the regression model receives an input of the pre-injection turbidity D2 of the raw water currently flowing into the water purification plant (current influent water, hereinafter sometimes referred to as "current raw water") ( a2) in FIG. Then, the chemical-feeding rate estimation system 100 estimates the chemical-feeding rate D3 using the input pre-injection turbidity D2 and the regression model (a3 in FIG. 1).

<Main configuration of chemical injection rate estimation system 100>
FIG. 2 is a block diagram showing an example of the main configuration of the chemical-feeding rate estimation system 100. As shown in FIG. The dosing rate estimation system 100 includes a model generation device 1 , an estimation device 2 , an input device 3 , an output device 4 and a storage device 5 .

The input device 3 receives a user's input and outputs an input signal based on the input to the model generating device 1 or the estimating device 2 .

The output device 4 outputs the information generated by the estimation device 2 . The output method by the output device 4 is not particularly limited. The output device 4 may be, for example, a display device that displays the information as an image, or an audio output device that outputs the information as audio. Also, the output device 4 may be a control signal output device that outputs the information as an electric signal to a coagulant injection device (not shown).

Storage device 5 holds programs and data used in chemical injection rate estimation system 100 . The data held (stored) by the storage device 5 will be described later.

(Principal configuration of model generation device 1)
The model generation device 1 generates a regression model for estimating the dosing rate of the coagulant to be injected into the raw water. The model generation device 1 has a control unit 10 . The control unit 10 centrally controls each unit of the model generation device 1 . The control unit 10 is realized by a processor and memory, for example. In this example, the processor accesses storage (not shown), loads a program (not shown) stored in the storage into memory, and executes a series of instructions contained in the program. Each unit included in the control unit 10 is thus configured.

As such units, the control unit 10 includes a dataset acquisition unit 101, a classification unit 102, an intermediate regression model generation unit 103 (identification unit, intermediate model generation unit), and a regression model generation unit 104 (model generation unit).

The data set acquisition unit 101 acquires a set of data sets (data set D1 in FIG. 1) including pre-injection turbidity, chemical dosing rate, consumption cost, sludge treatment cost, and filtration treatment cost. The data set acquisition unit 101 receives the set from the input device 3, for example. As another example, the set may be stored in the storage of the model generation device 1 or in the storage device 5 . In this example, the data set acquisition unit 101 reads the set from the storage or the storage device 5 according to the read instruction for the set received from the input device 3 . The dataset acquisition unit 101 outputs the acquired set to the classification unit 102 .

Classifier 102 classifies the set of data sets into a plurality of groups based on pre-injection turbidity. In other words, the classifier 102 generates groups for the set of data sets based on pre-injection turbidity. As an example, the classification unit 102 generates a plurality of groups, with a plurality of data sets having the same pre-injection turbidity as one group. As another example, the classification unit 102 presets a plurality of numerical ranges that do not overlap each other, and sets a plurality of data sets in which the pre-injection turbidity is within the same numerical range as one group. to generate The classification unit 102 associates group identification information for identifying each group with a set of data sets included in each group, and outputs the information to the intermediate regression model generation unit 103 .

The intermediate regression model generation unit 103 generates an intermediate regression model, which is a regression model having the injection rate and the total cost value as variables, for each of the plurality of groups. As an example, the intermediate regression model generation unit 103 generates a nonlinear regression model. The regression model generator 104 generates a nonlinear regression model using, for example, random forest (RF), gradient boosted tree regression (GBR), or neural network (NN).

(details of intermediate regression model)
FIG. 3 is a diagram showing an example of a regression model. FIG. 3 shows, by way of example, intermediate regression models S1, S2, S3 corresponding to three groups of pre-injection turbidity P1, P2, P3, respectively. The pre-injection turbidity P1, P2, P3 may be a single value as described above, or may be a range of values.

C1 shown in FIG. 3 is a curve showing the relationship between the total value of consumption cost and sludge treatment cost and the chemical feeding rate, and C2 is a curve showing the relationship between the filtration treatment cost and the chemical feeding rate. Regarding the curve C1, when the chemical feeding rate increases, the amount of coagulant injected into the raw water and the amount of sludge generated also increase, so the total value of the consumption cost and the sludge treatment cost increases as shown in FIG.

Regarding the curve C2, as shown in FIG. 3, the filtration cost decreases as the chemical feeding rate increases, and the filtration cost increases as the chemical feeding rate decreases. This relationship will be explained in detail.

