Sludge moisture content estimation system, sludge moisture content estimation device, sludge moisture content estimation method, sludge moisture content estimation program, database generator, database generation method, and optimal model database

Abstract

To provide a sludge moisture content estimation system used to estimate sludge moisture content with high accuracy after dehydration treatment under arbitrary treatment conditions.SOLUTION: A sludge moisture content estimation system 1 has a first file storage unit 621, an optimal model information storage unit 623, and a first estimation unit 614. The first file storage unit 621 stores files of multiple types of learning models for estimating the sludge moisture content after the sludge dehydration machine has performed the sludge dehydration process. The optimal model information storage unit 623 stores an optimal model database that contains identification information of optimal learning models for each sludge dehydration treatment condition. When an estimation process of sludge water content is requested, the first estimation unit 614 acquires the identification information of the optimal learning model to be used for this estimation process from the optimal model database, acquires the file of the optimal learning model corresponding to the acquired identification information, and executes the estimation process by inputting the dehydration processing conditions into this optimal learning model so as to obtain the estimated value of sludge moisture content.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a sludge moisture content estimation system, a sludge moisture content estimation device, a sludge moisture content estimation method, a sludge moisture content estimation program, a database creation device, a database creation method, and an optimum model database. [Background technology]

[0002] In sewage treatment plants, sludge is dehydrated using a dehydrator to reduce its volume and make subsequent treatment easier. In order to reduce the sludge to a desired volume, it is necessary to predict the water content of the sludge after dehydration based on the treatment conditions and control the dehydrator based on this prediction result.

[0003] In order to predict the moisture content of sludge after dehydration, there is a technology that constructs a calculation model to predict the moisture content of sludge after dehydration from the properties of the sludge and the operation data of the dehydrator. By using this calculation model, it is possible to accurately predict the moisture content of sludge when dehydration is performed under specified conditions (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 7187644 Summary of the Invention [Problem to be solved by the invention]

[0005] However, the above-mentioned calculation model was constructed for a specific dehydrator, and there was a problem in that it was not possible to predict the sludge moisture content when dehydration treatment was performed under arbitrary treatment conditions, such as when using another dehydrator.

[0006] Therefore, an object of the present invention is to provide a sludge moisture content estimation system, a sludge moisture content estimation device, a sludge moisture content estimation method, a sludge moisture content estimation program, a database generation device, a database generation method, and an optimal model database that are used to estimate with high accuracy the sludge moisture content after dewatering treatment is performed under any treatment conditions. [Means for solving the problem]

[0007] A sludge moisture content estimation system according to one embodiment of the present invention includes a first file memory unit that stores files of multiple types of learning models for estimating the sludge moisture content after a sludge dewatering process is performed by a sludge dewatering machine under arbitrary dewatering conditions; an optimal model information memory unit that stores an optimal model database that stores identification information of an optimal learning model identified from the multiple types of learning models for each sludge dewatering condition; an estimation request processing unit that, when a sludge moisture content estimation process is requested, acquires dewatering conditions for the sludge dewatering process that are the execution conditions of the estimation process; a first estimation unit that acquires identification information of an optimal learning model to be used in the sludge moisture content estimation process from the optimal model database based on the dewatering conditions acquired by the estimation request processing unit, acquires a file of the optimal learning model corresponding to the acquired identification information from the first file memory unit, and executes the sludge moisture content estimation process by inputting the dewatering conditions acquired by the estimation request processing unit into the optimal learning model to acquire an estimated value of the sludge moisture content; and an output control unit that outputs the estimated value acquired by the first estimation unit from an output device.

[0008] A sludge moisture content estimation device according to one aspect of the present invention is communicatively connected to a first file storage unit that stores files of multiple types of learning models for estimating the moisture content of sludge after a sludge dewatering process is performed by a sludge dewatering machine under arbitrary dewatering conditions, and an optimum model information storage unit that stores an optimum model database that stores identification information of an optimum learning model specified from the multiple types of learning models for each sludge dewatering condition, and when a sludge moisture content estimation process is requested, an estimation device that acquires the dewatering process conditions of the sludge dewatering process, which are the execution conditions of the estimation process. The system comprises a request processing unit, a first estimation unit that acquires, from the optimal model database, identification information of an optimal learning model to be used in the estimation process of the sludge moisture content based on the dewatering treatment conditions acquired by the estimation request processing unit, acquires a file of the optimal learning model corresponding to the acquired identification information from the first file memory unit, and executes the estimation process of the sludge moisture content by inputting the dewatering treatment conditions acquired by the estimation request processing unit into the optimal learning model to acquire an estimated value of the sludge moisture content, and an output control unit that outputs the estimated value acquired by the first estimation unit from an output device.

[0009] In addition, a sludge moisture content estimation method according to one embodiment of the present invention includes an estimation device communicatively connected to a first file memory unit that stores files of multiple types of learning models for estimating the sludge moisture content after a sludge dewatering process is performed by a sludge dewatering machine under arbitrary dewatering process conditions, and an optimal model information memory unit that stores an optimal model database that stores identification information of an optimal learning model identified from the multiple types of learning models for each sludge dewatering process condition, the estimation device, when a sludge moisture content estimation process is requested, acquires dewatering process conditions for the sludge dewatering process that are the execution conditions of the estimation process, acquires identification information of the optimal learning model to be used in the sludge moisture content estimation process from the optimal model database based on the acquired dewatering process conditions, acquires a file of the optimal learning model corresponding to the acquired identification information from the first file memory unit, and executes the sludge moisture content estimation process by inputting the acquired dewatering process conditions into the optimal learning model to obtain an estimated value of the sludge moisture content, and outputs the acquired estimated value from an output device.

[0010] In addition, a sludge moisture content estimation program according to one embodiment of the present invention has an estimation device communicatively connected to a first file memory unit that stores files of multiple types of learning models for estimating the sludge moisture content after a sludge dewatering process is performed by a sludge dewatering machine under arbitrary dewatering process conditions, and an optimal model information memory unit that stores an optimal model database that stores identification information of an optimal learning model identified from the multiple types of learning models for each sludge dewatering process condition, and executes the following functions when an estimation process of a sludge moisture content is requested: a function to acquire dehydration conditions for the sludge dehydration process, which are the execution conditions of the estimation process, when an estimation process of a sludge moisture content is requested, from the optimal model database based on the acquired dehydration process conditions, an optimal learning model file corresponding to the acquired identification information, an optimal learning model file corresponding to the acquired identification information, and an estimation process of the sludge moisture content by inputting the acquired dehydration process conditions into the optimal learning model to acquire an estimated value of the sludge moisture content, and a function to output the acquired estimated value from an output device.

[0011] Moreover, a database generation device according to one embodiment of the present invention includes a learning model generation unit that acquires, for each of a plurality of sludge dewatering machines, dewatering machine information, operating data when sludge dewatering treatment is performed within a specified period, sludge type information, and sludge moisture content after sludge dewatering treatment is performed, and generates a plurality of types of learning models based on the acquired information, and an optimal model information generation unit that is communicatively connected to an estimation accuracy calculation unit that calculates, for each of the learning models generated by the learning model generation unit, an estimation accuracy of the sludge moisture content for each combination of dewatering machine information related to the sludge dewatering machine and sludge type information, and identifies, for each combination of dewatering machine information related to the sludge dewatering machine and sludge type information, the learning model with the highest estimation accuracy calculated by the estimation accuracy calculation unit as an optimal learning model, and generates an optimal model database that stores identification information of the identified learning model.

