Prediction system, temperature control system, prediction method and program

Abstract

This invention provides a technology for accurately predicting pre-cooling time when other operations are intervening during cooling or heating operations. [Solution] The prediction system is a prediction system that predicts the pre-cooling time until the target temperature of a target space is achieved by temperature control operation when other operating conditions intervene during temperature control operation of the target space, and comprises an acquisition unit that acquires explanatory variables which are factors that affect temperature control of the target space, a model that shows the relationship between the explanatory variables and the time constant of dead time and the first-order lag based on time series data of the temperature of the target space from the start of temperature control operation when other operating conditions intervene in the middle until the temperature of the target space is controlled to the target temperature, and a prediction unit that calculates the dead time and the first-order lag time based on the acquired explanatory variables and predicts the pre-cooling time by multiplying the time constant by a coefficient and adding the dead time.

Description

Technical Field

[0004] , , ,

[0001] The present disclosure relates to a prediction system, a temperature control system, a prediction method, and a program.

Background Art

[0002] When cooling a room, it takes time for the room temperature to reach the target temperature after starting the cooling operation. A function is provided to start the cooling operation before the user's presence time so that the room becomes comfortable when the user is present. Hereinafter, not limited to cooling and refrigeration, including heating and warming, a function of starting the operation in advance so as to reach the target temperature at the target time is called precooling, a precooling function, etc. Also, the time required to achieve the target temperature by precooling is called the precooling time. The appropriate precooling time varies depending on the size of the room, the outside air temperature, etc. Therefore, when the time to start precooling is fixed, the precooling time may be insufficient and the target temperature may not be achieved by the target time, or the precooling time may be excessive and power consumption may be wasted. Patent Document 1 discloses a technique for predicting the indoor temperature Tc(t + ts) in the next control step by a heat characteristic model of a first-order lag system using the heat quantity Qm(t) supplied to the room, the time constant T, the dead time L, and the gain K. By using such a room temperature prediction technique and calculating how much time is required for precooling and starting precooling before the calculated time, it is considered possible to solve the above problems.

[0003] By the way, when performing operations such as cooling, refrigeration, heating, or warming with an air conditioner or a refrigerator, during those operations, there may be intervening operating states different from the purpose, such as a defrost operation or an operation stop when an engine-driven refrigerator mounted on a refrigerated vehicle stops the engine. In such a case, if the room temperature change is predicted and the precooling time is set without considering the influence of the intervening defrost operation, etc., there is a possibility of a decrease in comfort due to insufficient precooling time.

Prior Art Documents

Patent Documents

[0004] [Patent Document 1] International Publication No. 2023 / 149259 [Overview of the project] [Problems that the invention aims to solve]

[0005] When other operations are intervening during cooling or heating operations, there is a need for technology that can accurately predict the pre-cooling time.

[0006] This disclosure provides a prediction system, a temperature control system, a prediction method, and a program that can solve the above-mentioned problems. [Means for solving the problem]

[0007] The prediction system of this disclosure is a prediction system that predicts the pre-cooling time, which is the time required to achieve a target temperature of a target space by a temperature control operation, when another operating state is intervened during a temperature control operation to cool or heat a target space for temperature control, and comprises: an acquisition unit that acquires explanatory variables which are factors that affect the temperature control of the target space; a model that shows the relationship between the explanatory variables and the dead time and first-order lag time constant based on time-series data of the temperature of the target space from the start of the temperature control operation when the other operating state is intervened in the middle until the temperature of the target space is controlled to the target temperature by the temperature control operation; and a prediction unit that calculates the dead time and the first-order lag time constant based on the acquired explanatory variables, and predicts the pre-cooling time by multiplying the time constant by a predetermined coefficient and adding the dead time.

[0008] Furthermore, the temperature control system of the present disclosure comprises the above-mentioned prediction system, a temperature control device for controlling the temperature of the target space, and a control device for the temperature control device. The control device, upon receiving the setting of a target temperature for the target space and a target time for achieving the target temperature, instructs the prediction system to predict the pre-cooling time, receives the pre-cooling time predicted by the prediction system, and causes the temperature control device to start temperature control by the amount of the pre-cooling time before the target time.

[0009] Furthermore, the prediction method of this disclosure is a prediction method for predicting a pre-cooling time, which is the time required to achieve a target temperature of a target space by a temperature control operation, when another operating state is intervened during a temperature control operation to cool or heat a target space for temperature control, and comprises the steps of: a computer acquiring explanatory variables which are factors that affect the temperature control of the target space; the computer calculating the dead time and the first-order lag time constant based on the explanatory variables, a model which shows the relationship between the dead time and the first-order lag time constant based on time-series data of the temperature of the target space from the start of the temperature control operation when another operating state is intervened in the middle of the operation, and the acquired explanatory variables, and predicting the pre-cooling time by multiplying the time constant by a predetermined coefficient and adding the dead time.

[0010] Furthermore, the program of this disclosure causes a computer to perform a process for predicting a pre-cooling time, which indicates the time required to achieve a target temperature of a target space by a temperature control operation, when another operating state is intervened during a temperature control operation to cool or heat a target space for temperature control, the process comprising: obtaining explanatory variables which are factors that affect the temperature control of the target space; a model which shows the relationship between the explanatory variables, a dead time and a first-order lag time constant based on time-series data of the temperature of the target space from the start of the temperature control operation when another operating state is intervened in the middle of the operation; and the obtained explanatory variables, calculating the dead time and the first-order lag time constant, and predicting the pre-cooling time by multiplying the time constant by a predetermined coefficient and adding the dead time. [Effects of the Invention]

[0011] According to the prediction system and method described above, pre-cooling time can be accurately predicted when other operations are intervening during cooling or heating operations. [Brief explanation of the drawing]

[0012] [Figure 1] This figure shows an example of a control system according to the embodiment. [Figure 2] This figure shows an example of time-series data of the internal temperature according to the embodiment. [Figure 3A] This figure illustrates a first method for estimating the pre-cooling time according to an embodiment. [Figure 3B] This figure illustrates a second method for estimating the pre-cooling time according to the embodiment. [Figure 3C] This figure illustrates a third method for estimating the pre-cooling time according to the embodiment. [Figure 4A] Figure 1 shows an example of a parameter calculation model according to the embodiment. [Figure 4B] Figure 2 shows an example of a parameter calculation model according to the embodiment. [Figure 4C]FIG. 3 showing an example of the parameter calculation model according to the embodiment. [Figure 4D] FIG. 4 showing an example of the parameter calculation model according to the embodiment. [Figure 5] A flowchart showing an example of the creation process of the parameter calculation model according to the embodiment. [Figure 6] A flowchart showing an example of the prediction process of the pre-cooling time according to the embodiment. [Figure 7] A diagram showing an example of the hardware configuration of the control system according to the embodiment.

