Method for predicting quality of ready-mixed concrete, quality prediction program, manufacturing method of ready-mixed concrete, device for predicting quality of ready-mixed concrete, and manufacturing system of ready-mixed concrete

Abstract

To improve the accuracy of prediction when predicting the quality of fresh concrete using a prediction model. [Solution] A method for predicting the quality of ready-mixed concrete, comprising: a construction step of constructing a predictive model by machine learning based on training data including multiple data sets; an acquisition step of acquiring multiple feature quantities obtained when producing ready-mixed concrete using a mixer; and a prediction step of utilizing the predictive model to acquire a predicted value of quality according to the multiple feature quantities acquired in the acquisition step, wherein each of the multiple data sets corresponds to a correct value of quality, and the multiple feature quantities include one or more load feature quantities related to the power load value of the mixer when mixing concrete materials or stirring ready-mixed concrete, and wherein in the construction step, a predictive model is constructed such that one or more feature quantities selected from the multiple feature quantities correspond to the predicted value of quality using a decision tree.

Description

[Technical Field]

[0001] The present disclosure relates to a quality prediction method for ready-mixed concrete, a quality prediction program, a manufacturing method for ready-mixed concrete, a quality prediction device for ready-mixed concrete, and a manufacturing system for ready-mixed concrete. [Background technology]

[0002] Patent Documents 1 and 2 disclose methods for predicting the quality of ready-mixed concrete using a prediction model. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2020-144099 [Patent Document 2] Patent Publication No. 2021-124304 Summary of the Invention [Problem to be solved by the invention]

[0004] The present disclosure provides a ready-mixed concrete quality prediction method, a quality prediction program, a ready-mixed concrete manufacturing method, a ready-mixed concrete quality prediction device, and a ready-mixed concrete manufacturing system that are useful for improving prediction accuracy when predicting the quality of ready-mixed concrete using a prediction model. [Means for solving the problem]

[0005] [1] A quality prediction method for predicting the quality of ready-mixed concrete, comprising: a construction step of constructing a prediction model by machine learning based on training data including a plurality of data sets; an acquisition step of acquiring a plurality of feature quantities obtained when producing ready-mixed concrete using a mixer; and a prediction step of using the prediction model constructed in the construction step to acquire a predicted value of the quality corresponding to the plurality of feature quantities acquired in the acquisition step, wherein, for each of the plurality of data sets, the plurality of feature quantities are associated with a correct value of the quality, and the plurality of feature quantities include one or more load feature quantities related to the power load value of the mixer when mixing concrete materials or stirring ready-mixed concrete, and in the construction step, the prediction model is constructed so that one or more feature quantities selected from the plurality of feature quantities are associated with the predicted value of the quality using a decision tree.

[0006] [2] The method for predicting the quality of ready-mixed concrete according to [1] above, wherein in the construction step, the prediction model is constructed by machine learning including ensemble learning.

[0007] [3] The method for predicting the quality of ready-mixed concrete according to [2] above, wherein in the construction step, the prediction model is constructed by machine learning using boosting.

[0008] [4] The method for predicting the quality of ready-mixed concrete according to [2] or [3] above, wherein in the construction step, the prediction model is constructed by machine learning using gradient boosting.

[0009] [5] The method for predicting the quality of ready-mixed concrete according to any one of [1] to [4] above, wherein the one or more load features include at least one of a value at a predetermined time point in time series data of the power load value of the mixer when mixing concrete materials or stirring ready-mixed concrete, and a statistical quantity obtained from the time series data, and the statistical quantity is one or more values ​​selected from the group consisting of the difference between the maximum value and the minimum value, the difference between the maximum value and the final value, the sum, the mean, the standard deviation, the coefficient of variation, the median, the first quartile, the third quartile, kurtosis, and skewness.

[0010] [6] The method for predicting the quality of ready-mixed concrete according to any one of [1] to [5] above, wherein the plurality of feature amounts further includes one or more time feature amounts relating to a time when the mixer mixes the concrete materials, and the one or more time feature amounts include at least one of the date and time when the mixer mixes the concrete materials, the cumulative number of batches mixed by the mixer in one day, the time from an arbitrarily set reference time point to the time when the mixer mixed the concrete materials, and the cumulative number of batches mixed by the mixer from an arbitrarily set reference time point.

[0011] [7] A quality prediction program that causes a computer to execute the quality prediction method according to any one of [1] to [6] above.

[0012] [8] A method for producing ready-mixed concrete, comprising: a production step of producing ready-mixed concrete; and a quality prediction step of predicting the quality of the ready-mixed concrete using the quality prediction method described in any one of [1] to [6] above after or during the production step.

[0013] [9] A quality prediction device for predicting the quality of ready-mixed concrete, comprising: a model construction unit that constructs a prediction model by machine learning based on training data including a plurality of data sets; an acquisition unit that acquires a plurality of feature quantities obtained when producing ready-mixed concrete using a mixer; and a prediction unit that uses the prediction model constructed by the model construction unit to acquire a predicted value of the quality according to the plurality of feature quantities acquired by the acquisition unit, wherein, for each of the plurality of data sets, the plurality of feature quantities are associated with a correct value of the quality, and the plurality of feature quantities include one or more load feature quantities related to the power load value of the mixer when mixing concrete materials or stirring ready-mixed concrete, and the model construction unit constructs the prediction model so that one or more feature quantities selected from the plurality of feature quantities are associated with the predicted value of the quality using a decision tree.

[0014]

[10] A ready-mixed concrete manufacturing system comprising a manufacturing apparatus for manufacturing ready-mixed concrete and the quality prediction device described in [9] above, wherein the quality prediction device predicts the quality of the ready-mixed concrete manufactured by the manufacturing apparatus. [Effects of the Invention]

[0015] According to the present disclosure, there are provided a ready-mixed concrete quality prediction method, a quality prediction program, a ready-mixed concrete manufacturing method, a ready-mixed concrete quality prediction device, and a ready-mixed concrete manufacturing system that are useful for improving the prediction accuracy when predicting the quality of ready-mixed concrete using a prediction model. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a schematic diagram showing an example of a ready-mix concrete manufacturing system. [Figure 2] FIG. 2 is a block diagram illustrating an example of a functional configuration of the control device. [Figure 3] FIG. 3 is a diagram illustrating an example of a prediction model. [Figure 4] 4(a) and 4(b) are diagrams for explaining an example of the process of building a prediction model. [Figure 5] FIG. 5 is a diagram illustrating an example of a process for constructing a prediction model. [Figure 6] 6(a) and 6(b) are diagrams for explaining an example of the process of building a prediction model. [Figure 7] 7(a) and 7(b) are diagrams for explaining an example of the process of building a prediction model. [Figure 8] 8(a) and 8(b) are diagrams illustrating data relating to power load values. [Figure 9] FIG. 9 is a diagram for explaining an example of a format for expressing dates and times. [Figure 10] FIG. 10 is a block diagram illustrating an example of a hardware configuration of the control device. [Figure 11]11(a) is a flowchart showing an example of a series of processes in the training phase, and FIG. 11(b) is a flowchart showing an example of a series of processes in the evaluation phase. [Figure 12] 12(a) and 12(b) are graphs showing the results of verifying the prediction accuracy of the prediction model. [Figure 13] 13(a) and 13(b) are graphs showing the results of verifying the prediction accuracy of the prediction model. DETAILED DESCRIPTION OF THE INVENTION

[0017] An embodiment will be described below with reference to the drawings. In the description, identical elements or elements having identical functions are given the same reference numerals, and duplicated explanations will be omitted. Figure 1 shows a schematic diagram of a ready-mixed concrete manufacturing system equipped with a quality prediction device according to one embodiment.

[0018] [Ready-mix concrete manufacturing system] First, an overview of a ready-mixed concrete manufacturing system will be described. Manufacturing system 1 (ready-mixed concrete manufacturing system) shown in Fig. 1 is a system for manufacturing ready-mixed concrete. Manufacturing system 1 manufactures ready-mixed concrete through at least a process of mixing concrete materials. The concrete materials used in manufacturing system 1 include cement, admixtures, coarse aggregate, fine aggregate, water, and admixtures.

[0019] Examples of coarse aggregate include natural coarse aggregate and artificial coarse aggregate. Examples of coarse aggregate include gravel, crushed stone, slag coarse aggregate, lightweight coarse aggregate, recycled coarse aggregate, recovered coarse aggregate, and mixtures thereof. Examples of gravel include mountain gravel, land gravel, river gravel, and sea gravel. Examples of slag coarse aggregate include blast furnace slag aggregate, ferronickel slag aggregate, electric furnace oxidizing slag aggregate, and coal gasification slag aggregate. Examples of lightweight coarse aggregate include natural lightweight aggregate, by-product lightweight aggregate, and artificial lightweight aggregate. Examples of coarse aggregate include crushed rock or limestone crushed stone.

[0020] Examples of fine aggregates include natural aggregates and artificial aggregates. Examples of fine aggregates include sand, crushed sand, slag fine aggregate, lightweight fine aggregate, recycled fine aggregate, recovered fine aggregate, and mixtures of these. Examples of sand include mountain sand, land sand, river sand, and sea sand. Examples of slag fine aggregates include blast furnace slag aggregate, ferronickel slag aggregate, copper slag aggregate, electric furnace oxidized slag aggregate, and coal gasification slag aggregate. Examples of lightweight fine aggregates include natural lightweight aggregate, by-product lightweight aggregate, and artificial lightweight aggregate.

[0021] Examples of rock types for crushed stone and crushed sand include igneous rocks, sedimentary rocks, metamorphic rocks, quartz, limestone, domalite, and peridotite. Igneous rocks include granite, diorite, gabbro, porphyrite, diabase, rhyolite, andesite, basalt, and serpentine. Sedimentary rocks include conglomerate, sandstone, shale, slate, and tuff. Metamorphic rocks include gneiss and crystalline schist. A mixture of two or more of the above-mentioned examples may be used as the coarse aggregate and fine aggregate.

[0022] At least a portion of the manufacturing system 1 is installed, for example, in a location (such as a factory) different from the location (site) where the ready-mixed concrete is used. After the ready-mixed concrete is loaded onto the transport vehicle 200, the transport vehicle 200 transports the ready-mixed concrete to the site (such as a construction site) where the ready-mixed concrete is to be used. Examples of the transport vehicle 200 include an agitator vehicle (mixer vehicle) or a dump truck. The manufacturing system 1 may manufacture ready-mixed concrete from concrete materials so as to satisfy a target quality (required quality) set for each site. For example, an operator of the manufacturing system 1 determines the mix of concrete materials so as to satisfy the target quality, and inputs operating instructions to the manufacturing system 1.

[0023] In one example, ready-mixed concrete is produced in the manufacturing system 1, and the quality of the ready-mixed concrete is controlled (inspected, etc.) before shipping so that the target quality of the ready-mixed concrete when in use, which is set for each site, is met. The time when ready-mixed concrete is used corresponds to the time when the ready-mixed concrete is received at the site. In order to control the quality of the ready-mixed concrete before shipping, the target quality of the ready-mixed concrete when shipped may be set based on the target quality of the ready-mixed concrete when in use. The quality of ready-mixed concrete includes fresh properties. Specific examples of fresh properties include slump, slump flow, and air content.

[0024] The manufacturing system 1 includes, for example, a manufacturing apparatus 100 and a quality prediction apparatus 10. The manufacturing apparatus 100 is an apparatus that mixes concrete materials to manufacture ready-mixed concrete. The manufacturing apparatus 100 may be installed in a factory or the like that is different from the site where the ready-mixed concrete is used. The manufacturing apparatus 100 includes, for example, a material storage area 101, a transport device 104, a storage bottle 111, a measuring bottle 112, a collecting hopper 113, a mixer 114, and a loading hopper 115.