[0012] In addition, a database generation method according to one embodiment of the present invention includes a learning model generation unit that acquires, for each of a plurality of sludge dewatering machines, dewatering machine information, operating data when sludge dewatering treatment is performed within a specified period, sludge type information, and sludge moisture content after sludge dewatering treatment is performed, and generates a plurality of types of learning models based on the acquired information, and a database generation device that is communicatively connected to an estimation accuracy calculation unit that calculates, for each of the learning models generated by the learning model generation unit, an estimation accuracy of the sludge moisture content for each combination of dewatering machine information related to the sludge dewatering machine and sludge type information, identifies the learning model with the highest calculated estimation accuracy for each combination of dewatering machine information related to the sludge dewatering machine and sludge type information as an optimal learning model, and generates an optimal model database that stores identification information of the identified learning model.

[0013] In addition, an optimal model database according to one embodiment of the present invention is a database used in a sludge moisture content estimation system that estimates the moisture content of sludge after sludge dewatering treatment has been performed by a sludge dewatering machine, and stores, for each combination of dewatering machine information regarding the sludge dewatering machine and sludge type information of the sludge, identification information of an optimal learning model with the highest estimation accuracy among multiple types of learning models that estimate the moisture content of sludge after sludge dewatering treatment has been performed. Effect of the Invention

[0014] According to the present invention, it is possible to provide a sludge moisture content estimation system, a sludge moisture content estimation device, a sludge moisture content estimation method, a sludge moisture content estimation program, a database generation device, a database generation method, and an optimal model database that are used to estimate with high accuracy the sludge moisture content after dewatering treatment is performed under any treatment conditions. [Brief description of the drawings]

[0015] [Figure 1] 1 is a block diagram showing a configuration of a sludge water content estimation system according to one embodiment. FIG. [Diagram 2] 2 is a block diagram showing detailed configurations of a data processing system and a learning processing system in a sludge moisture content estimation system according to one embodiment. FIG. [Diagram 3] 1 is a block diagram showing a detailed configuration of a learning model management system in a sludge moisture content estimation system according to one embodiment. FIG. [Figure 4] 1 is a sequence diagram showing a learning model generation process and an optimal model database generation process executed by a sludge moisture content estimation system according to an embodiment. FIG. [Diagram 5] 3 is a sequence diagram showing a process of estimating a sludge moisture content executed by a sludge moisture content estimation system according to an embodiment of the present invention; FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] An exemplary embodiment of a sludge moisture content estimation system will be described below with reference to the drawings. In this embodiment, the sludge moisture content estimation system estimates the sludge moisture content when the sludge dewatering process is performed under any processing conditions, using information on the sludge dewatering process performed in the multiple sludge dewatering machines X1 to Xn.

[0017] Configuration of a sludge moisture content estimation system according to one embodiment 1 is a block diagram showing the configuration of a sludge moisture content estimation system 1 according to this embodiment. The sludge moisture content estimation system 1 includes a torque sensor 11, a temperature sensor 12, an ammeter 13, a pressure sensor 14, a moisture content meter 15, an operation data management device 20, and a moisture content management device 30, which are installed for each of n sludge dewatering machines X1 to Xn, a data processing system 40, a learning processing system 50, a learning model management system 60, an estimation device 70 as a sludge moisture content estimation device, and an output device 80. The data processing system 40, the learning processing system 50, and the learning model management system 60 are constructed by using, for example, a cloud C provided by a cloud service provider.

[0018] The sludge dewatering machines X1 to Xn have the same configuration. Hereinafter, when it is not specified which of the sludge dewatering machines X1 to Xn it is, it will be referred to as the sludge dewatering machine X. The sludge dewatering machine X includes a cylinder T for supplying the sludge P to be dewatered, a sludge supply port E for supplying the sludge P to the cylinder T, a screw SC which is a rotor for dewatering the sludge P supplied to the cylinder T, and a screw motor M for rotating the screw SC. The sludge dewatering machine X compresses and dewaters the sludge P supplied into the cylinder T by applying a predetermined torque to the screw motor M to rotate it. A dewatered cake D which is the dewatered sludge is discharged from the sludge dewatering machines X1 to Xn.

[0019] The torque sensor 11 measures the rotational torque of the screw motor M. The temperature sensor 12 measures the temperature of the sludge P to be treated. The ammeter 13 measures the current value of the screw motor M. The pressure sensor 14 measures the pressure value applied to the sludge P when the sludge P is input from the sludge supply port E. The moisture content meter 15 measures the moisture content of the dehydrated cake D.

[0020] The operation data management device 20 has an operation data input unit 21, an operation data temporary storage unit 22, and an operation data output unit 23.

[0021] The operation data input unit 21 inputs various measurement values ​​measured by the torque sensor 11, temperature sensor 12, ammeter 13, and pressure sensor 14 at predetermined time intervals, for example, one-minute intervals, as operation data of the sludge dewatering machine X, together with measurement time information. The operation data temporary storage unit 22 stores the operation data input by the operation data input unit 21. The operation data output unit 23 outputs the operation data stored in the operation data temporary storage unit 22 to the first upload device 20a. The first upload device 20a adds identification information of the sludge dewatering machine X to be measured to the operation data output from the operation data output unit 23, and uploads the data to the data processing system 40 at predetermined time intervals.

[0022] The moisture content management device 30 has a moisture content input unit 31 , a moisture content temporary storage unit 32 , and a moisture content output unit 33 .

[0023] The moisture content input unit 31 adds measurement time information to the moisture content data measured by the moisture content meter 15 and inputs it. The moisture content temporary storage unit 32 stores the moisture content data input by the moisture content input unit 31. The moisture content output unit 33 outputs the moisture content data stored by the moisture content temporary storage unit 32 to the second upload device 30a. The second upload device 30a adds identification information of the sludge dewatering machine X to be measured to the moisture content data output from the moisture content output unit 33, and uploads the data to the data processing system 40 at predetermined time intervals.

[0024] FIG. 2 is a block diagram showing a detailed configuration of the data processing system 40 and the learning processing system 50. As shown in FIG.

[0025] The data processing system 40 includes an operating data receiving unit 41, a moisture content receiving unit 42, a first CPU 43, and a first storage unit 44.

[0026] The driving data receiving unit 41 receives the driving data uploaded from the first upload device 20a. The moisture content receiving unit 42 receives the moisture content data uploaded from the second upload device 30a.

[0027] The first CPU 43 is, for example, a CPU (Central Processing Unit) provided in a general-purpose microcomputer, and configures one or more information processing units described below by installing and executing a predetermined data processing program. The first CPU 43 has an operation data acquisition unit 431, a moisture content acquisition unit 432, and a dehydration treatment information generation unit 433.

[0028] The operation data acquisition unit 431 acquires the operation data received by the operation data receiving unit 41, and stores it in an operation data storage unit 443 in the first storage unit 44, which will be described later. The moisture content acquisition unit 432 acquires the moisture content data received by the moisture content receiving unit 42, and stores it in a moisture content storage unit 444 in the first storage unit 44.

[0029] The dehydration treatment information generating unit 433 generates dehydration treatment information linking the operation data and moisture content data corresponding to each measurement time information using the information stored in the first storage unit 44. The dehydration treatment information generating unit 433 adds the dehydrator information of the corresponding sludge dehydrator X stored in a dehydrator information storage unit 441 described later and the sludge type information of the sludge to be dehydrated stored in a sludge type storage unit 442 to the generated dehydration treatment information.

[0030] The first storage unit 44 has a dehydrator information storage unit 441 , a sludge type storage unit 442 , an operation data storage unit 443 , and a moisture content storage unit 444 .

[0031] The dehydrator information storage unit 441 stores dehydrator information, which is identification information for each managed sludge dehydrator X. The dehydrator information includes the "dehydrator model," "size of the dehydrator," "diameter of the sludge supply port E," "serial number," "year of manufacture," etc. The diameter of the sludge supply port E varies depending on the size of the sludge dehydrator X, the inner diameter of the cylinder T, etc.

