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

Abstract

To improve the prediction accuracy when using predictive models to forecast the quality of ready-mixed concrete. [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 datasets; an acquisition step of acquiring multiple feature quantities obtained when manufacturing ready-mixed concrete using a mixer; and a prediction step of using the predictive model to acquire a predicted quality value corresponding to the multiple feature quantities acquired in the acquisition step, wherein each of the multiple datasets is associated with multiple feature quantities and the correct quality value, 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 in the construction step, a predictive model is constructed such that one or more feature quantities selected from the multiple feature quantities are associated with the predicted quality value by a decision tree.

Description

Technical Field

[0001] The present disclosure relates to a method for predicting the quality of fresh concrete, a quality prediction program, a method for manufacturing fresh concrete, an apparatus for predicting the quality of fresh concrete, and a manufacturing system for fresh concrete.

Background Art

[0002] Patent Documents 1 and 2 disclose methods for predicting the quality of fresh concrete using a prediction model.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] The present disclosure provides a method for predicting the quality of fresh concrete, a quality prediction program, a method for manufacturing fresh concrete, an apparatus for predicting the quality of fresh concrete, and a manufacturing system for fresh concrete, which are useful for improving the prediction accuracy when predicting the quality of fresh concrete using a prediction model.

Means for Solving the Problems

[0005] [1] A 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 multiple datasets; an acquisition step of acquiring multiple feature quantities obtained when manufacturing 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 multiple feature quantities acquired in the acquisition step, wherein each of the multiple datasets is associated with the multiple feature quantities and the correct value of the quality, 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 in the construction step, the prediction model is constructed such that one or more feature quantities selected from the multiple feature quantities are associated with the predicted value of the quality by 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 predictive model is constructed by machine learning including ensemble learning.

[0007] [3] The method for predicting the quality of ready-mixed concrete as described in [2] above, wherein in the construction step, the predictive 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 predictive 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 the values ​​at a predetermined time in the time-series data of the power load of the mixer when mixing concrete materials or stirring ready-mixed concrete, and a statistic obtained from the time-series data, wherein the statistic 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 features further include one or more time features relating to the timing when the mixer performs mixing of the concrete material, the one or more time features including at least one of the date and time when the mixer performs mixing of the concrete material, the cumulative number of batches of mixing 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 of mixing by the mixer from an arbitrarily set reference time.

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

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

[0013] [9] A quality prediction device for predicting the quality of ready-mixed concrete, comprising: a model building unit that constructs a prediction model by machine learning based on training data including multiple datasets; an acquisition unit that acquires multiple feature quantities obtained when ready-mixed concrete is manufactured using a mixer; and a prediction unit that uses the prediction model constructed by the model building unit to acquire a predicted value of the quality corresponding to the multiple feature quantities acquired by the acquisition unit, wherein in each of the multiple datasets, the multiple feature quantities and the correct value of the quality are associated, 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 the model building unit constructs the prediction model such that one or more feature quantities selected from the multiple feature quantities and the predicted value of the quality are associated by a decision tree.

[0014]

[10] A production system for ready-mixed concrete, comprising a production apparatus for producing ready-mixed concrete and the quality prediction apparatus described in [9] above, wherein the quality prediction apparatus predicts the quality of the ready-mixed concrete produced by the production apparatus.

Advantages of the Invention

[0015] According to the present disclosure, there are provided a quality prediction method for ready-mixed concrete, a quality prediction program, a production method for ready-mixed concrete, a quality prediction apparatus for ready-mixed concrete, and a production system for ready-mixed concrete, which are useful for improving the prediction accuracy when predicting the quality of ready-mixed concrete using a prediction model.

Brief Description of the Drawings

[0016] [Figure 1] FIG. 1 is a schematic diagram showing an example of a production system for ready-mixed concrete. [Figure 2] FIG. 2 is a block diagram showing an example of the functional configuration of a control device. [Figure 3] FIG. 3 is a diagram for explaining an example of a prediction model. [Figure 4] FIGS. 4(a) and 4(b) are diagrams for explaining an example of the construction process of a prediction model. [Figure 5] FIG. 5 is a diagram for explaining an example of the construction process of a prediction model. [Figure 6] FIGS. 6(a) and 6(b) are diagrams for explaining an example of the construction process of a prediction model. [Figure 7] FIGS. 7(a) and 7(b) are diagrams for explaining an example of the construction process of a prediction model. [Figure 8] FIGS. 8(a) and 8(b) are diagrams illustrating data related to power load values. [Figure 9] FIG. 9 is a diagram for explaining an example of the expression format of a date and time. [Figure 10] FIG. 10 is a block diagram showing an example of the hardware configuration of a control device. [Figure 11]FIG. 11(a) is a flowchart showing an example of a series of processes in the training phase. FIG. 11(b) is a flowchart showing an example of a series of processes in the evaluation phase. [Figure 12] FIGS. 12(a) and 12(b) are graphs showing the results of verifying the prediction accuracy by the prediction model. [Figure 13] FIGS. 13(a) and 13(b) are graphs showing the results of verifying the prediction accuracy by the prediction model.

MODE FOR CARRYING OUT THE INVENTION

[0017] Hereinafter, an embodiment will be described with reference to the drawings. In the description, the same reference numerals are given to the same elements or elements having the same function, and redundant descriptions are omitted. FIG. 1 schematically shows a fresh concrete manufacturing system including a quality prediction device according to an embodiment.

[0018] [Fresh Concrete Manufacturing System] First, the outline of the fresh concrete manufacturing system will be described. The manufacturing system 1 (fresh concrete manufacturing system) shown in FIG. 1 is a system for manufacturing fresh concrete. The manufacturing system 1 manufactures fresh concrete through at least a process of kneading concrete materials. The concrete materials used in the manufacturing system 1 include cement, admixtures, coarse aggregates, fine aggregates, water, and admixtures.

[0019] Examples of the coarse aggregate include natural coarse aggregate or artificial coarse aggregate. Also, examples of the coarse aggregate include gravel, crushed stone, slag coarse aggregate, lightweight coarse aggregate, recycled coarse aggregate, recovered coarse aggregate, or a mixture of these. The gravel is mountain gravel, land gravel, river gravel, or sea gravel. The slag coarse aggregate is blast furnace slag aggregate, ferronickel slag aggregate, electric furnace oxidized slag aggregate, or coal gasification slag aggregate. The lightweight coarse aggregate is natural lightweight aggregate, by-product lightweight aggregate, or artificial lightweight aggregate. The coarse aggregate may include crushed rock or lime crushed stone.

[0020] Examples of fine aggregate include natural aggregate and artificial aggregate. Other examples of fine aggregate include sand, crushed sand, slag fine aggregate, lightweight fine aggregate, recycled fine aggregate, recovered fine aggregate, or fine aggregates that are mixtures of these. Sand can be mountain sand, land sand, river sand, or sea sand. Slag fine aggregate can be blast furnace slag aggregate, ferronickel slag aggregate, copper slag aggregate, electric furnace oxidized slag aggregate, or coal gasification slag aggregate. Lightweight fine aggregate can be natural lightweight aggregate, by-product lightweight aggregate, or artificial lightweight aggregate.

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

[0022] At least a portion of the manufacturing system 1 is installed in a location different from the site where the ready-mixed concrete will be used (for example, a factory). The transport vehicle 200 transports the ready-mixed concrete to the site where it will be used (for example, a construction site) after it has been loaded. Examples of transport vehicles 200 include agitator trucks (mixer trucks) or dump trucks. The manufacturing system 1 may produce ready-mixed concrete from concrete materials to meet the target quality (required quality) set for each site. For example, an operator of the manufacturing system 1 determines the concrete mix to meet the target quality and inputs operation instructions to the manufacturing system 1.

[0023] In one example, ready-mixed concrete is manufactured in manufacturing system 1 to ensure that the target quality for use of the ready-mixed concrete, which is set for each site, is met, and the quality of the ready-mixed concrete is controlled (inspected, etc.) before shipment. Use of the ready-mixed concrete refers to the time when the ready-mixed concrete is received at the site. For quality control of the ready-mixed concrete before shipment, a target quality for shipment of the ready-mixed concrete may be determined based on the target quality for use of the ready-mixed concrete. The quality of ready-mixed concrete includes fresh properties. Specific examples of fresh properties include slump, slump flow, and air content.