[0025] The material storage yard 101 is a place where concrete materials are stored. The material storage yard 101 includes a plurality of silos 102. The plurality of silos 102 are containers that store at least a portion of the concrete materials by material type. The plurality of silos 102 include, for example, a silo 102 that stores coarse aggregate, a silo 102 that stores fine aggregate, and a silo 102 that stores cement.

[0026] The transporting device 104 is a device that transports the concrete materials stored in the multiple silos 102 to the storage bins 111. The transporting device 104 includes, for example, a belt conveyor that transports the concrete materials. The transporting device 104 may transport the concrete materials by type at different times. In one example, based on an operation instruction from a control device provided in the manufacturing system 1, a specific material from among the various concrete materials is transferred to the transporting device 104 and transported to the storage bins 111.

[0027] The storage bottles 111 temporarily store various types of concrete materials. The various types of concrete materials are transported (conveyed) to the storage bottles 111 from the material storage area 101 by the transport device 104. The storage bottles 111 are configured to individually store various types of concrete materials. Hereinafter, "concrete materials" may be simply referred to as "materials." The various materials stored in the storage bottles 111 are supplied to the measuring bottles 112 as needed.

[0028] The measuring bottle 112 is disposed below the storage bottle 111. The measuring bottle 112 operates based on operational instructions from a control device provided in the manufacturing system 1, and individually measures various materials. When the measuring bottle 112 detects the target amount of material instructed by the control device, it supplies the material to the collecting hopper 113. When water is supplied to the measuring bottle 112, an admixture may be mixed into the water. The collecting hopper 113 is disposed below the measuring bottle 112. The collecting hopper 113 collects the various materials discharged from the measuring bottle 112 and supplies the collected various materials to the mixer 114. Note that the manufacturing apparatus 100 does not necessarily have to be provided with the collecting hopper 113, and the various materials may be supplied to the mixer 114 from the measuring bottle 112.

[0029] The mixer 114 is disposed below the collecting hopper 113. The mixer 114 is a device that mixes concrete materials. The mixer 114 produces ready-mixed concrete by mixing (kneading) aggregate, cement, water, admixtures, etc. The ready-mixed concrete is discharged from the bottom of the mixer 114 into a loading hopper 115. The mixer 114 may be a tilting mixer, a horizontal single-shaft mixer, a horizontal twin-shaft mixer, or a pan-type mixer. The mixer 114 includes, for example, two stirring members 114a and a mixer drive unit 114b.

[0030] The agitating member 114a is a member that agitates various materials supplied to the mixer 114. The two agitating members 114a are arranged side by side inside the main body (container portion) of the mixer 114 and are rotatable. Each of the two agitating members 114a includes a rotation shaft that extends horizontally in one direction. The mixer driving unit 114b rotates the rotation shaft of each of the two agitating members 114a based on an operation instruction from a control device provided in the manufacturing system 1. The mixer driving unit 114b includes, for example, a driving source such as a motor that applies driving force to the agitating members 114a. An opening and closing port is provided at the bottom of the main body of the mixer 114 for discharging the manufactured ready-mixed concrete into the loading hopper 115.

[0031] The loading hopper 115 is disposed below the mixer 114 and temporarily stores the ready-mixed concrete. The loading hopper 115 supplies the temporarily stored ready-mixed concrete to the transport vehicle 200.

[0032] The manufacturing apparatus 100 described above is an example of a ready-mixed concrete manufacturing apparatus, and the ready-mixed concrete manufacturing apparatus may be configured in any manner as long as it is capable of manufacturing ready-mixed concrete through a process of mixing concrete materials with a mixer. In the present disclosure, the manufacturing of ready-mixed concrete may include not only obtaining ready-mixed concrete by mixing concrete materials, but also transporting the obtained ready-mixed concrete and stirring the ready-mixed concrete during transportation. In one example, the manufacturing system 1 also includes a transport vehicle 200.

[0033] In this disclosure, a unit of ready-mixed concrete produced by one mixing in the mixer 114 and loaded onto the transport vehicle 200 is defined as "one batch." The process executed by the manufacturing system 1 to produce one batch of ready-mixed concrete is defined as "batch processing." The number of times a batch processing has been executed since a certain reference point in time (cumulative number of executions) is referred to as the "cumulative batch number." At the reference point in time, the cumulative batch number is reset to zero. For example, the cumulative batch number is reset to zero each day when the manufacturing apparatus 100 starts operating. In this case, the cumulative batch number for the batch processing executed the kth time in a day (k is an integer greater than or equal to 1) is k.

[0034] One to three batches of ready-mixed concrete may be loaded onto one transport vehicle 200. For example, when two batches of ready-mixed concrete are loaded onto one transport vehicle 200, two batch processes according to the same production conditions are carried out at different times (in sequence). Multiple batch processes according to the same production conditions may be carried out consecutively in sequence during a certain period of the day.

[0035] <Quality prediction device> The quality prediction device 10 is a device that predicts the quality of ready-mixed concrete produced by the manufacturing apparatus 100. The quality prediction device 10 is configured with one or more computers. When the quality prediction device 10 is configured with multiple computers, these computers are connected to each other so that they can communicate with each other. The quality prediction device 10 may have a function to control the manufacturing apparatus 100 in addition to the function of predicting quality. For example, the quality prediction device 10 is configured as a part of a control device that controls the manufacturing apparatus 100. The quality prediction device 10 may control the manufacturing apparatus 100 according to set operating conditions. At least a part of the operating conditions may be determined by instructions from an operator such as a worker.

[0036] An input device 12 and a monitor 14 may be connected to the quality prediction apparatus 10. The input device 12 is a device that inputs information indicating instructions from a worker or the like to the quality prediction apparatus 10. The input device 12 may be any device that can input desired information, such as a keyboard (keypad), an operation panel, or a mouse. The monitor 14 is a device that displays information from the quality prediction apparatus 10 to a worker or the like. The monitor 14 may be any device that can display graphics, such as a liquid crystal display. The input device 12 and the monitor 14 may be integrated into one device, such as a touch panel. The quality prediction apparatus 10, the input device 12, and the monitor 14 may be integrated into one device, such as a tablet computer (tablet terminal).

[0037] The quality of ready-mixed concrete to be predicted by the quality prediction device 10 includes not only the quality of the ready-mixed concrete itself but also the quality of the ready-mixed concrete after hardening (hardened concrete). Predicting the quality of ready-mixed concrete means calculating a predicted value of an index value that represents the quality.

[0038] The quality prediction device 10 predicts, for example, the fresh properties of ready-mixed concrete as the quality of ready-mixed concrete. The quality prediction device 10 may predict, as the quality of ready-mixed concrete, the fresh properties of ready-mixed concrete after it has been obtained by mixing in the manufacturing apparatus 100 and before it is shipped to a construction site from a factory or the like in which the manufacturing apparatus 100 is installed. Alternatively, the quality prediction device 10 may predict, as the quality of ready-mixed concrete, the fresh properties of ready-mixed concrete after it has been shipped from a factory or the like and before it is poured at a construction site. In other words, the quality prediction device 10 may predict the fresh properties of ready-mixed concrete during transportation, or may predict the fresh properties of ready-mixed concrete at the time of unloading.

[0039] The quality prediction device 10 may predict one or more index values ​​of slump, slump flow, and air content as fresh properties of ready-mixed concrete. The quality prediction device 10 may predict two or more index values ​​of slump, slump flow, and air content as fresh properties of ready-mixed concrete. The quality prediction device 10 may predict index values ​​other than slump, slump flow, and air content as fresh properties of ready-mixed concrete. Indicators of fresh properties other than slump, slump flow, and air content include, for example, 500 mm flow time, flow stop time, presence or absence of material segregation, temperature, bleeding amount, bleeding rate, setting time, unit water content, plastic viscosity, yield value, funnel flow time, compactibility, deformability, packability, and gap passability.

[0040] The quality prediction device 10 may predict at least one of an index representing strength and an index representing durability as the quality of hardened fresh concrete. Examples of strength indexes include compressive strength, flexural strength, splitting tensile strength, bond strength, static modulus of elasticity, resilience, and ductility coefficient. Examples of durability indexes include length change rate, expansion rate, mass reduction rate, carbonation depth, chloride ion penetration depth, chloride ion diffusion coefficient, air permeability coefficient, water permeability coefficient, freeze-thaw resistance, electrical resistivity, dynamic modulus of elasticity, Poisson's ratio, and creep coefficient. Hereinafter, "quality of fresh concrete" may be simply referred to as "quality."

[0041] To provide an overview of the functions of the quality prediction device 10, the quality prediction device 10 is configured to at least construct a prediction model by machine learning based on training data including multiple data sets, and to acquire multiple feature quantities obtained when ready-mixed concrete is produced using a mixer 114. The quality prediction device 10 is further configured to use the prediction model to acquire a predicted value of quality according to the acquired multiple feature quantities. The quality prediction device 10 constructs the prediction model such that one or more feature quantities selected from the multiple feature quantities are associated with the predicted value of quality using a decision tree.

[0042] 2 shows an example of functional components (hereinafter referred to as "functional blocks") included in the quality prediction device 10. The quality prediction device 10 has, for example, the following functional blocks: an operation control unit 22, a feature acquisition unit 24, a model construction unit 26, a model storage unit 28, a prediction unit 30, and an output unit 32. The processes executed by these functional blocks correspond to the processes executed by the quality prediction device 10. Below, each functional block will be explained using an example in which the quality to be predicted is the fresh properties of ready-mixed concrete (more specifically, the fresh properties of ready-mixed concrete after it has been obtained by mixing in the mixer 114 and before it is shipped).

[0043] The operation control unit 22 controls the manufacturing apparatus 100 to manufacture ready-mixed concrete in accordance with predetermined operating conditions. At least some of the operating conditions may be determined by an operator, such as a worker, each time ready-mixed concrete is manufactured. The operation control unit 22 may control the mixer driving unit 114b of the mixer 114 so that the rotation speed of the mixer driving unit 114b follows a target rotation speed defined in the operating conditions. When controlling the mixer driving unit 114b, the operation control unit 22 may adjust the power (e.g., current value) supplied to the mixer driving unit 114b. If the ready-mixed concrete to be manufactured is hard, the power load value tends to be large, and if the ready-mixed concrete to be manufactured is soft, the power load value tends to be small.

[0044] The feature acquisition unit 24 (acquisition unit) acquires a plurality of feature amounts obtained when ready-mixed concrete is produced using the mixer 114. Hereinafter, each of the plurality of feature amounts acquired by the feature acquisition unit 24 may be referred to as feature amount F1, feature amount F2, ..., and feature amount FN (N is an integer of 2 or more). The feature amounts F1 to FN are input data to a prediction model for predicting fresh properties. At least some of the feature amounts F1 to FN (plurality of feature amounts) may be feature amounts that represent the results of the operation of the mixer 114 or feature amounts that represent the state of the mixer 114 while it is operating. The feature amounts F1 to FN may include feature amounts that represent the conditions of production using the mixer 114 (for example, the operating conditions described above).

[0045] Each of the feature quantities F1 to FN is data represented by a numerical value. The feature quantities F1 to FN may include data representing a classification (type), in which case a different numerical value (for example, 0 or 1) is assigned to each classification. For example, a value of 0 can be used when an object does not fall into that classification, and a value of 1 can be used when an object falls into that classification. The feature quantities F1 to FN do not include image data formed by combining coordinate information and pixel values. However, the feature quantities F1 to FN may include feature quantities (one-dimensional numerical data) obtained from an image related to such image data.