When the amount of coagulant injected into the raw water (i.e., the chemical injection rate) increases, the amount of coagulation of suspended solids in the raw water increases, resulting in an increase in the amount of sludge generated and the amount of sludge that settles in the sedimentation basin. The amount of suspended solids remaining in the raw water that has passed through the pond (ie after desludging) is reduced. Conversely, when the amount of coagulant injected into the raw water is reduced, the amount of coagulation of suspended solids in the raw water is reduced. The amount of suspended solids remaining in the raw water that has passed through is increased. As the amount of suspended solids remaining in the raw water that has passed through the sedimentation tank increases, the frequency and time required for backwashing also increase, resulting in an increase in treatment costs (electricity costs) for backwashing. Conversely, when the amount of suspended solids remaining in the raw water that has passed through the sedimentation basin decreases, the frequency and time required for backwashing also decrease, resulting in a reduction in treatment costs (electricity costs) for backwashing. As the amount of sludge generated increases, the processing cost (electricity cost) of the dehydrator that operates to dehydrate the sludge increases, and the amount of dehydrated sludge discharged from the dehydrator also increases. The processing cost (disposal fee) for transporting the waste to the outside and disposing of it by landfilling or the like also increases. Conversely, when the amount of generated sludge is reduced, the operation time of the dehydrator and the amount of dehydrated sludge are reduced, and as a result, the treatment cost (electricity cost + disposal cost) required for sludge treatment is reduced.

Since the relationship between the total consumption cost and sludge treatment cost and the chemical charging rate and the relationship between the filtration cost and the chemical charging rate are as described above, the relationship between the total cost and the chemical charging rate, that is, Intermediate regression models S1, S2, and S3 can be expressed as functions whose vertices are the minimum total cost values, as in the example shown in FIG.

Returning to the description of the intermediate regression model generating unit 103 with reference to FIG. 2 again. The intermediate regression model generation unit 103 specifies the chemical injection rate that minimizes the total cost value for each of the generated intermediate regression models. In the example shown in FIG. 3, for each of the intermediate regression models S1, S2, and S3, the chemical feeding rates M1, M2, and M3 when the total cost value is minimized are specified.

The intermediate regression model generation unit 103 creates a data set in which the specified chemical injection rate and the representative value of the pre-injection turbidity in each group, in other words, the representative value of the pre-injection turbidity in each intermediate regression model are associated with each other. is output to the regression model generation unit 104 . When the pre-injection turbidity in each group is one value, the representative value is the one value. Moreover, when the pre-injection turbidity in each group is within a numerical range, the representative value may be, for example, the median value of the numerical range, or the minimum or maximum value.

The regression model generation unit 104 generates a regression model 51 with the representative value as the explanatory variable and the chemical feeding rate as the objective variable from the acquired data set including the representative value and the chemical feeding rate. The regression model generation unit 104 generates a linear regression model 51 as an example. The regression model 51 corresponds to the regression line L1 with the minimum distance to each vertex of the intermediate regression model shown in FIG. The regression model generation unit 104 generates a linear regression model using, for example, generalized linear regression (GLR) or partial least squares regression (PLS). The regression model generator 104 stores the generated regression model 51 in the storage device 5 .

(Principal configuration of estimation device 2)
The estimation device 2 uses the regression model generated by the model generation device 1 to estimate the dosing rate of the coagulant to be injected into the current raw water. The estimation device 2 includes a control section 20 . The control unit 20 centrally controls each unit of the estimation device 2 . The control unit 20 is implemented by a processor and memory, for example. In this example, the processor accesses storage (not shown), loads a program (not shown) stored in the storage into memory, and executes a series of instructions contained in the program. Each unit included in the control unit 20 is thus configured.

As the respective units, the control unit 20 includes a pre-injection turbidity acquisition unit 201 and a chemical injection rate estimation unit 202 (estimation unit).

The pre-injection turbidity acquisition unit 201 acquires the pre-injection turbidity of the current raw water (the pre-injection turbidity D2 in FIG. 1). As an example, the pre-injection turbidity acquisition unit 201 receives the pre-injection turbidity of the current raw water from the input device 3 . The pre-injection turbidity acquisition unit 201 outputs the acquired pre-injection turbidity to the chemical injection rate estimation unit 202 .

The chemical dosing rate estimation unit 202 estimates the dosing rate of the coagulant for the current raw water based on the pre-injection turbidity acquired from the pre-injection turbidity acquisition unit 201 and the regression model 51 . After acquiring the pre-injection turbidity, the chemical injection rate estimating unit 202 reads out the regression model 51 from the storage device 5 . Then, the chemical-feeding rate estimating unit 202 acquires the chemical-feeding rate output from the regression model 51 by inputting the pre-injection turbidity to the read regression model 51 . The chemical dosing rate is a chemical dosing rate that minimizes the sum of consumption cost, sludge treatment cost, and filtration treatment cost in water purification treatment for removing suspended solids in the current raw water with the input turbidity before injection. The chemical-feeding-rate estimation part 202 outputs the acquired chemical-feeding rate to the output device 4 as an estimation result of the chemical-feeding rate. When the output device 4 is the control signal output device described above, the coagulant injection device can inject the coagulant into the current raw water based on the chemical injection rate.

