GB2563508A - Refrigerant shortage prediction apparatus, refrigerant shortage prediction method, and program - Google Patents

Abstract

A refrigerant shortage prediction apparatus according to the present invention has: a reception unit that receives a plurality of operation data including the outside air temperature, the operation frequency of the compressor of an air-conditioning cooling and heating device, the evaporation temperature or the condensing temperature of a heat exchanger, refrigerant shortage detection information, and a data collection date, and a refrigerant filling date, from a centralized monitoring apparatus that collects the operation data from the air-conditioning cooling and heating device via a network; an input unit that receives input data including predicted values of the outside air temperature, the operation frequency, and the evaporation temperature or the condensing temperature on a prediction date; an arithmetic unit that calculates an elapsed period after refrigerant filling for each operation data and an expected period after the refrigerant filling corresponding to the prediction date, groups a plurality of past data including the operation frequency, the outside air temperature, the evaporation temperature or the condensing temperature, and the elapsed period after the refrigerant filling by executing multi-variable analysis, and calculates a refrigerant shortage probability on the prediction date on the basis of the refrigerant shortage detection information about a group to which prediction data including the input data and the expected period after the refrigerant filling belongs; and a display unit that outputs the refrigerant shortage probability calculated by the arithmetic unit.

Description

In refrigerating and air-conditioning apparatuses, a shortage of refrigerant necessary for operation is caused by, for example, age deterioration or poor construction work in some cases. Patent Literature 1 discloses one example of a method for detecting a shortage of refrigerant in a refrigerating and air-conditioning apparatus.

A refrigeration cycle apparatus disclosed in Patent Literature 1 includes a refrigerant circuit that includes devices, such as a compressor and a variable expansion mechanism, and a control circuit that detects a state of refrigerant in the refrigerant circuit or an operating state of a device in the refrigerant circuit and repeatedly determines, on the basis of a detected value, whether an amount of refrigerant in the refrigerant circuit is insufficient.

Patent Literature 1 discloses that, when the control circuit determines that an amount of refrigerant in the refrigerant circuit is insufficient, a refrigerant shortage determination process is terminated, and execution of a refrigeration cycle operation control process is also inhibited.

Citation List

Patent Literature [0003]

Patent Literature 1: Japanese Unexamined Patent Application Publication No. 2015-140961 (claim 1) Summary of Invention

Technical Problem [0004]

In the method disclosed in Patent Literature 1, a refrigerant shortage is detected at a point in time when refrigerant necessary for operation actually runs short, and thus the operation of the refrigeration cycle apparatus has to be suspended until refrigerant has been charged by a maintenance service provider. Although it can be considered that the refrigeration cycle apparatus is operated with airconditioning capacity being reduced, the load on the compressor of the refrigeration cycle apparatus increases.

[0005]

The present invention has been accomplished to solve the above-described issue, and is directed to a refrigerant shortage prediction apparatus capable of predicting the occurrence of a refrigerant shortage before the refrigerant shortage actually occurs in a refrigerating and air-conditioning apparatus, a refrigerant shortage prediction method, and a program for causing a computer to execute the method. Solution to Problem [0006]

A refrigerant shortage prediction apparatus according to an embodiment of the present invention is a refrigerant shortage prediction apparatus configured to be connected, via a network, to a centralized monitoring apparatus configured to collect operating data that is data on an operating state from a refrigerating and airconditioning apparatus including a heat source unit including a compressor and a heat exchanger. The refrigerant shortage prediction apparatus includes a reception unit configured to receive, from the centralized monitoring apparatus, multiple pieces of the operating data including an operating frequency of the compressor, an outside air temperature that is a temperature of outside air with which the heat exchanger exchanges heat, an evaporating temperature or a condensing temperature in the heat exchanger, refrigerant shortage detection information that is information on whether a refrigerant shortage is detected, and a data collection date that is a date on which these pieces of data are collected, and a refrigerant charging date that is a date on which refrigerant is charged into the refrigerating and air-conditioning apparatus, an input unit configured to receive, as information for refrigerant shortage prediction, input data including a prediction date, a prediction value of the outside air temperature on the prediction date, a target setting value of the operating frequency of the compressor on the prediction date, and a target setting value of the evaporating temperature or the condensing temperature on the prediction date, a time period calculation unit configured to calculate, on the basis of the data collection date, the prediction date, and the refrigerant charging date, an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the data collection date, using the data collection date as a reference, for each piece of the operating data, and an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference, a probability calculation unit configured to, by multivariate analysis on multiple pieces of past data including the operating frequency, the outside air temperature, and the evaporating temperature or the condensing temperature that are included in each piece of the operating data, and the elapsed time period since refrigerant charging, classify the multiple pieces of past data into a plurality of groups, configured to identify, among the plurality of groups, a group to which prediction data including the input data and the estimated time period since refrigerant charging belongs, and configured to calculate, on the basis of the refrigerant shortage detection information of a piece of the operating data corresponding to a piece of the multiple pieces of past data belonging to the identified group, a refrigerant shortage probability that is an index indicating whether a refrigerant shortage occurs on the prediction date, and a display unit configured to output the refrigerant shortage probability calculated by the probability calculation unit.