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

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

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

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

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

[0029] The mixer 114 is located below the collection hopper 113. The mixer 114 is a device for mixing concrete materials. The mixer 114 produces ready-mixed concrete by mixing aggregate, cement, water, and admixtures. The ready-mixed concrete is discharged from the bottom of the mixer 114 into the loading hopper 115. The mixer 114 may be a tilting mixer, a horizontal single-axis mixer, a horizontal double-axis mixer, or a pan-type mixer. The mixer 114 includes, for example, two stirring members 114a and a mixer drive unit 114b.

[0030] The stirring members 114a are components that stir the various materials supplied to the mixer 114. Two stirring members 114a are arranged side by side inside the main body (container) of the mixer 114 and are rotatably mounted. Each of the two stirring members 114a includes a rotation axis extending in one horizontal direction. The mixer drive unit 114b rotates the rotation axes of each of the two stirring members 114a based on operation instructions from a control device provided in the manufacturing system 1. The mixer drive unit 114b includes a drive source, such as a motor, that provides driving force to the stirring members 114a. An opening 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 located below the mixer 114 and temporarily stores the ready-mix concrete. The loading hopper 115 then supplies the temporarily stored ready-mix concrete to the transport vehicle 200.

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

[0033] In this disclosure, a unit of ready-mixed concrete produced in one mixing cycle in the mixer 114 and loaded onto the transport vehicle 200 is defined as "1 batch". The process performed by the manufacturing system 1 to produce 1 batch of ready-mixed concrete is defined as "batch processing". The number of batch processing cycles performed from a certain reference point (cumulative number of executions) is called the "cumulative batch count". At the reference point, the cumulative batch count is reset to zero. For example, each day, when the manufacturing equipment 100 starts operation, the cumulative batch count is reset to zero. In this case, the cumulative batch count for the kth batch processing cycle (where k is an integer greater than or equal to 1) performed on a given day will be k.

[0034] One to three batches of ready-mix concrete may be loaded onto one transport vehicle 200. For example, if two batches of ready-mix concrete are loaded onto one transport vehicle 200, two batch processes, each following the same manufacturing conditions, will be carried out at different times (sequentially). Multiple batch processes, each following the same manufacturing conditions, may be carried out consecutively and sequentially during a certain period of time within a 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 device 100. The quality prediction device 10 is composed of one or more computers. If the quality prediction device 10 is composed of multiple computers, these computers are connected to each other in a way that allows them to communicate with one another. In addition to the function of predicting quality, the quality prediction device 10 may also have a function of controlling the manufacturing device 100. For example, the quality prediction device 10 may be configured as part of a control device that controls the manufacturing device 100. The quality prediction device 10 may control the manufacturing device 100 according to set operating conditions. At least a portion of the operating conditions may be determined by instructions from an operator such as a worker.

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

[0037] The quality of the ready-mixed concrete 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 predicted values ​​for indicator values ​​that represent quality.

[0038] The quality prediction device 10 predicts, for example, the fresh properties of ready-mixed concrete as the quality of the ready-mixed concrete. The quality prediction device 10 may predict the fresh properties of the ready-mixed concrete as the quality of the ready-mixed concrete after it has been mixed in the manufacturing device 100 and before it is shipped from the factory where the manufacturing device 100 is installed to the site. Alternatively, the quality prediction device 10 may predict the fresh properties of the ready-mixed concrete after it has been shipped from the factory and before it is placed at the site. In other words, the quality prediction device 10 may predict the fresh properties of the ready-mixed concrete during transport, or it may predict the fresh properties of the ready-mixed concrete at the time of unloading.

[0039] The quality prediction device 10 may predict one or more indicator values ​​from among slump, slump flow, and air content as fresh concrete properties. The quality prediction device 10 may predict two or more indicator values ​​from among slump, slump flow, and air content as fresh concrete properties. The quality prediction device 10 may predict indicator values ​​other than slump, slump flow, and air content as fresh concrete properties. Examples of fresh concrete property indicators other than slump, slump flow, and air content include 500 mm flow arrival 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, compaction, deformability, filling ability, and void passage ability.

[0040] The quality prediction device 10 may predict at least one of the following as indicators of strength and durability as the quality of the fresh concrete after hardening. Indicators of strength include compressive strength, flexural strength, splitting tensile strength, bond strength, static elastic modulus, rebound degree, and toughness coefficient. Indicators of durability include length change rate, expansion rate, mass loss rate, carbonation depth, chloride ion penetration depth, chloride ion diffusion coefficient, air permeability coefficient, water permeability coefficient, freeze-thaw resistance, electrical resistivity, dynamic elastic modulus, Poisson's ratio, and creep coefficient. Hereinafter, "quality of fresh concrete" may be simply referred to as "quality."

[0041] To give an overview of the functions of the quality prediction device 10, the quality prediction device 10 is configured to perform at least the following: construct a prediction model using machine learning based on training data including multiple datasets, and acquire multiple features obtained when producing ready-mixed concrete using the mixer 114. The quality prediction device 10 is further configured to use the above prediction model to obtain a predicted quality value corresponding to the acquired multiple features. The quality prediction device 10 constructs the prediction model such that one or more features selected from the above multiple features are associated with the predicted quality value using a decision tree.

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

[0043] The operation control unit 22 controls the manufacturing apparatus 100 to produce ready-mixed concrete according to predetermined operating conditions. At least a portion of the above operating conditions may be determined by an operator, such as a worker, each time ready-mixed concrete is produced. The operation control unit 22 may control the mixer drive unit 114b of the mixer 114 so that its rotational speed follows the target rotational speed specified in the operating conditions. When controlling the mixer drive unit 114b, the operation control unit 22 may adjust the power (e.g., current value) supplied to the mixer drive unit 114b. The power load tends to be higher when the ready-mixed concrete is hard, and lower when the ready-mixed concrete is soft.

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

[0045] Each of the features F1 to FN is data represented by a numerical value. Feature quantities F1 to FN may include data representing 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 could be used if the item does not belong to that classification, and a value of 1 if it does belong to that classification. Feature quantities F1 to FN do not include image data composed of a combination of coordinate information and pixel values. However, feature quantities F1 to FN may include features (one-dimensional numerical data) obtained from such image data.

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

[0047] One or more load features FL include, for example, the value at a predetermined point in time in the time-series data of the power load value of the mixer 114 when mixing concrete materials, and at least one of the statistics obtained from said time-series data. The time-series data of the power load value may be a data set consisting of continuous power load values ​​for at least a portion of the period during which the mixer 114 is operating during processing for one batch (a data set showing the change in power load value over time). Specific examples of features F1 to FN and specific examples of load features FL will be described later.

[0048] The model building unit 26 constructs a predictive model using machine learning based on training data that includes multiple datasets. Hereinafter, the training data that includes multiple datasets will be referred to as "training data TD," and the predictive model constructed by the model building unit 26 will be referred to as "predictive model M." Predictive model M is a model for predicting the fresh properties of ready-mixed concrete when features F1 to FN are obtained. The fresh properties predicted by predictive model M are, for example, slump, slump flow, or air content.

[0049] The model building unit 26 constructs a prediction model M using machine learning that shows the relationship between feature quantities F1 to FN and freshness characteristics (predicted values). The prediction model M is configured to output predicted values ​​of freshness characteristics in response to the input of feature quantities F1 to FN. Feature quantities F1 to FN can be called explanatory variables, and freshness characteristics can be called the target variable.

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

[0051] The model building unit 26 constructs a prediction model M such that one or more features selected from the feature vectors F1 to FN are associated with the predicted values ​​of fresh characteristics using a decision tree (decision tree algorithm). The one or more features used for associating with fresh characteristics in the prediction model M are selected from the feature vectors F1 to FN by machine learning. LightGBM (Light Gradient Boosting Machine) may be used as the decision tree algorithm. Alternatively, XGBoost (eXtreme Gradient Boosting) or RandomForest may be used instead of LightGBM as the decision tree algorithm.

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

[0053] The stage in which the predictive model M is autonomously constructed corresponds to the training phase (learning phase). The training phase may be performed before the production phase in which ready-mixed concrete is manufactured, or it may be performed at the beginning of the production phase. The stage in which the fresh concrete properties are predicted using the predictive model M from feature quantities F1 to FN (combinations of feature values ​​F1 to FN) whose fresh concrete properties are unknown corresponds to the evaluation phase. In the following explanation, terms may be labeled "for training" or "for evaluation" to distinguish whether it is the training phase or the evaluation phase, or which phase the data is used in. The same type of feature quantities F1 to FN are used between the training phase and the evaluation phase.

