KR102437297B1 - Apparatus and method for providing a hybrid model for prediction and management of building operational information - Google Patents

Abstract

A providing device is provided. The providing device includes: a first acquisition unit for acquiring first operation information of a specific building predicted through a first model; a second acquisition unit for acquiring second operation information of the specific building predicted through a second model; and a prediction unit for selectively outputting one of the first management information and the second management information. The present invention provides the device and method for providing a hybrid model which interconnects a physical model with a data-based model which utilizes artificial functions and big data.

Description

Apparatus and method for providing a hybrid model for prediction and management of building operational information

The present invention relates to an apparatus and method for providing a hybrid model for predicting operational information of a building.

Countries around the world are making various efforts to reduce greenhouse gas emissions and use of fossil energy.

Korea is also participating in this movement under the low-carbon green growth policy, but there are still few tangible results.

Building energy accounts for the largest share of energy consumption and carbon emissions.

Despite various efforts such as strengthening design standards and building energy rationalization projects, building energy consumption over the past five years has been surveyed to account for 56% of total energy consumption.

One of the reasons why such efforts to reduce building energy is not achieving practical results despite increasing efforts to reduce building energy is that the approach to building energy is different in the design stage and operation stage.

For example, even if a building design that optimizes building energy is made in the design stage, many values may change during the construction and process of the building, and actual operation may be performed differently from assumptions made in the design.

If it is designed to use energy-saving materials but other materials are used in the construction stage, or if the expected climate or operating time at the design stage is different from the actual climate and operating time, the building control method should be changed accordingly. is the reality

Reasons for this include that it is difficult and takes too much time to build a suitable model in the operation stage, and when a precise model cannot be built, the accuracy is lowered and the reliability of the results is lowered.

Korean Patent Publication No. 1461081 discloses a technique for constructing a model closest to an actual building while maximizing the model built in the design stage.

Korean Patent Publication No. 1461081

An object of the present invention is to provide an apparatus and method for providing a hybrid model that interconnects a physical model with a data-based model utilizing artificial functions and big data.

The providing apparatus of the present invention includes: a first acquisition unit for acquiring first operation information of a specific building predicted through a first model; a second acquisition unit for acquiring second operation information of the specific building predicted through a second model; and a prediction unit selectively outputting one of the first operation information and the second operation information.

The providing apparatus of the present invention may provide a hybrid model that selectively outputs one of the first predicted value of the first model and the second predicted value of the second model according to an error value compared with the actual measured value of the building operation information .

The providing method of the present invention includes an acquiring step of acquiring first operation information of a specific building predicted through a first model, and acquiring second operating information of the specific building predicted through a second model; a collecting step of collecting actual measured values of the operation information of the specific building; a comparison step of comparing a first predicted value included in the first operation information with the actual value, and comparing the measured value with a second predicted value included in the second operation information; a selection step of selecting a model outputting a predicted value having a small error between the measured value from among the first model and the second model; It may include; a prediction step of outputting the operation information predicted through the model selected in the selection step from among the first operation information and the second operation government.

According to the present invention, a system that interconnects a physical model together with a data-based model utilizing artificial intelligence and big data is implemented, and the stability of prediction performance can be secured.

According to the present invention, operation information of a building including energy consumption and the like can be accurately predicted.

The providing apparatus of the present invention can focus on maintaining various buildings efficiently and comfortably rather than predicting and responding to disasters and disaster situations.

The providing apparatus of the present invention may present a hybrid model in which a first model including a physical model and a second model including a data-based model are interconnected.

The hybrid model monitors in real time the prediction accuracy of each model, which varies from moment to moment, when the first model and the second model for predicting the building's operational information (including the same type of prediction value) are prepared, and has high prediction accuracy at the current time. You can output the model's predicted values.

The providing apparatus of the present invention collects and analyzes the physical characteristics of a building and information generated in the building operation stage, and predicts and manages operation information including at least one of energy usage (consumption), indoor environment state, and occupancy state. can provide

In comparison with a physical model having a constant prediction accuracy regardless of the passage of time, the data-based model may have improved prediction accuracy as various types of data are accumulated. At some point (intersection) over time, the prediction accuracy of a data-driven model can exceed that of a physical model. In such an environment, it may be evident that it is advantageous to use the prediction values of the first model in the beginning and the prediction values of the second model in the latter half. However, the problem is that it is necessary to understand at what point in time the first half and the second half are divided. The providing apparatus of the present invention may provide a specific method for identifying a corresponding time point (intersecting time point). As a result, according to the present invention, a variable model or hybrid model that outputs the predicted value of the first model in the early stage and outputs the predicted value of the second model in the latter half based on the crossing point may be provided.