[0007]

A refrigerant shortage prediction method according to another embodiment of the present invention is a refrigerant shortage prediction method performed by a refrigerant shortage prediction apparatus configured to be connected, via a network, to a centralized monitoring apparatus configured to collect operating data that is data on an operating state from a refrigerating and air-conditioning apparatus including a heat source unit including a compressor and a heat exchanger. The refrigerant shortage prediction method includes receiving, from the centralized monitoring apparatus, multiple pieces of the operating data including an operating frequency of the compressor, an outside air temperature that is a temperature of outside air with which the heat exchanger exchanges heat, an evaporating temperature or a condensing temperature in the heat exchanger, refrigerant shortage detection information that is information on whether a refrigerant shortage is detected, and a data collection date that is a date on which these pieces of data are collected, and a refrigerant charging date that is a date on which refrigerant is charged into the refrigerating and air-conditioning apparatus, in response to input of, as information for refrigerant shortage prediction, input data including a prediction date, a prediction value of the outside air temperature on the prediction date, a target setting value of the operating frequency of the compressor on the prediction date, and a target setting value of the evaporating temperature or the condensing temperature on the prediction date, calculating, on the basis of the data collection date, the prediction date, and the refrigerant charging date, an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the data collection date, using the data collection date as a reference, for each piece of the operating data, and an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference, by multivariate analysis on multiple pieces of past data including the operating frequency, the outside air temperature, and the evaporating temperature or the condensing temperature that are included in each piece of the operating data, and the elapsed time period since refrigerant charging, classifying the multiple pieces of past data into a plurality of groups, identifying, among the plurality of groups, a group to which prediction data including the input data and the estimated time period since refrigerant charging belongs, and calculating, on the basis of the refrigerant shortage detection information of a piece of the operating data corresponding to a piece of the multiple pieces of past data belonging to the identified group, a refrigerant shortage probability that is an index indicating whether a refrigerant shortage occurs on the prediction date, and outputting the refrigerant shortage probability.

[0008]

A program according to still another embodiment of the present invention is a program for causing a computer connected, via a network, to a centralized monitoring apparatus configured to collect operating data that is data on an operating state from a refrigerating and air-conditioning apparatus including a heat source unit including a compressor and a heat exchanger to execute steps including receiving, from the centralized monitoring apparatus, multiple pieces of the operating data including an operating frequency of the compressor, an outside air temperature that is a temperature of outside air with which the heat exchanger exchanges heat, an evaporating temperature or a condensing temperature in the heat exchanger, refrigerant shortage detection information that is information on whether a refrigerant shortage is detected, and a data collection date that is a date on which these pieces of data are collected, and a refrigerant charging date that is a date on which refrigerant is charged into the refrigerating and air-conditioning apparatus, in response to input of, as information for refrigerant shortage prediction, input data including a prediction date, a prediction value of the outside air temperature on the prediction date, a target setting value of the operating frequency of the compressor on the prediction date, and a target setting value of the evaporating temperature or the condensing temperature on the prediction date, calculating, on the basis of the data collection date, the prediction date, and the refrigerant charging date, an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the data collection date, using the data collection date as a reference, for each piece of the operating data, and an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference, by multivariate analysis on multiple pieces of past data including the operating frequency, the outside air temperature, and the evaporating temperature or the condensing temperature that are included in each piece of the operating data, and the elapsed time period since refrigerant charging, classifying the multiple pieces of past data into a plurality of groups, identifying, among the plurality of groups, a group to which prediction data including the input data and the estimated time period since refrigerant charging belongs, and calculating, on the basis of the refrigerant shortage detection information of a piece of the operating data corresponding to a piece of the multiple pieces of past data belonging to the identified group, a refrigerant shortage probability that is an index indicating whether a refrigerant shortage occurs on the prediction date, and outputting the refrigerant shortage probability.

Advantageous Effects of Invention [0009]

In an embodiment of the present invention, pieces of data including operating data on an operating state of the refrigerating and air-conditioning apparatus are subjected to multivariate analysis to thereby reduce disturbances included in the pieces of data, the pieces of data are classified into groups by the likelihood of a refrigerant shortage, a refrigerant shortage probability of a group to which a prediction date belongs is calculated, and thus the occurrence of a refrigerant shortage can be predicted before the refrigerant shortage actually occurs in the refrigerating and airconditioning apparatus.

Brief Description of Drawings [0010] [Fig. 1] Fig. 1 illustrates an example of the configuration of an air-conditioning system in Embodiment 1 of the present invention.

[Fig. 2] Fig. 2 is a functional block diagram illustrating an example of the configuration of a refrigerating and air-conditioning apparatus illustrated in Fig. 1.

[Fig. 3] Fig. 3 is a functional block diagram illustrating an example of the configuration of a centralized monitoring apparatus illustrated in Fig. 1.

[Fig. 4] Fig. 4 is a functional block diagram illustrating an example of the configuration of a refrigerant shortage prediction apparatus illustrated in Fig. 1.

[Fig. 5] Fig. 5 is a flowchart illustrating an operation procedure performed by the centralized monitoring apparatus illustrated in Fig. 3.

[Fig. 6] Fig. 6 is a flowchart illustrating an operation procedure performed by the refrigerant shortage prediction apparatus illustrated in Fig. 4.

[Fig. 7] Fig. 7 is a flowchart illustrating an operation procedure performed by a time period calculation unit illustrated in Fig. 4.

[Fig. 8] Fig. 8 schematically illustrates an elapsed time period since refrigerant charging calculated in step 702 and an estimated time period since refrigerant charging calculated in step 704.

[Fig. 9] Fig. 9 is a flowchart illustrating an operation procedure performed by a probability calculation unit illustrated in Fig. 4.

[Fig. 10] Fig. 10 illustrates a specific example of the procedure illustrated in Fig. 9.

[Fig. 11 A] Fig. 11A illustrates an example of a table output from a display unit illustrated in Fig. 4.

[Fig. 11B] Fig. 11B illustrates an example of a graph output from the display unit illustrated in Fig. 4.

Description of Embodiments [0011]

Embodiment 1 (Configuration of Air-conditioning System)

The configuration of an air-conditioning system in Embodiment 1 will be described.

Fig. 1 illustrates an example of the configuration of an air-conditioning system in Embodiment 1 of the present invention.

The air-conditioning system includes a plurality of refrigerating and airconditioning apparatuses 10Ato 10D, a centralized monitoring apparatus 1 that collects data on an operating state from the refrigerating and air-conditioning apparatuses 10Ato 10D, and a refrigerant shortage prediction apparatus 2. In Embodiment 1, each of the refrigerating and air-conditioning apparatuses 10Ato 10D is a refrigeration machine.