<Regression model generation process flow>
FIG. 4 is a flowchart showing an example of the flow of regression model generation processing executed by the model generation device 1. As shown in FIG.

The data set acquisition unit 101 calculates the turbidity before injection of raw water, the chemical injection rate of the coagulant actually injected into the raw water, the consumption cost of the coagulant injected at the chemical injection rate, and the injection of the coagulant. Acquire a set of data sets including the sludge treatment cost of generated sludge and the filtration treatment cost required for backwashing in the filtration treatment of the raw water after sludge removal (step S1, hereinafter description of "step" is omitted). The dataset acquisition unit 101 outputs the acquired set to the classification unit 102 .

Figure 5 shows the turbidity before injection and the turbidity of the raw water at the outlet of the sedimentation basin of the water purification plant after injecting the coagulant into the raw water with the turbidity before injection (turbidity at the sedimentation basin outlet), which is included in the above data set. It is a figure which shows an example of the relationship with. The sedimentation basin outlet turbidity is the turbidity of raw water after sludge is removed in the sedimentation basin.

As an example, the data set acquisition unit 101 uses only data sets in which the sedimentation pond outlet turbidity is equal to or less than a predetermined value (within a predetermined range) as data sets to be used for the regression model generation process. For example, the data set acquisition unit 101 uses only the data sets in which the sedimentation pond outlet turbidity is equal to or less than the threshold value T1 shown in FIG. 5 to generate the regression model. A specific value of the threshold T1 may be "1 (degrees)" shown in FIG. 5, but is not limited to this example.

As an example, the data set acquisition unit 101 may discard data sets in which the sedimentation pond outlet turbidity exceeds the threshold T1 from the set input from the input device 3 based on the threshold T1 stored in the storage. In addition, as an example, the data set acquisition unit 101 reads only data sets in which the sedimentation tank outlet turbidity is equal to or less than the threshold T1 from the storage (or the storage device 5) based on the threshold T1 stored in the storage. good. For these examples, the data set includes the settling basin outlet turbidity.

Further, as an example, the user of the model generating device 1 may input to the model generating device 1 only data sets in which the sedimentation pond outlet turbidity is equal to or less than the threshold value T1.

Referring back to FIG. 4, the description returns to the flow of the regression model generation process. The classification unit 102 classifies the data sets included in the acquired set into a plurality of groups based on the pre-injection turbidity (S2, classification step). The classification unit 102 associates group identification information for identifying each group with a set of data sets included in each group, and outputs the information to the intermediate regression model generation unit 103 .

The intermediate regression model generation unit 103 generates an intermediate regression model using the chemical injection rate and the total cost value as variables for each group (S3). Then, the intermediate regression model generation unit 103 identifies the chemical injection rate that minimizes the total cost value in each of the generated intermediate regression models (S4, identification step). Intermediate regression model generation unit 103 outputs to regression model generation unit 104 a set of data sets in which the identified chemical injection rate and the representative value of pre-injection turbidity in each group are associated with each other.

The regression model generation unit 104 generates a regression model 51 from the set obtained from the intermediate regression model generation unit 103, with the representative value as the explanatory variable and the chemical injection rate specified by the intermediate regression model generation unit 103 as the objective variable. (S5, model generation step). The regression model generator 104 stores the regression model 51 in the storage device 5 .

(Update of regression model 51)
It is preferable that the control unit 10 periodically update (automatically update) the regression model 51 . In this example, datasets are stored in the storage or storage device 5 of the model generation device 1 , and the dataset acquisition unit 101 acquires a set of datasets from the storage or storage device 5 .

Even after the regression model 51 is generated, the data set is stored in the storage or storage device 5 each time the raw water is purified.

As an example, when the data set acquisition unit 101 acquires a set of data sets and outputs them to the classification unit 102, it starts a timer (not shown) and measures the elapsed time from the execution of the above acquisition and output. . Then, when the elapsed time reaches a predetermined value, the dataset acquisition unit 101 acquires a set of datasets again and outputs them to the classification unit 102 . Since the subsequent processing is the same as the processing of S2 to S5 of the regression model generation processing described above, the description will not be repeated here.

<Flow of chemical dosing rate estimation processing>
FIG. 6 is a flowchart showing an example of the flow of the chemical-feeding-rate estimation process executed by the estimation device 2. As shown in FIG.

The pre-injection turbidity acquisition unit 201 acquires the pre-injection turbidity of the current raw water from the input device 3 (S11). The pre-injection turbidity acquisition unit 201 outputs the acquired pre-injection turbidity to the chemical injection rate estimation unit 202 .