[0046] The feature quantities F1 to FN include one or more feature quantities related to the power load value of the mixer 114. Hereinafter, the one or more feature quantities related to the power load value among the feature quantities F1 to FN will be referred to as "load feature quantity FL" to distinguish them from other types of feature quantities. The load feature quantity FL related to the power load value of the mixer 114 is, for example, a feature quantity related to the power load value of the mixer 114 when mixing concrete materials. The power load value of the mixer 114 may be a value indicating the power (W) itself supplied to the mixer 114, or may be a value indicating the current value (A) supplied to the mixer 114. Alternatively, the power load value of the mixer 114 may be a value indicating the load hydraulic pressure (Pa) or a value indicating the torque (N m) instead of the power or current value.

[0047] The one or more load feature values ​​FL include, for example, at least one of a value at a predetermined time point in time-series data of the power load value of the mixer 114 when mixing concrete materials and a statistic obtained from the time-series data. The time-series data of the power load value may be a data group (a data group showing changes in the power load value over time) made up of consecutive power load values ​​for at least a part of the period during which the mixer 114 is operating during the processing of one batch. Specific examples of the features F1 to FN and the load feature value FL will be described later.

[0048] The model construction unit 26 constructs a prediction model by machine learning based on training data including multiple data sets. Hereinafter, the training data including multiple data sets will be referred to as "training data TD", and the prediction model constructed by the model construction unit 26 will be referred to as "prediction model M". The prediction model M is a model for predicting the fresh properties of ready-mixed concrete when the feature quantities F1 to FN are obtained. The fresh properties predicted by the prediction model M are, for example, slump, slump flow, or air content.

[0049] The model construction unit 26 constructs a prediction model M that indicates the relationship between the features F1 to FN and the fresh property (predicted value) through machine learning. The prediction model M is configured to output a predicted value of the fresh property in response to the input of the features F1 to FN. The features F1 to FN can be referred to as explanatory variables, and the fresh property can be referred to as a target variable.

[0050] Machine learning is a method in which a machine (computer) autonomously discovers laws or rules by repeatedly learning based on given information. A predictive model M can be constructed using algorithms and data structures. A predictive model M is realized using a decision tree, which is a method of analyzing data using a tree structure (tree diagram). A predictive model M outputs a predicted value by sequentially setting conditions for given input data and branching to expected results.

[0051] The model construction unit 26 constructs a prediction model M so that one or more feature quantities selected from the feature quantities F1 to FN correspond to the predicted values ​​of the fresh properties using a decision tree (decision tree algorithm). The one or more feature quantities used for the correspondence with the fresh properties in the prediction model M are selected from the feature quantities F1 to FN by machine learning. As the decision tree algorithm, LightGBM (Light Gradient Boosting Machine) may be used. Note that, instead of LightGBM, XGBoost (eXtreme Gradient Boosting) or RandomForest may be used as the decision tree algorithm.

[0052] The model construction unit 26 may autonomously construct a prediction model M by performing machine learning using data provided as input for machine learning and correct data (correct values ​​of fresh properties) output from the machine learning. The input for the machine learning is the features F1 to FN. The output of the machine learning is data (numerical values) indicating the fresh properties of ready-mixed concrete. The model construction unit 26 iteratively learns a model that outputs predicted values ​​of the fresh properties using multiple combinations of the features F1 to FN and the correct values ​​of the fresh properties (the training data TD). In each of the multiple data sets that make up the training data TD, the features F1 to FN are associated with the correct values ​​of the fresh properties.

[0053] The stage of autonomously constructing the prediction model M corresponds to the training phase (learning phase). The training phase may be performed before the production phase in which ready-mix concrete is manufactured, or may be performed at the beginning of the production phase. The stage of predicting the fresh properties using the prediction model M from the features F1 to FN (combinations of the values ​​of the feature values ​​F1 to FN) for which the fresh properties are unknown corresponds to the evaluation phase. In the following description, the terms "for training" and "for evaluation" may be used to distinguish between the training phase and the evaluation phase, or in which phase the data is used. The same types of features F1 to FN are used in the training phase and the evaluation phase.

[0054] Here, an example of the analysis process in the prediction model M will be described with reference to FIG. 3. FIG. 3 shows a tree diagram visualizing an example of the analysis process in the prediction model M. Note that a simplified example will be used for easier understanding of the contents of the present disclosure. In the algorithm using the decision tree exemplified below, one data group is divided into two data groups at each stage according to a certain condition, and the stages are referred to in descending order as "first stage," "second stage," and "third stage." Furthermore, the point where division or branching occurs based on a condition is referred to as a "node (condition node)." In FIG. 3, a combination of "G" and a numerical value, such as "G1," represents the name of a data group. Furthermore, a combination of "Th" and a numerical value, such as "Th1," represents a threshold.

[0055] The model construction unit 26 prepares training data TD consisting of a plurality of data sets each having a different combination of values ​​of feature quantities F1 to FN. The data group G1 is, for example, all of the data sets included in the training data TD. In the first stage, the data group G1 is divided into two data groups, data group G21 and data group G22, by one feature quantity (feature quantity F2 in the example shown in FIG. 3) selected by machine learning from the feature quantities F1 to FN. In the first stage, the data group G1 is divided into two data groups based on the condition that the feature quantity F2 is smaller than a threshold value Th1 or greater than or equal to the threshold value Th1.

[0056] The data group G21 is composed of a plurality of data sets from the data group G1 in which the feature F2 is smaller than the threshold Th1, and the data group G22 is composed of the remaining plurality of data sets from the data group G1 in which the feature F2 is equal to or greater than the threshold Th1. The number of data sets in the data group G21 and the number of data sets in the data group G22 may be the same as or different from each other. In the first stage, the feature F2 is used to separate the data into two data groups, but the values ​​of features other than the feature F2 are also retained in the data group G21 and the data group G22. The threshold Th1 is also set autonomously by machine learning.

[0057] In the second stage, similar to the first stage, each of the data groups G21 and G22 is divided into two data groups using a feature selected by machine learning and a threshold. The feature selected in the second stage may be different from or the same as the feature selected in the first stage. The feature selected to divide each of the data groups G21 and G22 into two may be different from or the same as the feature selected in the first stage. In the example shown in FIG. 3, a feature F1 is selected to divide the data group G21 into two, and a feature F2 is selected to divide the data group G22 into two. In the second stage, the data group G21 is divided into the data group G31 and the data group G32, and the data group G22 is divided into the data group G33 and the data group G34.

[0058] In the example shown in FIG. 3, in the third stage, data group G31, data group G32, and data group G34 are not divided into two data groups. That is, data group G31, data group G32, and data group G34 each correspond to a "leaf" (or "end node") in the tree diagram. In the third stage, data group G33, which is not a leaf data group, is divided into two data groups using a feature selected by machine learning and a threshold value, as in the previous stage. A feature F10 is selected to divide data group G33 into two data groups, and data group G33 is divided into data group G41 and data group G42. Data group G41 and data group G42 each correspond to a leaf. As exemplified in FIG. 3, the leaf depths (levels) may differ among at least some of the multiple leaves in the tree diagram. Alternatively, unlike the example shown in FIG. 3, the depths of the leaves may be the same.

[0059] For each data group corresponding to a leaf, a predicted value of the fresh property can be determined based on the correct value of the fresh property contained in that data group. For example, the arithmetic mean of the correct value of the fresh property contained in the data group corresponding to the leaf is set as the predicted value for that leaf. In one example, if the data group G41 includes m data sets (m is an integer equal to or greater than 2), the arithmetic mean of the m correct values ​​of the fresh property is set as the predicted value for the leaf corresponding to the data group G41. The predicted values ​​of the fresh property set for multiple leaves differ.

[0060] The model construction unit 26 may perform division (branching) up to the data groups corresponding to leaves, while selecting features to be associated with predicted values ​​of fresh properties and setting thresholds to be used for branching, based on any construction conditions. As described above, in the training phase, the training data TD is divided into two data groups according to the conditions for each stage, and a predicted value of fresh properties is set for each leaf (end node).

[0061] In the evaluation phase, where a prediction model M is used to obtain a predicted value of fresh properties from the evaluation features F1 to FN, the given evaluation features F1 to FN (combinations of the values ​​of the features F1 to FN) are evaluated to determine which leaf they fall into according to the branching conditions for each stage. Each leaf has a predicted value of fresh properties, so a predicted value of fresh properties can be obtained according to the combination of the values ​​of the features F1 to FN.

[0062] In the prediction model M, there are cases where only a selected portion of the features F1 to FN are used as branching conditions. That is, in the prediction model M, there are cases where only a selected portion of the features F1 to FN are associated with the predicted value of the fresh property. Even in such cases, the portion of the features associated with the predicted value of the fresh property is selected from the features F1 to FN, so the prediction model M still indicates the relationship between the features F1 to FN and the predicted value of the fresh property.

[0063] The model construction unit 26 may construct a corresponding prediction model M for each type of fresh property (a prediction model M that predicts only one type of fresh property). For example, the model construction unit 26 may separately construct a prediction model M for obtaining a predicted value of slump, a prediction model M for obtaining a predicted value of slump flow, and a prediction model M for obtaining a predicted value of air volume.

[0064] 2, the model holding unit 28 holds (stores) the prediction model M constructed by the model construction unit 26. The prediction model M, which is a trained model, may be transferable between computers. Therefore, the prediction model M constructed in the quality prediction device 10 may be used in a manufacturing system other than the manufacturing system 1.

[0065] The prediction unit 30 uses the prediction model M constructed by the model construction unit 26 to obtain predicted values ​​of fresh properties according to the evaluation feature quantities F1 to FN obtained by the feature quantity acquisition unit 24. For example, the prediction unit 30 inputs the values ​​of the evaluation feature quantities F1 to FN into the prediction model M and obtains predicted values ​​of fresh properties output from the prediction model M. The prediction unit 30 may obtain predicted values ​​for each type of fresh property (for example, for each of slump, slump flow, and air content) using the corresponding prediction model M.

[0066] 3 is constructed, in the prediction model M, first, the value of the evaluation feature F2, of the evaluation features F1 to FN, is compared with a threshold value Th1. If the value of the evaluation feature F2 is equal to or greater than the threshold value Th1, the calculation step proceeds to a node corresponding to the data group G22, where the value of the evaluation feature F2 is compared with the threshold value Th22. If the value of the evaluation feature F2 is smaller than the threshold value Th22, the calculation step proceeds to a node corresponding to the data group G33, where the value of the evaluation feature F10 is compared with the threshold value Th31.

[0067] If the value of the evaluation feature F10 is smaller than the threshold value Th31, the prediction model M outputs the predicted value of the fresh properties set in the leaf (end node) corresponding to the data group G41. By performing the above calculations in the prediction model M, the prediction unit 30 can obtain predicted values ​​of the fresh properties of the ready-mixed concrete to be evaluated from the combination of the values ​​of the evaluation feature F1 to FN.

[0068] The output unit 32 outputs the predicted values ​​of the fresh properties acquired by the prediction unit 30 to the monitor 14. As a result, the predicted values ​​of the fresh properties of the ready-mixed concrete to be evaluated are displayed on the monitor 14, and an operator such as a worker can grasp the predicted values ​​of the fresh properties of the ready-mixed concrete.

[0069] <Example of decision tree algorithm> Next, a specific example of machine learning executed by the model construction unit 26 when constructing a prediction model M based on a decision tree algorithm will be described with reference to Figures 4 to 7. The model construction unit 26 may construct the prediction model M by machine learning including ensemble learning. In machine learning including ensemble learning, a single prediction model M may be constructed by combining multiple models.

[0070] FIG. 4(a) schematically illustrates an analytical process for obtaining predicted values ​​using bagging, a method of ensemble learning. In one example, a first model M1 (weak learner) is generated using a portion of data randomly extracted from training data TD, and a second model M2 (weak learner) is generated using another portion of data randomly extracted from training data TD. A portion of the data used to generate the first model M1 may overlap with a portion of the data used to generate the second model M2. In the prediction model M constructed by bagging, for example, the arithmetic mean of the predicted values ​​of the first model M1 and the second model M2 is output as the final predicted value.