MODE FOR CARRYING OUT THE INVENTION

[0013] <Embodiment> Hereinafter, a method for predicting the pre-cooling time when other operating states intervene during the cooling operation or the heating operation will be described with reference to the drawings. (System Configuration) FIG. 1 is a diagram showing an example of a control system according to the embodiment. The control system 100 is a system that controls the temperature inside the cold storage 2 provided in the refrigerated vehicle 1 to a desired temperature by the refrigerator 3. The refrigerator 3 cools or heats the temperature inside the cold storage 2 to a predetermined target temperature. The refrigerator 3 has a pre-cooling function. The control system 100 predicts the pre-cooling time when other operations such as a defrost operation are performed during the cooling operation or the heating operation of the refrigerator 3, and starts the pre-cooling operation at a time retrogressed by the pre-cooling time from the target time. Hereinafter, the case where the other operation occurring during the cooling operation or the heating operation is a defrost operation will be described as an example, but the other operating states intervening during the cooling operation or the heating operation may be operating states other than the defrost operation (for example, an operating state in which the refrigerator 3 is driven by the engine of the refrigerated vehicle 1 and the cooling or heating operation is stopped when the refrigerated vehicle 1 stops the engine). Further, the processing described below is not limited to the refrigerator 3 of the refrigerated vehicle 1, and can also be applied to the pre-cooling operation by an air conditioner that performs air conditioning for general houses, stores, offices, etc.

[0014] The control system 100 includes a control device 10 of the refrigerator 3 and a server 20. The control device 10 and the server 20 are communicably connected by a network NW. The refrigerated vehicle 1 includes an outside air temperature sensor 4 that measures the outside air temperature outside the cold storage 2, an inside temperature sensor 5 that measures the inside temperature of the cold storage 2, and a rotation speed sensor 6 that measures the rotation speed of the engine of the refrigerated vehicle 1.

[0015] (Configuration of the control device) The control device 10 includes a sensor value acquisition unit 11, a prediction unit 12, a control unit 13, a storage unit 14, and a communication unit 15.

[0016] The sensor value acquisition unit 11 acquires the outside air temperature measured by the outside air temperature sensor 4, the inside temperature measured by the inside temperature sensor 5, and the rotation speed of the engine measured by the rotation speed sensor 6.

[0017] The prediction unit 12 predicts the pre-cooling time of the cooling operation or the heating operation in consideration of the influence of the defrost operation that occurs during operation. For example, when cooling is performed for pre-cooling, if a defrost operation occurs during the process, the inside temperature will rise during that time, and when the cooling operation is restarted after the end of the defrost operation, the increased amount also needs to be cooled. In this case, the pre-cooling time becomes longer compared to the case where there is no defrost operation in between. The prediction unit .....

[0018] The control unit 13 controls the refrigerator 3 to cool or heat the cold storage 2. Also, the control unit 13 receives a pre-cooling setting from the user. When receiving the pre-cooling setting, the control unit 13 predicts the pre-cooling time using the prediction unit 12 and starts the cooling operation or the heating operation pre-cooling time before the target time.

[0019] The memory unit 14 stores various information necessary for predicting the pre-cooling time. For example, the memory unit 14 stores the internal temperature, external temperature, engine speed, and learned models and lookup tables (referred to as parameter calculation models) used to calculate thermal response parameters used for predicting the pre-cooling time, which are acquired by the sensor value acquisition unit 11. The parameter calculation models are created on the server 20.

[0020] The communication unit 15 communicates data with the server 20. For example, the communication unit 15 transmits the internal temperature, outside temperature, engine speed, and operating mode of the refrigerator 3, acquired by the sensor value acquisition unit 11, to the server 20 at predetermined intervals. The operating mode represents the operating state of the refrigerator 3 and can be one of the following: cooling operation, heating operation, defrost operation, thermo-off operation, or stop. Thermo-off operation is an operating state in which only air blowing occurs after the target temperature has been reached. Thermo-off operation continues until the internal temperature deviates from the target temperature by a predetermined temperature and cooling or heating operation is resumed. The communication unit 15 also receives a parameter calculation model from the server 20.

[0021] (Server configuration) The server 20 includes a performance data acquisition unit 21, an extraction unit 22, a division unit 23, an estimation unit 24, a learning unit 25, a storage unit 26, and a communication unit 27.

[0022] The performance data acquisition unit 21 acquires time-series data of the internal temperature of the refrigerator 2 when the refrigerator 3 is performing a cooling or heating operation, time-series data of the outside temperature, time-series data of the engine speed of the refrigerated vehicle 1, time-series data of the operating mode, the target temperature of the refrigerator 2, and other data.

[0023] The extraction unit 22 extracts data on the internal temperature from the time-series data of the internal temperature, specifically the data for the period from the start to the end of the cooling or heating operation (this period is referred to as the effective period) when a defrost operation is performed in the middle of the cooling or heating operation.

[0024] The division unit 23 divides the time-series data extracted by the extraction unit 22 into two parts: the time period during which cooling or heating operation is performed, and the time period during which defrosting operation is performed. For the time period during which cooling operation is performed, the division unit 23 divides the period from the point in which the internal temperature begins to decrease into the first-order lag response portion, and the period other than the first-order lag response into the dead time portion. For the time period during which heating operation is performed, the division unit 23 divides the period from the point in which the internal temperature begins to rise into the first-order lag response portion, and the period other than the first-order lag response into the dead time portion.

[0025] The estimation unit 24 estimates the length of the dead time portion as the dead time L, and also estimates a time constant τ that minimizes the error between the internal temperature in the first-order lag response portion and an approximation curve showing the change in internal temperature when the internal temperature is simulated while changing the time constant of the internal temperature change. The estimation unit 24 also estimates the length of time μ during which the defrost operation is performed. The first-order lag time constant τ, dead time L, and the time μ during which the defrost operation is performed are also referred to as thermal response parameters.

[0026] The learning unit 25 creates a parameter calculation model (a trained model constructed using machine learning or a lookup table, etc.) that associates explanatory variables, which are factors that affect the control of the internal temperature of the storage chamber, with the dead time L and the time constant τ. Alternatively, the learning unit 25 creates a parameter calculation model that associates explanatory variables with the dead time L, the time constant τ and the defrost operation time μ.

[0027] The memory unit 26 stores time-series data acquired by the performance data acquisition unit 21 and parameter calculation models created by the learning unit 25.

[0028] The communication unit 27 communicates data with the control device 10. For example, the communication unit 27 receives time-series data such as the internal temperature from the control device 10 and transmits a parameter calculation model to the control device 10.