After acquiring the pre-injection turbidity, the chemical injection rate estimating unit 202 reads out the regression model 51 from the storage device 5 . Then, the chemical-feeding-rate estimating unit 202 estimates the chemical-feeding rate using the pre-injection turbidity and the regression model 51 (S12, estimation step). Specifically, the chemical injection rate estimating unit 202 acquires the chemical injection rate output from the regression model 51 by inputting the acquired pre-injection turbidity to the read regression model 51 . The chemical-feeding rate estimation unit 202 causes the output device 4 to output the acquired chemical-feeding rate as the chemical-feeding rate estimation result (S13).

<effect>
As described above, the model generation device 1 according to the present embodiment includes the turbidity before injection of raw water, the chemical dosing rate of the coagulant actually injected into the raw water, and the chemical dosing rate injected at the chemical dosing rate. Data set including consumption cost of coagulant, sludge treatment cost of sludge generated by injecting the coagulant, and filtration cost for backwashing when filtering the raw water after sludge removal The population is sorted into groups based on pre-injection turbidity. Then, the model generation device 1 specifies the chemical injection rate that minimizes the total cost value for each group. Then, the model generation device 1 uses the representative value as an explanatory variable from a set of data sets including the representative value of the pre-injection turbidity and the specified chemical injection rate for each group, and the specified chemical injection Generate a regression model 51 with rate as the objective variable.

Also, the estimation device 2 according to the present embodiment estimates the dosing rate of the coagulant for the current raw water based on the regression model 51 and the pre-injection turbidity of the current raw water.

According to this configuration, the regression model 51 generated by the model generation device 1 defines the correspondence relationship between the pre-injection turbidity and the chemical dosing rate that minimizes the total cost value in the water purification treatment of the raw water with the pre-injection turbidity. The regression model shown is As a result, the estimating device 2 estimates the chemical dosing rate using the regression model 51, thereby estimating the chemical dosing rate that minimizes the total cost value in the current raw water purification treatment according to the pre-injection turbidity. be able to. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration treatment costs.

In addition, the model generation device 1 generates an intermediate model, which is a regression model with the chemical injection rate and the total cost value as variables, for each group, and uses the intermediate model to generate the chemical injection that minimizes the total cost value. It is preferable to estimate the rate. As a result, it is possible to estimate the chemical-feeding rate that minimizes the total cost value by considering the correspondence relationship between the chemical-feeding rate and the total cost value. can be improved.

Moreover, it is preferable that the model generation device 1 generates the regression model 51 using only data sets in which the sedimentation basin outlet turbidity is equal to or less than the threshold value T1. As a result, the regression model 51 can estimate the chemical dosing rate at which the sedimentation tank outlet turbidity is equal to or less than the threshold value T1. By estimating the chemical dosing rate using the regression model 51, it is possible to estimate the chemical dosing rate that will make the turbidity of the raw water at the sedimentation basin outlet less than or equal to the threshold value T1. By injecting the flocculant at the chemical dosing rate, the possibility of highly turbid water flowing out from the sedimentation basin and further from the water purification plant is reduced. As a result, the turbidity of the filtered water in the rapid filtration basin in the latter stage has been reduced, the filtration cost has been further reduced by reducing the frequency of backwashing, and the risk of infection with cryptosporidium, etc. due to drinking the discharged water has been reduced. can do.

[Embodiment 2]
Other embodiments of the invention are described below. For convenience of description, members having the same functions as those of the members described in the above embodiments are denoted by the same reference numerals, and description thereof will not be repeated.

<Outline of chemical dosing rate estimation system>
Drawing 7 is a figure showing an example of an outline of chemical injection rate presumption system 200 concerning this embodiment. The chemical dosing rate estimation system 200 estimates the chemical dosing rate of the coagulant for the raw water, specifically the chemical dosing rate that minimizes the total cost value, according to the water quality of the raw water before the coagulant is injected.

(Generation of learning model)
The chemical feeding rate estimation system 200 receives the water quality data of the raw water as input, and calculates the chemical feeding rate and the total cost value in order to estimate the chemical feeding rate that minimizes the total cost value according to the quality of the raw water. Generate a learning model that outputs correspondences. The chemical dosing rate estimation system 200 receives inputs of a plurality of data sets D11 including raw water quality data, chemical dosing rate, consumption cost, sludge treatment cost and filtration cost (a11 in FIG. 7). The data set D11 includes data on the quality of raw water that has flowed into the water purification plant in the past, the chemical dosing rate of the coagulant actually injected into the raw water, the consumption cost of the coagulant injected at the chemical dosing rate, the It includes the sludge treatment cost of the sludge generated by the injection of the coagulant and the filtration treatment cost required for backwashing in the filtration treatment of the raw water after sludge removal. The water quality data is data including pre-injection turbidity. However, data included in water quality data is not limited to pre-injection turbidity. Water quality data may include pH, water temperature, etc., for example, as shown in FIG.