In other words, according to the present invention, a hybrid model in which a physical model using physical parameters of a building, such as a construction year and surrounding climate, and a data-based model using building operation information, are complemented may be provided.

By using the hybrid model, a maintenance system can be built that simplifies the series of processes from diagnosing machine coupling to completing repairs.

According to the present invention, since a building shape can be visualized using the first model, an intuitive and user-friendly interface can be implemented.

Looking at it from another point of view, the problems caused by the lack of data in the initial stage can be solved in the second model using big data and artificial intelligence. In the initial stage when the training dataset is insufficient, the initial prediction accuracy of the operation information may be improved by using a physical model such as BIM (Building Information Modeling) instead of the second model having low prediction accuracy of the operation information. When the training dataset is accumulated and machine learning is sufficiently performed, the prediction accuracy of the second model exceeds that of the physical model, and from this point on, the second operation information predicted by the second model becomes the first operation of the physical model. It can be used in place of information.

1 is a schematic diagram showing a providing apparatus of the present invention.
2 is a schematic diagram illustrating prediction accuracy of a first model and a second model.
3 is a schematic diagram illustrating a selection point of the second model.
4 is a schematic diagram illustrating a hybrid model.
5 is a schematic diagram showing the operation of the providing device.
6 is a flowchart illustrating a method of providing the present invention.
7 is a diagram illustrating a computing device according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art to which the present invention pertains can easily implement them. However, the present invention may be embodied in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

In this specification, duplicate descriptions of the same components will be omitted.

Also, in this specification, when a component is referred to as 'connected' or 'connected' to another component, it may be directly connected or connected to the other component, but other components in the middle It should be understood that there may be On the other hand, in the present specification, when it is mentioned that a certain element is 'directly connected' or 'directly connected' to another element, it should be understood that the other element does not exist in the middle.

In addition, the terms used herein are used only to describe specific embodiments, and are not intended to limit the present invention.

Also in this specification, the singular expression may include the plural expression unless the context clearly dictates otherwise.

Also, in this specification, terms such as 'include' or 'have' are only intended to designate that the features, numbers, steps, operations, components, parts, or combinations thereof described in the specification exist, and one or more It should be understood that the existence or addition of other features, numbers, steps, acts, components, parts, or combinations thereof, is not precluded in advance.

Also in this specification, the term 'and/or' includes a combination of a plurality of listed items or any of a plurality of listed items. In this specification, 'A or B' may include 'A', 'B', or 'both A and B'.

Also, in this specification, detailed descriptions of well-known functions and configurations that may obscure the gist of the present invention will be omitted.

1 is a schematic diagram showing a providing apparatus 100 of the present invention. 4 is a schematic diagram illustrating a hybrid model. 5 is a schematic diagram illustrating an operation of the providing apparatus 100 .

The providing apparatus 100 illustrated in FIG. 1 includes a first acquisition unit 110 , a second acquisition unit 120 , a collection unit 130 , a comparison unit 150 , a selection unit 170 , and a prediction unit 190 . may include.

The first acquisition unit 110 may acquire first operation information of a specific building predicted through the first model.

The second acquisition unit 120 may acquire second operation information of a specific building predicted through the second model.

The first acquisition unit 110 and the second acquisition unit 120 may acquire future operation information for the same specific building.

The operation information may include at least one of energy information, indoor environment information, and occupancy state information.

The energy information may include at least one of information on heating and cooling loads and energy consumption (usage).

Indoor environment information may include temperature, air quality, and the like.

The occupancy state information may include the number of people present in the building.

Operational information can be predicted by various methods. As an example, the operation information may be output through the first model and the second model.

The data input to the first model and the second model includes at least one of a survey on the building, the size of the building, the facility system in the building, the envelope characteristics of the building, indoor environment information, energy measurement information, outdoor environment information, and occupancy information. may include The first model and the second model to which these information are input can predict future operation information. The survey may be obtained from users who use or manage the building. Building scale, equipment system, and envelope characteristics can be obtained through specifications, blueprints, administrative documents, etc. in which the characteristics of the building are recorded. Indoor environment information, energy measurement information, outdoor environment information, occupancy information, etc. may be acquired through various sensors. At least some of the outdoor environment information 'Weather info' may be provided from a separate weather server.