[0032] The sludge type storage unit 442 stores sludge type information indicating the type of property of the sludge P to be treated by each sludge dewatering machine X. The types of the properties of the sludge P are classified into about five types including raw sludge, thickened sludge, digested sludge, etc., and sludge type information corresponding to these types is set in advance.

[0033] The operation data storage unit 443 stores the operation data acquired by the operation data acquisition unit 431. The moisture content storage unit 444 stores the moisture content data acquired by the moisture content acquisition unit 432.

[0034] The learning processing system 50 has a second CPU 51 and a second storage unit 52. The second CPU 51 is, for example, a CPU (Central Processing Unit) provided in a general-purpose microcomputer, and configures one or more information processing units described below by installing and executing a predetermined learning processing program. The second CPU 51 has a dehydration processing information acquisition unit 511, a learning model generation unit 512, and an estimation accuracy calculation unit 513.

[0035] The dehydration treatment information acquisition unit 511 acquires the dehydration treatment information generated by the dehydration treatment information generation unit 433 of the data processing system 40, and stores the information in a dehydration treatment information storage unit 521 in the second storage unit 52 described later.

[0036] The learning model generation unit 512 generates multiple types of learning models for estimating the sludge moisture content after the sludge dewatering machine X performs sludge dewatering under any dewatering conditions, based on the dewatering treatment information for a predetermined period of time stored in the dewatering treatment information memory unit 521.

[0037] The estimation accuracy calculation unit 513 calculates the estimation accuracy of the sludge water content for each combination of the dehydrator information of the sludge dehydrator X and the sludge type information of the sludge P for each learning model generated by the learning model generation unit 512.

[0038] 3 is a block diagram showing detailed configurations of the learning model management system 60 and the estimation device 70. The learning model management system 60 has a third CPU 61 and a third storage unit 62.

[0039] The third CPU 61 is, for example, a CPU (Central Processing Unit) provided in a general-purpose microcomputer, and configures one or more information processing units described below by installing and executing a predetermined learning model management program. The third CPU 61 has a learning model acquisition unit 611, an estimation accuracy acquisition unit 612, an optimal model information generation unit 613, and a first estimation unit 614.

[0040] The learning model acquisition unit 611 acquires a file of the learning model generated by the learning model generation unit 512 of the learning processing system 50, and stores it in the first file memory unit 621 in the third memory unit 62 described later.

[0041] The estimation accuracy acquisition unit 612 acquires information on the estimation accuracy of the sludge moisture content for each combination of the learning model, dehydrator information of the sludge dehydrator X, and sludge type information of the sludge P calculated by the estimation accuracy calculation unit 513 of the learning processing system 50, and stores the information in the model attribute information memory unit 622 in the third memory unit 62 as attribute information for each learning model.

[0042] The optimal model information generating unit 613 specifies, for each sludge dehydration treatment condition, as the optimal learning model, the learning model having the highest estimation accuracy calculated by the estimation accuracy calculating unit 513 among the learning models generated by the learning model generating unit 512, based on the estimation accuracy information stored in the model attribute information storing unit 622. In this embodiment, the optimal model information generating unit 613 specifies, for each combination of the dehydrator information of the sludge dehydrator X, which is the sludge dehydration treatment condition, and the sludge type information of the sludge P, the learning model having the highest estimation accuracy calculated by the estimation accuracy calculating unit 513, as the optimal learning model.

[0043] The optimal model information generation unit 613 generates an optimal model database that stores identification information of the identified learning model for each combination of dewatering machine information of the sludge dewatering machine X and sludge type information of the sludge P, and stores the generated optimal model database in the optimal model information memory unit 623 in the third memory unit 62.

[0044] When the first estimation unit 614 receives an estimation processing request together with the estimation processing conditions for the sludge moisture content from the estimation device 70 as described below, the first estimation unit 614 receives identification information of an optimal learning model to be used in the estimation processing of the sludge moisture content from the optimal model information storage unit 623 based on the information of the estimation processing conditions included in the received request. The first estimation unit 614 receives from the first file storage unit 621 a file of the optimal learning model corresponding to the received identification information.

[0045] The first estimation unit 614 inputs the estimation processing conditions for the sludge moisture content acquired from the estimation device 70 into the acquired optimal learning model, and thereby acquires an estimated value of the sludge moisture content based on these estimation processing conditions. The first estimation unit 614 also checks whether the file of the acquired learning model has been transmitted to the estimation device 70, and if not, transmits this file to the estimation device 70 together with the dehydrator information of the corresponding sludge dehydrator X and the sludge type information of the sludge P.

[0046] The third storage unit 62 has a first file storage unit 621, a model attribute information storage unit 622, and an optimal model information storage unit 623. The first file storage unit 621 stores a file of the learning model acquired by the learning model acquisition unit 611. The model attribute information storage unit 622 stores attribute information for each learning model, including information on the estimation accuracy acquired by the estimation accuracy acquisition unit 612. The optimal model information storage unit 623 stores an optimal model database generated by the optimal model information generation unit 613.

[0047] The estimation device 70 has an input unit 71, a fourth CPU 72, and a fourth memory unit 73. The input unit 71 inputs dehydrator information, operation data, and sludge type information, which are processing conditions for estimating the sludge water content, and an estimation processing request based on these conditions.

[0048] The fourth CPU 72 is, for example, a CPU (Central Processing Unit) provided in a general-purpose microcomputer, and configures one or more information processing units described below by installing and executing a predetermined estimation processing program. The fourth CPU 72 has an estimation request processing unit 721, a second estimation unit 722, and an output control unit 723.

[0049] The estimation request processing unit 721 acquires an estimation processing request with the dehydrator information, operation data, and sludge type information input from the input unit 71 as estimation processing conditions for the sludge water content, and transmits the request to the learning model management system 60. The estimation request processing unit 721 sends the estimated value of the sludge water content acquired from the learning model management system 60 by transmitting the estimation processing request to the output control unit 723. When the estimation request processing unit 721 acquires a learning model file from the learning model management system 60, it stores the file in a second file storage unit 731 of the fourth storage unit 73 described later together with the corresponding dehydrator information of the sludge dehydrator X and the sludge type information of the sludge P.

[0050] When a sludge moisture content estimation process is requested while communication between the estimation device 70 and the cloud C is disconnected, the second estimation unit 722 obtains a learning model file corresponding to the sludge moisture content estimation process conditions input from the input unit 71 from the second file storage unit 731, and obtains an estimated value of the sludge moisture content by inputting the estimation process conditions. The second estimation unit 722 sends the obtained estimated value of the sludge moisture content to the output control unit 723.

[0051] The output control unit 723 causes the output device 80 to output the estimated value of the sludge water content sent from the estimation request processing unit 721 or the second estimation unit 722.

[0052] The fourth storage unit 73 includes a second file storage unit 731. The second file storage unit 731 stores a file of the learning model acquired from the estimation request processing unit 721.

[0053] The output device 80 is composed of a display device or a speaker device.

[0054] <Operation of the sludge moisture content estimation system according to one embodiment> 4 is a sequence diagram showing a learning model generation process and an optimum model database generation process executed by the sludge water content estimation system 1 according to this embodiment. As an example, the learning model generation process and the optimum model database generation process executed for the sludge dewatering process executed by the sludge dewatering machine X1 will be described here.