[0054] Here, with reference to Figure 3, we will explain an example of the analysis process in the prediction model M. Figure 3 shows a tree diagram that visualizes an example of the analysis process in the prediction model M. For the sake of easier understanding of the contents of this disclosure, we will use a simplified example. In the decision tree algorithm exemplified below, one data set is divided into two data sets in stages according to certain conditions, and each stage is referred to as "Stage 1," "Stage 2," and "Stage 3," in descending order from the highest level. The points where division or branching occurs based on conditions are called "nodes (condition nodes)." In Figure 3, combinations of "G" and numerical values, such as "G1," represent the names of the data sets. Also, combinations of "Th" and numerical values, such as "Th1," represent thresholds.

[0055] The model building unit 26 prepares training data TD, which consists of multiple datasets in which the combinations of feature values ​​F1 to FN are different from each other. Data group G1 is, for example, all the datasets included in training data TD. In the first stage, data group G1 is divided into two data groups, data group G21 and data group G22, based on one feature (feature F2 in the example shown in Figure 3) selected by machine learning from among the features F1 to FN. In the first stage, data group G1 is divided into two data groups based on the condition that feature F2 is less than or equal to the threshold Th1.

[0056] Dataset G21 consists of multiple datasets from dataset G1 where the feature F2 is less than the threshold Th1, and dataset G22 consists of the remaining multiple datasets from dataset G1 where the feature F2 is greater than or equal to the threshold Th1. The number of datasets in dataset G21 and the number of datasets in dataset G22 may be the same or different. In the first stage, the data is divided into two sets using the feature F2, but the values ​​of features other than feature F2 are also retained in datasets G21 and G22. The threshold Th1 is also set autonomously by machine learning.

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

[0058] In the example shown in Figure 3, in the third stage, data groups G31, G32, and G34 are not divided into two data groups. That is, each of data groups G31, G32, and G34 corresponds to a "leaf" (or "end node") in the tree diagram. In the third stage, with respect to data group G33, which is not a leaf, it is divided into two data groups by a feature and threshold selected by machine learning, as in the previous stage. 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. Each of data groups G41 and G42 corresponds to a leaf. As illustrated in Figure 3, the leaf depth (level) may differ among at least some of the leaves in the tree diagram. Alternatively, unlike the example shown in Figure 3, the depth of each leaf may be the same.

[0059] For each data set corresponding to a leaf, the predicted value of the fresh characteristics can be determined by the ground truth values ​​of the fresh characteristics included in that data set. For example, the arithmetic mean of the ground truth values ​​of the fresh characteristics included in the data set corresponding to a leaf is set as the predicted value for that leaf. In one example, if data set G41 contains m data sets (where m is an integer greater than or equal to 2), the arithmetic mean of the m ground truth values ​​of the fresh characteristics is set as the predicted value for the leaf corresponding to data set G41. The predicted values ​​of the fresh characteristics set for each leaf will differ among multiple leaves.

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

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

[0062] In the prediction model M, only a selection of features from features F1 to FN may be used in the branching conditions. That is, in the prediction model M, only a selection of features from features F1 to FN may be associated with the predicted values ​​of fresh characteristics. Even in such cases, the features associated with the predicted values ​​of fresh characteristics are selected from features F1 to FN, so the prediction model M still represents the relationship between features F1 to FN and the predicted values ​​of fresh characteristics.

[0063] The model building unit 26 may construct a corresponding prediction model M (a prediction model M that predicts only one type of freshness) for each type of freshness. For example, the model building unit 26 may individually 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] Returning to Figure 2, the model holding unit 28 holds (stores) the prediction model M constructed by the model building unit 26. The trained prediction model M may be reusable 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 freshness characteristics corresponding to the evaluation features F1 to FN obtained by the feature acquisition unit 24. For example, the prediction unit 30 inputs the values ​​of the evaluation features F1 to FN into the prediction model M and obtains the predicted values ​​of freshness characteristics output from the prediction model M. The prediction unit 30 may also obtain predicted values ​​for each type of freshness characteristic (for example, for slump, slump flow, and air volume) using the corresponding prediction model M.

[0066] Assuming that the prediction model M illustrated in Figure 3 is constructed, the prediction model M first compares the value of evaluation feature F2 (one of the evaluation features F1-FN) with the threshold Th1. If the value of evaluation feature F2 is greater than or equal to the threshold Th1, the calculation step proceeds to the node corresponding to data group G22, where the value of evaluation feature F2 is compared with the threshold Th22. If the value of evaluation feature F2 is less than the threshold Th22, the calculation step proceeds to the node corresponding to data group G33, where the value of evaluation feature F10 is compared with the threshold Th31.

[0067] Furthermore, if the value of the evaluation feature F10 is smaller than the threshold Th31, the prediction model M outputs a predicted value for the fresh properties set for 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 a predicted value for 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 fresh concrete properties obtained by the prediction unit 30 to the monitor 14. As a result, the predicted fresh concrete properties of the concrete being evaluated are displayed on the monitor 14, allowing operators such as workers to understand the predicted fresh concrete properties.

[0069] <Specific examples of decision tree algorithms> Next, with reference to Figures 4 to 7, we will describe specific examples of machine learning performed by the model building unit 26 when constructing a prediction model M based on a decision tree algorithm. The model building unit 26 may construct the prediction model M using machine learning, including ensemble learning. In machine learning including ensemble learning, multiple models may be merged to construct a single prediction model M.

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

[0071] Figure 4(a) illustrates, for simplicity, how a predictive model M is constructed by fusing two models (weak learners). However, a predictive model M that ultimately outputs a single predicted value may be constructed from three or more models (weak learners). Alternatively, a predictive model M that ultimately outputs a single predicted value may be constructed by using a Random Forest, which is generated by a randomly selected set of features from models (weak learners) such as the first model M1, which are then generated from these models. Note that Random Forest can also be considered a type of bagging technique.

[0072] The model building unit 26 may construct a predictive model M using machine learning that utilizes boosting, as a form of machine learning that includes ensemble learning. In machine learning that utilizes boosting, intermediate models Mtk (weak learners) are sequentially generated using a decision tree algorithm. Here, k is an integer from 1 to K, and K is an integer greater than or equal to 2. If intermediate models Mtk are generated in the order of k from 1 to K, then in machine learning that utilizes boosting, intermediate model Mt1 is generated first. Then, focusing on the error between the predicted value and the correct value in intermediate model Mt1, the next intermediate model, intermediate model Mt2, is generated. The single predicted value that is ultimately output by the predictive model M is a value that is aggregated by some calculation from the outputs of each intermediate model Mt1 to MtK.

[0073] The model building unit 26 may, for example, construct a prediction model M consisting of intermediate models Mt1 to MtK using "Adaboost," a boosting technique. Figure 4(b) schematically shows the process of building 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 actual value. For example, the weights are updated to be larger for datasets with larger errors. Subsequently, the weight updates and the generation of intermediate model Mtk are repeated for a set number of times, for example, K times.

[0074] The model building unit 26 may construct a predictive model M using machine learning that utilizes boosting, specifically by using gradient boosting. In other words, the model building unit 26 may construct a predictive model M using machine learning that utilizes a gradient boosting decision tree. In a gradient boosting tree, boosting, gradient descent, and a decision tree are combined to construct the predictive model M.

[0075] Figure 5 schematically illustrates the process of building a model using gradient boosting. In gradient boosting, an intermediate model Mt1 is first generated, and the error between the predicted value (1) from this intermediate model Mt1 and the actual value is calculated. Then, an intermediate model Mt2 is generated that can predict the error in the predicted value (1), and a correction value (2) is calculated from the intermediate model Mt2 to correct the predicted value (1). Subsequently, the calculation of the correction value (k) to correct the predicted value (k) at each point in time is repeated, for example, up to a set number of times K. The correction value (k) is also called the gradient (k).

[0076] Examples of decision tree algorithms that include gradient boosting include LightGBM and XGBoost mentioned above, as well as Catboost. In XGBoost, as shown in Figure 6(a), node branching (splitting) is performed level-wise. That is, all nodes are branched from left to right. On the other hand, in LightGBM, as shown in Figure 6(b), node branching (splitting) is performed leaf-wise. That is, branching is performed only on nodes that should be branched (for example, prioritizing nodes that result in a smaller loss). No further calculations are performed on nodes that no longer need to be branched.