The first model is EnergyPlus, which simulates the energy use of a building, and based on a three-dimensional information model, the shape, properties, etc. of the facility are converted into information so that all information that occurs throughout the life cycle of the facility can be integrated and utilized. It may include at least one of Building Information Modeling (BIM), which means an expressed digital model. Alternatively, the first model may include a digital twin including at least one of Energy Plus and BIM.

The second model may include a model in which prediction accuracy is initially lower than that of the first model, but in which prediction accuracy is higher than that of the first model as time passes. As an example, the second model may include an artificial intelligence model generated based on machine learning. The artificial intelligence model may use artificial intelligence algorithms such as regression, NN-based, and decision trees based. The second model using these artificial intelligence algorithms tends to improve the output accuracy as the amount of input data or training dataset increases and accumulates. In the present invention, the output accuracy of the second model may mean the prediction accuracy of future operation information for a building.

The prediction unit 190 may selectively output one of the first operation information and the second operation information.

When the providing apparatus 100 including the first acquiring unit 110 , the second acquiring unit 120 , and the predicting unit 190 is viewed from the outside, the output value of the first model is output in some situations, and in other situations, the output value of the first model is output. An output value of the second model may be output. Such an appearance may be similar to a hybrid model in which the first model and the second model are mutually transformed or interrelated.

The comparison unit 150 may compare the first predicted value included in the first operation information with the actual value measured for a specific building. The actual value may include a measurement value for each item of operation information directly measured in a specific building.

The comparison unit 150 may compare the second predicted value included in the second operation information with the actual value measured for a specific building.

The selector 170 may select a model outputting a predicted value having a small error from the measured value from among the first model and the second model.

For example, the comparator 150 may calculate a first error value including an error value between the first predicted value and the measured value. The comparator 150 may calculate a second error value including an error value between the second predicted value and the actual measured value.

The selector 170 may analyze a smaller value among the first predicted value and the second predicted value, and select a model that is a source of the corresponding value. For example, when the first predicted value is smaller than the second predicted value, the selector 170 may select the first model. When the second predicted value is smaller than the first predicted value, the selector 170 may select the second model.

The prediction unit 190 may output future operation information predicted through the model selected by the selection unit 170 .

The collection unit 130 may collect the measured values. As an example, the collection unit 130 may include various sensors for measuring an actual value. In this case, the collection unit 130 may directly generate and collect the measured values. Alternatively, the collection unit 130 may include a communication module for wired/wireless communication with various sensors that output measured values. In this case, the collection unit 130 may receive the measured value from the sensor.

The collection unit 130 may also collect a measurement time point at which a specific building is actually measured together with an actual measurement value. The collection unit 130 may provide the measured value and the measurement time to the comparison unit 150 .

The first acquisition unit 110 may additionally acquire a first target viewpoint targeted by the first operation information.

The second acquirer 120 may additionally acquire a second target viewpoint targeted by the second operation information.

In this case, the comparator 150 may synchronize the measurement time, the first target time, and the second target time.

For example, the measurement time may be a current time. The first model may predict a future or future first operational information that has not yet arrived. Likewise, the second model may also predict the future or future second operational information.

Therefore, it is unreasonable to compare the first operation information or the second operation information output at the current time with the measured values measured at the current time.

The measured value measured at the present time should be compared with the value predicted in the past for the operation information at the current time.

Accordingly, the comparator 150 may track the first target time and the second target time, which correspond to a time point when each operation information is a prediction target. In addition, the comparator 150 may extract the first operation information having a first target time that matches the measurement time of the actual value. The comparator 150 may extract second operation information having a second target time point matching the measurement time point of the actual value.

Specifically, the comparison unit 150 may obtain an actual value of the operation information of a specific building through the collection unit 130 .

The comparator 150 may extract a first matching value and a second matching value.

The first matching value may correspond to a first prediction value included in the first operation information predicted for the same time point as the measurement time of the actual value.

The second matching value may correspond to the second predicted value included in the second operation information predicted for the same time point as the measurement time of the actual value.