[0055] During operation of the sludge dewatering machine X1, at predetermined time intervals, for example one minute intervals, a torque sensor 11 measures the torque of the screw motor M, a temperature sensor 12 measures the temperature of the sludge P to be treated, an ammeter 13 measures the current value of the screw motor M, and a pressure sensor 14 measures the pressure value applied to the sludge P when it is fed through the sludge supply port E. An operation data input unit 21 of the operation data management device 20 inputs these various measurement values, together with measurement time information, as the corresponding operation data of the sludge dewatering machine X1, and stores them in an operation data temporary storage unit 22.

[0056] The operation data output unit 23 outputs to the first upload device 20a the operation data stored in the operation data temporary storage unit 22. The first upload device 20a adds identification information of the sludge dewatering machine X1 to the operation data output from the operation data output unit 23, and uploads the data to the data processing system 40 at predetermined time intervals, for example, at one-minute intervals (S1).

[0057] In the data processing system 40, the driving data receiving unit 41 receives the driving data uploaded from the first upload device 20a and sends it to the driving data acquiring unit 431. The driving data acquiring unit 431 stores the driving data acquired from the driving data receiving unit 41 in the driving data storage unit 443 of the first storage unit 44 (S2).

[0058] Furthermore, in the sludge dewatering machine X1, an operator uses the moisture content meter 15 to measure the moisture content of the dehydrated cake D at a predetermined timing, for example, at a preset time about twice a day. The operator adds measurement time information to the moisture content data measured by the moisture content meter 15, inputs the data to the moisture content management device 30, and stores the data in the moisture content temporary storage unit 32. The process of measuring the moisture content of the dehydrated cake at a predetermined timing may be automatically performed by the moisture content meter 15. Furthermore, the process of inputting the moisture content data to the moisture content management device 30 may be automatically performed by the moisture content meter 15.

[0059] The moisture content output unit 33 outputs the moisture content data stored in the moisture content temporary storage unit 32 to the second upload device 30a. The second upload device 30a adds the identification information of the sludge dewatering machine X1 to the moisture content data output from the moisture content output unit 33, and uploads the data to the data processing system 40 at predetermined time intervals (S3).

[0060] In the data processing system 40, the moisture content receiving unit 42 receives the moisture content data uploaded from the second upload device 30a and sends it to the moisture content acquiring unit 432. The moisture content acquiring unit 432 stores the moisture content data acquired from the moisture content receiving unit 42 in the moisture content memory unit 444 of the first memory unit 44 (S4).

[0061] When the dewatering treatment information generating unit 433 detects that new moisture content data has been stored in the moisture content memory unit 444, it acquires the operation data of the sludge dewatering machine X1 corresponding to the measurement time information of this new moisture content data from the operation data memory unit 443. The "operation data of the sludge dewatering machine X1 corresponding to the measurement time information of the new moisture content data" is the operation data of the sludge dewatering machine X1 having measurement time information whose difference from the measurement time information of the new moisture content data is within a predetermined time. The dewatering treatment information generating unit 433 generates dewatering treatment information linking this new moisture content data with the acquired operation data.

[0062] The dehydration treatment information generating unit 433 associates the generated dehydration treatment information for each measurement time information with predetermined information in the dehydrator information of the corresponding sludge dehydrator X1, for example, information on the aperture of the sludge supply port E, based on the information stored in the dehydrator information storage unit 441. The dehydration treatment information generating unit 433 also associates the generated dehydration treatment information with sludge type information of the sludge P to be dehydrated, based on the information stored in the sludge type storage unit 442. The dehydration treatment information generating unit 433 transmits the generated dehydration treatment information to the learning processing system 50 (S5).

[0063] In the learning processing system 50, the dehydration processing information acquisition unit 511 acquires the dehydration processing information transmitted from the data processing system 40, and stores it in the dehydration processing information storage unit 521 in the second storage unit 52 (S6).

[0064] The learning model generating unit 512 generates multiple types of learning models for estimating the sludge water content after sludge dehydration, using the diameter of the sludge supply port E of the sludge dehydrator X, the sludge type information, and the measurement value of the torque sensor 11 in the operation data as essential dehydration conditions, based on the dehydration information stored in the dehydration information storage unit 521. The learning model generating unit 512 may further add at least one of the measurement value of the temperature sensor 12, the measurement value of the ammeter, and the measurement value of the pressure sensor in the operation data to be used as the dehydration conditions.

[0065] In this embodiment, the learning model generation unit 512 generates three learning models M1 to M3 using different algorithms. When generating these learning models M1 to M3, for example, an algorithm such as linear regression, logistic regression, decision tree, random forest, or a combination of these is used. The learning model generation unit 512 transmits files of the generated learning models M1 to M3 to the learning model management system 60 (S7).

[0066] In the learning model management system 60, the learning model acquisition unit 611 acquires the files of the learning models M1 to M3 transmitted from the learning processing system 50, and stores them in the first file storage unit 621 in the third storage unit 62 (S8).

[0067] Furthermore, in the learning processing system 50, the estimation accuracy calculation unit 513 calculates, for each of the learning models generated by the learning model generation unit 512, the estimation accuracy of the sludge water content for each combination of the diameter of the sludge supply port E of the sludge dewatering machine X and the sludge type information of the sludge P. The estimation accuracy calculation unit 513 transmits the calculated estimation accuracy of each sludge water content to the learning model management system 60 (S9).

[0068] In the learning model management system 60, the estimation accuracy acquisition unit 612 acquires the estimation accuracy of each sludge moisture content transmitted from the learning processing system 50, and stores it in the model attribute information storage unit 622 as attribute information for each learning model (S10).

[0069] Based on the information on the estimation accuracy stored in the model attribute information storage unit 622, the optimal model information generation unit 613 identifies the learning model with the highest estimation accuracy calculated by the estimation accuracy calculation unit 513 as the optimal learning model for each combination of the diameter of the sludge supply port E of the sludge dewatering machine X and the sludge type information of the sludge P. The optimal model information generation unit 613 stores identification information of the learning model identified for each combination of the diameter of the sludge supply port E of the sludge dewatering machine X and the sludge type information of the sludge P to generate an optimal model database and store it in the third storage unit 62 (S11).

[0070] If the identified optimal learning model differs from the optimal learning model previously identified by a similar method and stored in the optimal model database for the combination of the diameter of the sludge supply port E of the sludge dewatering machine X1 currently being processed and the sludge type information of the sludge P, the optimal model database is updated with the identification information of the newly identified learning model. In other words, every time new moisture content data is measured by the sludge dewatering machine X1, the learning model in the first file storage unit 621 is updated as appropriate by the learning model generation unit 512, the estimation accuracy of the sludge moisture content in the model attribute information storage unit 622 is updated by the estimation accuracy calculation unit 513, and the optimal model database in the optimal model information storage unit 623 is further updated.

[0071] The optimum model information generating unit 613 may generate an optimum model database by storing, for each identification information of the sludge dehydrator X, the identification information of the optimum learning model specified for each combination of the diameter of the sludge supply port E of the sludge dehydrator X and the sludge type information. In this case, when new moisture content data is acquired for any sludge dehydrator X, the estimation accuracy calculating unit 513 updates the estimation accuracy of the sludge moisture content, and the optimum model information generating unit 613 specifies an optimum learning model corresponding to the combination of the diameter of the sludge supply port E of the sludge dehydrator X and the sludge type information based on the estimation accuracy of the updated sludge moisture content. Then, the optimum model information generating unit 613 updates the optimum model database by changing the identification information of the optimum learning model of all sludge dehydrators X having the same combination of the diameter of the sludge supply port E of the sludge dehydrator X and the sludge type information of the sludge dehydrator X using the identification information of the specified optimum learning model. This completes the learning model generation process and the optimum model database generation process.