Subsequently, the chemical-feeding rate estimation system 200 generates the learning model by performing machine learning using the plurality of data sets D11 as teacher data (b11 in FIG. 7). The correspondence between the chemical charge rate and the total cost value is, for example, the chemical charge rate is the dependent variable and the total cost value is the independent variable (or the chemical charge rate is the independent variable and the total cost value is the dependent variable). It can be expressed as a function.

(Estimation of chemical injection rate)
The chemical dosing rate estimation system 200 that has generated the learning model receives the input of the current raw water quality data D12 (a12 in FIG. 7). Then, the chemical-feeding rate estimation system 200 uses the input water quality data D12 and the learning model to estimate the correspondence relationship between the chemical-feeding rate and the total cost value (b12 in FIG. 7).

Finally, the chemical-feeding-rate estimation system 200 specifies the chemical-feeding rate D13 with the minimum total cost value from the estimated correspondence (a3 in FIG. 7).

<Main configuration of chemical injection rate estimation system 200>
FIG. 8 is a block diagram showing an example of the main configuration of the chemical-feeding rate estimation system 200. As shown in FIG. In the following, among the members of the chemical-feeding-rate estimation system 200, only the differences from the same-named members of the chemical-feeding-rate estimation system 100 will be described.

The chemical-feeding rate estimation system 200 differs from the chemical-feeding rate estimation system 100 in that it includes a model generation device 11 , an estimation device 12 and a storage device 15 .

(Principal Configuration of Model Generating Device 11)
The control unit 110 of the model generation device 11 includes a dataset acquisition unit 111 and a learning model generation unit 112 (model generation unit).

Data set acquisition unit 111 acquires a set of data sets (data set D11 in FIG. 7) including water quality data, chemical dosing rate, consumption cost, sludge treatment cost, and filtration treatment cost, and outputs them to learning model generation unit 112. .

The learning model generation unit 112 performs machine learning based on the set acquired from the data set acquisition unit 111 to generate a learning model 510 . The learning model generation unit 112 stores the generated learning model 510 in the storage device 15 . When the current raw water quality data is input to the learning model 510, the correspondence relationship between the chemical dosing rate and the total cost value is output. Note that the learning model generation unit 112 generates the learning model 510 using a known machine learning method such as a neural network (NN), for example.

(Principal configuration of estimation device 12)
The control unit 210 of the estimation device 12 includes a correspondence estimation unit 211 (estimation unit) and a chemical injection rate identification unit 212 .

Based on the input water quality data of the raw water (water quality data D12 in FIG. 7) and the learning model 510, the correspondence estimation unit 211 calculates the dosing rate of the coagulant for the raw water and the cost of water purification of the raw water. Estimate the correspondence with the total value. As an example, the correspondence estimation unit 211 receives an input of current raw water quality data from the input device 3 . The correspondence estimation unit 211 reads the learning model 510 from the storage device 15 upon receiving the input of the water quality data. Then, the correspondence estimation unit 211 acquires the correspondence between the chemical dosing rate and the total cost value output from the learning model 510 by inputting the acquired water quality data to the read learning model 510 . The correspondence estimation unit 211 outputs the obtained correspondence to the chemical-feeding rate identification unit 212 .

The chemical-feeding rate identification unit 212 identifies the chemical-feeding rate that minimizes the total cost value based on the correspondence relationship between the chemical-feeding rate and the total cost value. The corresponding relationship is represented by curves having the same shapes as those of the intermediate regression models S1, S2, and S3 shown in FIG. 3, for example. Therefore, the chemical dosing rate that minimizes the total cost value is uniquely determined. When the chemical charging rate specifying unit 212 specifies the chemical charging rate that minimizes the total cost value in the correspondence acquired from the correspondence estimating unit 211, the chemical charging rate is output to the output device 4 as the chemical charging rate estimation result. output to

<Flow of learning model generation processing>
FIG. 9 is a flowchart showing an example of the flow of learning model generation processing executed by the model generation device 11. As shown in FIG.

Data set acquisition unit 111 acquires a set of data sets including water quality data, chemical dosing rate, consumption cost, sludge treatment cost, and filtration treatment cost (S21), and outputs them to learning model generation unit 112 . As an example, the data set acquisition unit 111 uses only data sets in which the sedimentation pond outlet turbidity is equal to or less than a predetermined value in the learning model generation process, similar to the data set acquisition unit 101 described in the first embodiment. do.