The centralized monitoring apparatus 1 is connected to the refrigerant shortage prediction apparatus 2 via a network 3. The network 3 is a public network. The network 3 may include the Internet. In this case, the centralized monitoring apparatus 1 and the refrigerant shortage prediction apparatus 2 may use TCP/IP (Transmission Control Protocol/lnternet Protocol) as communications protocols. [0012]

The refrigerating and air-conditioning apparatuses 10Ato 10D respectively include heat source units 4A to 4D and load-side units 5A to 5D. The heat source units 4Ato 4D are connected to the load-side units 5Ato 5D via refrigerant pipes 7A to 7D, respectively. The load-side units 5Ato 5D are installed in a cold storage warehouse 8.

The heat source units 4A to 4D and the load-side units 5A to 5D are connected to the centralized monitoring apparatus 1 via signal lines 6Ato 6E.

The configuration of each apparatus illustrated in Fig. 1 will be described in detail below.

[0013] (Configurations of Refrigerating and Air-conditioning Apparatuses 10Ato 10D)

Configurations of the refrigerating and air-conditioning apparatuses 10Ato 10D illustrated in Fig. 1 will be described.

The refrigerating and air-conditioning apparatuses 10Ato 10D have the same configuration. Here, the configuration of the refrigerating and air-conditioning apparatus 10Awill be thus described, and detailed descriptions of the refrigerating and air-conditioning apparatuses 10B to 10D are omitted. In Embodiment 1, of the refrigerating and air-conditioning apparatus 10A, the configuration related to a refrigerant shortage prediction method will be described in detail, and a detailed description of a typical refrigeration cycle is omitted.

Fig. 2 is a functional block diagram illustrating an example of the configuration of a refrigerating and air-conditioning apparatus illustrated in Fig. 1.

As illustrated in Fig. 2, the refrigerating and air-conditioning apparatus 10A includes the heat source unit 4A and the load-side unit 5A.

[0014]

The heat source unit 4A includes a compressor 11, a heat exchanger 12 that operates as a condenser, a fan 13 that sends air to the heat exchanger 12, a temperature sensor 14 that measures an outside air temperature that is a temperature of outside air with which the heat exchanger 12 exchanges heat, an electronic expansion valve 15, and a controller 16. The controller 16 includes an operation unit (not illustrated) via which a manager including a user inputs an instruction, a timer (not illustrated) that measures time, a memory (not illustrated), and a microcomputer (not illustrated) that executes a process to follow a program. The memory (not illustrated) is, for example, a nonvolatile memory. The refrigerating and air-conditioning apparatuses 10Ato 10D are given respective different identifiers in advance, and the memory (not illustrated) stores an identifier of the refrigerating and air-conditioning apparatus 10A including the memory. In the operation unit (not illustrated), a refrigerant shortage warning light (not illustrated) for notifying the manager of the necessity to add refrigerant is provided.

The load-side unit 5A includes a heat exchanger 17 that operates as an evaporator, a temperature sensor 18 that measures an evaporating temperature in the heat exchanger 17, a fan 19 that sends air to the heat exchanger 17, and a temperature sensor 20 that measures a temperature in the cold storage warehouse 8.

The compressor 11, the heat exchanger 12, the electronic expansion valve 15, and the heat exchanger 17 are connected with the refrigerant pipe 7A.

[0015]

The controller 16 is connected, via the signal lines 6A, to the compressor 11, the electronic expansion valve 15, a motor (not illustrated) of the fan 13, and the temperature sensor 14, and also to the temperature sensors 18 and 20, and a motor (not illustrated) of the fan 19 that are included in the load-side unit 5A. The controller 16 is also connected to the centralized monitoring apparatus 1 via the signal line 6E.

When the manager inputs an instruction for refrigerating operation via the operation unit (not illustrated), the controller 16 starts the compressor 11, and the fans 13 and 19, and controls the opening degree of the electronic expansion valve 15, an operating frequency of the compressor 11, and rotation speeds of the fans 13 and 19 so that a measurement value of the temperature sensor 20 reaches an indicated temperature.

[0016]

When it becomes necessary to add refrigerant to the refrigerant pipe 7A, the controller 16 switches the refrigerant shortage warning light (not illustrated) from an off state to an on state. As a method for detecting a refrigerant shortage, there is a method disclosed in Patent Literature 1, for example. When a refrigerant shortage is detected, the controller 16 refers to a date and time measured by the timer (not illustrated) and records information on a date on which the refrigerant shortage has been detected into the memory (not illustrated).

The controller 16 monitors a date and time measured by the timer (not illustrated) and acquires, on a predetermined date and time of each month, an operating frequency from the compressor 11, an evaporating temperature from the temperature sensor 18, and an outside air temperature from the temperature sensor 14. Then, the controller 16 transmits operating data including the acquired operating frequency, evaporating temperature, and outside air temperature, and a data collection date that is a date on which these pieces of data have been collected to the centralized monitoring apparatus 1 via the signal line 6E.

[0017]

Furthermore, when the controller 16 transmits the operating data that is data on an operating state to the centralized monitoring apparatus 1, the controller 16 refers to information stored in the memory (not illustrated) and checks whether a refrigerant shortage has been detected in a most recent time period between data collection dates that is a time period from a previous data collection date to this data collection date. As a consequence of the check, when a refrigerant shortage has been detected in the most recent time period between data collection dates, the controller 16 generates refrigerant shortage detection information including information of refrigerant shortage detected. On the other hand, when no refrigerant shortage has been detected in the most recent time period between data collection dates, the controller 16 generates refrigerant shortage detection information including information of refrigerant shortage not detected. Then, the controller 16 transmits the generated refrigerant shortage detection information to the centralized monitoring apparatus 1 via the signal line 6E.

Refrigerant shortage detection information for the case where a refrigerant shortage has been detected is used as learning data for a refrigerant shortage prediction process to be described.