[0072] FIG. 5 is a sequence diagram showing the process of estimating the sludge water content executed by the sludge water content estimating system 1 according to this embodiment.

[0073] When performing the sludge moisture content estimation process, for example, a user inputs, into the input unit 71 of the estimation device 70, the following estimation process conditions: the diameter of the sludge supply port E, which is the dehydrator information of the sludge dehydrator X that is the entity performing the sludge dehydration process; the torque value of the screw motor M, which is the operating data of the sludge dehydrator X; and sludge type information of the sludge P to be dehydrated, and performs an operation to request the estimation process of the sludge moisture content based on these conditions.

[0074] As the estimation processing conditions for the sludge moisture content, instead of the diameter of the sludge supply port E of the sludge dewatering machine X to be estimated and the sludge type information, identification information such as the serial number of the sludge dewatering machine X may be input. In this case, the temperature of the sludge P, the current value of the screw motor M, and the pressure value applied to the sludge P when the sludge P is input from the sludge supply port E may also be input as the estimation processing conditions.

[0075] When an operation is performed on the input unit 71 to request estimation processing of the sludge moisture content ("YES" in S21), if the estimation device 70 is communicatively connected to the cloud C ("YES" in S22), the estimation request processing unit 721 sends an estimation processing request based on the estimation processing conditions for the sludge moisture content input from the input unit 71 to the learning model management system 60 (S23).

[0076] In the learning model management system 60, when an estimation processing request is acquired from the estimation device 70, the first estimation unit 614 acquires identification information of the optimal learning model to be used in the estimation processing of the sludge water content from the optimal model information storage unit 623 based on the information included in the acquired request (S24). The first estimation unit 614 acquires a file of the learning model corresponding to the acquired identification information from the first file storage unit 621 (S25).

[0077] The first estimation unit 614 inputs the estimation processing conditions for the sludge moisture content acquired from the estimation device 70 into the acquired learning model, thereby acquiring an estimated value of the sludge moisture content based on these estimation processing conditions. The first estimation unit 614 transmits the acquired estimated value of the sludge moisture content to the estimation device 70 (S26).

[0078] In the estimation device 70, the estimation request processing unit 721 acquires the estimated value of the sludge moisture content transmitted from the learning model management system 60, and sends it to the output control unit 723. The output control unit 723 causes the output device 80 to output the acquired estimated value of the sludge moisture content (S27).

[0079] The first estimation unit 614 also checks whether the file of the learning model acquired in step S25 has been sent to the estimation device 70, and if not sent ("YES" in S28), adds the diameter of the sludge supply port E of the corresponding sludge dewatering machine X and the sludge type information of the sludge P and sends this file to the estimation device 70 (S29). Here, the diameter of the sludge supply port E of the sludge dewatering machine X and the sludge type information of the sludge P to be added to the acquired learning model file are the diameter of the sludge supply port E of the sludge dewatering machine X and the sludge type information of the sludge P that have been identified by the optimal model information generation unit 613 as the learning model that is to be the optimal learning model.

[0080] In the estimation device 70, the estimation request processing unit 721 acquires the learning model file transmitted from the learning model management system 60 and stores it in the second file memory unit 731 of the fourth memory unit 73 described later together with the added diameter of the sludge supply port E of the sludge dewatering machine X and sludge type information of the sludge P (S30).

[0081] Furthermore, in step S22, if the estimation device 70 is not communicatively connected to the cloud C ("NO" in S22), the second estimation unit 722 acquires the estimation processing conditions for the sludge moisture content input from the input unit 71. The second estimation unit 722 acquires a learning model file corresponding to the acquired estimation processing conditions for the sludge moisture content from among those stored in the second file storage unit 731 by past estimation processing of the sludge moisture content, and inputs the estimation processing conditions to acquire an estimated value of the sludge moisture content.

[0082] The second estimation unit 722 sends the acquired estimated value of the sludge water content to the output control unit 723. The output control unit 723 causes the acquired estimated value of the sludge water content to be output from the output device 80 (S31).

[0083] This estimation process of the sludge moisture content may be performed at a predetermined time interval, for example, by the input unit 71 of the estimation device 70 acquiring and inputting operation data of a predetermined sludge dewatering machine X from the operation data output unit 23 of the corresponding operation data management device 20. In this case, the operation data output unit 23 adds identification information of the sludge dewatering machine X to the operation data, and inputs it from the input unit 71 of the estimation device 70.

[0084] The estimation request processing unit 721 transmits the input operation data and the identification information of the sludge dewatering machine X to the learning model management system 60. In the learning model management system 60, the first estimation unit 614 acquires the identification information of the optimal learning model to be used in the estimation process of the sludge moisture content, using an optimal model database generated using the identification information of each sludge dewatering machine X in the optimal model information storage unit 623. Thereafter, the estimation process of the sludge moisture content is executed in the same manner as the above-mentioned process.

[0085] [Effects of the embodiment] As described above in detail, the sludge moisture content estimation system according to this embodiment includes a first file storage unit, an optimum model information storage unit, an estimation request processing unit, a first estimation unit, and an output control unit. The first file storage unit stores files of multiple types of learning models for estimating the sludge moisture content after the sludge dewatering machine executes a sludge dewatering process under any dewatering process conditions. The optimum model information storage unit stores an optimum model database that stores identification information of an optimum learning model specified from multiple types of learning models for each sludge dewatering process condition. When a sludge moisture content estimation process is requested, the estimation request processing unit acquires the dewatering process conditions of the sludge dewatering process, which are the execution conditions of the estimation process. The first estimation unit acquires, from the optimum model database, identification information of an optimum learning model to be used in the estimation process of the sludge moisture content based on the dehydration treatment conditions acquired by the estimation request processing unit, acquires a file of the optimum learning model corresponding to the acquired identification information from the first file storage unit, and executes the estimation process of the sludge moisture content by inputting the dehydration treatment conditions acquired by the estimation request processing unit into the optimum learning model to acquire an estimated value of the sludge moisture content. The output control unit causes the output device to output the estimated value acquired by the first estimation unit.

[0086] In this way, by identifying and using an optimal learning model from among multiple types of learning models for each sludge dehydration treatment condition, it is possible to perform an estimation process of the sludge moisture content with high accuracy. Multiple types of learning models can be generated using various algorithms, but since the type of optimal learning model varies depending on the dehydration treatment condition, it is possible to perform an estimation process of the sludge moisture content according to the dehydration condition by performing the process as described above.

[0087] In addition, in this embodiment, the optimal model information memory unit stores an optimal model database that stores identification information of an optimal learning model among multiple types of learning models for each combination of dehydrator information regarding a sludge dehydrator set as a sludge dehydration treatment condition and sludge type information indicating the type of sludge properties, and the estimation request processing unit acquires dehydrator information regarding the sludge dehydrator that is the entity executing the sludge dehydration treatment, operating data regarding the sludge dehydrator, and sludge type information of the sludge to be dehydrated as dehydration treatment conditions for the sludge dehydration treatment, which are execution conditions for the estimation treatment, and the first estimation unit acquires identification information of the optimal learning model to be used in the estimation process of the sludge water content from the optimal model database based on the dehydrator information and sludge type information in the dehydration treatment conditions acquired by the estimation request processing unit, acquires a file of the optimal learning model corresponding to the acquired identification information from the first file memory unit, and executes the estimation process of the sludge water content by inputting the dehydration treatment conditions acquired by the estimation request processing unit into this optimal learning model to acquire an estimated value of the sludge water content.

[0088] This enables more accurate estimation of the sludge water content by taking into account dehydrator information related to the sludge dehydrator, which is a dehydration treatment condition that affects the estimation accuracy, and sludge type information indicating the type of sludge properties.