[0077] Furthermore, LightGBM uses a histogram-based algorithm when performing branching. Figure 7(a) schematically shows the decision-making process when determining branching using a pre-sorted algorithm, which is different from the histogram-based algorithm. Figure 7(b) schematically shows the decision-making process when determining branching using a histogram-based algorithm. In Figures 7(a) and 7(b), "d1" to "d6" each represent one data point (or one dataset). In the pre-sorted algorithm, the decision of where to branch (split) is made after examining each individual value in data d1 to d6. On the other hand, in the histogram-based algorithm, the values ​​in data d1 to d6 are grouped by a histogram, and the decision of where to branch (split) is made on a group (cluster) basis.

[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 it is used as untrained data, while for data with small errors, only a portion is used as it is considered to have been trained to some extent. In addition, LightGBM uses a technique called FEB (Exclusive Feature Bundling) to combine multiple features of different types that do not simultaneously become zero (low correlation) and can be combined without problems into a single feature.

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

[0080] In the graph representing the time-series data in Figure 8(a), "Time t1" indicates the point in time when the mixer 114 started mixing, and "Time t2" indicates the point in time when the mixer 114 finished mixing. Time t1 corresponds, for example, to the point in time when the operation control unit 22 started driving the stirring member 114a. Time t2 corresponds, for example, to the point in time when the operation control unit 22 determined that the conditions for ending the mixing by the mixer 114 were met. In one example, the operation control unit 22 determines that the above conditions are met when a predetermined time has elapsed since the start of driving the stirring member 114a, and stops driving the stirring member 114a.

[0081] One or more load features FL may include the value of the power load at a predetermined point in time in the time series data. The predetermined point in time is a time set in advance by an operator or the like. One or more load features FL may include one or more of the following values: initial value P1, minimum value P2, maximum value P3, and 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 in time after the time when the initial value P1 was obtained (time t1). 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 time in the time series data. The final value P4 is the power load value at the end of time in the time series data.

[0082] The period of time series data used to obtain the load feature FL is not limited to the period from time t1 to time t2. The period of time series data used to obtain the load feature FL may be the period from time t1 to an arbitrary elapsed time tp. In the example shown in Figure 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 mixer 114 is completed.

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

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

[0085] Statistical quantities other than the range of variation and the magnitude of decline may be calculated from a data set of power load values ​​included in at least a portion of the time series data (a collection of power load values ​​obtained for each sampling period). For example, the sum can be obtained by accumulating the power load values ​​for each sampling period in at least a portion of the time series data. In addition, statistical quantities such as the sum may be calculated after some of the power load value data for each sampling period has been decimated. The time series data used to obtain statistical quantities may represent a moving average of the power load values ​​instead of data representing the time change of the power load values ​​themselves. One or more load feature quantities FL may include features obtained from the waveform of the time series data, and / or features obtained from an image displaying the time series data.

[0086] The sampling period when measuring the power load value may be 0.01 to 1 second (for example, 0.1 seconds). Measurement of the power load value may start immediately after mixing of concrete materials other than water begins, or immediately after adding water to the concrete materials other than water and starting to mix after dry mixing (mixing of concrete materials other than water). The number of power load values ​​measured will vary depending on the power load measurement interval and mixing time, but from the viewpoint of predicting 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 supply as the mixer 114.

[0087] One or more load features FL may include the change in power load value, and may also include information (numerical data) representing the pattern of change in power load value. From the viewpoint of predicting fresh properties with higher accuracy and at an earlier stage, data on the power load value after the time when the power load value of mixer 114 stabilizes (when the change in power load value becomes small) after mixing of water and other concrete materials has started may be used as the load feature FL. In the mixing of concrete materials, the timing for determining when the power load value of mixer 114 stabilizes varies depending on the water-cement ratio of the ready-mixed concrete, etc. However, for example, the power load value of the mixer may be considered stable after any of the following conditions (1) to (3) are met. (1) During a period in which the power load value is decreasing, if the rate of change of the power load value at predetermined intervals (e.g., 1 second) that are longer than the sampling period has been within ±1% for a predetermined set time (e.g., 3 seconds) or longer. The rate of change of the above 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 point, t is the current measurement point, and t-1 is the previous measurement point. (2) If the water-cement ratio of the ready-mixed concrete is around 50-70%, then approximately 30 seconds have elapsed since the start of mixing the water and other concrete materials (for example, when the other concrete materials are put into the mixer and dry-mixed, and then water is added to the mixer and mixing begins). (3) If the water-cement ratio is reduced (to less than 50%) from the viewpoint of strength development, the above timing will be delayed, and in the case of high-strength concrete, it will be about 1 to 10 minutes after the start of mixing of water and other concrete materials.

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

[0089] The date on which mixer 114 performs mixing is represented by a combination of month and day, indicating the specific day on which mixer 114 produced the ready-mixed concrete. In this case, the specific month may be one feature, and the specific day may be another feature. The time on which mixer 114 performs mixing is represented by a combination of hours, minutes, and seconds, or hours and minutes, indicating the specific time on which mixer 114 produced the ready-mixed concrete. In this case, the specific time may be one feature, and the specific minute (or each of the specific minute and second) may be another feature.

[0090] The above-mentioned time as a feature (the specific point in time when the mixer 114 produced ready-mixed concrete) may be the time when concrete material was supplied to the mixer 114, or the time when ready-mixed concrete was discharged from the mixer 114. The above-mentioned time as a feature (the specific point in time when the mixer 114 produced ready-mixed concrete) may be the time when the mixing member 114a was started to operate in the mixer 114, or the time when the mixing member 114a was stopped to operate.

[0091] The month and day when mixer 114 performs mixing, as well as the hour, minute, and second when mixer 114 performs mixing, may be specified using trigonometric functions. First, with reference to Figure 9, we will explain how to specify (represent) the "hour" of the time using trigonometric functions. Normally, the "hour" of a day is specified (represented) by 24 numbers from 0:00 to 23:00, assuming that it resets to 0:00 at 24:00. If we specify using such normal numbers, for example, the difference between 0:00 and 23:00 is only one hour if we do not consider the day, but numerically it will appear to be far apart. Therefore, the "hour" may be specified (represented) using trigonometric functions in a way that exhibits periodicity.

[0092] In one example, time h (where h is any integer between 0 and 23) can be represented by a combination of cos{2π×(h / 24)} and sin{2π×(h / 24)}, as shown in Figure 9. In this case, 0:00 is (1,0), 6:00 is (0,1), 12:00 is (-1,0), and 18:00 is (0,-1). In the input to the prediction model M, cos{2π×(h / 24)} may be one feature and sin{2π×(h / 24)} may be the other feature.

[0093] m minutes (where m is any integer between 0 and 59) and s seconds (where s is any integer between 0 and 59) may also be represented by the following combinations. Furthermore, in the input to the prediction model M, similar to "hours," each of the two numerical values ​​may be a single 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 months (where M is any integer between 1 and 12) and the "day" is D days (where D is any integer between 1 and Dm), then M months and D days may be represented by the following combinations. Dm is determined for each month and is one of the integers 28, 29, 30, and 31. In the input to the prediction model M, similar to the "hour," each of the two numbers may be a single 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 mixed by Mixer 114 in a day is information that identifies which batch process was performed on that day (which batch process produced the ready-mixed concrete). The batch process in which the ready-mixed concrete of interest was produced also indicates when the mixing by Mixer 114 was performed.

[0096] When the time from an arbitrarily set reference point to the time when the mixer 114 performs mixing is used as a feature, the reference point may be set by an operator such as a worker. The reference point may be set to the time after the inside of the mixer 114 has been cleaned during maintenance of the manufacturing equipment 100, and before manufacturing by the mixer 114 is resumed. In the case where the mixer 114 is cleaned every day after being shut down, the time before manufacturing by the mixer 114 is started for the first time in a day is also included. The time between the reference point and the time when the mixer 114 performs mixing (production of ready-mixed concrete) may be expressed in minutes or in seconds.