The learning model generation unit 112 performs machine learning based on the set acquired from the data set acquisition unit 111, inputs the current raw water quality data, and outputs a learning model 510 that outputs the correspondence relationship between the chemical dosing rate and the total cost value. (S22, model generation step). The learning model generation unit 112 stores the generated learning model 510 in the storage device 15 .

(Update of learning model 510)
Control unit 110 preferably updates (automatically updates) learning model 510 on a regular basis. In this example, datasets are stored in the storage of the model generation device 11 or the storage device 15 , and the dataset acquisition unit 111 acquires a set of datasets from the storage or storage device 15 .

Even after the learning model 510 is generated, the data set is stored in the storage or storage device 15 each time the raw water is purified.

As an example, when the data set acquisition unit 111 acquires a set of data sets and outputs them to the learning model generation unit 112, it starts a timer (not shown), and the elapsed time from the execution of the acquisition and output is measure. Then, when the elapsed time reaches a predetermined value, the data set acquisition unit 111 acquires a set of data sets again and outputs them to the learning model generation unit 112 . The learning model updating process executed by the learning model generation unit 112 is the same as the process of S22 in FIG. 9, and thus description thereof will not be repeated here.

<Flow of estimation process>
FIG. 10 is a flowchart showing an example of the flow of estimation processing executed by the estimation device 12. As shown in FIG.

The correspondence estimation unit 211 acquires the water quality data of the current raw water (S31), and reads the learning model 510 from the storage device 15. FIG. Then, based on the acquired water quality data and the learning model 510, the correspondence estimation unit 211 estimates the correspondence between the dosing rate of the coagulant for the current raw water and the total cost value in the water purification treatment of the current raw water ( S32, estimation step). Specifically, the correspondence estimation unit 211 inputs the acquired water quality data to the read learning model 510, thereby acquiring the correspondence between the chemical dosing rate and the total cost value output from the learning model 510. . The correspondence estimation unit 211 outputs the obtained correspondence to the chemical-feeding rate identification unit 212 .

The chemical-feeding rate identification unit 212 identifies the chemical-feeding rate that minimizes the total cost value in the acquired correspondence (S33). For example, when the obtained correspondence relationship can be expressed as a quadratic function with the minimum total cost value as a vertex, the chemical-feeding-rate identifying unit 212 identifies the chemical-feeding-rate value at the vertex. And the chemical-feeding-rate specific|specification part 212 outputs the specified chemical-feeding rate to the output device 4 as an estimation result of a chemical-feeding rate (S34).

<effect>
As described above, the model generation device 11 according to the present embodiment includes water quality data including pre-injection turbidity of raw water, the dosing rate of the coagulant actually injected into the raw water, and Consumption cost of the injected flocculant, sludge treatment cost of the sludge generated by injecting the flocculant, and filtration processing cost for backwashing when filtering the raw water after sludge removal. perform machine learning as Then, the model generation device 11 receives the water quality data and generates a learning model 510 that outputs the correspondence relationship between the chemical dosing rate and the total cost value.

In addition, the estimation device 12 according to the present embodiment calculates, based on the learning model 510 and the water quality data of the current raw water, the chemical dosing rate of the coagulant for the current raw water and the total cost value in the water purification treatment of the current raw water. Estimate correspondence.

According to this configuration, using the correspondence estimated by the estimation device 12, it is possible to specify the chemical injection rate that minimizes the total cost value. As a result, it is possible to realize water purification treatment with reduced consumption costs, sludge treatment costs, and filtration treatment costs.

Moreover, it is preferable that the estimating device 12 according to the present embodiment specifies the chemical injection rate that minimizes the total cost value based on the estimated correspondence. This saves the user of the estimation device 12 the trouble of specifying the chemical injection rate that minimizes the total cost value.

[Modification]
The estimating devices 2 and 12 may calculate the consumption cost, sludge treatment cost, and filtration treatment cost required for purification of the current raw water based on the estimated chemical dosing rate, and cause the output device 4 to output them. In this example, the estimating devices 2 and 12 acquire data necessary for calculating the consumption cost (for example, the unit price of the coagulant used, the amount of raw water into which the coagulant is injected, etc.), and the data and the estimated chemical feeding rate Calculate the consumption cost based on Also, the estimation devices 2 and 12 calculate the sludge treatment cost from the estimated chemical dosing rate. In addition, the estimation devices 2 and 12 specify the total cost value corresponding to the estimated chemical dosing rate, and subtract the calculated consumption cost and sludge treatment cost value from the total cost value, thereby calculating the filtration treatment cost. calculate.

The data set acquired by the model generating device 11 includes precipitation data on precipitation (rainfall) in a predetermined area before the raw water flows into the water purification plant, that is, during the period when the raw water was present in the water intake target (river, etc.). You can stay. Here, the predetermined area is an area where there is a possibility that precipitation will flow into the water purification plant. ). A plurality of predetermined areas may be provided. Also, rainfall data will be described as being the total amount of rainfall (per hour) in a predetermined area, that is, the total amount of rainfall. Note that the definition of total precipitation is not limited to this example. For example, total precipitation may be the total amount of precipitation per day.