[0071] For simplicity, FIG. 4(a) illustrates the construction of a prediction model M from the fusion of two models (weak learners), but a prediction model M that ultimately outputs one predicted value may also be constructed from three or more models (weak learners). A prediction model M that ultimately outputs one predicted value may also be constructed by Random Forest, in which models (weak learners) generated from a portion of data, such as a first model M1, are generated using a portion of features randomly selected for each model. Note that Random Forest can also be considered a bagging technique.

[0072] The model construction unit 26 may construct the prediction model M by machine learning using boosting as machine learning including ensemble learning. In machine learning using boosting, intermediate models Mtk (weak learners) are generated sequentially using a decision tree algorithm. Note that k is an integer from 1 to K, where K is an integer equal to or greater than 2. If the intermediate models Mtk are generated in order of k from 1 to K, in machine learning using boosting, an intermediate model Mt1 is generated first. Then, focusing on the error between the predicted value in the intermediate model Mt1 and the correct value, an intermediate model Mt2, which is the next intermediate model, is generated. The single predicted value finally output by the prediction model M is a value obtained by aggregating the outputs of the intermediate models Mt1 to MtK through some kind of calculation.

[0073] The model construction unit 26 may construct a prediction model M consisting of intermediate models Mt1 to MtK, for example, using "Adaboost," a boosting technique. FIG. 4(b) schematically shows the process of constructing a model using Adaboost. The weights of the training data TD used to generate the next intermediate model Mt2 are updated based on the error between the predicted value (1) of the previous intermediate model Mt1 and the correct value. For example, the weights are updated so that they are larger for data sets with larger errors. Thereafter, the weight updating and generation of the intermediate model Mtk are repeated, for example, up to a set number of times, K times.

[0074] The model construction unit 26 may construct the prediction model M by machine learning using gradient boosting as the machine learning using boosting. That is, the model construction unit 26 may construct the prediction model M by machine learning using a gradient boosting tree. In the gradient boosting tree, the prediction model M is constructed by combining boosting, gradient descent, and a decision tree.

[0075] FIG. 5 schematically shows the process of constructing a model using gradient boosting. In gradient boosting, an intermediate model Mt1 is first generated, and the error between a predicted value (1) by the intermediate model Mt1 and a correct value is calculated. Then, an intermediate model Mt2 capable of predicting the error in the predicted value (1) is generated, and a correction value (2) for correcting the predicted value (1) is calculated from the intermediate model Mt2. Thereafter, calculation of a correction value (k) for correcting the current predicted value (k) is repeated up to, for example, a set number of times, K. The correction value (k) is also referred to as a gradient (k).

[0076] In addition to the above-mentioned LightGBM and XGBoost, CatBoost is also an example of a decision tree algorithm that includes gradient boosting. In XGBoost, node branching (splitting) is performed level-wise, as shown in FIG. 6(a). That is, all nodes are branched from the left. On the other hand, in LightGBM, node branching (splitting) is performed leaf-wise, as shown in FIG. 6(b). That is, branching is performed by narrowing down to nodes that should be branched (for example, by prioritizing nodes that result in smaller losses). For nodes that no longer require branching, no further calculations are performed.

[0077] Furthermore, LightGBM uses a histogram-based algorithm when branching. FIG. 7(a) schematically shows the decision process when determining branching using a pre-sorted algorithm, which is different from the histogram-based algorithm. FIG. 7(b) schematically shows the decision process when determining branching using the histogram-based algorithm. In FIGS. 7(a) and 7(b), each of "d1" to "d6" represents one piece of data (or one data set). The pre-sorted algorithm checks each value in the data d1 to d6 and then determines where to branch (split). On the other hand, the histogram-based algorithm groups the values ​​in the data d1 to d6 using a histogram, and determines where to branch (split) for each group (group).

[0078] Furthermore, LightGBM uses a technique called GOSS (Gradient-based One-Side Sampling) to reduce the amount of data used during the training process. Specifically, for data with large errors, all of the data is used as data that has not yet been trained, while for data with small errors, only a portion of the data is used as it has been trained to a certain extent. In addition, LightGBM uses a technique called FEB (Exclusive Feature Bundling) to calculate multiple features of different types that do not simultaneously become zero (have low correlation) and can be combined into one.

[0079] <Specific example of load feature FL> Next, specific examples of one or more load feature quantities FL included in the feature quantities F1 to FN will be described with reference to Fig. 8. The feature quantities F1 to FN include, for example, one or more load feature quantities FL among the multiple types of load feature quantities FL exemplified below. Fig. 8(a) schematically shows time-series data relating to the power load values ​​of the mixer 114 in one batch of processing. The time-series data relating to the power load values ​​is, for example, data obtained by repeatedly measuring the power (kW) supplied to the mixer 114 at a predetermined sampling period.

[0080] In the graph showing the time series data in FIG. 8(a), "time t1" indicates the time when the mixer 114 starts mixing, and "time t2" indicates the time when the mixer 114 finishes mixing. Time t1 corresponds to, for example, the time when the operation control unit 22 starts driving the agitating member 114a. Time t2 corresponds to, for example, the time when the operation control unit 22 determines that the condition for finishing mixing by the mixer 114 is met. In one example, when a predetermined time has elapsed since the start of driving the agitating member 114a, the operation control unit 22 determines that the above condition is met and stops driving the agitating member 114a.

[0081] The one or more load feature quantities FL may include values ​​at predetermined time points in the time series data of power load values. The predetermined time points are time points set in advance by an operator or the like. The one or more load feature quantities FL may include one or more values ​​from an initial value P1, a minimum value P2, a maximum value P3, and a final value P4. The initial value P1 is the power load value at the start of mixing in the time series data. The minimum value P2 is the power load value at the minimum point after the time (time t1) when the initial value P1 is obtained. Note that the power load value at the start of mixing (initial value P1) may also be the minimum. The maximum value P3 is the power load value at the maximum point in the time series data. The final value P4 is the power load value at the end of the time series data.

[0082] The period of time series data that is the target for obtaining the load feature FL is not limited to the period from time t1 to time t2. The period of time series data that is the target for obtaining the load feature FL may be a period starting from time t1 and ending at a point when an arbitrary elapsed time tp has elapsed since time t1. In the example shown in FIG. 8(a), the elapsed time tp corresponds to the difference between time t2 and time t1. In this case, the final value P4 corresponds to the power load value at the time when mixing by the mixer 114 is completed.

[0083] Unlike the time series data shown in FIG. 8(a), as shown in FIG. 8(b), one or more values ​​of an initial value P1, a minimum value P2, a maximum value P3, and a final value P4 may be acquired as the load feature FL from the time series data for a period from time t1 to a time before time t2. The start point of the period of time series data that is the target for obtaining the load feature FL may be a time after time t1. For example, the start point and end point of the period are set so that the period of time series data that is the target for obtaining the load feature FL is the same between the training phase and the evaluation phase.

[0084] The one or more load features FL may include statistics obtained from time-series data of power load values ​​instead of or in addition to the power load values ​​at the predetermined time points. The statistics as the load features FL are one or more values ​​selected from the group consisting of a fluctuation range, a decline width, a total value, a mean value, a standard deviation, a coefficient of variation, a median, a first quartile, a third quartile, kurtosis, and skewness. The fluctuation range is calculated by the difference between the maximum value P3 and the minimum value P2. The decline width is calculated by the difference between the maximum value P3 and the final value P4.

[0085] Statistical quantities other than the fluctuation range and the decline width may be calculated from a data group of power load values ​​included in at least a portion of the time series data (a set of power load values ​​obtained for each sampling period). For example, the total value may be found by accumulating the power load values ​​for each sampling period in at least a portion of the time series data. Note that some of the data for the power load values ​​for each sampling period may be thinned out before calculating statistical quantities such as the total value. The time series data for obtaining statistical quantities may be data representing a moving average of the power load values ​​instead of data representing the time change of the power load values ​​themselves. The one or more load feature quantities FL may include feature quantities obtained from the waveform of the time series data and / or feature quantities obtained from an image displaying the time series data.

[0086] The sampling period for measuring the power load value may be 0.01 to 1 second (for example, 0.1 second). Measurement of the power load value may be started immediately after mixing of the concrete materials other than water begins, or immediately after dry mixing (mixing of the concrete materials other than water) and adding water to the concrete materials other than water and starting mixing. The number of power load values ​​measured varies depending on the measurement interval of the power load value and the mixing time, but from the perspective of predicting the fresh properties with higher accuracy, it is preferably 1,000 or more, more preferably 2,000 or more, and particularly preferably 2,500 or more. The power load value can be measured by a device installed on the same power source as the mixer 114.

[0087] The one or more load features FL may include a change in the power load value, or may include information (numerical data) that indicates a change pattern in the power load value. From the viewpoint of predicting the fresh properties with higher accuracy and at an earlier stage, data on the power load value after the start of mixing water and concrete materials other than water and the point at which the power load value of the mixer 114 stabilizes (the point at which the change in the power load value becomes small) may be used as the load feature FL. When mixing concrete materials, the determination of when the power load value of the mixer 114 becomes stable varies depending on the water-cement ratio of the ready-mixed concrete, etc., but for example, the power load value of the mixer may be determined to be stable after any of the following conditions (1) to (3) is satisfied. (1) During a period when the power load value is on a downward trend, the rate of change of the power load value at a predetermined interval (e.g., 1 second) longer than the sampling period is within the range of ±1% for a predetermined set time (e.g., 3 seconds) or more. The rate of change of the power load value is calculated, for example, by [P(t)-P(t-1)] / P(t-1), where P() is the power load value at each time, t is the current measurement time, and t-1 is the measurement time immediately before that. (2) If the water-cement ratio of fresh concrete is about 50 to 70%, about 30 seconds after the start of mixing water and concrete ingredients other than water (for example, after the concrete ingredients other than water are poured into the mixer and mixed dry, water is poured into the mixer and mixing begins). (3) If the water-cement ratio is reduced (less than 50%) from the viewpoint of strength development, the above period will be delayed. In the case of high-strength concrete, this period will occur after 1 to 10 minutes have passed since the start of mixing the water and concrete materials other than water.

[0088] <Specific examples of features other than the load feature FL> The features F1 to FN may include, as information other than the load feature FL, one or more features related to the timing when the mixer 114 mixes the concrete materials (hereinafter referred to as "timing feature"). For example, the one or more timing feature includes at least one of the date and time when the mixer 114 mixes the concrete materials, the cumulative number of batches mixed by the mixer 114 in one day, the time from an arbitrarily set reference time point to the time when the mixer 114 mixed, and the cumulative number of batches mixed by the mixer 114 from the arbitrarily set reference time point.

[0089] The date when the mixer 114 performs mixing represents the specific day when the mixer 114 produced ready-mixed concrete, as a combination of month and day. In this case, the specific month may be one feature, and the specific day may be another feature. The time when the mixer 114 performs mixing represents the specific point in time when the mixer 114 produced ready-mixed concrete, as a combination of hour, minute, and second, or a combination of hour and minute. In this case, the specific hour may be one feature, and the specific minute (or each of the specific minute and second) may be another feature.

[0090] The time (the specific time when mixer 114 produces ready-mixed concrete) as a feature quantity may be the time when concrete materials are supplied to mixer 114, or the time when ready-mixed concrete is discharged from mixer 114. The time (the specific time when mixer 114 produces ready-mixed concrete) as a feature quantity may be the time when drive of agitator 114a in mixer 114 is started, or the time when drive of agitator 114a is stopped.