[0097] The time when mixer 114 performs mixing (production of ready-mixed concrete) may be the time when concrete materials are supplied to mixer 114, or the time when the ready-mixed concrete is discharged from mixer 114. The time when mixer 114 performs mixing (production of ready-mixed concrete) may be the time when the agitator 114a is started to drive, or the time when the agitator 114a is stopped from driving. The quality prediction device 10 may measure the time from the reference time to the time when mixer 114 performs mixing.

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

[0099] Feature quantities F1 to FN may include information related to mixing by the mixer 114 (information other than the power load value). As information related to mixing, in addition to the elapsed time tp mentioned above, examples of information related to mixing include the amount of concrete mixed for one batch, the mixing time, the time when the power load value reached its maximum value, the time from when the power load value reached its maximum value until the fresh concrete was discharged from the mixer 114, and the time from when the power load value reached its maximum value until the measurement time mp (final value P4) described later. Feature quantities F1 to FN may include information indicating the nominal strength, the type of cement, the type of admixture, the type of fine aggregate, the type of coarse aggregate, and the amount of additives added. In the information indicating the type of cement, etc., the types of cement, etc. may be classified in any way before each type is identified. Feature quantities F1 to FN may include numerical data obtained from images of fresh concrete being manufactured in or immediately after manufacturing in the mixer 114, or numerical data obtained from images of fresh concrete after it has been discharged from the mixer 114. The feature quantities F1 to FN may include at least one of the target quality of the ready-mixed concrete when it is used and the target quality of the ready-mixed concrete when it is shipped.

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

[0101] (Data on mix design conditions for ready-mixed concrete) The feature quantities F1 to FN include data related to the mix design conditions of ready-mix concrete, such as (i) the type and origin (or name) of materials used, (ii) cement, water, fine aggregate, coarse aggregate, various admixtures (blast furnace slag powder, fly ash, silica fume, expansive agent, volcanic glass powder, crushed stone powder, metakaolin, etc.), various admixtures (AE agent, water-reducing agent, AE water-reducing agent, high-performance water-reducing agent, high-performance AE water-reducing agent, fluidizer, 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 following: (iii) the amount (mass, weight, volume) and density of aramid fibers, nylon fibers, vinylon fibers, polyethylene fibers, polypropylene fibers, etc.; (iii) the water-cement ratio, water-binder ratio, air content, fine aggregate ratio, coarse aggregate bulk volume, maximum coarse aggregate size, coarse particle ratio, particle shape determination actual volume ratio, total alkali content in concrete, sludge solids content ratio, and recovered aggregate replacement rate; and (iv) the amount of stabilizers used in the adhering mortar and sludge water.

[0102] (Data regarding the quality of the target ready-mixed concrete) The feature quantities F1 to FN include data related to the quality of the target ready-mixed concrete, such as (i) target slump, slump flow, air content, 500mm flow arrival time, flow stop time, appearance (still or moving image), temperature, unit water content, plastic viscosity, yield value, funnel flow time, compaction properties, deformability, filling properties, and void passability, and (ii) concrete strength (design standard strength, durability design standard strength). The following data may be included: (iii) the degree of strength, quality standard strength, structural strength correction value (S value), mix design strength, nominal strength, etc., and one or more data selected from (iii) the static modulus of design, rebound coefficient, toughness coefficient, chloride content in concrete, length change rate, expansion rate, mass loss rate, carbonation depth, chloride ion penetration depth, chloride ion diffusion coefficient, void ratio, air permeability coefficient, water permeability coefficient, spacing coefficient, electrical resistivity, dynamic modulus, Poisson's ratio, creep coefficient, and color tone. The strength, slump, slump flow, air content, and chloride content of concrete can be measured by the test methods described in JIS A 5308:2024 (Ready-Mixed Concrete).

[0103] (Data related to cement) The feature quantities F1 to FN may include one or more data selected from the following as data related to cement, which is a concrete material: (a) data related to cement as a whole, (b) data related to the raw materials of cement clinker, (c) data related to the firing conditions of cement clinker, (d) data related to the crushing conditions of cement, and (e) data related to cement clinker.

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

[0105] The feature quantities F1 to FN include (b) data related to the raw materials of cement clinker, (b1) chemical composition, hydraulic coefficient, sieve test residue, Blaine specific surface area (particle density), ignition loss, supply amount, supply amount of auxiliary materials (special materials such as waste), storage volume (remaining amount) in the blending silo, and storage volume (remaining amount) in the storage silo, (b2) current value of the cyclone located between the raw material mill and the blending silo of the blended raw materials (representing the rotation speed of the cyclone and correlated with the speed of the raw materials passing through the cyclone), and (b3) at the time of loading into the kiln. (b4) The data may include one or more data selected from the following: (b4) the chemical composition and hydraulic coefficient of the raw materials for cement clinker (raw materials for cement clinker from which fine particles, etc., have been removed by countercurrent airflow during transport; hereafter referred to as raw materials for cement clinker kiln entry) at a predetermined time prior to (for example, one time point 5 hours prior, or multiple time points such as 3 hours prior, 4 hours prior, 5 hours prior, and 6 hours prior); and (b4) the chemical composition, hydraulic coefficient, Blaine specific surface area, sieve test residue, decarboxylation rate, and moisture content of the raw material obtained by mixing the raw materials for cement clinker kiln entry with auxiliary materials.

[0106] The feature quantities F1 to FN include (c) data related to the firing conditions of cement clinker, (c1) the amount of cement clinker raw material inserted into the kiln, kiln rotation speed, discharge temperature, firing zone temperature, cement clinker temperature, average kiln torque, O2 concentration, and NO during the firing of cement clinker. X The data may include one or more selected from (c2) concentration, (c3) clinker cooler temperature, and (c4) preheater gas flow rate (which correlates with the preheater temperature).

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

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

[0109] (Data on concrete materials other than cement) The characteristic quantities F1 to FN include data on concrete materials other than cement, such as (i) aggregate (at least one of fine aggregate and coarse aggregate), including type, oven-dry density, surface-dry density, water absorption rate, moisture content, surface moisture content, mass loss fraction in stability tests, abrasion loss, maximum size, particle size, coarse particle ratio, mass fraction of material remaining between consecutive sieves, particle shape determination volume, fine particle content, clay clump content, organic impurities, chloride content, and classification by alkali-silica reactivity; (ii) admixture type, density, specific surface area, residue on a 45 μm sieve, residue on a 1.2 mm sieve, flow value ratio, activity index, moisture content, expansion (length change rate), and various chemical components (silicon dioxide: SiO2, magnesium oxide). (iii) Content of admixtures (e.g., MgO, aluminum oxide: Al2O3, sulfur trioxide: SO3, free calcium oxide: f.CaO, free silicon: f.Si, etc.), (iii) type, composition, density, appearance (color), chloride ion content, total alkali content, water loss rate, flow value ratio, bleeding rate ratio, bleeding rate difference, setting time difference (start time, end time), compressive strength ratio, length change ratio, freeze-thaw resistance (relative dynamic modulus), and change over time, and (iv) type, density, nominal diameter, nominal length, shape, fineness, tensile strength, tensile modulus, adhering moisture content, melting temperature, and alkali resistance (strength retention rate).

[0110] (Data on concrete mixing equipment) Feature quantities F1 to FN may include one or more data selected from the following as data relating to concrete mixing equipment: (i) type, model, product name, manufacturer name, year of manufacture, rated capacity, total output, production capacity, main body mass, operating no-load mass, external dimensions, inclination angle (in the case of drum type), rotation speed (drum, stirring blades, stirring shaft), and mixing performance (deviation rate of air content in concrete, deviation rate of mortar content in concrete, deviation rate of coarse aggregate in concrete, deviation rate of consistency (slump), deviation rate of compressive strength), and (ii) information relating to the material input sequence, mixing amount, input time, mixing time, discharge time, re-input time, and cycle time.

[0111] (Environmental data) The feature quantities F1 to FN may include one or more data selected from the following as data relating to the environment during the production (including during transportation) of concrete: (i) outdoor temperature and humidity, atmospheric pressure, solar radiation, sunshine duration, rainfall, wind speed, wind direction, weather, climate, and climate; (ii) temperature of each concrete material, temperature and humidity of the place where the material is stored (in containers such as silos and storage bottles), and 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 on the vehicle.

[0112] <Control device hardware configuration> As shown in Figure 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 composed of one or more non-volatile memory devices such as flash memory or a hard disk. The storage 53 stores at least a quality prediction program that causes a computer to execute the construction process, acquisition process, and prediction process, which will be described later. The storage 53 stores quality prediction programs for configuring each functional block of the quality prediction device 10.