FIG. 11 is a schematic diagram showing an example of a river 99 from which the water purification plant 90 takes in water, and is a diagram showing the predetermined area. In addition, in FIG. 11, in order to make the explanation easier to understand, it is assumed that there are three predetermined regions (regions R1, R2, and R3). A data set generation device (not shown) that generates a data set, for example, communicates with a weather information server (not shown) based on an operation input by a worker, to get As another example, the data set generation device obtains hourly precipitation amounts for various locations, and then determines whether the hourly precipitation amounts for regions R1, R2, and R3 are included in the precipitation amounts. You can judge. The local weather information server is the server of the Japan Meteorological Agency or a company that provides weather forecasts.

FIG. 12 is a diagram illustrating an example of calculation of total precipitation. The example of FIG. 12 shows a calculation example of the total precipitation contained in the raw water that flowed into the water purification plant 90 at 10:30 on a certain day. The time at which the raw water was flowing through regions R1, R2 and R3 is before 10:30. In other words, there is a time lag between the time the raw water entered the water purification plant 90 and the time it flowed through the areas R1, R2 and R3. The data set generator calculates the total precipitation considering this time lag.

Specifically, in each of the regions R1, R2, and R3, the time required for the raw water to reach the water purification plant 90 from each region (that is, the time lag) is specified in advance. In the example of FIG. 12, the time is 60 minutes for region R1, 300 minutes for region R2, and 120 minutes for region R3.

The data set generation device calculates the amount of rainfall in each region from 10:30, the time when raw water flowed into the water purification plant 90, going back in time by the time lag. In FIG. 12, this point in time is referred to as "precipitation confirmation time". In the example of FIG. 12, the data set generation device acquires the amount of precipitation in region R1 at 9:30. Also, the data set generation device acquires the amount of precipitation in region R2 at 5:30. Also, the data set generation device acquires the amount of precipitation in region R3 at 8:30. In FIG. 12, these precipitation amounts are referred to as "regional precipitation amounts". In the example of FIG. 12, the regional precipitation amount for region R1 is 6 (mm/h), the regional precipitation amount for region R2 is 5 (mm/h), and the regional precipitation amount for region R3 is 4 (mm/h). was assumed to be

The data set generator sums up the obtained regional precipitation amounts to calculate the total precipitation amount. In the example of FIG. 12, the total precipitation is 6+5+4=15 (mm/h).

The dataset generator includes the calculated total rainfall in the dataset for the incoming raw water at 10:30. As a result, the learning model generated by the model generating device 11 takes into consideration the precipitation contained in the raw water.

The data set may also include data on the effluent from the dam that may flow into the raw water (eg, dam effluent). In this example, the time required for the water discharged from the dam to reach the water purification plant 90 (that is, the time lag) is specified in advance. The data set generation device identifies the amount of water discharged at the time lag before the time when the raw water flowed into the water purification plant 90, and includes it in the water quality data of the raw water that flowed in at 10:30.

In addition, the water quality data includes the alkalinity of the raw water, the injection rate of sodium hypochlorite actually injected into the raw water before the coagulant injection, and the caustic soda actually injected into the raw water before the coagulant injection. may include the injection rate of

In Embodiment 1, the model generating device 1 and the estimating device 2 may be configured as an integrated device. Similarly, in the second embodiment, the model generating device 11 and the estimating device 12 may be configured as an integrated device.

In the first embodiment, the estimation device 2 may receive the regression model 51 generated by the model generation device 1 from the model generation device 1 and store it in its own storage. Similarly, in the second embodiment, the estimation device 12 may receive the learning model 510 generated by the model generation device 11 from the model generation device 11 and store it in its own storage.

In generating the regression model 51 and the learning model 510, the model generators 1 and 11 may include data sets in which the sedimentation tank outlet turbidity exceeds the threshold value T1 in the set of data sets.

In Embodiment 2, the model generation device 11 may generate a program that specifies the chemical injection rate that minimizes the total cost value from the generated learning model 510 . Each unit of the control unit 210 of the estimation device 12 is configured by executing the program in the estimation device 12 .

[Example of realization by software]
Control blocks (especially control units 10 and 110 and control units 20 and 210) of the model generating devices 1 and 11 and the estimating devices 2 and 12 are logic circuits (hardware) formed in integrated circuits (IC chips) or the like. software) or software.