[0091] The month and day when the mixer 114 performs mixing, as well as the hour, minute, and second when the mixer 114 performs mixing, may be specified using trigonometric functions. First, specifying (representing) the "hour" of time using trigonometric functions will be described with reference to FIG. 9. Normally, the "hour" of a day is specified (represented) using 24 numerical values ​​from 0:00 to 23:00, assuming that it returns to 0:00 at midnight. When specifying using such normal numerical values, for example, even though there is only a one-hour difference between 0:00 and 23:00 without taking the date into consideration, there is a large numerical difference between them. Therefore, the "hour" may be specified (represented) to have periodicity using trigonometric functions.

[0092] In one example, hour h (h is any integer between 0 and 23) is expressed as a combination of cos{2π×(h / 24)} and sin{2π×(h / 24)} as shown in FIG. 9. In this case, midnight is (1,0), 6 o'clock is (0,1), 12 o'clock is (-1,0), and 18 o'clock is (0,-1). In the input of the prediction model M, cos{2π×(h / 24)} may be one feature and sin{2π×(h / 24)} may be another feature.

[0093] The m minutes (m is any integer between 0 and 59) and s seconds (s is any integer between 0 and 59) may also be represented by the following combinations. In addition, in the input of the prediction model M, similar to the "hour," each of the two numerical values ​​may be one feature. ·m minutes:cos{2π×(m / 60)},sin{2π×(m / 60)} ·s seconds: cos{2π×(s / 60)}, sin{2π×(s / 60)}

[0094] If the "month" of a date is M (M is any integer between 1 and 12) and the "day" is D (D is any integer between 1 and Dm), month M and day D may each be represented by the following combinations. Dm is determined for each month and is an integer of 28, 29, 30, or 31. In the input of the prediction model M, as with the "hour," each of the two numerical values ​​may be one feature. ·M month: cos[2π×{(M-1) / 12)], sin[2π×((M-1) / 12)] ·D day: cos[2π×{(D-1) / Dm)], sin[2π×((M-1) / Dm)]

[0095] The cumulative number of batches per day related to mixing by the mixer 114 is information that identifies the number of batch processes that were performed on that day (the number of batch processes that were performed to produce the ready-mixed concrete). The timing at which mixing by the mixer 114 was performed can also be indicated by the number of batch processes per day that the ready-mixed concrete of interest was produced by.

[0096] When the time from an arbitrarily set reference time point to the time point when the mixer 114 executes mixing is used as a feature, the reference time point may be set by an operator such as a worker. The reference time point may be set to a time point after the interior of the mixer 114 is cleaned during maintenance of the manufacturing apparatus 100 and before production by the mixer 114 is resumed. Note that the time point before the restart also includes a time point before production by the mixer 114 is started for the first time in a day in the case where the mixer 114 is cleaned after a daily shutdown. The time between the reference time point and the time point when the mixer 114 executes mixing (production of ready-mixed concrete) may be expressed in minutes or seconds.

[0097] The time when the mixer 114 executes mixing (production of ready-mixed concrete) may be the time when concrete materials are supplied to the mixer 114, or the time when the produced ready-mixed concrete is discharged from the mixer 114. The time when the mixer 114 executes mixing (production of ready-mixed concrete) may be the time when the mixer 114 starts driving the agitating member 114a, or the time when the mixer 114 stops driving the agitating member 114a. The quality prediction device 10 may measure the time from the reference time point to the time when the mixer 114 executes mixing.

[0098] The cumulative batch number from an arbitrarily set reference point in time is information that specifies which batch process is to be executed since the reference point in time (which batch process from the reference point in time the ready-mixed concrete was produced in). In the cumulative batch number from the reference point in time, the cumulative batch number is reset to 0 at the reference point in time. The reference point in time for counting the cumulative batch number may also be set to a point in time after the interior of the mixer 114 is cleaned during maintenance of the manufacturing apparatus 100 and before production by the mixer 114 is resumed (including a point in time before production by the mixer 114 is started for the first time in a day).

[0099] The feature quantities F1 to FN may include information related to mixing by the mixer 114 (information other than the power load value). Examples of information related to mixing include, in addition to the elapsed time tp, the amount of concrete mixed for one batch, the mixing time, the time when the power load value reaches its maximum value, the time from when the power load value reaches its maximum value until the ready-mixed concrete is discharged from the mixer 114, and the time from when the power load value reaches its maximum value until a measurement time mp (final value P4) described below. The feature quantities F1 to FN may also include nominal strength, information indicating the type of cement, information indicating the type of admixture, information indicating the type of fine aggregate, information indicating the type of coarse aggregate, and information indicating the amount of additive added. The information indicating the type of cement, etc. may include how the types of cement, etc. are classified and each type identified. The feature quantities F1 to FN may also include numerical data obtained from images of ready-mixed concrete being produced by the mixer 114 or immediately after production, or numerical data obtained from images of ready-mixed concrete after it has been discharged from the mixer 114. The feature amounts F1 to FN may include at least one of a target quality of the ready-mixed concrete when it is used and a target quality of the ready-mixed concrete when it is shipped.

[0100] Further examples of features other than the load feature FL included in the features F1 to FN are given below. Note that some of the features overlap with the examples described above. In addition to the load feature FL, the features F1 to FN may include one or more types of data selected from data on the mix conditions of ready-mixed concrete, data on the quality of the desired ready-mixed concrete, data on cement, which is a concrete material, data on concrete materials other than cement, data on concrete material mixing equipment, and data on the environment during the production (including transportation) of ready-mixed concrete.

[0101] (Data on mix conditions for ready-mix concrete) The feature quantities F1 to FN include data on the mix conditions of ready-mixed concrete, such as (i) the type and origin (or product name) of materials used, (ii) cement, water, fine aggregate, coarse aggregate, various admixtures (ground granulated blast furnace slag, fly ash, silica fume, expansive agent, fine volcanic glass powder, crushed stone powder, metakaolin, etc.), various admixtures (AE agent, water reducer, AE water reducer, high-performance water reducer, high-performance AE water reducer, superplasticizer, setting retarder, hardening accelerator, shrinkage reducer, etc.), and various fibers (steel fiber, glass fiber, carbon fiber, The data may include at least one of the amount (mass, weight, volume) and density of fibers (aramid fiber, nylon fiber, vinylon fiber, polyethylene fiber, polypropylene fiber, etc.) used, (iii) water-cement ratio, water-binder ratio, air content, fine aggregate rate, coarse aggregate bulk volume, maximum size of coarse aggregate, coarse particle rate, particle shape determination actual volume rate, total amount of alkali in concrete, sludge solids rate, and recovered aggregate replacement rate, and (iv) amount of stabilizer used in the adhering mortar and sludge water.

[0102] (Data on the quality of the desired ready-mix concrete) The features F1 to FN are used to store data on the quality of the target ready-mixed concrete, including (i) the target slump, slump flow, air content, time to reach 500 mm flow, time to stop flowing, appearance (still or moving image), temperature, unit water content, plastic viscosity, yield value, funnel flow time, compactibility, deformability, filling property, and void passability at the time of shipping, transportation, unloading, or pouring of the ready-mixed concrete, (ii) the concrete strength (design standard strength, durability design standard strength), The concrete strength, slump, slump flow, air content, and chloride content of concrete may be measured by the test methods described in JIS A 5308:2024 (Ready-Mixed Concrete).

[0103] (Cement data) The features F1 to FN may include one or more types of data related to cement, which is a concrete material, selected from (a) data related to cement as a whole, (b) data related to the raw materials of cement clinker, (c) data related to the burning conditions of cement clinker, (d) data related to the grinding conditions of cement, and (e) data related to cement clinker.

[0104] The features F1 to FN may include one or more types of data selected from (a) data relating to the entire cement, such as (a1) the type, chemical composition, mineral composition, wet f.CaO, loss on ignition, Blaine specific surface area, particle size distribution, sieve test residue, and color of the cement used as concrete material, (a2) the mineralogical properties and crystallographic properties of each mineral contained in the cement, and (a3) ​​the hemihydration rate of gypsum contained in the cement.

[0105] The feature quantities F1 to FN include (b) data on the raw materials of cement clinker, such as (b1) the chemical composition, hydraulic hardness, sieve test residue, Blaine specific surface area (fineness), ignition loss, supply amount, supply amount of auxiliary materials (special raw materials such as waste), amount stored in the blending silo (remaining amount), and amount stored in the storage silo (remaining amount), of the raw materials mixed for cement clinker, (b2) the current value of the cyclone located between the raw material mill and the blending silo for the mixed raw materials (representing the rotation speed of the cyclone, which is correlated with the speed of the raw materials passing through the cyclone), and (b3) the time of feeding into the kiln. The data may include one or more types of data selected from the chemical composition and hydraulic hardness of the cement clinker raw materials (mixed raw materials for cement clinker from which fine particles and the like have been removed by a countercurrent airflow during transportation; hereinafter referred to as the raw materials for cement clinker) at a point in time a predetermined time before (for example, one point in time 5 hours before, or multiple points in time such as four points in time 3 hours, 4 hours, 5 hours, and 6 hours before), and (b4) the chemical composition, hydraulic hardness, Blaine specific surface area, sieve test residue, decarbonation rate, and moisture content of the raw materials obtained by mixing the raw materials for cement clinker and auxiliary materials.

[0106] The feature quantities F1 to FN are: (c) data on the cement clinker burning conditions, including (c1) the amount of cement clinker raw materials inserted into the kiln, the kiln rotation speed, the outlet temperature, the burning zone temperature, the cement clinker temperature, the kiln average torque, the O2 concentration, and the NO X The data may include one or more types of data selected from (c1) the concentration, (c2) the temperature of the clinker cooler, and (c3) the flow rate of the gas in the preheater (which is correlated with the temperature of the preheater).

[0107] The features F1 to FN may include (d) data relating to the cement grinding conditions, such as one or more types of data selected from the grinding temperature, the amount of water sprayed in the finishing mill, the separator air volume, the type of gypsum, the amount of gypsum added, the amount of cement clinker added, the rotation speed of the finishing mill, the temperature of the powder discharged from the finishing mill, the amount of powder discharged from the finishing mill, and the amount of powder not discharged from the finishing mill.

[0108] The feature quantities F1 to FN may include, as data on (e) cement clinker, one or more types of data selected from (e1) the mineral composition, chemical composition, wet f.CaO (free lime), and volume gravity of the cement clinker, (e2) the crystallographic properties (such as lattice constant and crystallite size) of each mineral contained in the cement clinker, and (e3) the ratio of two or more mineral compositions contained in the cement clinker.

[0109] (Data on concrete materials other than cement) The feature quantities F1 to FN include data on concrete materials other than cement, such as: (i) the type of aggregate (at least one of fine aggregate and coarse aggregate), bone dry density, surface dry density, water absorption rate, water content, surface water rate, mass loss in stability test, abrasion loss, maximum size, particle size, coarse particle rate, mass fraction of material remaining between successive sieves, particle shape determination actual volume rate, fine particle content, clay mass content, organic impurities, chloride content, and classification by alkali-silica reactivity; (ii) the type of admixture, density, specific surface area, 45 μm sieve residue, 1.2 mm sieve residue, flow value ratio, activity index, moisture content, expansion (length change rate), and various chemical components (silicon dioxide: SiO2, magnesium oxide, etc.); The data may include one or more types of data selected from (i) the content of inorganic fillers (e.g., magnesium oxide: MgO, aluminum oxide: Al2O3, sulfur trioxide: SO3, free calcium oxide: f.CaO, free silicon: f.Si, etc.), (ii) the type, components, density, appearance (color), chloride ion content, total alkali content, water reduction rate, flow value ratio, bleeding amount ratio, bleeding amount difference, difference in setting time (initial time, final setting time), compressive strength ratio, length change ratio, resistance to freezing and thawing (relative dynamic modulus of elasticity), and change over time, and (iii) the type, components, density, appearance (color), chloride ion content, total alkali content, water reduction rate, flow value ratio, bleeding amount ratio, difference in setting time (initial time, final setting time), compressive strength ratio, length change ratio, resistance to freezing and thawing (relative dynamic modulus of elasticity), and change over time of the admixtures, and (iv) the type, density, nominal diameter, nominal length, shape, fineness, tensile strength, tensile modulus of elasticity, adhered moisture content, melting temperature, and alkali resistance (strength retention rate).