The prediction unit 190 may output future operation information predicted from the current time point through a model outputting a matching value having a small error with the actual measurement value. In other words, if the error value between the measured value and the first matching value is less than the error value between the measured value and the second matching value, the prediction unit 190 performs the first operation of the future predicted through the first model from the current time point onwards. information can be printed. If the error value between the measured value and the second matching value is less than the error value between the measured value and the first matching value, the prediction unit 190 may generate second operation information for the future predicted through the second model from the current time point onwards. can be printed out.

As described above, the second model initially has lower prediction accuracy than the first model, but the prediction accuracy may be improved as time goes by. Accordingly, the prediction accuracy of the second model may exceed that of the first model after a certain point in time. In this case, a time point at which the prediction accuracy of the second model applies for the first model may be defined as a crossover time tx.

2 is a schematic diagram illustrating prediction accuracy of a first model and a second model.

The second model m2 may have lower prediction accuracy than the first model m1 before the crossing time tx, and may have higher prediction accuracy than the first model m1 after the crossing time tx. In this case, the prediction unit 190 may output the first operation information from among the first operation information and the second operation information before the crossing time tx.

The prediction unit 190 may output second operation information from among the first operation information and the second operation information after the crossing time tx. In this case, it may be important to determine the intersection point tx in the relationship between the first model m1 and the second model m2. Due to the characteristics of the first model or the second model for outputting information predicted as a target at a point in the future rather than a specific point in time at a specific point in time, the crossing point tx may correspond to a point in the past with respect to the current point in time.

3 is a schematic diagram illustrating a selection point of the second model.

Theoretically, it may be advantageous for the second model to be selected directly at the intersection point tx. However, the crossing point tx can be determined only after a certain time has elapsed.

For example, it is assumed that the accuracy of the second operation information predicted by the second model is higher than the first operation information predicted by the first model after the crossing point tx.

At the intersection time tx, the first operational information may be predicted from the first model, and the second operational information may be predicted from the second model. In this case, the first operation information and the second operation information may target the first time point t1 at which a predetermined time has elapsed based on the crossing time point tx. The first operation information and the second operation information predicted with the first time point t1 as a target should be compared with the measured values measured at the first time point t1.

Accordingly, the prediction unit 190 cannot know that the prediction accuracy of the second model is higher than the prediction accuracy of the first model until the first time point t1 that can be compared with the actual value.

As a result, even if the prediction performance of the second model is better than that of the first model from the intersection time tx, the prediction unit 190 may output the first operation information of the first model until the first time point t1. It is only after the first time t1 is reached and comparison with the actual value is performed, the prediction unit 190 may output the second operation information of the second model among the first operation information and the second operation information.

According to the above description, the comparator 150 may detect the crossing point tx that has already passed from the current time point. The comparator 150 may detect a past intersection time tx by comparing the measured values measured for a specific building with each operation information.

The prediction unit 190 may determine whether to output the second operation information instead of the first operation information according to the detection result of the crossing time tx.

For example, when the past crossing time tx is detected at a specific time point, the prediction unit 190 may output second operation information from the specific time point at which the crossing time tx is detected.

In summary, whenever a measured value is obtained, the comparator 150 may determine a measurement time of the obtained measured value.

The comparator 150 may extract a first matching value corresponding to the first prediction value included in the first operation information generated at a time in the past than the measurement time and predicted with the measurement time as a target, from among the plurality of first operation information. have.

The comparator 150 may extract a second matching value corresponding to a second prediction value included in the second operation information generated at a past time point and predicted with the measurement time point as a target, from among the plurality of second operation information.

Whenever an actual measured value is obtained, the comparator 150 may calculate a first error value and a second error value by using the obtained measured value. The first error value may include an error value between the measured value and the first matching value. The second error value may include an error value between the measured value and the second matching value.

The comparator 150 maintains a state in which the second error value is greater than the first error value, and when the second error value is smaller than the first error value due to the acquisition of a specific measured value, the measurement point of the specific measured value is set as a target. The generation time of the predicted first matching value or the second matching value may be detected as the crossing time tx.

The second model discussed above may include a model having lower prediction accuracy than the first model initially, but having higher prediction accuracy than the first model as time goes by. In this case, the comparator 150 may check the prediction accuracy by comparing the measured values measured for a specific building with each operation information.

The prediction unit 190 may initially output the first operation information.