[0018]

Furthermore, when the manager charges refrigerant into the refrigerant pipe 7A, the controller 16 refers to a date and time measured by the timer (not illustrated) and transmits information on a refrigerant charging date that is a date on which the refrigerant has been charged to the centralized monitoring apparatus 1 via the signal line 6E. When the refrigerant is charged into the refrigerant pipe 7A, the controller 16 switches the refrigerant shortage warning light (not illustrated) from the on state to the off state.

[0019]

When the controller 16 each provided to the refrigerating and air-conditioning apparatuses 10Ato 10D transmits operating data, a refrigerant charging date, and refrigerant shortage detection information to the centralized monitoring apparatus 1, the controller 16 attaches an identifier of the refrigerating and air-conditioning apparatus including the controller 16 to these pieces of data.

Furthermore, although, in Embodiment 1, the case where each refrigerating and air-conditioning apparatus collects data on the same date and time of each month is described, a data collection time interval is a certain time interval. The data collection time interval is not limited to a one-month interval and may be a one11 week interval or a ten-day interval. Furthermore, a value of data on an operating state transmitted to the centralized monitoring apparatus 1 is not limited to one measurement value measured on a predetermined date and time, and may be an average value of a plurality of measurement values measured in a certain time period from the predetermined date and time. For example, an operating frequency of the compressor 11 and measurement values of the temperature sensor 14 and the temperature sensor 18 are each measured ten times at one-minute intervals from a predetermined date and time, and the controller 16 may include an average value of ten measurement values obtained for each measurement in operating data.

[0020] (Configuration of Centralized Monitoring Apparatus 1)

Fig. 3 is a functional block diagram illustrating an example of the configuration of the centralized monitoring apparatus illustrated in Fig. 1.

As illustrated in Fig. 3, the centralized monitoring apparatus 1 includes an input unit 101, a display unit 102, a memory 106, a transmission unit 107, a reception unit 108, and a controller 110 that controls each unit. The controller 110 includes a memory (not illustrated) that stores a program, and a CPU (Central Processing Unit) (not illustrated) that executes a process to follow the program. The memory (not illustrated) is, for example, a nonvolatile memory.

[0021]

The memory 106 stores various pieces of data collected from the refrigerating and air-conditioning apparatuses 10Ato 10D. The memory 106 is, for example, a hard disk drive.

The reception unit 108 includes an operating state reception unit 103, a refrigerant charging date reception unit 104, and a detection information reception unit 105.

When the operating state reception unit 103 receives pieces of operating data of the refrigerating and air-conditioning apparatuses 10Ato 10D from the heat source units 4Ato 4D via the signal line 6E, the operating state reception unit 103 passes the pieces of operating data to the controller 110.

When the refrigerant charging date reception unit 104 receives pieces of data on refrigerant charging dates from the heat source units 4Ato 4D via the signal line 6E, the refrigerant charging date reception unit 104 passes the pieces of data on refrigerant charging dates to the controller 110.

When the detection information reception unit 105 receives pieces of refrigerant shortage detection information from the heat source units 4Ato 4D via the signal line 6E, the detection information reception unit 105 passes the pieces of refrigerant shortage detection information to the controller 110.

[0022]

The controller 110 stores various pieces of data received from the reception unit 108 and information input from the input unit 101 into the memory 106. The controller 110 includes pieces of refrigerant shortage detection information ofthe refrigerating and air-conditioning apparatuses 10Ato 10D in respective pieces of operating data ofthe refrigerating and air-conditioning apparatuses 10Ato 10D among pieces of information stored by the memory 106, and passes the pieces of operating data to the transmission unit 107. Furthermore, the controller 110 reads respective pieces of information on refrigerant charging dates of the refrigerating and air-conditioning apparatuses 10Ato 10D from the memory 106 and passes the pieces of information on refrigerant charging dates to the transmission unit 107. The controller 110 may cause the display unit 102 to display the pieces of information read from the memory 106.

The transmission unit 107 transmits pieces of information received from the controller 110 to the refrigerant shortage prediction apparatus 2 via the network 3.

The display unit 102 displays pieces of information received from the controller 110.

[0023] (Configuration of Refrigerant Shortage Prediction Apparatus 2)

Fig. 4 is a functional block diagram illustrating an example of the configuration of the refrigerant shortage prediction apparatus illustrated in Fig. 1.

As illustrated in Fig. 4, the refrigerant shortage prediction apparatus 2 includes a reception unit 201, an input unit 211, a memory 206, a display unit 209, a calculation unit 212, and a controller 210. The refrigerant shortage prediction apparatus 2 includes a memory (not illustrated) that stores a program, and a CPU (not illustrated) that executes a process to follow the program. The memory (not illustrated) is, for example, a nonvolatile memory. The CPU (not illustrated) executing a process to follow the program corresponds to the controller 210 and the calculation unit 212 in the refrigerant shortage prediction apparatus 2.

Although, in the functional block diagram illustrated in Fig. 4, the controller 210 and the calculation unit 212 are configured separately, the controller 210 may have a function of the calculation unit 212. The refrigerant shortage prediction apparatus 2 is, for example, an information processing apparatus including a computer and a server.

[0024]

The memory 206 stores various pieces of data received from the centralized monitoring apparatus 1, and input data including a prediction date that is a date on which the manager wants to predict the probability of occurrence of a refrigerant shortage. The memory 206 is, for example, a hard disk drive. The term prediction date does not refer to a date on which a refrigerant shortage prediction process is performed, but refers to a future date on which the manager wants to check the possibility of occurrence of a refrigerant shortage for each of the refrigerating and airconditioning apparatuses 10Ato 10D.

When the reception unit 201 receives data from the centralized monitoring apparatus 1, the reception unit 201 passes the received data to the controller 210.

The input unit 211 receives input data for refrigerant shortage prediction. The input unit 211 includes a prediction outside air temperature acquisition unit 202, a prediction frequency input unit 203, a prediction evaporating temperature input unit 204, and a prediction date input unit 205.

[0025]

When a prediction date is input, the prediction date input unit 205 passes information on the prediction date to the controller 210.