[0110] (Data on concrete material mixing equipment) The features F1 to FN may include one or more types of data selected from the following as data related to concrete material mixing equipment: (i) the type, model, product name, manufacturer name, manufacturing year, rated capacity, total output, production capacity, main body mass, unloaded mass during operation, external dimensions, tilt angle (if drum type), rotation speed (drum, mixing blade, mixing shaft), and mixing performance (deviation rate of air amount in concrete, deviation rate of mortar amount in concrete, deviation rate of coarse aggregate in concrete, deviation rate of consistency (slump), deviation rate of compressive strength) of the mixer 114; and (ii) information related to the order in which materials are added, mixing amount, addition time, mixing time, discharge time, re-addition time, and cycle time.

[0111] (Environmental data) The features F1 to FN may include one or more types of data selected from the following as data related to the environment during concrete production (including transportation): (i) outdoor temperature and humidity, atmospheric pressure, amount of solar radiation, hours of sunshine, rainfall, wind speed, wind direction, weather, climate, and climate; (ii) the temperature of each concrete material, the temperature and humidity of the place where the materials are stored (inside containers such as silos and storage jars), and the temperature of the mixer or inside the mixer; and (iii) vehicle information of the truck agitator, load capacity, transportation time, transportation distance, temperature and humidity of the drum or inside the drum, information on road traffic conditions, and information on vibrations applied to the vehicle.

[0112] <Control device hardware configuration> 10, the quality prediction device 10 includes a circuit 50. The circuit 50 includes a processor 51, a memory 52, a storage 53, an input / output port 54, and a timer 55. The storage 53 is configured with one or more nonvolatile memory devices such as a flash memory or a hard disk. The storage 53 stores at least a quality prediction program that causes a computer to execute a construction step, an acquisition step, and a prediction step, which will be described later. The storage 53 stores a quality prediction program for configuring each functional block of the quality prediction device 10.

[0113] The memory 52 is composed of one or more volatile memory devices such as a random access memory. The memory 52 temporarily stores a quality prediction program loaded from the storage 53. The processor 51 is composed of one or more arithmetic devices such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). The processor 51 configures each functional block of the quality prediction device 10 by executing the quality prediction program loaded into the memory 52. ​​The calculation results by the processor 51 are temporarily stored in the memory 52. ​​The input / output port 54 inputs and outputs information to and from the input device 12, the monitor 14, the mixer 114, etc. in response to a request from the processor 51.

[0114] The timer 55 measures the elapsed time by, for example, counting reference pulses at a fixed interval. The circuit 50 is not necessarily limited to one in which each function is configured by a program. For example, the circuit 50 may have at least some of its functions configured by a dedicated logic circuit or an ASIC (Application Specific Integrated Circuit) that integrates such dedicated logic circuits. The quality prediction program may be provided by being permanently recorded on a tangible recording medium such as a CD-ROM, a DVD-ROM, or a semiconductor memory. Alternatively, the quality prediction program may be provided via a communication network as a data signal superimposed on a carrier wave.

[0115] [Ready-mix concrete manufacturing method] Next, an example of a method for producing ready-mixed concrete executed in the production system 1 will be described. The method for producing ready-mixed concrete includes a production process and a quality prediction process. The production process is a process for producing ready-mixed concrete. The quality prediction process is a process for predicting the fresh properties of the ready-mixed concrete after or during the production process. The quality prediction process may be executed during a period that overlaps with at least a portion of the period during which the production process is repeatedly executed.

[0116] The manufacturing process includes, for example, a transporting process, a weighing process, a feeding process, a mixing process, a discharging process, and a loading process. In the transporting process, various types of concrete materials are transported to storage bottles 111 by a transporting device 104, and the various materials are individually supplied to the storage bottles 111. In the weighing process, the various materials are individually supplied from the storage bottles 111 to measuring bottles 112, and the various materials are weighed in the measuring bottles 112. In the weighing process, when the measured amount of each material reaches a predetermined set amount, the material is discharged into a collecting hopper 113. In the feeding process, after all types of materials have been collected in the collecting hopper 113, the materials in the collecting hopper 113 are fed (supplied) into a mixer 114.

[0117] In the mixing process, multiple types of concrete materials are mixed in the mixer 114. In the mixing process, the quality prediction device 10, which has a function of controlling the manufacturing apparatus 100, may control the mixer driving unit 114b in accordance with predetermined operating conditions. In the mixing process, the power supplied from the quality prediction device 10 to the mixer driving unit 114b may be adjusted so that the rotation speed of the mixer driving unit 114b follows a target rotation speed.

[0118] In the discharging process, after mixing of the concrete materials in the mixer 114 is completed, the ready-mixed concrete is discharged from the mixer 114 into the loading hopper 115. In the loading process, the ready-mixed concrete discharged into the loading hopper 115 is loaded onto the transport vehicle 200.

[0119] The quality prediction process (method for predicting the quality of ready-mixed concrete) includes a model construction process in a learning phase and a quality evaluation process in an evaluation phase. In the quality prediction process, the model construction process is executed before the quality evaluation process.

[0120] The model construction process includes a construction step. The construction step is a step of constructing a prediction model M by machine learning based on training data TD including multiple data sets. In the construction step, the prediction model M is constructed so that one or more feature quantities selected from multiple feature quantities F1 to FN correspond to predicted values ​​of fresh properties using a decision tree. The construction step may be performed by the model construction unit 26 of the quality prediction device 10. In each of the multiple data sets in the training data TD, the training feature quantities F1 to FN correspond to the correct value of the fresh property.

[0121] The quality evaluation process includes an acquisition process and a prediction process. The acquisition process is a process of acquiring evaluation feature quantities F1 to FN obtained when ready-mixed concrete is manufactured using the mixer 114. The evaluation feature quantities F1 to FN are the same type of data as the training feature quantities F1 to FN. The feature quantities F1 to FN include at least one or more load feature quantities FL related to the power load value of the mixer 114. The acquisition process may be performed by the feature quantity acquisition unit 24 of the quality prediction device 10.

[0122] The prediction step is a step of acquiring a predicted value of the fresh property according to the evaluation features F1 to FN acquired in the acquisition step, using the prediction model M constructed in the construction step. The acquisition step may be performed by the prediction unit 30 of the quality prediction device 10. The quality evaluation step may include an output step. The output step is a step of outputting the predicted value of the fresh property acquired in the prediction step to the monitor 14. The output step may be performed by the output unit 32 of the quality prediction device 10.

[0123] (Model building process) 11(a) is a flowchart showing an example of a series of processes executed in the model construction process. This model construction process is executed before the production phase including the above-mentioned manufacturing process is executed in the manufacturing apparatus 100, or at the beginning of the start of the above-mentioned production phase. In this model construction process, for example, ready-mixed concrete actually manufactured in the manufacturing apparatus 100 is used as ready-mixed concrete for training.

[0124] In the model building process, step S11 is executed first. In step S11, training data TD for machine learning is prepared by an operator such as a worker. As described above, the training data TD is made up of multiple data sets. Each of the multiple data sets in the training data TD includes feature quantities F1 to FN (training input information) obtained when ready-mixed concrete for training is produced, and correct values ​​of fresh properties (for example, slump) associated with the feature quantities F1 to FN.

[0125] The correct values ​​of the fresh properties may be values ​​obtained by actually measuring the fresh properties of the training ready-mixed concrete. In one example, after the training ready-mixed concrete is loaded onto the transport vehicle 200, a portion of the ready-mixed concrete is extracted by a worker or the like. The slump, slump flow, air content, etc. of the extracted ready-mixed concrete are then measured by the worker or the like, and at least a portion of these measured values ​​are used as the correct values ​​in the training data TD.

[0126] Next, step S12 is executed. In step S12, for example, the model construction unit 26 of the quality prediction device 10 constructs a prediction model M by performing machine learning using the training data TD prepared in step S11. The model construction unit 26 may construct the prediction model M by machine learning using a decision tree algorithm. The model construction unit 26 may construct the prediction model M by performing machine learning involving ensemble learning in machine learning using a decision tree algorithm. The model construction unit 26 may construct the prediction model M by performing machine learning using boosting (e.g., gradient boosting) in machine learning involving ensemble learning. In one example, the model construction unit 26 constructs the prediction model M using LightGBM developed by Microsoft (registered trademark).

[0127] The model construction unit 26 may construct one or more models from among a prediction model M that outputs a predicted value of slump, a prediction model M that outputs a predicted value of slump flow, and a prediction model M that outputs a predicted value of air volume. The model construction unit 26 may construct a prediction model M that outputs the ratio of slump flow to slump as a fresh property.

[0128] Next, step S13 is executed. In step S13, for example, the model holding unit 28 stores the prediction model M constructed in step S12. This completes the model construction process.

[0129] (Quality evaluation process) 11(b) is a flowchart showing an example of a series of processes executed in the quality evaluation process. This quality evaluation process is carried out, for example, during a period overlapping with at least a part of the period during which the above-mentioned production process for the ready-mixed concrete to be evaluated is executed.

[0130] In the quality evaluation process, first, the quality prediction device 10 executes step S21. In step S21, for example, the quality prediction device 10 waits until the evaluation timing, which is the timing to evaluate the quality of the ready-mixed concrete to be evaluated, arrives. The evaluation timing may be predetermined to a certain time period in a day, or may be predetermined to the timing of executing a certain number of batch processes in a day. The evaluation timing may also be the timing at which an instruction to execute the evaluation is received from an operator such as a worker.

[0131] Next, the quality prediction device 10 executes step S22. In step S22, for example, the feature acquisition unit 24 acquires feature quantities F1 to FN when the ready-mixed concrete to be evaluated is produced. The feature quantities F1 to FN acquired in step S22 are input data for evaluation in which the fresh properties (for example, slump) are unknown.

[0132] Next, the quality prediction device 10 executes step S23. In step S23, for example, the prediction unit 30 predicts the fresh properties of the ready-mixed concrete to be evaluated based on the evaluation feature quantities F1 to FN acquired in step S22 and the prediction model M held in the model holding unit 28. In one example, the prediction unit 30 inputs the evaluation feature quantities F1 to FN acquired in step S22 into the prediction model M, and acquires predicted values ​​of the fresh properties output from the prediction model M.

[0133] Next, the quality prediction device 10 executes step S24. In step S24, for example, the output unit 32 displays the predicted values ​​of the fresh properties acquired in step S23 on the monitor 14. This allows an operator such as a worker to check the prediction results regarding the fresh properties of the ready-mixed concrete to be evaluated.

[0134] This completes the quality evaluation process. The quality prediction device 10 may execute the series of processes from steps S21 to S24 each time one batch of ready-mixed concrete is produced (each time a batch process is performed). The quality prediction device 10 may execute the series of processes from steps S21 to S24 each time multiple batches of ready-mixed concrete are produced (each time multiple batch processes are performed).

[0135] [Variations] The series of processes shown in Figures 11(a) and 11(b) are examples and can be modified as appropriate. In the series of processes, one step and the next step may be executed in parallel, or some steps may be executed in an order different from that of the example described above. Steps different from those of the example described above may be executed instead of or in addition to at least some of the steps of the series of processes described above.

[0136] In addition to predicting the quality of ready-mixed concrete by the quality prediction device 10, the manufacturing apparatus 100 may measure the fresh properties of ready-mixed concrete periodically (for example, once to 100 times a day). In this case, the prediction model M may be updated based on the actual measured values ​​of the fresh properties of the ready-mixed concrete and the feature quantities F1 to FN when the actual measured values ​​were obtained. In the prediction step, the fresh properties of the ready-mixed concrete may be predicted using the updated prediction model M. Note that even when the updated prediction model M is used, the step of predicting the fresh properties of the ready-mixed concrete is still performed based on the prediction model M and the feature quantities F1 to FN acquired in the acquisition step.