[0089] In addition, in this embodiment, the device may further include a learning model generation unit that acquires, for each of the multiple sludge dewatering machines, dewatering machine information, operating data when sludge dewatering treatment is performed within a specified period, sludge type information, and sludge moisture content after sludge dewatering treatment is performed, and generates multiple types of learning models based on the acquired information and stores them in the first file memory unit.

[0090] This makes it possible to appropriately generate a learning model used in the estimation process of sludge moisture content, taking into account the dewatering treatment conditions that affect the estimation accuracy.

[0091] In addition, in this embodiment, the learning model generating unit may further include an estimation accuracy calculation unit that calculates an estimation accuracy of the sludge water content for each combination of dehydrator information related to the sludge dehydrator and sludge type information for each learning model generated by the learning model generation unit, and an optimal model information generating unit that identifies the learning model with the highest estimation accuracy calculated by the estimation accuracy calculation unit for each combination of dehydrator information related to the sludge dehydrator and sludge type information as an optimal learning model, stores identification information of the identified learning model to generate an optimal model database, and stores it in the optimal model information storage unit.

[0092] This makes it possible to appropriately generate an optimal model database used in the process of identifying the learning model to be used in the process of estimating the sludge moisture content, depending on the estimation accuracy of each learning model.

[0093] In addition, in this embodiment, each time a new sludge moisture content is measured in any of the multiple sludge dehydrators, the learning model generation unit acquires dehydrator information for the sludge dehydrator from which the sludge moisture content was measured, operating data for the sludge dehydrator when the sludge dehydration process was performed, sludge type information of the sludge, and the sludge moisture content after the sludge dehydration process was performed, and updates the multiple types of learning models based on the acquired information.When the learning model generation unit updates the multiple types of learning models, the estimation accuracy calculation unit updates the estimation accuracy of the sludge moisture content using the updated learning models.When the estimation accuracy calculation unit updates the estimation accuracy of the sludge moisture content, the optimal model information generation unit identifies an optimal learning model for each combination of dehydrator information and sludge type information for the sludge dehydrator based on the estimation accuracy of the updated sludge moisture content, and updates the optimal model database using the identification information of the identified optimal learning model.

[0094] As a result, even if the sludge moisture content for each dehydration treatment condition changes over time, it is possible to execute a process for estimating the sludge moisture content in accordance with the latest situation.

[0095] In addition, in this embodiment, the optimal model information generation unit generates an optimal model database by storing identification information of the optimal learning model identified for each combination of dehydrator information and sludge type information related to the sludge dehydrator, and when the estimation accuracy calculation unit updates the estimation accuracy of the sludge moisture content, it may identify an optimal learning model corresponding to the combination of the dehydrator information and sludge type information related to the sludge dehydrator based on the updated estimation accuracy of the sludge moisture content, and update the optimal model database by using the identification information of the identified optimal learning model to change the identification information of the optimal learning models of all sludge dehydrators that correspond to the combination of the dehydrator information and sludge type information related to this sludge dehydrator.

[0096] This makes it possible to collectively and appropriately manage information on optimal learning models for a plurality of sludge dewatering machines having the same combination of dewatering machine information and sludge type information relating to the sludge dewatering machine.

[0097] In addition, in this embodiment, the device further includes a dehydrator information storage unit that stores dehydrator information for each sludge dehydrator and information on the sludge type to be treated, and the learning model generation unit acquires, from any one of the multiple sludge dehydrators, operating data for the sludge dehydrator when the sludge dehydration process is performed, the sludge water content after the sludge dehydration process is performed, and dehydrator information for this sludge dehydrator, and may acquire dehydrator information and sludge type information for the sludge dehydrator corresponding to the acquired information based on the information stored in the dehydrator information storage unit and use these for the generation process of multiple types of learning models.

[0098] This makes it possible to reduce the amount of information when transmitting operating data related to the sludge dewatering machines and the sludge moisture content after performing sludge dewatering processing from multiple sludge dewatering machines to the sludge moisture content estimation system, thereby reducing the communication load within the system.

[0099] In this embodiment, the system further includes a second file memory unit that acquires and stores the file of the optimal learning model acquired by the first estimation unit, and a second estimation unit that, when a sludge moisture content estimation process is requested when communication between the estimation request processing unit and the first estimation unit is disconnected, executes the sludge moisture content estimation process by inputting dehydration treatment conditions corresponding to the requested estimation process into the file of the optimal learning model stored in the second file memory unit, and acquires an estimated value of the sludge moisture content, and the output control unit may output the estimated value acquired by the second estimation unit from an output device.

[0100] This makes it possible to execute the estimation process for sludge moisture content even when an estimation device having an estimation request processing unit is in an environment where it cannot connect to a cloud in which a learning model management system having a first estimation unit is installed.

[0101] The sludge moisture content estimation device of this embodiment is communicatively connected to a first file memory unit that stores files of multiple types of learning models for estimating the sludge moisture content after a sludge dewatering machine performs sludge dewatering under any dewatering treatment conditions, and an optimal model information memory unit that stores an optimal model database that stores identification information of an optimal learning model identified from among the multiple types of learning models for each sludge dewatering treatment condition, and is equipped with an estimation request processing unit, a first estimation unit, and an output control unit.

[0102] In this way, by identifying and using the most suitable learning model from among multiple types of learning models for each sludge dewatering treatment condition, it is possible to perform an estimation process for the sludge moisture content with high accuracy.

[0103] The sludge moisture content estimation method according to the present embodiment is executed by an estimation device communicably connected to a first file storage unit that stores files of multiple types of learning models for estimating the sludge moisture content after the sludge dehydrator executes the sludge dehydration process under any dehydration process condition, and an optimal model information storage unit that stores an optimal model database that stores identification information of an optimal learning model specified from multiple types of learning models for each sludge dehydration process condition. When a sludge moisture content estimation process is requested, the estimation device acquires dehydration process conditions for the sludge dehydration process that are execution conditions of the estimation process, acquires identification information of the optimal learning model used in the sludge moisture content estimation process from the optimal model database based on the acquired dehydration process conditions, acquires a file of the optimal learning model corresponding to the acquired identification information from the first file storage unit, executes the sludge moisture content estimation process by inputting the acquired dehydration process conditions into the optimal learning model, acquires an estimated value of the sludge moisture content, and causes the acquired estimated value to be output from an output device.

[0104] In this way, by identifying and using the most suitable learning model from among multiple types of learning models for each sludge dewatering treatment condition, the estimation device can accurately perform the estimation process of the sludge moisture content.

[0105] The sludge moisture content estimation program according to this embodiment causes an estimation device communicatively connected to a first file memory unit that stores files of multiple types of learning models for estimating the sludge moisture content after a sludge dewatering machine has performed sludge dewatering under any dewatering conditions, and an optimal model information memory unit that stores an optimal model database that stores identification information of an optimal learning model identified from multiple types of learning models for each sludge dewatering condition, to perform functions related to sludge moisture content estimation. The sludge moisture content estimation program realizes the following functions: when a sludge moisture content estimation process is requested, the estimation device acquires dehydration process conditions for the sludge dehydration process, which are the execution conditions for the estimation process; acquires identification information of an optimal learning model to be used for the sludge moisture content estimation process from an optimal model database based on the acquired dehydration process conditions, acquires a file of the optimal learning model corresponding to the acquired identification information from a first file memory unit, executes the sludge moisture content estimation process by inputting the acquired dehydration process conditions into the optimal learning model, and acquires an estimated value of the sludge moisture content; and outputs the acquired estimated value from an output device.

[0106] In this way, by identifying and using the most suitable learning model from among multiple types of learning models for each sludge dewatering treatment condition, it is possible to cause the estimation device to perform the estimation process of the sludge moisture content with high accuracy.