When a target setting value of an operating frequency of the compressor 11 on the prediction date is input, the prediction frequency input unit 203 passes the target setting value of the operating frequency to the controller 210.

When a target setting value of an evaporating temperature in the heat exchanger 17 on the prediction date is input, the prediction evaporating temperature input unit 204 passes the target setting value of the evaporating temperature to the controller 210.

A prediction date, a target setting value of an operating frequency, and a target setting value of an evaporating temperature may be input by the manager operating the input unit 211, or may be input by the manager via an information processing terminal (not illustrated) connected to the refrigerant shortage prediction apparatus 2 via the network 3.

[0026]

When a prediction value of an outside air temperature for the refrigerating and air-conditioning apparatuses 10Ato 10D on the prediction date is input, the prediction outside air temperature acquisition unit 202 passes the prediction value of the outside air temperature to the controller 210. Examples of a prediction value of the outside air temperature on a prediction date include values provided by the Meteorological Agency and a weather forecasting company. A prediction value of the outside air temperature may be input by the manager via the input unit 211, or may be provided from a server (not illustrated) of a weather forecasting company via the network 3.

For example, the memory 206 stores information on a geographical location of the cold storage warehouse 8 in advance. When the prediction outside air temperature acquisition unit 202 transmits pieces of information on the geographical location of the cold storage warehouse 8 and a prediction date to the server (not illustrated) of the weather forecasting company via the network 3, the server (not illustrated) of the weather forecasting company returns a prediction value of an outside air temperature corresponding to the received prediction date and geographical location to the prediction outside air temperature acquisition unit 202. In this case, a prediction value of the outside air temperature may be a prediction value of an average daily temperature on the prediction date, or may be a prediction value of an outside air temperature during a predetermined time period.

[0027]

Input data received by the input unit 211 is information including a prediction date, a target setting value of an operating frequency of the compressor 11 on the prediction date, a target setting value of an evaporating temperature in the heat exchanger 17 on the prediction date, and a prediction value of an outside air temperature on the prediction date.

A refrigerating and air-conditioning apparatus for which a refrigerant shortage is to be predicted may be any of the refrigerating and air-conditioning apparatuses 10A to 10D, and, among the refrigerating and air-conditioning apparatuses 10Ato 10D, a refrigerating and air-conditioning apparatus for which the manager wants to predict a refrigerant shortage may be specified by the manager operating the input unit 211. In this case, input data includes information on an identifier of a refrigerating and airconditioning apparatus for which a refrigerant shortage is to be predicted.

[0028]

The controller 210 stores data including operating data received from the reception unit 201 and input data received from the input unit 211 into the memory 206. When the controller 210 stores the input data received from the input unit 211 into the memory 206, the controller 210 instructs the calculation unit 212 to perform a refrigerant shortage prediction process.

The calculation unit 212 includes a time period calculation unit 207 and a probability calculation unit 208. When the calculation unit 212 receives an instruction from the controller 210 to perform the refrigerant shortage prediction process, the calculation unit 212 starts the time period calculation unit 207 and the probability calculation unit 208.

[0029]

The time period calculation unit 207 calculates, for each of the refrigerating and air-conditioning apparatuses 10Ato 10D, an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to a data collection date of operating data, using the data collection date as a reference, and notifies the probability calculation unit 208 of the calculated elapsed time period since refrigerant charging. Furthermore, the time period calculation unit 207 calculates, for each of the refrigerating and air-conditioning apparatuses 10Ato 10D, an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to a prediction date, using the prediction date as a reference, and notifies the probability calculation unit 208 of the calculated estimated time period since refrigerant charging.

Hereinafter, information including an operating frequency, an outside air temperature, and an evaporating temperature that are included in operating data, and an elapsed time period since refrigerant charging is referred to as past data. Furthermore, information including input data including a prediction date, a prediction value of an outside air temperature, a target setting value of an operating frequency, and a target setting value of an evaporating temperature, and an estimated time period since refrigerant charging is referred to as prediction data. [0030]

The probability calculation unit 208, by multivariate analysis on multiple pieces of past data, classifies the multiple pieces of past data into a plurality of groups, and identifies, among the plurality of groups, a group to which prediction data belongs. Then, the probability calculation unit 208 calculates, on the basis of refrigerant shortage detection information of a piece of past data belonging to the identified group, a refrigerant shortage probability that is an index indicating whether a refrigerant shortage has occurred on a prediction date.

As a method of, by multivariate analysis, classifying multiple pieces of data into a plurality of groups, for example, a method obtained by combining principal component analysis and clustering can be considered. Clustering includes hierarchical clustering and non-hierarchical clustering. In the case where clustering is non-hierarchical clustering, an example of a clustering method is a k-means method.

The probability calculation unit 208 notifies the controller 210 of the calculated refrigerant shortage probability and causes the display unit 209 to display the calculated refrigerant shortage probability.

The display unit 209 displays information on the refrigerant shortage probability received from the controller 210.

[0031] (Descriptions of Operations Performed in Air-conditioning System)

Next, operations performed in the centralized monitoring apparatus 1 and the refrigerant shortage prediction apparatus 2 in Embodiment 1 will be described. [0032] (Operation Performed by Centralized Monitoring Apparatus 1)

Fig. 5 is a flowchart illustrating an operation procedure performed by the centralized monitoring apparatus illustrated in Fig. 3.

When the centralized monitoring apparatus 1 receives pieces of operating data of the refrigerating and air-conditioning apparatuses 10Ato 10D from the heat source units 4Ato 4D via the signal line 6E and the operating state reception unit 103, the centralized monitoring apparatus 1 stores the pieces of operating data of the refrigerating and air-conditioning apparatuses 10Ato 10D into the memory 106 (step

501) .