[0029] Figure 2 shows an example of time-series data of the internal temperature during cooling operation. Based on the operating mode transmitted from the control device 10, the extraction unit 22 identifies the start time ts and end time te of the cooling operation in the time-series data of the internal temperature transmitted from the control device 10, and extracts the time-series data for this period. The end time te is, for example, the time when the target temperature is achieved, or the time when the operation changes from cooling operation to thermo-off operation. The period from start time ts to end time te is considered the valid interval. The data in Figure 2 is the time-series data after extraction. Based on the operating mode transmitted from the control device 10, the division unit 23 identifies the start time t0 of the defrost operation in the time-series data after extraction. The division unit 23 also identifies the time t1 when the internal temperature returns to the same temperature as the temperature at time t0 after the defrost operation (the time when the temperature first becomes below the temperature at time t0 after the defrost operation). The division unit 23 divides the effective period into a first cooling operation period (ts~t0), a defrosting operation period (t0~t1), and a second cooling operation period (t1~te). (In reality, it is thought that the cooling operation resumes when the internal temperature drops in the latter half of the defrosting operation period shown in the diagram, but for convenience it is referred to as the "defrosting operation period.") Furthermore, the division unit 23 divides the cooling operation period into a first-order lag response portion and a dead time portion. Taking the first cooling operation period (ts~t0) as an example, the division unit 23 divides it into a dead time portion (ts~ta) and a first-order lag response portion (ta~t0).

[0030] The estimation unit 24 estimates the time constant τ of the first-order lag and the dead time L, or the time constant τ of the first-order lag, the dead time L, and the defrosting time μ. In this embodiment, three estimation methods are provided.

[0031] (Method 1 for estimating τ and L) Figure 3A is a diagram illustrating a first method for estimating the pre-cooling time according to an embodiment. The estimation unit 24 estimates the dead time L and the first-order lag time constant τ before and after the defrost operation. For example, the dead time L and the like are estimated by the following process. (1-1) The divided section 23 determines the start time t0 of the defrost operation based on the operating mode. (1-2) The division section 23 divides the effective period into before and after time t0. In Figure 3A, D1 is the time series data before the start of defrost operation, and D2 is the time series data after the start of defrost operation. (1-3) The splitting unit 23 estimates the dead time L1 and L2 for each of the time series data D1 and D2 after splitting. The splitting unit 23 identifies the times ta and tb when the internal temperature of the storage unit begins to decline. The estimation unit 24 estimates the dead time L1 from the start time ts of the time series data D1 to time ta, and estimates the dead time L2 from the start time t0 of the time series data D2 to time tb. More specifically, the splitting unit 23 searches for an extreme value (maximum value) from the time series data D1, and determines that the internal temperature of the storage unit has begun to decline when it has fallen below a predetermined temperature (for example, 1°C) from the extreme value. At this time, the splitting unit 23 considers the extreme value immediately before the decline as a boundary point, and the elements after the boundary point as the first-order lag response portion. In the example in Figure 3A, after time ta, the internal temperature of the storage unit has fallen below the maximum value by a predetermined temperature or more. Therefore, the division unit 23 uses the element at time ta, just before the internal temperature begins to decline, as a boundary point, considering the period from time ts to time ta as the dead time portion, and the period from time ta to time t0 as the first-order lag response portion. The same applies to time series data D2. The division unit 23 detects that the internal temperature has fallen by a predetermined temperature or more from the maximum value in time series data D2 after time tb, considering the period from time t0 to time tb as the dead time portion, and considering the period from time tb to time te as the first-order lag response portion. The estimation unit 24 then estimates the length of the time that the division unit 23 considered as the dead time portion for time series data D1 and D2 as dead time L1 and L2, respectively.

[0032] For the sake of clarity, let me explain the case of heating operation. During heating operation, the division unit 23 detects extreme values ​​(local minimums) from the time-series data. When the temperature rises by a predetermined amount (for example, 1°C) or more from the extreme value, it determines that the room temperature has started to rise. The extreme value immediately before the temperature starts to rise is considered the boundary point, the period after the boundary point is considered the first-order lag response portion, and the period up to the extreme point is considered the dead time portion.

[0033] (1-4) The estimation unit 24 estimates the first-order time constants τ1 and τ2 for each of the time-series data D1 and D2 after division. The estimation unit 24 estimates the first-order response time constant τ based on the curve of the first-order response portion identified in (1-3). For example, the estimation unit 24 simulates the time change of the internal temperature for each assumed value of the time constant, sequentially changing it from a lower limit to an upper limit in increments of a fixed time (e.g., 1 minute). Various known techniques may be used for this simulation. For example, the time change of the internal temperature can be represented by a differential equation of a first-order system, this differential equation can be approximated by a difference equation, and the internal temperature can be simulated using the approximated difference equation. Then, while varying the assumed value of the time constant representing the magnitude of the temperature change, the estimation unit 24 fits the curve showing the change in internal temperature indicated by the difference equation to the measured value (the first-order response portion divided by the division unit 23), and calculates an approximate curve of the measured value. The estimation unit 24 calculates the error between the approximation curve for each assumed value of the time constant and the actual value, and sets the assumed value of the time constant that minimizes the error as the time constant τ. The estimation unit 24 performs this process for each of the time series data D1 and D2, estimating the first-order lag time constant τ1 of the time series data D1 and the first-order lag time constant τ2 of the time series data D2.

[0034] (1-5) Using the estimated values ​​of the dead time L1 and L2 and the first-order lag time constants τ1 and τ2, the pre-cooling time D can be calculated by the following equations (1) to (3). D = D1 + D2 ... (1) D1 = L1 + c1·τ1···(2) D2 = L2 + c2·τ2···(3) c1 and c2 are coefficients of the time constant. For example, if the target temperature is 95% of the set temperature, c1 and c2 can be set to 3, and if the target temperature is 90% of the set temperature, c1 and c2 can be set to 2.3, etc. c1 and c2 may have different values. As will be described later, L1, L2, τ1, and τ2 estimated by the estimation unit 24 become training data for training the parameter calculation model.

[0035] (Method 2 for estimating τ and L) Figure 3B illustrates a second method for estimating the pre-cooling time according to the embodiment. The estimation unit 24 estimates the dead time L and the first-order lag time constant τ after interpolating the defrosting operation time period with other data. Interpolation methods include missing value interpolation and linear interpolation.