[0137] In the above example, the quality prediction device 10 predicts the quality of ready-mixed concrete after it has been obtained by mixing using the mixer 114 of the manufacturing apparatus 100 and before it is shipped to the construction site. The timing at which the quality prediction device 10 predicts the quality is not limited to this example. The quality prediction device 10 may also predict the quality of ready-mixed concrete while it is being mixed using the mixer 114 (ready-mixed concrete that has already been obtained before mixing is complete). The quality prediction device 10 may also predict the quality of ready-mixed concrete while it is being transported to a construction site where it will be used, or the quality of ready-mixed concrete upon arrival at the construction site (at the time of unloading). For example, the transport vehicle 200 is provided with a mixer 214 that mixes the ready-mixed concrete (see FIG. 1). The mixer 214 in the transport vehicle 200, such as an agitator vehicle, is also called a drum.

[0138] If the production system 1 includes a transport vehicle 200, at least some of the features F1 to FN may include features obtained when producing (mixing) ready-mixed concrete using the mixer 214. The one or more load features FL may be one or more feature values ​​related to the power load value of the mixer 214 when mixing the ready-mixed concrete. The feature acquisition unit 24 of the quality prediction device 10 may calculate the one or more load features FL from time-series data up to a measurement time point mp during the mixing operation by the mixer 214. The prediction unit 30 may predict the quality of the ready-mixed concrete after it is discharged from the mixer 214 while the mixing operation by the mixer 214 is continuing.

[0139] The feature acquisition unit 24 of the quality prediction device 10 may acquire, as at least a part of the feature quantities F1 to FN (one or more load feature quantities FL), feature quantities related to the power load value of the mixer 214 when mixing ready-mixed concrete. The feature quantities F1 to FN may include, as data related to the transportation of ready-mixed concrete, one or more types of data selected from (i) the volume, thickness, mass, and temperature of the ready-mixed concrete in the mixer 214 of the transport vehicle 200, (ii) the outside air temperature during transportation, and (iii) transportation time (the time from the end of mixing to the end of transportation (unloading)) and transportation distance. The feature quantities F1 to FN may also include batcher data (type of heating / cooling device, stored water temperature, stored aggregate temperature, stored cement temperature, measuring bottle type, discharge volume, maximum discharge pressure, etc.). The quality prediction device 10 does not necessarily have a function to control the manufacturing apparatus 100.

[0140] In the above example, ready-mixed concrete is produced at a location (manufacturing apparatus 100) separate from the construction site. The location where ready-mixed concrete is produced is not limited to this example. Concrete materials (e.g., concrete materials excluding water) may be transported to the construction site by a transport vehicle, and ready-mixed concrete may be produced at the construction site. At this time, a mixer provided on the transport vehicle may add water to the concrete materials transported by the transport vehicle, and the materials may be mixed to obtain ready-mixed concrete. When ready-mixed concrete is produced at the construction site, the quality prediction device 10 may predict the quality of the ready-mixed concrete, such as its fresh properties, at the construction site. The feature acquisition unit 24 of the quality prediction device 10 may acquire, as at least a part of one or more load feature values ​​FL, a feature value related to the power load value of the mixer provided on the transport vehicle when ready-mixed concrete is produced at the construction site.

[0141] In one example of the various examples described above, at least some of the features described in other examples may be combined.

[0142] [Verification of prediction results using the prediction model] Next, we will explain the results of verifying the prediction results of freshness properties by a prediction model M constructed with a decision tree algorithm using a dataset for which the correct values ​​are known and inputting the features F1 to FN. In this verification, we used the same training data TD and compared the prediction results with those of a comparison model constructed with a deep neural network (DNN) different from the decision tree algorithm.

[0143] A training dataset of 120,552 items was prepared as training data TD, and a test dataset of 13,383 items, for which the correct values ​​were known, was prepared for evaluating the prediction results. In each of the training dataset and the test dataset, the features F1 to FN were linked to the correct values ​​of the fresh properties. Data obtained from a ready-mixed concrete manufacturing system configured in the same way as manufacturing system 1 was used as the correct values ​​(measured values) of the features F1 to FN and the fresh properties. In addition, the fresh properties of ready-mixed concrete before shipping were used as the target for prediction by the prediction model.

[0144] The data items shown in Table 1 below were used as feature quantities F1 to FN. In Table 1, "manufacturing month (sin, cos)" means data expressing the manufacturing month using the trigonometric functions sin and cos, and has the same meaning for manufacturing date, etc. "cement type (a1, a2, a3, a4, a5)" means data indicating which cement of the cement types a1 to a5 is included, and has the same meaning for admixture type, etc. [Table 1]

[0145] (Example 1: LightGBM) A prediction model M was constructed using LightGBM, a type of decision tree algorithm. Specifically, the prediction model M was constructed using "LightGBM (version 4.3.0)" provided by Microsoft. A model for predicting slump, a model for predicting slump flow, and a model for predicting air content were constructed separately. In constructing the prediction model M according to Example 1, the main hyperparameters were set as shown in Table 2 below. [Table 2]

[0146] (Example 2: XGBoost) A prediction model M was constructed using XGBoost, a type of decision tree algorithm. Specifically, the prediction model M was constructed using the well-known library "XGBoost (version 1.5.2)." A model for predicting slump, a model for predicting slump flow, and a model for predicting air content were constructed separately. In constructing the prediction model M according to Example 2, the main hyperparameters were set as shown in Table 3 below. [Table 3]

[0147] (Comparative example: DNN) A comparative prediction model was constructed using a deep neural network. In constructing the comparative prediction model, the main conditions (network configuration) were set as follows: - We did not use a Batch Normalization layer to normalize the input data to the model. Dropout, which removes some connections in the intermediate layers of the neural network, was not used. The intermediate layer of the neural network was a fully connected layer with eight layers. Adam was used as the weight update formula and HuberLoss was used as the loss function. Machine learning was performed so that three types of data - slump, slump flow, and air volume - could be output as freshness data from a single prediction model.

[0148] (Verification results) For each of Example 1, Example 2, and Comparative Example, a test dataset was used to compare the predicted values ​​obtained by the prediction model with the correct values ​​contained in the dataset. Specifically, the correct value was defined as the case where the predicted value obtained by the model was within a range obtained by adding a predetermined tolerance to the correct value in the test dataset, and the accuracy rate, which represents the ratio of the number of correct datasets to the total number of test data, was calculated as an evaluation index.

[0149] FIG. 12(a) shows the evaluation results of the accuracy rate for slump. In FIG. 12(a), when the tolerance is "±0.5 cm," data sets whose predicted values ​​by the model are within a range of ±0.5 cm of the correct value are considered to be correct, and the percentage of data sets that are considered to be correct is shown as the accuracy rate (%). Accuracy rates (%) were also calculated in the same way when the tolerance is "±1.0 cm," "±1.5 cm," "±2.0 cm," and "±2.5 cm." The results shown in FIG. 12(a) reveal that the accuracy rates when predicting slump using the prediction models M of Examples 1 and 2 are significantly higher than those of the comparative example.

[0150] FIG. 12(b) shows the evaluation results of the accuracy rate for slump flow. In FIG. 12(b), when the tolerance is "±2.5 cm," data sets whose predicted values ​​by the model are within a range of ±2.5 cm of the correct value are considered to be correct, and the percentage of data sets that are considered to be correct is shown as the accuracy rate (%). Similarly, accuracy rates (%) were calculated when the tolerance is "±3.0 cm," "±4.0 cm," "±5.0 cm," "±6.0 cm," and "±7.5 cm." The results shown in FIG. 12(b) reveal that the accuracy rates when predicting slump flow using the prediction model M according to Examples 1 and 2 are significantly higher than those of the comparative example.

[0151] FIG. 13(a) shows the evaluation results of the accuracy rate for the air volume. In FIG. 13(a), when the tolerance is "±0.3%," the data set whose predicted value by the model is within a range of ±0.3% of the correct value is considered to be the correct answer, and the percentage of the data set that is considered to be the correct answer is shown as the accuracy rate (%). The accuracy rates (%) are also calculated in the same way when the tolerance is "±0.5%, "±0.7%, "±1.0%, "±1.2%," and "±1.5%". From the results shown in FIG. 13(a), it can be seen that the accuracy rates when predicting the air volume using the prediction model M according to Examples 1 and 2 are significantly higher than in the comparative example.

[0152] As an evaluation index separate from the accuracy rate, we calculated the "sum of the mean and standard deviation" regarding the error between the predicted value and the correct value. Using the sum of the mean and standard deviation as an evaluation index means that the error is evaluated taking into account variability. In other words, even if the error itself (the mean of the errors) is small, if the standard deviation is large, it means that the variability of the errors is large, which is not desirable. Also, even if the standard deviation of the errors is small, if the mean value is large, it means that the error itself is large, which is not desirable. In calculating the sum of the mean and standard deviation, we calculated the error between the model's predicted value and the correct value for each test dataset, and then found the mean and standard deviation from these errors.

[0153] Fig. 13(b) shows the results of calculating the average error and the sum of the standard deviation for each of the fresh properties of slump, slump flow, and air content. The results shown in Fig. 13(b) show that the average error and the sum of the standard deviation are significantly smaller when predictions are made using the prediction model M of Examples 1 and 2 than in the comparative example, for all of the fresh properties of slump, slump flow, and air content.

[0154] From a different perspective than prediction accuracy, we evaluated the "learning time," which is the time from when the computer starts training to when it completes building a predictive model, after the preparation of the training data TD and the setting of various hyperparameters are complete. Learning time can be described as the time required for the computer to build a predictive model. The evaluation results for learning time are shown in Table 4 below. [Table 4]

[0155] In Table 4, Examples 1 and 2 represent the learning time required to construct a prediction model M that predicts only slump, while Comparative Example represents the learning time required to construct a prediction model that simultaneously outputs slump, slump flow, and air volume. The results shown in Table 4 demonstrate that the learning time can be significantly reduced in Examples 1 and 2 compared to the Comparative Example. It is assumed that the learning time in the Comparative Example is not significantly different from the learning time required to construct a comparative model that predicts only slump using the same method as the Comparative Example. However, even if the learning time in Examples 1 and 2 is tripled, it is still significantly reduced compared to the Comparative Example. Furthermore, it can be seen that the learning time can be further reduced in Example 1 compared to Example 2. For example, by shortening the learning time, it is possible to construct or update the prediction model M while still producing ready-mixed concrete (the production of ready-mixed concrete can continue without being restricted by the construction or update of the prediction model M).

[0156] (Consideration of factors that improved prediction accuracy compared to the comparative example) [Prediction model based on decision trees] The prediction models M constructed in Examples 1 and 2 above are both constructed using a decision tree algorithm. In a decision tree, branching is performed by focusing on the part of each node where the feature values ​​differ most. Therefore, as long as there is a relationship between the feature values ​​and the data to be predicted, a certain degree of accuracy can be ensured even with a limited amount of training data. Furthermore, because the model structure is simple, overlearning is unlikely to occur even with a small amount of data. On the other hand, the DNN in the comparative example performs complex learning using all feature values ​​in the training data, and excels at capturing detailed features and patterns in the data. A DNN with such characteristics requires a huge amount of data to ensure accuracy. Furthermore, because the model structure is complex, excessive fitting to the training data is likely to occur when the amount of data is small. It is practically difficult to prepare training data with a huge amount of data in machine learning. Therefore, it is believed that a decision tree can construct a more accurate prediction model when predictions are made using limited data.