The prediction unit 190 repeats the setting number of times successively when the prediction accuracy of the second model is higher than the prediction accuracy of the first model through the comparison unit 150 (eg, n times (here, n is a natural number of 2 or more) )), the second operation information may be output instead of the first operation information after the setting number of times is repeated.

According to the present embodiment, it is possible to prevent a phenomenon in which the second operation information is output instead of the first operation information irrespective of prediction accuracy due to various noises or errors in measured values.

The above providing device 100 selectively outputs one of the first predicted value of the first model m1 and the second predicted value of the second model m2 according to the error value compared with the measured value of the building operation information hybrid model ms can provide

The hybrid model ms may predict at least one of indoor environment information, energy usage, and occupancy information, or may diagnose a malfunction. These operation information may be provided to the management server 90 in charge of energy demand management, facility maintenance management, and indoor environment management through the prediction unit 190 .

6 is a flowchart illustrating a method of providing the present invention.

The providing method of FIG. 6 may be performed by the providing apparatus 100 illustrated in FIG. 6 .

The providing method may include an obtaining step (S510), a collecting step (S520), a comparison step (S530), a selection step (S540), and a prediction step (S550).

In the acquiring step ( S510 ), first operation information of the specific building predicted through the first model may be acquired, and second operation information of the specific building predicted through the second model may be acquired. The obtaining step ( S510 ) may be performed by the first obtaining unit 110 and the second obtaining unit 120 .

In the collecting step ( S520 ), actual measured values obtained by measuring operation information of a specific building may be collected. The collecting step ( S520 ) may be performed by the collecting unit 130 .

The comparison step ( S530 ) may compare the first predicted value and the actual value included in the first operation information, and compare the second predicted value and the actual value included in the second operation information. The comparison step ( S530 ) may be performed by the comparison unit 150 .

In the selecting step ( S540 ), a model outputting a predicted value having a small error between the measured value and the first model may be selected from among the first model and the second model. The selection step ( S540 ) may be performed by the selection unit 170 .

In the prediction step (S550), the operation information predicted through the model selected by the selection step (S540) from among the first operation information and the second operation government may be output. The prediction step S550 may be performed by the prediction unit 190 .

7 is a diagram illustrating a computing device according to an embodiment of the present invention. The computing device TN100 of FIG. 7 may be a device (eg, the providing device 100 ) described herein.

In the embodiment of FIG. 7 , the computing device TN100 may include at least one processor TN110 , a transceiver device TN120 , and a memory TN130 . Also, the computing device TN100 may further include a storage device TN140 , an input interface device TN150 , an output interface device TN160 , and the like. Components included in the computing device TN100 may be connected by a bus TN170 to communicate with each other.

The processor TN110 may execute a program command stored in at least one of the memory TN130 and the storage device TN140. The processor TN110 may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which methods according to an embodiment of the present invention are performed. The processor TN110 may be configured to implement procedures, functions, and methods described in connection with an embodiment of the present invention. The processor TN110 may control each component of the computing device TN100 .

Each of the memory TN130 and the storage device TN140 may store various information related to the operation of the processor TN110 . Each of the memory TN130 and the storage device TN140 may be configured as at least one of a volatile storage medium and a nonvolatile storage medium. For example, the memory TN130 may include at least one of a read only memory (ROM) and a random access memory (RAM).

The transceiver TN120 may transmit or receive a wired signal or a wireless signal. The transceiver TN120 may be connected to a network to perform communication.

On the other hand, the embodiment of the present invention is not implemented only through the apparatus and/or method described so far, and a program for realizing a function corresponding to the configuration of the embodiment of the present invention or a recording medium in which the program is recorded may be implemented. And, such an implementation can be easily implemented by those skilled in the art from the description of the above-described embodiments.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept of the present invention as defined in the following claims are also present. It belongs to the scope of the invention.