[0036] Refer to Figure 3B(a) to explain the process for interpolating missing values. (2a-1) The segmented section 23 identifies the time t0 immediately before the defrost operation begins. (2a-2) The divided section 23 identifies the time t1 (the time when the temperature inside the chamber first becomes below the temperature at time t0) after the defrosting operation. (2a-3) The estimation unit 24 interpolates the internal temperature between time t0 and time t1 with information that represents missing or blank values, such as NULL values. (2a-4) The estimation unit 24 estimates the dead time L and the time constant τ using the interpolated data. The method for estimating the dead time L and the time constant τ is the same as in (1-3) and (1-4) of "Method 1 for τ and L". The estimation unit 24 tentatively sets the time constant τ while the period between t0 and t1 is missing, and estimates the time constant τ by simulating the internal temperature and calculating an approximation curve. (2a-5) Using the estimated dead time L and the time constant τ of the first-order lag, the pre-cooling time D can be calculated by the following equation, where c is the coefficient of the time constant. D = L + c·τ ···(4a)

[0037] Refer to Figure 3B(b) to explain the process when performing linear value interpolation. (2b-1) The segmented section 23 identifies the time t0 immediately before the defrost operation begins. (2b-2) The divided section 23 identifies the time t1 (the time when the temperature inside the chamber first becomes below the temperature at time t0) after the defrosting operation. (2a-3) The estimation unit 24 connects time t0 and time t1 with a straight line and interpolates the data in between. (2a-4) The estimation unit 24 estimates the dead time L and the time constant τ using the interpolated data. The method for estimating the dead time L and the time constant τ is the same as in (1-3) and (1-4) of "Method 1 for estimating τ and L". The estimation unit 24 tentatively sets the time constant τ based on a curve drawn by a straight line connecting t0 and t1, and then estimates the time constant τ by simulating the internal temperature and calculating an approximate curve. (2a-5) Using the estimated dead time L and the time constant τ of the first-order lag, the pre-cooling time D can be calculated by the following equation, where c is the coefficient of the time constant. D = L + c·τ ···(4b) In the second method for estimating τ and L, it becomes possible to estimate the time constant τ without considering the temperature changes during defrosting.

[0038] (Method 3 for estimating τ and L) Figure 3C illustrates a third method for estimating the pre-cooling time according to the embodiment. The estimation unit 24 sets the time period for defrosting operation and the associated recooling separately as μ, and estimates the dead time L and the first-order lag time constant τ for the time periods other than μ. (3-1) The divided section 23 identifies the time t0 immediately before the defrost operation begins. (3-2) The divided section 23 identifies the time t1 (the time when the temperature inside the chamber first becomes below the temperature at time t0) after the defrosting operation. (3-3) The estimation unit 24 combines the time series data before time t0 and the time series data after time t1. (These two time series data are combined at time t0 and time t1.) (3-4) The estimation unit 24 estimates the dead time L and the time constant τ from the combined data. The method for estimating the dead time L and the time constant τ is the same as in (1-3) and (1-4) of "Method for Estimating τ and L 1". (3-5) The estimation unit 24 takes the time from t0+1 to t1-1 as μ. (3-6) Using the estimated dead time L and the first-order lag time constant τ and μ, the pre-cooling time D can be calculated by the following equation, where c is the coefficient of the time constant. D = L + c·τ + μ ···(5) In the third method for estimating τ and L, separating defrosting operation time from other times makes it easier to interpret the data for each operation.

[0039] The estimation unit 24 estimates L, τ, μ, etc., using one of the estimation methods 1 to 3 described above, and stores the estimated values ​​in the storage unit 26, associating them with explanatory variables that affect the control of the internal temperature of the storage unit. Explanatory variables include, for example, the outside temperature at the start of pre-cooling, the internal temperature at the start of pre-cooling, the target temperature of the storage unit 2, the engine speed (for example, an estimated average rotation speed during pre-cooling), and the operating mode (cooling or heating). Note that the explanatory variables exemplified here are just examples and are not limited to these. For example, if the refrigeration unit 3 is powered by a battery mounted on the refrigeration vehicle 1, the engine speed may be excluded from the explanatory variables. Humidity and weather (sunny, rainy, snowy, cloudy), etc., may also be added as explanatory variables.

[0040] The learning unit 25 learns the relationships between the explanatory variables stored in the memory unit 26 by the estimation unit 24 and the thermal response parameters L, τ, etc., and creates a parameter calculation model that can calculate the thermal response parameters from the explanatory variables. Figures 4A to 4C show examples of parameter calculation models.

[0041] The parameter calculation model in Figure 4A is an example of a parameter calculation model when the dead time L1 and L2 and the first-order lag time constants τ1 and τ2, estimated by estimation method 1 for τ and L, are stored in the memory unit 26. The learning unit 25 uses the explanatory variables and L1 and τ1 as training data set 1, and the explanatory variables and L2 and τ2 as training data set 2, and uses all the data from set 1 and set 2 as training data to learn the relationship between the explanatory variables and L and τ to create the parameter calculation model in Figure 4A. For example, the ambient temperature and internal temperature at the start of pre-cooling time prediction, the target temperature of the refrigerator 2, the engine speed during pre-cooling, and the operating mode (cooling or heating) are input to the parameter calculation model in Figure 4A. Then, the parameter calculation model outputs the dead time L1 and the first-order lag time constant τ1. Furthermore, the parameter calculation model in Figure 4A is input with the following values: the ambient temperature at the start of the pre-cooling time prediction, the predicted internal temperature at the start of defrost operation (t0) (for example, a predetermined fixed value such as the average internal temperature at the start of defrost operation based on past performance), the target temperature of the refrigerator 2, the engine speed during pre-cooling, and the operating mode (cooling or heating). The parameter calculation model then outputs the dead time L2 and the first-order lag time constant τ2. By substituting the output dead time L1, L2, and first-order lag time constants τ1 and τ2 into equations (1) to (3) above, the pre-cooling time D can be calculated (predicted). In this method, compared to estimation method 1 for τ and L, the number of intervals obtained doubles, so the amount of training data increases, and an improvement in prediction accuracy can be expected.

[0042] Furthermore, if the dead time L1 and L2 and the first-order lag time constants τ1 and τ2 estimated by estimation method 1 for τ and L are stored in the memory unit 26, the learning unit 25 may learn the relationship between the explanatory variables and L1, L2, τ1, and τ2 to create a parameter calculation model. For example, the ambient temperature at the start of the pre-cooling time prediction, the internal temperature, the target temperature of the refrigerator 2, the engine speed during pre-cooling, and the operating mode (cooling or heating) are input to this parameter calculation model. The parameter calculation model then outputs the dead time L1 and L2 and the first-order lag time constants τ1 and τ2. By substituting the output dead time L1 and L2 and the first-order lag time constants τ1 and τ2 into equations (1) to (3) above, the pre-cooling time D can be calculated (predicted).