[0107] The database generation device according to this embodiment includes a learning model generation unit that acquires, for each of a plurality of sludge dewatering machines, dewatering machine information, operating data when sludge dewatering treatment is performed within a specified period, sludge type information, and sludge moisture content after sludge dewatering treatment is performed, and generates a plurality of types of learning models based on the acquired information, and an optimal model information generation unit that is communicatively connected to an estimation accuracy calculation unit that calculates, for each learning model generated by the learning model generation unit, an estimation accuracy of the sludge moisture content for each combination of dewatering machine information related to the sludge dewatering machine and sludge type information, and identifies, for each combination of dewatering machine information related to the sludge dewatering machine and sludge type information, the learning model with the highest estimation accuracy calculated by the estimation accuracy calculation unit as an optimal learning model, and generates an optimal model database that stores identification information of the identified learning model.

[0108] This makes it possible to appropriately generate an optimal model database to be used in the process of identifying the optimal learning model to be used in the process of estimating the sludge moisture content.

[0109] The database generation method according to this embodiment includes a learning model generation unit that acquires, for each of a plurality of sludge dewatering machines, dewatering machine information, operating data when sludge dewatering treatment is performed within a specified period, sludge type information, and sludge moisture content after sludge dewatering treatment is performed, and generates a plurality of types of learning models based on the acquired information, and a database generation device that is communicatively connected to an estimation accuracy calculation unit that calculates, for each of the learning models generated by the learning model generation unit, the estimation accuracy of the sludge moisture content for each combination of dewatering machine information related to the sludge dewatering machine and sludge type information, identifies the learning model with the highest calculated estimation accuracy for each combination of dewatering machine information related to the sludge dewatering machine and sludge type information as an optimal learning model, and generates an optimal model database that stores identification information of the identified learning model.

[0110] This enables the database generation device to appropriately generate an optimal model database to be used in the process of identifying the optimal learning model to be used in the process of estimating the sludge moisture content.

[0111] The optimal model database according to this embodiment is a database used in a sludge moisture content estimation system that estimates the moisture content of sludge after sludge dehydration processing is performed by a sludge dehydrator, and stores, for each combination of dehydrator information regarding the sludge dehydrator and sludge type information of the sludge, identification information of the optimal learning model with the highest estimation accuracy among multiple types of learning models that estimate the moisture content of sludge after sludge dehydration processing is performed.

[0112] This makes it possible to provide information on a learning model suitable for use when performing an estimation process for sludge moisture content.

[0113] As described above, a learning model is constructed using the sludge dewatering results from multiple sludge dewatering machines, and by using this, even when a new sludge dewatering machine that has no operational history is introduced, it is possible to accurately confirm the estimated sludge moisture content after sludge dewatering is performed by inputting the dewatering machine information, sludge type information, and expected operating data for the sludge dewatering machine.

[0114] Although some embodiments have been described, the embodiments can be modified or modified based on the above disclosure. All components of the above embodiments and all features described in the claims may be individually extracted and combined as long as they are not mutually inconsistent. [Explanation of symbols]

[0115] 1. Sludge moisture content estimation system 11 Torque sensor 12 Temperature Sensor 13 Ammeter 14 Pressure Sensor 15 Moisture content meter 20 Operational data management device 20a First upload device 21 Operation data input section 22 Temporary operating data storage unit 23 Operation data output section 30 Moisture content control device 30a Second upload device 31 Moisture content input section 32 Moisture content temporary storage section 33 Moisture content output section 40 Data Processing System 41 Operation data receiver 42 Moisture content receiver 43 1st CPU 44 1st memory section 50 Learning Processing System 51 2nd CPU 52 2nd memory section 60 Learning Model Management System 61 3rd CPU 62 Third memory section 70 Estimation device 71 Input section 72 4th CPU 73 4th memory section 80 Output Device 431 Operation Data Acquisition Unit 432 Moisture content acquisition part 433 Dehydration Processing Information Generation Unit 441 Dehydrator information storage section 442 Sludge type memory unit 443 Operation data storage unit 444 Moisture content storage section 511 Dehydration Processing Information Acquisition Unit 512 Learning model generation unit 513 Estimation Accuracy Calculation Unit 521 Dehydration processing information storage unit 611 Learning Model Acquisition Unit 612 Estimation accuracy acquisition unit 613 Optimal Model Information Generation Unit 614 1st estimation part 621 First file storage unit 622 Model attribute information storage unit 623 Optimal Model Information Storage Unit 721 Estimation Request Processing Unit 722 Second estimation part 723 Output control section 731 Second File Storage Unit

Claims

1. a first file storage unit that stores a plurality of types of learning model files for estimating a sludge water content after a sludge dewatering process is performed by a sludge dewatering machine under arbitrary dewatering conditions; an optimal model information storage unit that stores an optimal model database that stores identification information of an optimal learning model identified from the plurality of types of learning models for each sludge dewatering treatment condition; an estimation request processing unit that, when a request is made for estimating a sludge moisture content, acquires dehydration conditions for a sludge dehydration process, which are conditions for executing the estimation process; a first estimation unit that acquires, from the optimal model database, identification information of an optimal learning model to be used in the estimation process of the sludge moisture content based on the dewatering treatment conditions acquired by the estimation request processing unit, acquires a file of the optimal learning model corresponding to the acquired identification information from the first file storage unit, and executes the estimation process of the sludge moisture content by inputting the dewatering treatment conditions acquired by the estimation request processing unit into the optimal learning model, thereby acquiring an estimated value of the sludge moisture content; an output control unit that outputs the estimated value acquired by the first estimation unit from an output device.

2. the optimum model information storage unit stores an optimum model database that stores identification information of an optimum learning model among the plurality of types of learning models for each combination of dehydrator information related to a sludge dehydrator set as a sludge dehydration treatment condition and sludge type information indicating a type of sludge property; the estimation request processing unit acquires, as dehydration treatment conditions for the sludge dehydration treatment which are execution conditions for the estimation process, dehydrator information on a sludge dehydrator which is an entity executing the sludge dehydration treatment, operation data on the sludge dehydrator, and sludge type information on the sludge to be dehydrated; The sludge moisture content estimation system of claim 1, wherein the first estimation unit acquires identification information of an optimal learning model to be used in the estimation process of the sludge moisture content from the optimal model database based on the dehydrator information and sludge type information in the dehydration treatment conditions acquired by the estimation request processing unit, acquires a file of the optimal learning model corresponding to the acquired identification information from the first file memory unit, and performs the estimation process of the sludge moisture content by inputting the dehydration treatment conditions acquired by the estimation request processing unit into this optimal learning model, thereby acquiring an estimated value of the sludge moisture content.

3. The sludge moisture content estimation system of claim 1, further comprising a learning model generation unit that acquires, for each of a plurality of sludge dewatering machines, dewatering machine information, operating data when sludge dewatering treatment is performed within a specified period, sludge type information, and sludge moisture content after sludge dewatering treatment is performed, and generates the plurality of types of learning models based on the acquired information and stores them in the first file memory unit.

4. an estimation accuracy calculation unit that calculates an estimation accuracy of a sludge water content for each combination of dehydrator information related to a sludge dehydrator and sludge type information for each of the learning models generated by the learning model generation unit; The sludge water content estimation system of claim 3, further comprising an optimal model information generation unit that identifies, for each combination of dewatering machine information related to a sludge dewatering machine and sludge type information, the learning model with the highest estimation accuracy calculated by the estimation accuracy calculation unit as an optimal learning model, stores identification information of the identified learning model to generate the optimal model database, and stores it in the optimal model information storage unit.