When the centralized monitoring apparatus 1 receives pieces of information on refrigerant charging dates from the heat source units 4Ato 4D via the signal line 6E and the refrigerant charging date reception unit 104, the centralized monitoring apparatus 1 stores the pieces of information on refrigerant charging dates of the refrigerating and air-conditioning apparatuses 10Ato 10D into the memory 106 (step

502) .

Subsequently, the centralized monitoring apparatus 1 transmits the pieces of information stored in the memory 106 to the refrigerant shortage prediction apparatus 2 via the transmission unit 107 and the network 3 (step 503).

[0033] (Operation Performed by Refrigerant Shortage Prediction Apparatus 2)

Fig. 6 is a flowchart illustrating an operation procedure performed by the refrigerant shortage prediction apparatus illustrated in Fig. 4.

When the refrigerant shortage prediction apparatus 2 receives pieces of operating data and refrigerant charging dates relating to the refrigerating and airconditioning apparatuses 10Ato 10D from the centralized monitoring apparatus 1 via the network 3 and the reception unit 201, the refrigerant shortage prediction apparatus 2 stores the pieces of operating data and the refrigerant charging dates of the refrigerating and air-conditioning apparatuses 10Ato 10D into the memory 206 (step 601).

When a prediction date is input as information for refrigerant shortage prediction via the prediction date input unit 205, the refrigerant shortage prediction apparatus 2 stores information on the prediction date into the memory 206 (step 602).

When the refrigerant shortage prediction apparatus 2 acquires, as information for refrigerant shortage prediction, a future outside air temperature announced by, for example, the Meteorological Agency as an outside air temperature on the prediction date via the prediction outside air temperature acquisition unit 202, the refrigerant shortage prediction apparatus 2 stores a prediction value of the outside air temperature on the prediction date into the memory 206 (step 603).

When a prediction value of an operating frequency of the compressor 11 on the prediction date is input as information for refrigerant shortage prediction via the prediction frequency input unit 203, the refrigerant shortage prediction apparatus 2 stores the prediction value of the operating frequency on the prediction date into the memory 206 (step 604).

[0034]

When a prediction value of an evaporating temperature is input as information for refrigerant shortage prediction via the prediction evaporating temperature input unit 204, the refrigerant shortage prediction apparatus 2 stores the prediction value of the evaporating temperature on the prediction date into the memory 206 (step 605).

The refrigerant shortage prediction apparatus 2 calculates elapsed time periods since refrigerant charging and estimated time periods since refrigerant charging of the refrigerating and air-conditioning apparatuses 10Ato 10D on the basis of the data collection dates and the refrigerant charging dates of the refrigerating and airconditioning apparatuses 10Ato 10D, and the prediction date that are stored in the memory 206 (step 606).

The refrigerant shortage prediction apparatus 2, by multivariate analysis on pieces of past data of the refrigerating and air-conditioning apparatuses 10Ato 10D, classifies the pieces of past data into a plurality of groups, identifies, among the plurality of groups, a group to which prediction data belongs, and calculates, on the basis of refrigerant shortage detection information of a piece of operating data corresponding to a piece of past data belonging to the identified group, a refrigerant shortage probability on the prediction date (step 607).

The refrigerant shortage prediction apparatus 2 outputs the refrigerant shortage probability calculated in step 607 to the display unit 209 (step 608).

[0035] (Operation Performed by Time Period Calculation Unit 207)

The process of step 606 illustrated in Fig. 6 will be described in detail.

Fig. 7 is a flowchart illustrating an operation procedure performed by the time period calculation unit illustrated in Fig. 4. Fig. 8 schematically illustrates an elapsed time period since refrigerant charging calculated in step 702 and an estimated time period since refrigerant charging calculated in step 704.

The time period calculation unit 207 checks whether an elapsed time period since refrigerant charging has been calculated for each of the pieces of operating data collected from the refrigerating and air-conditioning apparatuses 10Ato 10D (step 701).

In step 701, when there is, among the already collected pieces of operating data, a piece of operating data for which an elapsed time period since refrigerant charging has not been calculated, the time period calculation unit 207 calculates an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to a data collection date of the piece of operating data, using the data collection date as a reference (step 702).

[0036]

On the other hand, in step 701, when an elapsed time period since refrigerant charging has been calculated for all of the pieces of operating data already collected from the refrigerating and air-conditioning apparatuses 10Ato 10D, the time period calculation unit 207 checks whether an estimated time period since refrigerant charging of each ofthe refrigerating and air-conditioning apparatuses 10Ato 10D has been calculated (step 703).

In step 703, when there is, among the refrigerating and air-conditioning apparatuses 10Ato 10D, a refrigerating and air-conditioning apparatus for which an estimated time period since refrigerant charging has not been calculated, the time period calculation unit 207 calculates, for the refrigerating and air-conditioning apparatus, an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference (step 704).

For the refrigerating and air-conditioning apparatuses 10Ato 10D, when respective most recent refrigerant charging dates are different, an elapsed time period since refrigerant charging illustrated in Fig. 8 differs for each of pieces of operating data of the refrigerating and air-conditioning apparatuses 10A to 10D. Also, for the refrigerating and air-conditioning apparatuses 10Ato 10D, when respective most recent refrigerant charging dates are different, an estimated time period since refrigerant charging illustrated in Fig. 8 differs for each of the refrigerating and airconditioning apparatuses 10Ato 10D. In step 703, when an estimated time period since refrigerant charging of each of the refrigerating and air-conditioning apparatuses 10Ato 10D has been calculated, the time period calculation unit 207 ends the calculation process for calculating a time period.

[0037] (Operation Performed by Probability Calculation Unit 208)

The processes of steps 607 and 608 illustrated in Fig. 6 will be described in detail.

Fig. 9 is a flowchart illustrating an operation procedure performed by the probability calculation unit illustrated in Fig. 4. Here, descriptions are given for the case where, as a method of, by multivariate analysis on multiple pieces of past data, classifying the multiple pieces of past data into a plurality of groups, a method obtained by combining principal component analysis and clustering is used.