[0113] Memory 52 is composed of one or more volatile memory devices, such as random access memory. Memory 52 temporarily stores the quality prediction program loaded from storage 53. Processor 51 is composed of one or more computing devices, such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). Processor 51 executes the quality prediction program loaded into memory 52 to configure each functional block of the quality prediction device 10. The calculation results by processor 51 are temporarily stored in memory 52. ​​Input / output ports 54 perform input and output of information to and from input devices 12, monitors 14, mixers 114, etc., in response to requests from processor 51.

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

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

[0116] The manufacturing process includes, for example, a conveying process, a weighing process, a loading process, a mixing process, a discharge process, and a loading process. In the conveying process, various concrete materials are conveyed to the storage bottle 111 by the conveying device 104, and each material is supplied to the storage bottle 111 individually. In the weighing process, each material is supplied individually from the storage bottle 111 to the weighing bottle 112, and each material is weighed in the weighing bottle 112. In the weighing process, when the measured amount of each material reaches a predetermined set amount, that material is discharged into the collection hopper 113. In the loading process, after all types of materials have been collected in the collection hopper 113, the materials in the collection hopper 113 are loaded (supplied) into the mixer 114.

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

[0118] In the discharge process, after the mixing of the concrete materials is completed in the mixer 114, the ready-mixed concrete is discharged from the mixer 114 to 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 building process in the learning phase and a quality evaluation process in the evaluation phase. In the quality prediction process, the model building process is performed before the quality evaluation process.

[0120] The model building process includes a construction process. The construction process is the process of building a predictive model M using machine learning based on training data TD, which includes multiple datasets. In the construction process, the predictive model M is built so that one or more features selected from among multiple features F1 to FN are associated with the predicted values ​​of freshness characteristics using a decision tree. The construction process may also be performed by the model building unit 26 of the quality prediction device 10. In each of the multiple datasets in the training data TD, training features F1 to FN are associated with the correct values ​​of freshness characteristics.

[0121] The quality evaluation process includes an acquisition process and a prediction process. The acquisition process is the process of acquiring evaluation feature quantities F1 to FN obtained when producing ready-mixed concrete 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 load feature quantity 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 process is a process of obtaining predicted values ​​of fresh properties according to the evaluation feature quantities F1 to FN obtained in the acquisition process, using the prediction model M constructed in the construction process. The acquisition process may be performed by the prediction unit 30 of the quality prediction device 10. The quality evaluation process may include an output process. The output process is a process of outputting the predicted values ​​of fresh properties obtained in the prediction process to the monitor 14. The output process may be performed by the output unit 32 of the quality prediction device 10.

[0123] (Model building process) Figure 11(a) is a flowchart showing an example of a series of processes performed in the model building process. This model building process is performed before the production phase, which includes the above-mentioned manufacturing process, is executed in the manufacturing apparatus 100, or in the early stages after the production phase has started. In this model building process, for example, the ready-mixed concrete actually produced 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, for example, an operator such as a worker prepares the training data TD for machine learning. As described above, the training data TD consists of multiple datasets. Each of the multiple datasets in the training data TD contains feature quantities F1 to FN (training input information) obtained when training ready-mix concrete was manufactured, and the correct values ​​of fresh properties (e.g., slump) associated with those feature quantities F1 to FN.

[0125] The correct values ​​for fresh concrete properties may be obtained by actually measuring the fresh concrete properties used for training. In one example, after the training concrete is loaded onto the transport vehicle 200, some of the concrete is extracted by workers. The slump, slump flow, and air content of the extracted concrete are then measured by the workers, and at least some 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 building 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 building unit 26 may construct the prediction model M by machine learning using a decision tree algorithm. The model building unit 26 may construct the prediction model M by performing machine learning with ensemble learning in machine learning using a decision tree algorithm. The model building unit 26 may construct the prediction model M by performing machine learning with boosting (for example, gradient boosting) in machine learning with ensemble learning. In one example, the model building unit 26 constructs the prediction model M using LightGBM developed by Microsoft®.

[0127] The model building unit 26 may build one or more of the following models: 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 building unit 26 may also build a prediction model M that outputs the ratio of slump flow to slump as freshness characteristics.

[0128] Next, step S13 is executed. In step S13, for example, the model storage unit 28 stores the prediction model M constructed in step S12. With this, the model construction process is completed.

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

[0130] In the quality evaluation process, the quality prediction device 10 first executes step S21. In step S21, for example, the quality prediction device 10 waits until it is time to evaluate the quality of the ready-mixed concrete to be evaluated. The evaluation timing may be predetermined to be a specific time period within the day, or it may be predetermined to be the timing of executing a certain batch process within the day. The evaluation timing may also be the timing when an operator, such as a worker, instructs the device to perform the evaluation.

[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 manufactured. The feature quantities F1 to FN acquired in step S22 are input data for evaluation in which the fresh properties (e.g., 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 obtains the predicted value 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 fresh concrete properties obtained in step S23 on the monitor 14. This allows operators, such as workers, to confirm the predicted results regarding the fresh concrete properties being evaluated.

[0134] The quality evaluation process is thus completed. The quality prediction device 10 may perform the series of processes from steps S21 to S24 each time a batch of ready-mixed concrete is manufactured (for each batch process). The quality prediction device 10 may also perform the series of processes from steps S21 to S24 each time multiple batches of ready-mixed concrete are manufactured (for each batch process).

[0135] [Differentiation] The series of processes shown in Figures 11(a) and 11(b) are examples and can be modified as appropriate. In the above series of processes, one step and the next step may be executed in parallel, and some steps may be executed in a different order than the example above. In place of at least some of the steps in the above series of processes, or in addition to the above series of processes, steps with content different from the example above may be executed.

[0136] In addition to predicting the quality of the ready-mixed concrete using the quality prediction device 10, the fresh properties of the ready-mixed concrete may be measured periodically (for example, once to 100 times per day) in the manufacturing device 100. In this case, the prediction model M may be updated based on the measured values ​​of the fresh properties of the ready-mixed concrete and the feature quantities F1 to FN obtained at the time the measured values ​​were acquired. In the above prediction process, the fresh properties of the ready-mixed concrete may be predicted using the updated prediction model M. Even when the updated prediction model M is used, the process of predicting the fresh properties of the ready-mixed concrete based on the prediction model M and the feature quantities F1 to FN acquired in the above acquisition process remains unchanged.

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

[0138] If the manufacturing system 1 includes a transport vehicle 200, at least a portion of the feature quantities F1 to FN may include feature quantities obtained when manufacturing (mixing) ready-mixed concrete using the mixer 214. One or more load feature quantities FL may be one or more feature quantities related to the power load value of the mixer 214 when mixing the ready-mixed concrete. The feature quantity acquisition unit 24 of the quality prediction device 10 may calculate one or more load feature quantities FL from time-series data up to measurement time mp during the mixing operation by the mixer 214. The prediction unit 30 may predict the quality of the ready-mixed concrete at the time 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 feature quantities related to the power load value of the mixer 214 when mixing the ready-mixed concrete, as at least a part of the feature quantities F1 to FN (one or more load feature quantities FL). The feature quantities F1 to FN may include one or more data selected from the following as data related to the transportation of ready-mixed concrete: (i) the volume of ready-mixed concrete in the mixer 214 of the transport vehicle 200, the thickness, mass, and temperature of the placed concrete, (ii) the ambient temperature during transportation, and (iii) the transportation time (time from the end of mixing to the end of transportation (when unloading)) and the transportation distance. The feature quantities F1 to FN may also include batcher data (type of heating / cooling device, storage water temperature, storage aggregate temperature, storage cement temperature, measuring bottle type, discharge amount, maximum discharge pressure, etc.). The quality prediction device 10 does not need to have a function to control the manufacturing device 100.

[0140] In the above example, the ready-mixed concrete is manufactured at a location other than the construction site (manufacturing equipment 100). The location where the ready-mixed concrete is manufactured is not limited to this example. Concrete materials (for example, concrete materials excluding water) may be transported to the construction site by a transport vehicle, and the ready-mixed concrete may be manufactured at the site. In this case, water may be added to the concrete materials transported by the transport vehicle using a mixer installed on the transport vehicle, and the materials may be mixed to obtain the ready-mixed concrete. The quality prediction device 10 may predict the quality of the ready-mixed concrete, such as its fresh properties, at the site when the ready-mixed concrete is manufactured there. The feature acquisition unit 24 of the quality prediction device 10 may acquire feature quantities related to the power load value of the mixer installed on the transport vehicle when the ready-mixed concrete is manufactured at the site, as at least a part of one or more load feature quantities FL.