5. the learning model generation unit acquires, each time a new sludge water content is measured in any of the plurality of sludge dewatering machines, dewatering machine information on the sludge dewatering machine in which the sludge water content was measured, operation data on the sludge dewatering machine when the sludge dewatering treatment was performed, sludge type information on the sludge, and the sludge water content after the sludge dewatering treatment was performed, and updates the plurality of types of learning models based on the acquired information; When the learning model generation unit updates the plurality of types of learning models, the estimation accuracy calculation unit updates the estimation accuracy of the sludge water content using the updated learning models; The sludge moisture content estimation system of claim 4, wherein when the estimation accuracy calculation unit updates the estimation accuracy of the sludge moisture content, the optimal model information generation unit identifies an optimal learning model for each combination of dehydrator information related to the sludge dehydrator and sludge type information based on the updated estimation accuracy of the sludge moisture content, and updates the optimal model database using identification information of the identified optimal learning model.

6. The sludge water content estimation system of claim 5, wherein the optimal model information generation unit generates the optimal model database by storing identification information of an optimal learning model identified for each combination of dehydrator information and sludge type information related to a sludge dehydrator, and when the estimation accuracy calculation unit updates the estimation accuracy of the sludge water content, it identifies an optimal learning model corresponding to the combination of dehydrator information and sludge type information related to the sludge dehydrator based on the updated estimation accuracy of the sludge water content, and updates the optimal model database by using the identification information of the identified optimal learning model to change the identification information of the optimal learning models of all sludge dehydrators corresponding to the combination of dehydrator information and sludge type information related to this sludge dehydrator.

7. Further comprising a dehydrator information storage unit that stores dehydrator information for each sludge dehydrator and information on the type of sludge to be treated, The sludge water content estimation system of claim 3, wherein the learning model generation unit acquires, from any one of the plurality of sludge dewatering machines, operating data related to the sludge dewatering machine when a sludge dewatering process is performed, the sludge water content after the sludge dewatering process is performed, and dewatering machine information related to the sludge dewatering machine, and acquires dewatering machine information and sludge type information related to the sludge dewatering machine corresponding to the acquired information based on the information stored in the dewatering machine information storage unit, and uses these in the generation process of the plurality of types of learning models.

8. a second file storage unit that acquires and stores a file of the optimal learning model acquired by the first estimation unit; a second estimation unit that, when a sludge moisture content estimation process is requested while communication between the estimation request processing unit and the first estimation unit is disconnected, executes the sludge moisture content estimation process by inputting dewatering treatment conditions corresponding to the requested estimation process into a file of the optimal learning model stored in the second file storage unit, and obtains an estimated value of the sludge moisture content, The sludge water content estimating system according to claim 1 , wherein the output control unit causes an output device to output the estimated value acquired by the second estimating unit.

9. a first file storage unit that stores a plurality of types of learning model files for estimating a sludge water content after a sludge dewatering process is performed by a sludge dewatering machine under arbitrary dewatering conditions; a communication-capable connection to an optimum model information storage unit that stores an optimum model database that stores identification information of an optimum learning model identified from the plurality of types of learning models for each sludge dewatering treatment condition; an estimation request processing unit that, when a request is made for estimating a sludge moisture content, acquires dehydration conditions for a sludge dehydration process, which are conditions for executing the estimation process; a first estimation unit that acquires, from the optimal model database, identification information of an optimal learning model to be used in the estimation process of the sludge moisture content based on the dewatering treatment conditions acquired by the estimation request processing unit, acquires a file of the optimal learning model corresponding to the acquired identification information from the first file storage unit, and executes the estimation process of the sludge moisture content by inputting the dewatering treatment conditions acquired by the estimation request processing unit into the optimal learning model, thereby acquiring an estimated value of the sludge moisture content; an output control unit that causes an output device to output the estimated value acquired by the first estimation unit.

10. a first file storage unit that stores a plurality of types of learning model files for estimating a sludge water content after a sludge dewatering process is performed by a sludge dewatering machine under arbitrary dewatering conditions; an estimation device communicably connected to an optimum model information storage unit that stores an optimum model database storing identification information of an optimum learning model specified from among the plurality of types of learning models for each sludge dewatering treatment condition; When a sludge moisture content estimation process is requested, a dehydration process condition for the sludge dehydration process, which is an execution condition for the estimation process, is obtained; acquiring, from the optimum model database, identification information of an optimum learning model to be used in the estimation process of the sludge moisture content based on the acquired dewatering treatment conditions, acquiring a file of the optimum learning model corresponding to the acquired identification information from the first file storage unit, and executing the estimation process of the sludge moisture content by inputting the acquired dewatering treatment conditions into the optimum learning model to acquire an estimated value of the sludge moisture content; The method for estimating sludge moisture content includes outputting the obtained estimated value from an output device.

11. a first file storage unit that stores a plurality of types of learning model files for estimating a sludge water content after a sludge dewatering process is performed by a sludge dewatering machine under arbitrary dewatering conditions; an estimation device communicably connected to an optimum model information storage unit that stores an optimum model database storing identification information of an optimum learning model specified from among the plurality of types of learning models for each sludge dewatering treatment condition; A function for acquiring, when a sludge moisture content estimation process is requested, dehydration process conditions for the sludge dehydration process, which are the execution conditions for the estimation process; a function of acquiring, from the optimum model database, identification information of an optimum learning model to be used in the estimation process of the sludge moisture content based on the acquired dewatering treatment conditions, acquiring a file of the optimum learning model corresponding to the acquired identification information from the first file storage unit, and inputting the acquired dewatering treatment conditions into the optimum learning model to execute the estimation process of the sludge moisture content and acquire an estimated value of the sludge moisture content; A function to output the obtained estimated value from an output device, and a sludge moisture content estimation program to achieve this.

12. a learning model generation unit that acquires, for each of the plurality of sludge dewatering machines, dewatering machine information, operation data when a sludge dewatering process is performed within a predetermined period, sludge type information, and a sludge water content after the sludge dewatering process is performed, and generates a plurality of types of learning models based on the acquired information; a communication-capable estimation accuracy calculation unit that calculates an estimation accuracy of a sludge water content for each combination of dehydrator information related to a sludge dehydrator and sludge type information for each of the learning models generated by the learning model generation unit; A database generation device comprising an optimal model information generation unit that identifies, for each combination of dewatering machine information related to a sludge dewatering machine and sludge type information, the learning model with the highest estimated accuracy calculated by the estimated accuracy calculation unit as an optimal learning model, and generates an optimal model database that stores identification information of the identified learning model.

13. a learning model generation unit that acquires, for each of the plurality of sludge dewatering machines, dewatering machine information, operation data when a sludge dewatering process is performed within a predetermined period, sludge type information, and a sludge water content after the sludge dewatering process is performed, and generates a plurality of types of learning models based on the acquired information; a database generating device communicably connected to an estimation accuracy calculating unit that calculates an estimation accuracy of a sludge water content for each combination of dehydrator information related to a sludge dehydrator and sludge type information for each learning model generated by the learning model generating unit, A database generation method in which, for each combination of dewatering machine information related to a sludge dewatering machine and sludge type information, the learning model with the highest calculated estimation accuracy is identified as an optimal learning model, and an optimal model database is generated that stores identification information for the identified learning model.

14. A database used in a sludge moisture content estimation system that estimates a sludge moisture content after a sludge dewatering process is performed by a sludge dewatering machine, An optimal model database that stores, for each combination of dehydrator information about a sludge dehydrator and sludge type information about the sludge, identification information for the optimal learning model with the highest estimation accuracy among multiple types of learning models that estimate the sludge water content after sludge dehydration processing is performed.

Images