[0043] The parameter calculation model in Figure 4B is an example of a parameter calculation model when the dead time L and the first-order lag time constant τ, estimated by estimation method 2 for τ and L, are stored in the memory unit 26. The learning unit 25 learns the relationship between the explanatory variables and L and τ to create the parameter calculation model in Figure 4B. For example, the ambient temperature at the start of the pre-cooling time prediction, the internal temperature, the target temperature of the refrigerator 2, the engine speed during pre-cooling, and the operating mode (cooling or heating) are input to the parameter calculation model in Figure 4B. The parameter calculation model then outputs the dead time L and the first-order lag time constant τ. By substituting the output dead time L and first-order lag time constant τ into equation (4a) or (4b) described above, the pre-cooling time D can be calculated (predicted).

[0044] The parameter calculation model in Figure 4C is an example of a parameter calculation model when the dead time L, the first-order lag time constant τ, and the defrost operation time μ, estimated by estimation method 3 for τ and L, are stored in the memory unit 26. The learning unit 25 learns the relationship between the explanatory variables and L, τ, and μ to create the parameter calculation model in Figure 4C. For example, the ambient temperature at the start of pre-cooling time prediction, the internal temperature, the target temperature of the refrigerator 2, the engine speed during pre-cooling, and the operating mode (cooling or heating) are input to the parameter calculation model in Figure 4C. The parameter calculation model then outputs the dead time L, the first-order lag time constant τ, and the defrost operation time μ. By substituting the output dead time L, first-order lag time constant τ, and defrost operation time μ into equation (5) described above, the pre-cooling time D can be calculated (predicted). In this method, the number of parameters increases, so especially in the control device 10 with limited resources, if the lookup table described next is stored, an increase in memory usage may become a problem.

[0045] Furthermore, the parameter calculation model may be configured as a lookup table. An example of a lookup table is shown in Figure 4D. A lookup table is a table that associates explanatory variables with values ​​of thermal response parameters. Figure 4D shows an example of a lookup table that defines the correspondence between two explanatory variables and the first-order lag time constant τ, as shown in the figure. For example, when both the outside temperature and the inside temperature are 18°C, the value of τ is a1. As in the example in Figure 4D, a range may be set for each explanatory variable, and the system may be configured to look up τ and L for each combination of ranges for each explanatory variable. For example, if the dead time L1, L2 and the first-order lag time constants τ1, τ2 estimated by the estimation method 1 for τ and L are stored in the memory unit 26, the learning unit 25 creates a lookup table associating explanatory variables with dead time L1, a lookup table associating explanatory variables with dead time L2, a lookup table associating explanatory variables with the first-order lag time constant τ1, and a lookup table associating explanatory variables with the first-order lag time constant τ2. The same applies to estimation methods 2 and 3 for τ and L.

[0046] In this embodiment, thermal response parameters are calculated using the parameter calculation model illustrated in Figures 4A to 4D, and the pre-cooling time is predicted using equations (1) to (5). For example, if it is necessary to load cargo onto the refrigerated vehicle 1 at 9:00 AM, the cooling of the cold storage compartment 2 can be started at a time prior to that time, equal to the loading time and pre-cooling time, thereby controlling the temperature of the cold storage compartment 2 to the target temperature by the time of loading. Furthermore, by setting a longer pre-cooling time, fuel consumption for driving the refrigeration unit 3 can be suppressed. As a result, unnecessary pre-cooling is reduced, and effects such as reduced exhaust emissions, energy saving, and suppression of equipment deterioration can be expected.

[0047] (operation) First, let's refer to Figure 5 to explain the process of creating the parameter calculation model. Figure 5 is a flowchart showing an example of the process for creating a parameter calculation model according to the embodiment. As a prerequisite, the storage unit 26 of the server 20 is assumed to store time-series data of the internal temperature of the refrigerator 2, time-series data of the outside temperature, time-series data of the engine speed, time-series data of the operating mode, and the target temperature of the refrigerator 2, all acquired by the performance data acquisition unit 21. Furthermore, it is assumed that the method of estimation of thermal response parameters using one of the estimation methods 1 to 3 for τ and L is predetermined. In Figures 5 and 6 below, the explanation is given using the case where the chiller 3 is performing a cooling operation as an example, but the same applies when the chiller 3 is performing a heating operation.

[0048] The extraction unit 22 reads time-series data of the internal temperature from the storage unit 26 and extracts the valid interval (step S1). For example, based on the time-series data of the operating mode, the extraction unit 22 detects the start time of the cooling operation and the time when the target temperature is first achieved after the start of the cooling operation and the system switches to thermo-off operation. Furthermore, based on the time-series data of the operating mode, the extraction unit 22 checks whether a defrost operation is performed between the start of the cooling operation and the start of the thermo-off operation. If a section is found where the system is in the sequence of cooling operation → defrost operation → thermo-off operation as illustrated in Figure 2, the extraction unit 22 extracts the data for that section as a valid interval and stores the time-series data of the extracted valid interval in the storage unit 26. The storage unit 26 stores time-series data for multiple valid intervals.

[0049] Next, the division unit 23 divides the effective interval (step S2). The division unit 23 reads out multiple effective intervals stored in the storage unit 26 and divides each of them into cooling operation and defrosting operation. Based on the time-series data of the operating mode, the division unit 23 divides the time period from the start of cooling operation (ts) to the start of defrosting operation (t0) and the time period thereafter (t0~te). For the time period thereafter (t0~te), it divides it again from the start of defrosting operation (t0) to the time when the internal temperature returns to the temperature at the start of defrosting operation (t1) and the time period after the return (t1~te). Furthermore, for the time periods ts~t0 and t1~te, it divides each into a dead time portion and a first-order delayed response portion. The division unit 23 stores the start time and end time of each divided time period in the storage unit 26.

[0050] Next, the estimation unit 24 estimates the thermal response parameters (step S3). The estimation unit 24 estimates the thermal response parameters by one of the following methods: The estimation unit 24 estimates the dead times L1 and L2 and the time constants τ1 and τ2 of the first-order lag response as thermal response parameters (method 1 for τ and L estimation). The estimation unit 24 estimates the dead time L and the time constant τ of the first-order lag response as thermal response parameters (method 2 for τ and L estimation). The estimation unit 24 estimates the dead time L, the time constant τ of the first-order lag response, and the defrost operation time μ as thermal response parameters (method 3 for τ and L estimation). The estimation unit 24 stores the estimated thermal response parameters in the storage unit 26.