[0157] [Ensemble learning using decision trees] The prediction model M constructed in both Examples 1 and 2 above is constructed using a decision tree algorithm that includes ensemble learning. Ensemble learning using decision trees includes a technique called "bagging," in which multiple decision trees with different characteristics are arranged in parallel and trained independently, and a technique called "boosting," in which multiple decision trees with different characteristics are arranged in series and trained sequentially. These techniques use multiple decision trees to construct a prediction model, which is effective in reducing the variability in prediction results and the error between predicted values ​​and correct values. Therefore, it is inferred that ensemble learning using multiple decision trees improved prediction accuracy compared to learning using a single DNN. However, DNN ensemble learning can have a disadvantage in that it requires a great deal of effort to construct a model, as the model structure becomes more complex and it can be difficult to appropriately combine models and adjust parameters.

[0158] [Gradient Boosting Decision Tree] The prediction model M constructed in the above-mentioned Example 1 is constructed by LightGBM, which is a type of machine learning that includes gradient boosting. (1) LightGBM excels at capturing the synergistic effects when combining numerical data between parameters (features). On the other hand, DNNs are not suited to capturing the mutual relationships between data. Considering that fresh properties, an example of quality, are determined by a combination of concrete mix, date, load value, etc., it is inferred that prediction accuracy improved by building a model using LightGBM, which can capture interactions. (2) LightGBM can select and learn important features of data. Specifically, it creates a predictive model by using all data for features with large prediction errors, assuming they have important features, and randomly extracting data and retraining features with small prediction errors, assuming they have already been trained. On the other hand, DNNs use all data related to each feature to perform complex training and excel at capturing the finer features and patterns of data. DNNs with these characteristics are effective when used with image, video, and audio data, but require massive amounts of data to improve accuracy. When making predictions using limited numerical data, it is believed that highly accurate predictive models can be created by selecting important features of the data and training them.

[0159] Summary of this disclosure The quality prediction method for ready-mixed concrete described above is a quality prediction method for predicting the quality of ready-mixed concrete. This quality prediction method includes a construction step of constructing a prediction model (M) by machine learning based on training data (TD) including multiple data sets; an acquisition step of acquiring multiple feature quantities (F1-FN) obtained when producing ready-mixed concrete using a mixer (114, 214); and a prediction step of using the prediction model (M) constructed in the construction step to acquire a predicted value of quality based on the multiple feature quantities (F1-FN) acquired in the acquisition step. In each of the multiple data sets, the multiple feature quantities (F1-FN) are associated with a correct value of quality. The multiple feature quantities (F1-FN) include one or more load feature quantities (FL) related to a power load value of the mixer (114, 214) when mixing concrete materials or stirring ready-mixed concrete. In the construction step, the prediction model (M) is constructed so that one or more feature quantities selected from the multiple feature quantities (F1-FN) are associated with a predicted value of quality using a decision tree.

[0160] As explained in the above verification results, it was confirmed that the prediction accuracy of the prediction model can be significantly improved by using a prediction model (M) constructed so that one or more feature quantities selected from a plurality of feature quantities (F1 to FN) correspond to the predicted quality value using a decision tree. In other words, the above quality prediction method in which the quality of ready-mixed concrete is predicted using the prediction model (M) is useful for improving prediction accuracy. Furthermore, since a model is constructed using a decision tree, the time required to construct the prediction model can be reduced.

[0161] In the above-described method for predicting the quality of ready-mixed concrete, the construction step may involve constructing a prediction model (M) by machine learning including ensemble learning. In this case, a prediction model (M) is constructed by integrating multiple models to obtain a single final predicted value, which is more useful for improving prediction accuracy.

[0162] In the above-described method for predicting the quality of ready-mixed concrete, the construction step may involve constructing a prediction model (M) by machine learning using boosting. In this case, multiple models generated in sequence with errors taken into account are combined to construct a prediction model (M) that provides a single final predicted value, thereby improving the prediction accuracy of a model using a decision tree.

[0163] In the above-described method for predicting the quality of ready-mixed concrete, the construction step may involve constructing a prediction model (M) by machine learning using gradient boosting. In this case, multiple models generated in sequence to reduce errors are combined to construct a prediction model (M) that provides a single final predicted value, thereby further improving the prediction accuracy of a model using a decision tree.

[0164] In the above-described method for predicting the quality of ready-mixed concrete, the one or more load features (FL) may include at least one of a value at a predetermined time point in time-series data of the power load value of the mixer (114, 214) when mixing concrete materials or mixing ready-mixed concrete, and a statistic obtained from the time-series data. The statistic may be one or more values ​​selected from the group consisting of the difference between the maximum and minimum values ​​(variation range), the difference between the maximum and final values ​​(decline range), the sum, the mean, the standard deviation, the coefficient of variation, the median, the first quartile, the third quartile, kurtosis, and skewness. The power load value when mixing or mixing using the mixer fluctuates over time. By using at least one of the power load value at a specific time point and the statistic obtained from the time-series data as a feature, as in the above method, information that better reflects the characteristics of the power load value fluctuating over time can be added to the input of the prediction model (M).

[0165] In the above-described method for predicting the quality of ready-mixed concrete, the plurality of features (F1 to FN) may further include one or more timing features related to the timing when the mixer (114) mixes the concrete materials. The one or more timing features may include at least one of the following: the date and time when the mixer (114) mixes the concrete materials; the cumulative number of batches mixed by the mixer (114) in a day; the time from an arbitrarily set reference point until the mixer (114) mixed the concrete materials; and the cumulative number of batches mixed by the mixer (114) from the arbitrarily set reference point. The inventors' verification confirmed that even if the production conditions, including the operating conditions of the mixer (114), are the same, the power load value itself can vary depending on the timing of the production of ready-mixed concrete. In the above-described method, by adding the timing features to the prediction model (M), the prediction model (M) can be constructed taking into account the fluctuations in the power load value depending on the timing of the production of ready-mixed concrete. As a result, it is possible to reduce variations in the prediction results of the prediction model (M) that may occur due to differences in the manufacturing time.

[0166] The quality prediction program described above is a program that causes a computer to execute the quality prediction method. Since the quality prediction program executes the quality prediction method, it is useful for improving prediction accuracy, just like the quality prediction method.

[0167] The method for producing ready-mixed concrete described above includes a production process for producing ready-mixed concrete, and a quality prediction process for predicting quality using the quality prediction method after or during the production process. In this method for producing ready-mixed concrete, the quality is predicted using the quality prediction method, and therefore, like the quality prediction method, it is useful for improving prediction accuracy.

[0168] The above-described ready-mixed concrete quality prediction device (10) is a device for predicting the quality of ready-mixed concrete. The quality prediction device (10) includes a model construction unit (26) that constructs a prediction model (M) by machine learning based on training data (TD) including multiple data sets; an acquisition unit (24) that acquires multiple feature quantities (F1-FN) obtained when producing ready-mixed concrete using a mixer (114, 214); and a prediction unit (30) that uses the prediction model (M) constructed by the model construction unit (26) to acquire a predicted value of quality based on the multiple feature quantities (F1-FN) acquired by the acquisition unit (24). In each of the multiple data sets, the multiple feature quantities (F1-FN) are associated with a correct value of quality. The multiple feature quantities (F1-FN) include one or more load feature quantities (FL) related to a power load value of the mixer (114, 214) when mixing concrete materials or stirring ready-mixed concrete. The model construction unit (26) constructs a prediction model (M) so that one or more feature quantities selected from the plurality of feature quantities (F1 to FN) correspond to the predicted quality values ​​using a decision tree. This quality prediction device is useful for improving prediction accuracy, similar to the above-mentioned quality prediction method.

[0169] The ready-mixed concrete manufacturing system (1) described above includes a manufacturing apparatus (100) that manufactures ready-mixed concrete and the quality prediction device (10). The quality prediction device (10) predicts the quality of ready-mixed concrete manufactured by the manufacturing apparatus (100). Since this manufacturing system includes the quality prediction device, it is useful for improving prediction accuracy, similar to the quality prediction device. [Explanation of symbols]

[0170] 1...manufacturing system, 10...quality prediction device, 24...feature acquisition unit, F1 to FN...feature, FL...load feature, 26...model construction unit, TD...training data, 30...prediction unit, M...prediction model, 100...manufacturing equipment, 114...mixer, 114a...agitating member, 114b...mixer drive unit, 214...mixer.

Claims

1. A quality prediction method for predicting fresh properties of ready-mixed concrete, comprising: A construction process of constructing a predictive model by machine learning including ensemble learning based on training data including multiple datasets; an acquisition step of acquiring a plurality of feature amounts obtained when producing ready-mixed concrete using a mixer; a prediction step of acquiring a predicted value of the fresh property according to the plurality of feature quantities acquired in the acquisition step, using the prediction model constructed in the construction step; In each of the plurality of data sets, the plurality of feature amounts are associated with a correct answer value of the fresh property, the plurality of feature quantities include one or more load feature quantities related to a power load value of the mixer when mixing concrete materials or mixing ready-mixed concrete, In the construction step, the prediction model is constructed so that one or more feature quantities selected from the plurality of feature quantities correspond to the predicted value of the fresh property using a decision tree; In the construction step, the prediction model is constructed by machine learning using boosting as machine learning including ensemble learning. Methods for predicting the quality of ready-mix concrete.

2. In the construction step, the prediction model is constructed by machine learning using gradient boosting. The method for predicting the quality of ready-mixed concrete according to claim 1.

3. the one or more load feature quantities include statistics obtained from time series data of power load values ​​of the mixer when mixing concrete materials or mixing fresh concrete, The statistical quantity is one or more values ​​selected from the group consisting of a difference between a maximum value and a minimum value, a difference between a maximum value and a final value, a sum, a mean value, a standard deviation, a coefficient of variation, a median, a first quartile, a third quartile, kurtosis, and skewness. The method for predicting the quality of ready-mixed concrete according to claim 1.

4. The plurality of feature quantities further include one or more time feature quantities related to a time when the mixer mixes the concrete material, The one or more time features are the date and time when the mixer performs mixing of the concrete materials; The cumulative number of batches mixed by the mixer in one day; The time from an arbitrarily set reference time to the time when the mixer performs mixing, and The cumulative number of batches mixed by the mixer from an arbitrarily set reference point in time, at least one of: The method for predicting the quality of ready-mixed concrete according to any one of claims 1 to 3.

5. A quality prediction program that causes a computer to execute the quality prediction method according to any one of claims 1 to 3.

6. A manufacturing process for producing ready-mix concrete; and a quality prediction step of predicting the fresh properties by the quality prediction method according to any one of claims 1 to 3 after or during the production step. Method for manufacturing ready-mix concrete.

7. A quality prediction device for predicting fresh properties of ready-mixed concrete, a model construction unit that constructs a predictive model by machine learning including ensemble learning based on training data including multiple data sets; an acquisition unit that acquires a plurality of feature amounts obtained when producing ready-mixed concrete using a mixer; a prediction unit that acquires a predicted value of the fresh property according to the plurality of feature quantities acquired by the acquisition unit, using the prediction model constructed by the model construction unit; In each of the plurality of data sets, the plurality of feature amounts are associated with a correct answer value of the fresh property, the plurality of feature quantities include one or more load feature quantities related to a power load value of the mixer when mixing concrete materials or mixing ready-mixed concrete, the model construction unit constructs the prediction model such that one or more feature quantities selected from the plurality of feature quantities correspond to the predicted value of the fresh property using a decision tree; the model construction unit constructs the prediction model by performing machine learning using boosting as machine learning including ensemble learning. Ready-mix concrete quality prediction device.

8. A manufacturing device for manufacturing ready-mix concrete; and the quality prediction device according to claim 7, The quality prediction device predicts the fresh properties of the ready-mixed concrete produced by the manufacturing device. Ready-mix concrete manufacturing system.

Images