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

[0142] [Verification of prediction results using predictive models] Next, we will explain the results of verifying the prediction results of fresh characteristics by a prediction model M constructed using a decision tree algorithm, with features F1 to FN as input, using a dataset where the correct values ​​are known. In this verification, we compared the prediction results with those of a comparison model constructed using a deep neural network (DNN), which is different from the decision tree algorithm, using the same training data TD.

[0143] A training dataset of 120,552 data points was prepared as training data TD, and a test dataset of 13,383 data points with known ground truth values ​​were prepared to evaluate the prediction results. In both the training and test datasets, the features F1-FN were linked to the ground truth values ​​of fresh concrete properties. Data obtained from a ready-mix concrete manufacturing system configured similarly to manufacturing system 1 was used as the ground truth values ​​(measured values) for the features F1-FN and fresh concrete properties. In addition, the fresh concrete properties before shipment were used as the target of prediction by the prediction model.

[0144] The data items shown in Table 1 were used as feature vectors F1 to FN. In Table 1, "Manufacturing month (sin, cos)" refers to data representing the manufacturing month using the trigonometric functions sin and cos, and the same meaning applies to manufacturing date, etc. "Cement type (a1, a2, a3, a4, a5)" refers to data indicating which of the cement types a1 to a5 is included, and the same meaning applies to 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. Models for predicting slump, slump flow, and air volume were constructed separately. In constructing the prediction model M for 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)". Models for predicting slump, slump flow, and air volume were constructed separately. In constructing the prediction model M for 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. The main conditions (network configuration) for constructing the comparative prediction model were set as follows: • A Batch Normalization layer was not used to normalize the input data to the model. • Dropout, which involves removing some connections in the intermediate layers of the neural network, was not used. The neural network's hidden layers consisted of eight fully connected layers. Adam was used as the weight update formula, and HuberLoss was used as the loss function. Machine learning was used to output three types of data as freshness characteristics from a single predictive model: slump, slump flow, and air volume.

[0148] (Verification results) For each of Example 1, Example 2, and Comparative Example, the predicted values ​​obtained by the prediction model were compared with the correct values ​​included in the test dataset using a test dataset. Specifically, a correct answer was defined as a predicted value obtained by the model that falls within a range obtained by adding a predetermined tolerance to the correct values ​​in the test dataset. 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 metric.

[0149] Figure 12(a) shows the evaluation results for the accuracy rate regarding slump. In Figure 12(a), when the tolerance is "±0.5cm", datasets in which the model's predicted value falls within the range of ±0.5cm of the correct value are considered correct, and the percentage of datasets considered correct is shown as the accuracy rate (%). The accuracy rate (%) is calculated similarly when the tolerance is "±1.0cm", "±1.5cm", "±2.0cm", and "±2.5cm". From the results shown in Figure 12(a), it can be seen that the accuracy rate when predicting slump using the prediction model M of Example 1 and Example 2 is significantly higher compared to the comparative example.

[0150] Figure 12(b) shows the evaluation results for the accuracy of slump flow. In Figure 12(b), when the tolerance is "±2.5cm", datasets in which the model's predicted value falls within ±2.5cm of the correct value are considered correct, and the percentage of these correct datasets is shown as the accuracy rate (%). The accuracy rate (%) is calculated similarly when the tolerance is "±3.0cm", "±4.0cm", "±5.0cm", "±6.0cm", and "±7.5cm". From the results shown in Figure 12(b), it can be seen that the accuracy rate when predicting slump flow using the prediction model M of Example 1 and Example 2 is significantly higher compared to the comparative example.

[0151] Figure 13(a) shows the evaluation results of the accuracy rate for air volume. In Figure 13(a), when the tolerance is "±0.3%", datasets in which the model's predicted value falls within ±0.3% of the correct value are considered correct, and the percentage of these correct datasets is shown as the accuracy rate (%). The accuracy rate (%) is calculated similarly when the tolerance is "±0.5%", "±0.7%", "±1.0%", "±1.2%", and "±1.5%". From the results shown in Figure 13(a), it can be seen that the accuracy rate when predicting air volume using the prediction model M of Example 1 and Example 2 is significantly higher compared to the comparative example.

[0152] As an evaluation metric separate from accuracy, we calculated the "sum of the mean and standard deviation" of the error between the predicted value and the correct value. Using the sum of the mean and standard deviation as an evaluation metric means that the error is evaluated while taking variability into account. In other words, even if the error itself (mean of the error) is small, if the standard deviation is large, it means that the variability of the error is large, which is undesirable. Also, even if the standard deviation of the error is small, if the mean is large, it means that the error itself is large, which is undesirable. 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 determined the mean and standard deviation from these errors.

[0153] Figure 13(b) shows the results of calculating the mean and standard deviation sum of errors for slump, slump flow, and air volume, respectively. From the results shown in Figure 13(b), it can be seen that for all fresh properties of slump, slump flow, and air volume, the sum of the mean and standard deviation of errors is significantly smaller when predicted by the prediction model M of Examples 1 and 2 compared to the comparative example.

[0154] In addition to prediction accuracy, we evaluated the "training time," which is the time from when the computer starts training after the preparation of the training data (TD) and the setting of various hyperparameters have been completed until the prediction model is completed. Training time can be defined as the time required for the computer to build the prediction model. The evaluation results of the training time are shown in Table 4 below. [Table 4]

[0155] Table 4 shows the training time for constructing a predictive model M that predicts only slump in Examples 1 and 2, and the training time for constructing a predictive model that simultaneously outputs slump, slump flow, and air volume in the Comparative Example. From the results shown in Table 4, it can be seen that the training time can be significantly reduced in Examples 1 and 2 compared to the Comparative Example. It is assumed that the training time in the Comparative Example is not significantly different from the training time when constructing a comparative model that predicts only slump using the same method as the Comparative Example, but even if the training time in Examples 1 and 2 is tripled, it is still significantly shorter than in the Comparative Example. Furthermore, it can be seen that the training time can be further reduced in Example 1 compared to Example 2. For example, by shortening the training time, it is possible to construct or update the predictive model M while manufacturing ready-mixed concrete (it is possible to continue manufacturing ready-mixed concrete without being constrained by the construction or updating of the predictive model M).

[0156] (Consideration of factors contributing to the improved prediction accuracy compared to the comparative example) [Predictive model based on decision trees] The prediction models M constructed in Examples 1 and 2 above are both built using the decision tree algorithm. In a decision tree, branching is performed by focusing on the part with the greatest difference in features at each node. Therefore, if there is a relationship between the features and the data to be predicted, a certain level of accuracy can be ensured even with a limited amount of training data. Also, because the model structure is simple, overfitting is less likely to occur even with a small amount of data. On the other hand, the DNN in the comparative example uses all the feature data in the training data to perform complex learning and excels at capturing fine features and patterns in the data. DNNs with such characteristics require a huge amount of data to ensure accuracy. Also, because the model structure is complex, it is easy to overfit to the training data when the amount of data is small. In reality, it is difficult to prepare training data with a huge amount of data in machine learning, so in predictions using limited data, it is thought that decision trees can construct prediction models with higher accuracy.

[0157] [Ensemble learning using decision trees] The prediction models M constructed in Examples 1 and 2 above are both built using decision tree algorithms that include ensemble learning. Ensemble learning using decision trees includes a method called "bagging," which involves arranging multiple decision trees with different characteristics in parallel and learning them independently, or a method called "boosting," which involves arranging multiple decision trees with different characteristics in series and learning them sequentially. Because these methods construct a prediction model using multiple decision trees, they have the effect of reducing the variability of prediction results and the error between predicted values ​​and correct values. Therefore, it is presumed that the prediction accuracy was improved by performing ensemble learning using multiple decision trees compared to learning with a single DNN. However, ensemble learning of DNNs can be a disadvantage because the model structure becomes more complex, and it can be difficult to find an appropriate combination of models and adjust the parameters, thus requiring a great deal of effort to build the model.