The probability calculation unit 208 prepares operating frequencies of compressors 11, outside air temperatures, and evaporating temperatures of pieces of operating data collected from the refrigerating and air-conditioning apparatuses 10A to 10D, elapsed time periods since refrigerant charging, and pieces of refrigerant shortage detection information (step 801).

Furthermore, for each of the refrigerating and air-conditioning apparatuses 10A to 10D, the probability calculation unit 208 prepares, as input data for calculating a refrigerant shortage probability, a target setting value of an operating frequency of the compressor 11, a prediction value of an outside air temperature, a target setting value of an evaporating temperature, and an estimated time period since refrigerant charging (step 802).

[0038]

The probability calculation unit 208 performs principal component analysis and clustering on the basis of, as input data, pieces of past data prepared in step 801 and pieces of prediction data prepared in step 802 to calculate a data distribution using a first principal component and a second principal component as references and to classify the pieces of past data into clusters (step 803).

The probability calculation unit 208 adds the pieces of refrigerant shortage detection information to the respective pieces of past data plotted in a coordinate system with the first principal component and the second principal component (step

804).

The probability calculation unit 208 calculates, for each cluster, the proportion of the number of pieces of data with refrigerant shortage detected to the number of all pieces of data in the cluster as a refrigerant shortage probability (step 805).

The probability calculation unit 208 causes the display unit 209 to output, as prediction results on a refrigerant shortage, refrigerant shortage probabilities of clusters to which the pieces of prediction data prepared in step 802 belong (step 806). [0039]

A specific example of principal component analysis and clustering illustrated in Fig. 9 will be described.

Fig. 10 illustrates a specific example ofthe procedure illustrated in Fig. 9. Here, the case where a refrigerant shortage probability is obtained for any one refrigerating and air-conditioning apparatus of the refrigerating and air-conditioning apparatuses 10Ato 10D will be described, and any one refrigerating and airconditioning apparatus is denoted by 10.

In Fig. 10, an upper portion illustrates an example of past data and prediction data, a middle portion illustrates an analysis method, and a lower portion illustrates a prediction result on a refrigerant shortage. In Fig. 10, an arrow from the upper portion to the middle portion refers to performing, on the basis of pieces of data illustrated in the upper portion, the analysis method illustrated in the middle portion, and an arrow from the middle portion to the lower portion refers to the obtainment of the prediction result illustrated in the lower portion by the analysis method illustrated in the middle portion.

[0040]

In the example illustrated in the upper portion of Fig. 10, there are nine pieces of past data, and data collection dates of the pieces of past data are the first days of months ranging from January to September in 2015. Fig. 10 illustrates pieces of past data collected from one refrigerating and air-conditioning apparatus, and nine pieces of past data are collected for each of the refrigerating and air-conditioning apparatuses 10Ato 10D. In a table illustrated in Fig. 10, as past data, in addition to an operating frequency of the compressor 11, an outside air temperature, an evaporating temperature, and an elapsed time period since refrigerant charging, refrigerant shortage detection information is described corresponding to a data collection date.

Prediction data illustrated in Fig. 10 refers to an example of prediction data of the any one refrigerating and air-conditioning apparatus 10 of the refrigerating and air-conditioning apparatuses 10Ato 10D. Here, for the prediction data, a prediction date is January 1,2016. As illustrated in Fig. 10, for the refrigerating and airconditioning apparatus 10, a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference, is calculated as 9 months.

In a table illustrated in Fig. 10, as prediction data, a target setting value of an operating frequency of the compressor 11, a prediction value of an outside air temperature, a target setting value of an evaporating temperature, and an estimated time period since refrigerant charging are described corresponding to the prediction date.

In the case where the manager does not specify a refrigerating and airconditioning apparatus, the time period calculation unit 207 calculates estimated time periods since refrigerant charging for the refrigerating and air-conditioning apparatuses 10Ato 10D and prepares pieces of prediction data for the refrigerating and air-conditioning apparatuses 10A to 10D. However, for the sake of simplicity of explanation, the case of one piece of prediction data will be described below. [0041]

The probability calculation unit 208 performs principal component analysis and clustering on the basis of the pieces of past data and the prediction data illustrated in the upper portion of Fig. 10 (see the middle portion of Fig. 10).

The probability calculation unit 208 performs principal component analysis on the pieces of past data to determine a first principal component and a second principal component. Although the number of principal components may be three or more, a reduction in the number of principal components to two makes it possible to prevent calculation processes including clustering from becoming complex.

The probability calculation unit 208 defines a two-dimensional coordinate system with the first principal component and the second principal component as axes and generates a data distribution obtained by plotting the pieces of past data in the two-dimensional coordinate system. Here, for the sake of simplicity of explanation, 10 pieces of past data among 36 pieces of past data are plotted. [0042]

Subsequently, the probability calculation unit 208 classifies the 10 pieces of past data into clusters. Here, as illustrated in the middle portion of Fig. 10, the 10 pieces of past data is classified into two clusters, that is, first and second clusters. In each of the first and second clusters, five pieces of past data are included (step 803 illustrated in Fig. 9).

Subsequently, the probability calculation unit 208 adds pieces of refrigerant shortage detection information to the respective 10 pieces of past data (step 804 illustrated in Fig. 9). In Fig. 10, data with refrigerant shortage detected is represented by diagonal lines, and data with refrigerant shortage not detected is represented by a pattern of dots.

The probability calculation unit 208 calculates refrigerant shortage probabilities for the respective first and second clusters. The first cluster has, among the five pieces of past data, four pieces of past data with refrigerant shortage not detected, and one piece of past data with refrigerant shortage detected, and thus the probability calculation unit 208 calculates a refrigerant shortage detection probability as (1/5) x 100 = 20%. On the other hand, the second cluster has, among the five pieces of past data, four pieces of past data with refrigerant shortage detected, and one piece of past data with refrigerant shortage not detected, and thus the probability calculation unit 208 calculates a refrigerant shortage detection probability as (4/5) x 100 = 80% (step 805 illustrated in Fig. 9).