[0051] Next, the learning unit 25 creates a parameter calculation model (step S4). The learning unit 25 reads from the storage unit 26 a set of thermal response parameters and explanatory variables corresponding to the time period (effective interval) of the time series data from which the thermal response parameters originated. For example, when reading the thermal response parameters for an effective interval A, the learning unit 25 reads from the storage unit 26 the thermal response parameters for effective interval A, along with the target temperature for the cooling operation related to effective interval A, the internal temperature at the start of the cooling operation, the ambient temperature, and the average engine speed in effective interval A. The learning unit 25 learns the relationship between the read explanatory variables and the thermal response parameters using machine learning and creates a parameter calculation model as illustrated in Figures 4A to 4D. The learning unit 25 saves the created parameter calculation model in the storage unit 26 and transmits the parameter calculation model to the control device 10 using the communication unit 27 (step S5). The control device 10 receives the parameter calculation model and saves the received parameter calculation model in the storage unit 14.

[0052] Next, referring to Figure 6, we will explain the flow of the pre-cooling time prediction process. Figure 6 is a flowchart showing an example of the pre-cooling time prediction process according to the embodiment. As a premise, the memory unit 14 of the control device 10 is assumed to store the parameter calculation model created by the server 20. Furthermore, it is assumed that defrosting operation is performed during pre-cooling.

[0053] First, the control device 10 of the refrigerated vehicle 1 accepts the pre-cooling settings (step S11). For example, the user inputs the target temperature of the refrigerated storage unit 2 and the target time by which that target temperature must be achieved to the control device 10. The control unit 13 accepts these pre-cooling settings. Next, the prediction unit 12 predicts the pre-cooling time (step S12). For example, when the control unit 13 accepts the pre-cooling settings, it instructs the prediction unit 12 to predict the pre-cooling time. The prediction unit 12 then reads the parameter calculation model from the storage unit 14, acquires the current outside temperature and internal temperature through the sensor value acquisition unit 11, and inputs the acquired outside temperature and internal temperature, the target temperature included in the pre-cooling settings, and a predetermined engine speed into the parameter calculation model. The parameter calculation model outputs thermal response parameters L, τ, etc., corresponding to these explanatory variables. The prediction unit 12 selects an appropriate calculation formula from equations (1) to (5) according to the output thermal response parameters and predicts the pre-cooling time. If thermal response parameters L1, L2, τ1, and τ2 are output, the prediction unit 12 substitutes the thermal response parameters into equations (1) to (3) to predict the pre-cooling time. If thermal response parameters L and τ are output, the prediction unit 12 substitutes the thermal response parameters into equation (4a) or (4b) to predict the pre-cooling time. If thermal response parameters L, τ, and μ are output, the prediction unit 12 substitutes the thermal response parameters into equation (5) to predict the pre-cooling time. The prediction unit 12 outputs the predicted value of the pre-cooling time to the control unit 13.

[0054] Next, the control unit 13 performs pre-cooling (step S13). The control unit 13 calculates a time by the amount of pre-cooling time predicted in step S12 from the target time set in step S11, and when that time arrives, it sets the set temperature of the refrigerator 3 to the target temperature set in step S11 and starts the cooling operation. This makes it possible to control the temperature of the refrigerator 2 to the target temperature by the target time. In addition, since the pre-cooling time is predicted taking into account the effect of the defrosting operation being performed during the cooling operation on the internal temperature, it is possible to suppress failure to reach the target temperature due to insufficient pre-cooling time, and waste of power consumption and fuel consumption due to excessive pre-cooling time.

[0055] (effect) As explained above, according to this embodiment, when other operating conditions intervene during cooling or heating operations, the pre-cooling time is predicted by considering the effect of those other operating conditions on the temperature of the controlled object, so that the pre-cooling time is neither too long nor too short can be predicted with high accuracy.

[0056] In the above embodiment, the parameter calculation model is transmitted from the server 20 to the control device 10. However, the control device 10 may not hold the parameter calculation model, and instead, when the control device 10 transmits explanatory variables to the server 20, the server 20 predicts the pre-cooling time and transmits the predicted pre-cooling time to the control device 10 (a configuration in which the server 20 is equipped with a prediction unit 12). Furthermore, although the above embodiment was described using the example of controlling the temperature of the refrigerated storage compartment 2 of the refrigerated vehicle 1, the pre-cooling time prediction method of this embodiment can also be applied when using an air conditioner to cool or heat a room such as an office. In the case of an air conditioner, the explanatory variables of the parameter calculation model may be the outside temperature, the temperature of the room to be air-conditioned, the airflow rate of the indoor unit's fan, and the target temperature of the room to be air-conditioned.

[0057] Figure 7 shows an example of the hardware configuration of the control system according to the embodiment. The computer 900 includes a CPU 901, main memory 902, auxiliary memory 903, input / output interface 904, and communication interface 905. The control device 10 and server 20 described above are implemented in the computer 900. The functions described above are stored in auxiliary storage device 903 in the form of programs. The CPU 901 reads the programs from auxiliary storage device 903, loads them into main memory 902, and executes the above processes according to the programs. The CPU 901 also allocates memory space in main memory 902 according to the programs. The CPU 901 also allocates memory space in auxiliary storage device 903 to store data being processed according to the programs.

[0058] Furthermore, a program to implement all or part of the functions of the control device 10 and server 20 may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be loaded into a computer system and executed to perform processing by each functional unit. Here, "computer system" includes hardware such as the OS and peripheral devices. Also, if a WWW system is used, "computer system" also includes the homepage provisioning environment (or display environment). Furthermore, "computer-readable recording medium" refers to portable media such as CDs, DVDs, USBs, and storage devices such as hard disks built into the computer system. In addition, if this program is distributed to computer 900 via a communication line, computer 900 that receives the distribution may load the program into main memory 902 and execute the above processing. Furthermore, the above program may be for implementing only a part of the functions described above, and may also be able to implement the above functions in combination with programs already recorded in the computer system.

[0059] As described above, several embodiments relating to this disclosure have been explained, but all of these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be carried out in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents.

[0060] <Note> The prediction system, temperature control system, prediction method, and program described in each embodiment can be understood, for example, as follows:

[0061] (1) A prediction system according to the first embodiment is a prediction system that predicts the pre-cooling time, which is the time required to achieve a target temperature of a target space by the temperature control operation, when another operating state is intervened in the middle of a temperature control operation that cools or heats a target space for temperature control, and comprises: an acquisition unit that acquires explanatory variables which are factors that affect the temperature control of the target space; a model that shows the relationship between the explanatory variables and the dead time and first-order lag time constant based on time series data of the temperature of the target space from the start of the temperature control operation when another operating state is intervened in the middle until the temperature of the target space is controlled to the target temperature by the temperature control operation; and a prediction unit that calculates the dead time and the first-order lag time constant based on the acquired explanatory variables, and predicts the pre-cooling time by multiplying the time constant by a predetermined coefficient and adding the dead time. This allows for accurate prediction of the appropriate pre-cooling time, even when other operations are intervening during cooling or heating.