In the latter case, the model generating devices 1 and 11 and the estimating devices 2 and 12 are provided with computers that execute program instructions, which are software that implements each function. This computer includes, for example, one or more processors, and a computer-readable recording medium storing the program. In the computer, the processor reads the program from the recording medium and executes it, thereby achieving the object of the present invention. As the processor, for example, a CPU (Central Processing Unit) can be used. As the recording medium, a "non-temporary tangible medium" such as a ROM (Read Only Memory), a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. In addition, a RAM (Random Access Memory) for developing the above program may be further provided. Also, the program may be supplied to the computer via any transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. Note that one aspect of the present invention can also be implemented in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

The present invention is not limited to the above-described embodiments, but can be modified in various ways within the scope of the claims, and can be obtained by appropriately combining technical means disclosed in different embodiments. is also included in the technical scope of the present invention.

Reference Signs List 1, 11 model generation device 2, 12 estimation device 51 regression model 102 classification unit 103 intermediate regression model generation unit (specification unit, intermediate model generation unit)
104 Regression model generator (model generator)
202 Chemical dosing rate estimation unit (estimation unit)
112 learning model generation unit (model generation unit)
211 correspondence estimation unit (estimation unit)
510 learning model

Claims (5)

A model generation device for generating a regression model for estimating the injection rate of a flocculant to be injected into influent water flowing into a water purification plant,
The turbidity of the influent before injecting the coagulant, the injection rate of the coagulant injected into the influent, the consumption cost of the coagulant injected into the influent, and the injection of the coagulant into the influent The turbidity a classifier that classifies into a plurality of groups based on
Using the intermediate model generated by the intermediate model generation unit that generates an intermediate model that is a regression model with the consumption cost, the total value of the sludge treatment cost and the filtration treatment cost, and the injection rate as variables, a specifying unit that specifies, for each group, the injection rate that minimizes the total value;
From a set of data sets containing the representative value of turbidity and the specified injection rate for each group, a regression model is generated with the representative value as an explanatory variable and the specified injection rate as an objective variable. A model generation device, comprising: a model generation unit for generating. 2. The model generation device according to claim 1, wherein said model generation unit generates said regression model from said set including only data sets in which turbidity of said influent at a sedimentation basin outlet is within a predetermined range. An estimating device for estimating the injection rate of a coagulant injected into influent water flowing into a water purification plant,
The turbidity of the influent before injecting the flocculant, the injection rate of the flocculant injected into the influent, the consumption cost of the flocculant injected into the influent, the cost generated by injecting the flocculant into the influent A plurality of data sets classified based on the turbidity, including the sludge treatment cost for treating the sludge and the filtration treatment cost for backwashing when filtering the influent after the sludge is removed An intermediate model that is a regression model with the representative value of the turbidity in each group as an explanatory variable, and the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost, and the injection rate as variables. A regression model with the injection rate that minimizes the total value as an objective variable, which is specified by the specifying unit using the intermediate model generated by the intermediate model generating unit, and the inflow currently flowing into the water purification plant An estimating device comprising an estimating unit for estimating an injection rate of a coagulant to be injected into the current influent based on the turbidity of the current influent, which is water. A model generation method executed by a model generation device for generating a regression model for estimating the injection rate of a coagulant injected into influent water flowing into a water purification plant,
The turbidity of the influent before injecting the coagulant, the injection rate of the coagulant injected into the influent, the consumption cost of the coagulant injected into the influent, and the injection of the coagulant into the influent The turbidity a classification step of classifying into a plurality of groups based on
For each group, using an intermediate model that is a regression model generated using the total value of the consumption cost, the sludge treatment cost and the filtration treatment cost, and the injection rate as variables, the total value is the minimum a determining step of determining the injection rate to be
From a set of data sets containing the representative value of turbidity and the specified injection rate for each group, a regression model is generated with the representative value as an explanatory variable and the specified injection rate as an objective variable. a model generation step to generate; and a model generation method. An estimation method performed by an estimation device for estimating the injection rate of a coagulant injected into influent water flowing into a water purification plant,
The turbidity of the influent before injecting the flocculant, the injection rate of the flocculant injected into the influent, the consumption cost of the flocculant injected into the influent, the cost generated by injecting the flocculant into the influent A plurality of data sets classified based on the turbidity, including the sludge treatment cost for treating the sludge and the filtration treatment cost for backwashing when filtering the influent after the sludge is removed An intermediate model that is a regression model with the representative value of the turbidity in each group as an explanatory variable, and the total value of the consumption cost, the sludge treatment cost, and the filtration treatment cost, and the injection rate as variables. A regression model with the injection rate that minimizes the total value as an objective variable, which is specified by the specifying unit using the intermediate model generated by the intermediate model generating unit, and the inflow currently flowing into the water purification plant an estimating step of estimating an injection rate of a flocculant to be injected into the current influent based on the turbidity of the current influent, which is water.

Images