90...management server 100...provider device
110...First Acquisition Unit 120...Second Acquisition Unit
130...Collection Division 150...Comparison Division
170...selection unit 190...prediction unit

Claims (10)

a first acquisition unit for acquiring first operation information of a specific building predicted through a first model;
a second acquisition unit for acquiring second operation information of the specific building predicted through a second model;
a prediction unit for selectively outputting one of the first operation information predicted through the first model and the second operation information predicted through the second model;
The second model includes a model in which the prediction accuracy is lower than that of the first model initially, but the prediction accuracy is improved than that of the first model as time goes by,
When the prediction accuracy of the second model overtakes the first model is defined as a crossover time point,
The second model has lower prediction accuracy than the first model before the crossover time, and has higher predictive accuracy than the first model after the crossover time,
The prediction unit outputs the first operation information predicted through the first model among the first operation information and the second operation information before the intersection time with respect to the intersection time,
The prediction unit outputs the second operation information predicted through the second model among the first operation information and the second operation information after the intersection time based on the intersection time,
A comparison unit is provided that compares the first predicted value included in the first operation information and the measured value directly measured for the specific building, and compares the second predicted value included in the second operation information with the actual value,
The comparison unit detects the crossing point that has already passed through the comparison of the measured value and each operation information at the current time point,
Each time the measured value is obtained, the comparator identifies a measurement time of the obtained measured value,
The comparison unit extracts a first matching value corresponding to a first prediction value included in the first operation information generated at a time in the past than the measurement time and predicted with the measurement time as a target, from among the plurality of first operation information,
The comparison unit extracts a second matching value corresponding to a second prediction value included in the second operation information generated at the past time point and predicted with the measurement time point as a target, from among a plurality of second operation information,
Each time the measured value is obtained, the comparator calculates a first error value and a second error value using the obtained measured value,
The first error value includes an error value between the measured value and the first matching value,
The second error value includes an error value between the measured value and the second matching value,
The comparator maintains a state in which the second error value is greater than the first error value, and when the second error value is smaller than the first error value due to the acquisition of a specific measured value, the measurement time of the specific measured value is determined. Detecting the generation time of the first matching value or the second matching value predicted as the target as the crossing time,
The comparison unit confirms the prediction accuracy by comparing the measured value with each operation information,
When a situation in which the prediction accuracy of the second model is higher than the prediction accuracy of the first model through the comparison unit is repeated a set number of times continuously, the prediction unit outputs the second operation information instead of the first operation information thereafter providing device.
delete delete delete delete delete delete delete delete A providing method performed by a providing device, comprising:
an acquisition step of acquiring first operation information of a specific building predicted through a first model, and acquiring second operating information of the specific building predicted through a second model;
a collecting step of collecting actual measured values of the operation information of the specific building;
a comparison step of comparing a first predicted value included in the first operation information with the actual value, and comparing the measured value with a second predicted value included in the second operation information;
a selection step of selecting a model outputting a predicted value having a small error between the measured value and the first model and the second model;
A prediction step of outputting the operation information predicted through the model selected in the selection step from among the first operation information and the second operation information;
The second model includes a model in which the prediction accuracy is initially lower than that of the first model, but the prediction accuracy is improved as time goes by,
When the prediction accuracy of the second model overtakes the first model is defined as a crossover time point,
The second model has lower prediction accuracy than the first model before the crossover time, and has higher predictive accuracy than the first model after the crossover time,
The prediction step is
Before the crossing point, the first operation information is output from among the first operation information and the second operation information,
After the crossing point, the second operation information is outputted from among the first operation information and the second operation information,
The comparison step is
By comparing the measured value with each operation information, the crossing point that has already passed is detected at the current time point,
Whenever the measured value is obtained, the measurement time of the obtained measured value is identified,
extracting a first matching value corresponding to a first prediction value generated at a time in the past from the measurement time from among the plurality of first operation information and included in the first operation information predicted with the measurement time as a target,
extracting a second matching value corresponding to a second prediction value included in the second operation information generated at the past time point and predicted by the measurement time point from among the plurality of second operation information,
Whenever the measured value is obtained, a first error value and a second error value are calculated using the obtained measured value,
The first error value includes an error value between the measured value and the first matching value,
The second error value includes an error value between the measured value and the second matching value,
The comparison step is
When the second error value maintains a state greater than the first error value and the second error value becomes smaller than the first error value due to the acquisition of a specific measured value, the measurement time of the specific measured value is predicted as a target Detecting the generation time of the first matching value or the second matching value as the crossing time,
Check the prediction accuracy by comparing the measured value with each operation information,
The prediction step is
When a situation in which the prediction accuracy of the second model is higher than the prediction accuracy of the first model through the comparison step is repeated a set number of times consecutively, thereafter, the second operation information is output instead of the first operation information Way.

Images