[0062] (2) A prediction system according to a second embodiment is the prediction system of (1), further comprising a learning unit that learns the relationship between the explanatory variables and the dead time and the time constant of the first-order lag to create the model, wherein the learning unit learns the relationship between the dead time and the time constant of the first-order lag based on the time series data for the first time period and the explanatory variables for the first time period, when the time series data is divided into a first time period from the start of the temperature control operation to the start of the other operating state and a second time period from the start of the other operating state until the temperature of the target space is controlled to the target temperature, and the relationship between the dead time and the time constant of the first-order lag based on the time series data for the second time period and the explanatory variables for the second time period. This allows us to create a model that calculates the dead time and the time constant of the first-order lag from the explanatory variables.

[0063] (3) A prediction system according to a third embodiment is the prediction system of (1), further comprising a learning unit that learns the relationship between the explanatory variables and the dead time and the first-order lag time constant to create the model, wherein the learning unit learns the relationship between the dead time and the first-order lag time constant calculated based on the interpolated time series data obtained by interpolating the temperature of the target space from the last time of the first time of the first time of the first time of the second time of the first This allows us to create a model that calculates the dead time and the time constant of the first-order lag from the explanatory variables.

[0064] (4) A prediction system according to a fourth embodiment is the prediction system of (1), further comprising a learning unit that learns the relationship between the explanatory variables and the dead time and the time constant of the first-order lag to create the model, wherein the learning unit learns the relationship between the dead time and the time constant of the first-order lag calculated based on time series data obtained by drawing a straight line connecting the temperature of the target space from the last time of the first time in the first time in the second time in the other operating state to the first time of the second time in the other operating state, when the time series data is divided into a first time period from the start of the temperature control operation to the start of the other operating state and a second time period from the time the temperature of the target space becomes the temperature at the start time of the other operating state for the first time after the other operating state has started until the temperature of the target space is controlled to the target temperature, and the explanatory variables. This allows us to create a model that calculates the dead time and the time constant of the first-order lag from the explanatory variables.

[0065] (5) A prediction system according to a fifth embodiment is the prediction system of (1), further comprising a learning unit that learns the relationship between the explanatory variables, the dead time, and the time constant of the first-order lag to create the model, wherein the learning unit learns the relationship between the explanatory variables and the explanatory variables, when the time series data is divided into a first time period from the start of the temperature control operation to the start of the other operating state, and a second time period from the time the temperature of the target space becomes the temperature at the start time of the other operating state for the first time after the other operating state has started until the temperature of the target space is controlled to the target temperature, the dead time calculated based on the connected time series data obtained by connecting the last time of the time series data for the first time period and the first time of the time series data for the second time period. This allows us to create a model that calculates the dead time and the time constant of the first-order lag from the explanatory variables.

[0066] (6) The prediction system according to the sixth embodiment is the prediction system according to (5), further comprising: a division unit that divides the time series data after connection into a first-order lag response portion, where the portion after the time when the temperature of the target space begins to decrease in the case of cooling, and the portion after the time when the temperature of the target space begins to rise in the case of heating, and the portion other than the first-order lag response is divided into a dead time portion; and an estimation unit that estimates the length of the dead time portion as dead time, and estimates the time constant that minimizes the error between the temperature in the first-order lag response portion and the simulated value of the temperature when the temperature transition is simulated while changing the time constant that represents the magnitude of the change in temperature, as the time constant of the first-order lag. This allows us to calculate the dead time and the time constant of the first-order lag.

[0067] (7) The prediction system relating to the seventh aspect is the prediction system of (1) to (6), wherein the other operating state is defrost operation. This makes it possible to predict the pre-cooling time when defrosting is performed during cooling or heating operations.

[0068] (8) The temperature control system according to the eighth embodiment comprises a prediction system described in any one of (1) to (7), a temperature control device for controlling the temperature of the target space, and a control device for the temperature control device, wherein when the control device receives a target temperature for the target space and a target time for achieving the target temperature, it instructs the prediction system to predict the pre-cooling time, receives the pre-cooling time predicted by the prediction system, and causes the temperature control device to start temperature control by the amount of the pre-cooling time before the target time. This allows the temperature of the target space to be controlled so that it reaches the target temperature by the target time.

[0069] (9) The temperature control system according to the ninth embodiment is the temperature control system described in (8), wherein the target space is the cold storage compartment of a refrigerated vehicle, the temperature control device is the refrigeration unit of the refrigerated vehicle, and the explanatory variables include the outside temperature, the internal temperature of the cold storage compartment, the target temperature, and the engine speed of the refrigerated vehicle. This allows for the estimation of the pre-cooling time for the refrigerated compartment of a refrigerated vehicle, and enables the cooling or heating process related to pre-cooling.

[0070] (10) A prediction method according to a tenth embodiment is a prediction method for predicting a pre-cooling time, which is the time required to achieve a target temperature of a target space by a temperature control operation, when another operating state is intervened in the middle of a temperature control operation for cooling or heating a target space for temperature control, and comprises the steps of: a computer acquiring explanatory variables which are factors that affect the temperature control of the target space; the computer calculating the dead time and the first-order lag time constant based on the explanatory variables, a model which shows the relationship between the dead time and the first-order lag time constant based on time series data of the temperature of the target space from the start of the temperature control operation when another operating state is intervened in the middle of the operation, until the temperature of the target space is controlled to the target temperature by the temperature control operation, and the acquired explanatory variables, and predicting the pre-cooling time by multiplying the time constant by a predetermined coefficient and adding the dead time.

[0071] (11) A program according to the eleventh embodiment causes a computer to perform a process for predicting a pre-cooling time, which indicates the time required to achieve a target temperature of a target space by a temperature control operation when another operating state is intervened in the middle of a temperature control operation for cooling or heating a target space for temperature control, the process comprising: acquiring explanatory variables which are factors that affect the temperature control of the target space; a model which shows the relationship between the explanatory variables and a time constant of dead time and a first-order lag based on time series data of the temperature of the target space from the start of the temperature control operation when another operating state is intervened in the middle of the temperature control operation, when the temperature of the target space is controlled to the target temperature by the temperature control operation; and the acquired explanatory variables which calculate the dead time and the time constant of a first-order lag, and predict the pre-cooling time by multiplying the time constant by a predetermined coefficient and adding the dead time. [Explanation of Symbols]