[0158] [Gradient Boosting Decision Tree] The predictive model M constructed in Example 1 above is built using LightGBM, a type of machine learning that includes gradient boosting. (1) LightGBM excels at capturing the synergistic effects when combining numerical data between parameters (between features). On the other hand, DNNs are not suitable for capturing the interrelationships between data. Considering that freshness, an example of quality, is determined by a combination of factors such as concrete mix, date, and load value, it is presumed that the prediction accuracy improved by constructing a model using LightGBM, which can capture these interactions. (2) LightGBM can select and learn important features of the data. Specifically, it uses all the data for features with large prediction errors, considering them to be important features, and randomly extracts data for features with small prediction errors, which are considered already learned, to create a predictive model. On the other hand, DNNs use all the data related to each feature to perform complex learning and are good at capturing fine features and patterns of the data. DNNs with such characteristics are effective for image, video, and audio data, but require a huge amount of data to improve accuracy. In predictions using limited numerical data, it is thought that a highly accurate predictive model can be created by selecting and learning important features of the data.

[0159] [Summary of this disclosure] The ready-mix concrete quality prediction method described above is a quality prediction method for predicting the quality of ready-mix concrete. This quality prediction method includes a construction step of building a prediction model (M) using machine learning based on training data (TD) containing multiple datasets, an acquisition step of acquiring multiple feature quantities (F1~FN) obtained when manufacturing ready-mix concrete using a mixer (114,214), and a prediction step of obtaining predicted quality values ​​according to the multiple feature quantities (F1~FN) acquired in the acquisition step, using the prediction model (M) built in the construction step. In each of the above multiple datasets, multiple feature quantities (F1~FN) are associated with the correct quality values. The multiple feature quantities (F1~FN) include one or more load feature quantities (FL) related to the power load value of the mixer (114,214) when mixing concrete materials or stirring ready-mix 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 the predicted quality values ​​using a decision tree.

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

[0161] In the ready-mix concrete quality prediction method described above, a prediction model (M) may be constructed during the construction process using machine learning, including ensemble learning. In this case, since multiple models are merged to construct a single final prediction model (M), it is even more useful for improving prediction accuracy.

[0162] In the concrete quality prediction method described above, a prediction model (M) may be constructed during the construction process using machine learning with boosting. In this case, multiple models generated sequentially, taking errors into account, are merged to construct a single final prediction model (M), thereby improving the prediction accuracy of the decision tree model.

[0163] In the concrete quality prediction method described above, a prediction model (M) may be constructed during the construction process using machine learning with gradient boosting. In this case, multiple models are sequentially generated to reduce the error and merge to construct a single prediction model (M) that obtains a final predicted value, thereby further improving the prediction accuracy of the decision tree model.

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

[0165] In the ready-mix concrete quality prediction method described above, the multiple feature quantities (F1 to FN) may further include one or more time feature quantities relating to the timing when the mixer (114) performs mixing of the concrete material. The one or more time feature quantities may include at least one of the following: the date and time when the mixer (114) performs mixing of the concrete material, the cumulative number of batches mixed by the mixer (114) in one day, the time from an arbitrarily set reference point to the time when the mixer (114) performs mixing, and the cumulative number of batches mixed by the mixer (114) from an arbitrarily set reference point. As a result of our verification, it has been confirmed that even if the manufacturing conditions, including the operating conditions of the mixer (114), are the same, the power load value itself can fluctuate depending on the timing of ready-mix concrete production. In the above method, by adding the above time feature quantities to the input of the prediction model (M), it is possible to construct a prediction model (M) that also takes into account the fluctuations in the power load value due to the timing of ready-mix concrete production. As a result, it is possible to reduce the variability in the prediction results of the predictive model (M) that may arise due to differences in manufacturing timing.

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

[0167] The ready-mix concrete manufacturing method described above includes a manufacturing process for manufacturing ready-mix concrete, and a quality prediction process for predicting the quality using the quality prediction method described above, either after or during the manufacturing process. In this ready-mix concrete manufacturing method, the quality is predicted using the quality prediction method described above, and is therefore useful for improving prediction accuracy, similar to the quality prediction method described above.

[0168] The ready-mixed concrete quality prediction device (10) described above is a device for predicting the quality of ready-mixed concrete. This quality prediction device (10) comprises a model construction unit (26) that constructs a prediction model (M) by machine learning based on training data (TD) including multiple datasets, an acquisition unit (24) that acquires multiple feature quantities (F1~FN) obtained when ready-mixed concrete is manufactured 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 quality value corresponding to the multiple feature quantities (F1~FN) acquired by the acquisition unit (24). In each of the multiple datasets, multiple feature quantities (F1~FN) are associated with the correct quality value. The multiple feature quantities (F1~FN) include one or more load feature quantities (FL) related to the power load value of the mixer (114,214) when mixing concrete materials or stirring ready-mixed concrete. The model building unit (26) constructs a prediction model (M) such that one or more features selected from among multiple features (F1 to FN) are associated with the predicted quality value using a decision tree. This quality prediction device is useful for improving prediction accuracy, similar to the quality prediction method described above.

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

[0170] 1...Manufacturing system, 10...Quality prediction device, 24...Feature acquisition unit, F1~FN...Features, FL...Load features, 26...Model construction unit, TD...Training data, 30...Prediction unit, M...Prediction model, 100...Manufacturing equipment, 114...Mixer, 114a...Agitation member, 114b...Mixer drive unit, 214...Mixer.

Claims

1. A quality prediction method for predicting the quality of ready-mixed concrete, The construction process involves building a predictive model using machine learning based on training data containing multiple datasets, An acquisition process for obtaining multiple characteristic quantities obtained when producing ready-mixed concrete using a mixer, The process includes a prediction step which uses the prediction model constructed in the construction step to obtain a prediction value of the quality corresponding to the plurality of features obtained in the acquisition step, In each of the aforementioned datasets, the aforementioned features and the ground truth values ​​of the quality are associated. The aforementioned 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 fresh concrete. In the construction step, the prediction model is constructed such that one or more features selected from the plurality of features are associated with the predicted quality values ​​using a decision tree. A method for predicting the quality of ready-mixed concrete.

2. In the aforementioned construction process, the predictive model is constructed using machine learning, including ensemble learning. The method for predicting the quality of ready-mixed concrete according to claim 1.

3. In the aforementioned construction process, the predictive model is constructed using machine learning with boosting. The method for predicting the quality of ready-mixed concrete according to claim 2.

4. In the aforementioned construction process, the predictive model is constructed using machine learning with gradient boosting. The method for predicting the quality of ready-mixed concrete according to claim 3.

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

6. The aforementioned plurality of feature quantities further include one or more time feature quantities relating to the timing when the mixer performs the mixing of the concrete material, The one or more time-specific features mentioned above are, The date and time when the mixer performs the mixing of the concrete materials, The cumulative number of batches mixed by the aforementioned mixer in one day, The time from an arbitrarily set reference point to the point in time when the mixer performs mixing, and The cumulative number of batches mixed by the mixer from an arbitrarily set reference point, Including at least one of the following: A method for predicting the quality of ready-mixed concrete according to any one of claims 1 to 4.

7. A quality prediction program that causes a computer to execute the quality prediction method described in any one of claims 1 to 4.

8. The manufacturing process for producing ready-mixed concrete, A quality prediction step includes predicting the quality after or during the execution of the manufacturing step using the quality prediction method described in any one of claims 1 to 4. A method for manufacturing ready-mixed concrete.

9. A quality prediction device for predicting the quality of ready-mixed concrete, A model building unit that constructs a predictive model using machine learning based on training data containing multiple datasets, An acquisition unit that acquires multiple characteristic quantities obtained when manufacturing ready-mixed concrete using a mixer, The system includes a prediction unit that uses the prediction model constructed by the model construction unit to acquire a prediction value of the quality corresponding to the plurality of feature quantities acquired by the acquisition unit, In each of the aforementioned datasets, the aforementioned features and the ground truth values ​​of the quality are associated. The aforementioned 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 fresh concrete. The model building unit constructs the prediction model such that one or more features selected from the plurality of features and the predicted quality values ​​are associated by a decision tree. A device for predicting the quality of ready-mixed concrete.

10. A manufacturing apparatus for producing ready-mixed concrete, The quality prediction device according to claim 9 comprises, The quality prediction device predicts the quality of the ready-mixed concrete produced by the manufacturing device. A ready-mix concrete manufacturing system.

Images