[0072] 1. Refrigerated vehicle 2. Cooler 3. Refrigeration unit 4. Outdoor temperature sensor 5. Internal temperature sensor 6. Rotation speed sensor 10. Control device 11. Sensor value acquisition unit 12.. Prediction Department 13. Control Unit 14...Storage section 15. Communications Department 20... Server 21. Performance Data Acquisition Department 22...Extraction part 23...Divided part 24...Estimation part 25. Learning Department 26...Storage section 27. Communications Department 100... Control System NW... Network 900... Computer 901···CPU 902...Main memory 903...Auxiliary storage device 904... Input / Output Interface 905...Communication Interface

Claims

1. A prediction system that predicts the pre-cooling time, which indicates the time required to achieve the target temperature of a space by the temperature control operation, when other operating conditions intervene during a temperature control operation to cool or heat a space to be temperature controlled, An acquisition unit that acquires explanatory variables which are factors that affect the temperature control of the target space, A model showing the relationship between the explanatory variables and the dead time and first-order lag time constant based on time-series data of the temperature of the target space from the start of the temperature control operation when other operating conditions are intervened in between, until the temperature of the target space is controlled to the target temperature by the temperature control operation; and a prediction unit that calculates the dead time and the first-order lag time constant based on the acquired explanatory variables, and predicts the pre-cooling time by multiplying the time constant by a predetermined coefficient and adding the dead time; A prediction system equipped with the following features.

2. A learning unit that learns the relationship between the explanatory variables, the dead time, and the time constant of the first-order lag to create the model. Furthermore, The learning unit learns the relationship between the dead time and the time constant of the first-order lag based on the time series data for the first time period and the explanatory variables for the first time period, when the time series data is divided into a first time period from the start of the temperature control operation to the start of the other operating state and a second time period from the start of the other operating state until the temperature of the target space is controlled to the target temperature, and the relationship between the dead time and the time constant of the first-order lag based on the time series data for the second time period and the explanatory variables for the second time period. The prediction system according to claim 1.

3. A learning unit that learns the relationship between the explanatory variables, the dead time, and the time constant of the first-order lag to create the model. Furthermore, The learning unit learns the relationship between the dead time and the first-order lag time constant, calculated based on the interpolated time series data obtained by interpolating the temperature of the target space from the last time of the first time period to the first time of the second time period with a predetermined value representing missing or blank data, when the time series data is divided into a first time period from the start of the temperature control operation to the start of the other operating state, and a second time period from the time the temperature of the target space reaches the temperature at the start time of the other operating state for the first time after the other operating state has started until the temperature of the target space is controlled to the target temperature, and the explanatory variables. The prediction system according to claim 1.

4. A learning unit that learns the relationship between the explanatory variables, the dead time, and the time constant of the first-order lag to create the model. Furthermore, The learning unit learns the relationship between the dead time and the first-order lag time constant, calculated based on the time series data obtained by drawing a straight line connecting the temperature of the target space from the last time of the first time period to the first time of the second time period, and the explanatory variables, when the time series data is divided into a first time period from the start of the temperature control operation to the start of the other operating state, and a second time period from when the temperature of the target space first reaches the temperature at the start time of the other operating state after the other operating state has started until the temperature of the target space is controlled to the target temperature. The prediction system according to claim 1.

5. A learning unit that learns the relationship between the explanatory variables, the dead time, and the time constant of the first-order lag to create the model. Furthermore, The learning unit learns the relationship between the explanatory variables and the following when the time series data is divided into a first time period from the start of the temperature control operation to the start of the other operating state, and a second time period from when the temperature of the target space first reaches the temperature at the start time of the other operating state after the other operating state has started until the temperature of the target space is controlled to the target temperature. The learning unit then learns the relationship between the dead time based on the connected time series data obtained by connecting the last time of the time series data related to the first time period and the first time of the time series data related to the second time period, the time constant of the first-order lag based on the connected time series data, and the length of time from the last time of the first time period to the first time of the second time period. The prediction system according to claim 1.

6. A division unit divides the time-series data after connection into a first-order lag response portion, where the portion from the time when the temperature of the target space begins to decrease in the case of cooling, and from the time when the temperature of the target space begins to rise in the case of heating, and the portion other than the first-order lag response is considered a dead time portion. An estimation unit estimates the length of the aforementioned dead time portion as dead time, and estimates the time constant that minimizes the error between the temperature in the first-order lag response portion and the simulated value of the temperature when the temperature transition is simulated while changing the time constant representing the magnitude of the temperature change, as the time constant of the first-order lag. The prediction system according to claim 5, further comprising:

7. The other operating condition mentioned above is defrost operation. A prediction system according to any one of claims 1 to 6.

8. A prediction system according to any one of claims 1 to 6, A temperature control device that controls the temperature of the target space, The control device of the temperature control device, Equipped with, When the control device receives the target temperature of the target space and the target time for achieving that target temperature, it instructs the prediction system to predict the pre-cooling time, receives the pre-cooling time predicted by the prediction system, and causes the temperature control device to start temperature control by the amount of the pre-cooling time before the target time. Temperature control system.

9. The aforementioned target space is the cold storage compartment of the refrigerated vehicle, and the temperature control device is the refrigeration unit installed in the refrigerated vehicle. The explanatory variables include the outside temperature, the internal temperature of the refrigerator, the target temperature, and the engine speed of the refrigerated vehicle. The temperature control system according to claim 8.

10. A prediction method for predicting the pre-cooling time, which indicates the time required to achieve the target temperature of a space by the temperature control operation, when other operating conditions intervene during a temperature control operation to cool or heat a space to be temperature controlled, The computer obtains explanatory variables which are factors that affect the temperature control of the target space, The computer calculates the dead time and the first-order lag time constant based on the explanatory variables, a model showing the relationship between the dead time and the first-order lag time constant based on time-series data of the temperature of the target space from the start of the temperature control operation when other operating conditions are intervened in between, until the temperature of the target space is controlled to the target temperature by the temperature control operation, and the acquired explanatory variables, and predicts the pre-cooling time by multiplying the time constant by a predetermined coefficient and adding the dead time. A prediction method having the following characteristics.

11. On the computer, A process for predicting the pre-cooling time, which indicates the time required to achieve the target temperature of a space by the temperature control operation, when another operating state intervenes during a temperature control operation to cool or heat a space to be temperature controlled, The steps include obtaining explanatory variables which are factors that affect the temperature control of the target space, A step of predicting the pre-cooling time by calculating the dead time and the first-order lag time constant based on the explanatory variables, a model showing the relationship between the dead time and the first-order lag time constant based on time-series data of the temperature of the target space from the start of the temperature control operation when other operating conditions are intervened in between, until the temperature of the target space is controlled to the target temperature by the temperature control operation, and the acquired explanatory variables, and adding the dead time by multiplying the time constant by a predetermined coefficient, A program that performs a process that includes the following.

Images