[0043]

As described above, the probability calculation unit 208 calculates refrigerant shortage probabilities of the respective first and second clusters, and then assigns the prediction data to the data distribution. In the example illustrated in Fig. 10, the prediction data belongs to the second cluster, and the probability calculation unit 208 causes the display unit 209 to output, as a prediction result of the prediction data, the refrigerant shortage probability of the second cluster (step 806 illustrated in Fig. 9).

The lower portion of Fig. 10 illustrates an example of the prediction result displayed by the display unit 209. In the example illustrated in the lower portion of Fig. 10, PROBABILITY OF REFRIGERANT SHORTAGE IS 80% is described in a field for REFRIGERANT SHORTAGE PROBABILITY of the prediction data illustrated in the upper portion of Fig. 10. In the case where there are multiple pieces of prediction data, the probability calculation unit 208 causes the display unit 209 to output the same number of prediction results, such as the prediction result illustrated in the lower portion of Fig. 10, as the number of the pieces of prediction data. [0044]

The manager can see a refrigerant shortage probability with reference to a prediction result output by the refrigerant shortage prediction apparatus 2 and determine whether to charge refrigerant. For example, the manager sees a prediction result indicating that a refrigerant shortage probability is 80% as illustrated in Fig. 10 and then determines that refrigerant is to be charged, thereby making it possible to prevent a refrigerant shortage from occurring.

On the other hand, for example, when the refrigerant shortage probability is not less than 50% but less than 80%, the manager may be unable to determine whether to charge refrigerant. At this stage, some managers may charge refrigerant, and some managers still may not charge refrigerant. In the case where the manager does not charge refrigerant with the refrigerant shortage probability being in the range from not less than 50% to less than 80%, even when a refrigerant shortage occurs, refrigerant shortage detection information for this case will be useful as learning data for a future refrigerant shortage prediction process. Consequently, the effect of further increasing the accuracy of refrigerant shortage prediction can be expected. [0045]

The probability calculation unit 208 may cause the display unit 209 to output, in tabular form and graph form, multiple pieces of operating data including, as a temperature in a room to be subjected to air-conditioning by the refrigerating and airconditioning apparatuses 10Ato 10D, a temperature in the cold storage warehouse 8, and diagnostic information including a plurality of prediction results on a refrigerant shortage on different prediction dates. An example of information output from the display unit 209 will be described.

[0046]

Fig. 11A illustrates an example of a table output from the display unit illustrated in Fig. 4.

In the table indicating operating data illustrated in Fig. 11 A, changes in the temperature in the cold storage warehouse 8 are described in place of the elapsed time periods since refrigerant charging in the pieces of past data illustrated in the upper portion of Fig. 10.

In diagnostic information illustrated in Fig. 11 A, November 1 and December 1, 2015 are added as prediction dates to the prediction data illustrated in the lower portion of Fig. 10, and refrigerant shortage probabilities on the added prediction dates are also described.

Fig. 11B illustrates an example of a graph output from the display unit illustrated in Fig. 4. In the graph illustrated in Fig. 11B, the horizontal axis represents time, and the vertical axis represents the temperature in the cold storage warehouse 8 of the operating data illustrated in Fig. 11Aand refrigerant shortage probability. Values indicated in the graph are not limited to the values illustrated in Fig. 11B.

As illustrated in Fig. 11A and Fig. 11B, multiple pieces of operating data and a plurality of prediction results are indicated chronologically in the table or the graph, thus making it easier for the manager to understand a history of an operating state and to identify a time when a refrigerant shortage occurs.

[0047]

Although, in Embodiment 1 described above, as data related to an operating state, evaporating temperatures in the heat exchangers 17 of the load-side units 5Ato

5D are included in pieces of operating data, condensing temperatures in the heat exchangers 12 of the heat source units 4Ato 4D may be used in place of the evaporating temperatures in the heat exchangers 17 of the load-side units 5Ato 5D. Each heat exchanger 12 exchanges heat with outside air and thus is closely related to an outside air temperature. Hence, a condensing temperature in the heat exchanger 12 can also be considered suitable for a parameter for detecting a refrigerant shortage.

[0048]

Furthermore, although, in Embodiment 1, the case where each of the refrigerating and air-conditioning apparatuses 10Ato 10D is a refrigeration machine has been described, each of the refrigerating and air-conditioning apparatuses 10Ato 10D is not limited to a refrigeration machine. Each of the refrigerating and airconditioning apparatuses 10Ato 10D may be an air-conditioning apparatus capable of performing both heating operation and cooling operation. In this case, among evaporating temperatures and condensing temperatures of the heat exchanger 12 of a heat source unit and the heat exchanger 17 of each load-side unit, any temperature may be used as data related to an operating state.

[0049]

Furthermore, although, in Embodiment 1, the case where pieces of operating data and pieces of information on refrigerant charging dates are collected from a plurality of refrigerating and air-conditioning apparatuses has been described, the number of refrigerating and air-conditioning apparatuses to be managed may be one. Even in the case of one refrigerating and air-conditioning apparatus, when many pieces of operating data are accumulated, the probability calculation unit 208 can calculate a refrigerant shortage probability by multivariate analysis as described above on the basis of the accumulated pieces of operating data.

[0050]

A refrigerant shortage prediction apparatus according to Embodiment 1 is the refrigerant shortage prediction apparatus 2 connected, via the network 3, to the centralized monitoring apparatus 1 configured to collect operating data that is data on an operating state from a refrigerating and air-conditioning apparatus including a heat source unit including the compressor 11 and the heat exchanger 12. The refrigerant shortage prediction apparatus includes the reception unit 201 configured to receive, from the centralized monitoring apparatus 1, multiple pieces of operating data including an operating frequency of the compressor 11, an outside air temperature that is a temperature of outside air with which the heat exchanger 12 exchanges heat, an evaporating temperature or a condensing temperature in the heat exchanger 12, refrigerant shortage detection information that is information on whether a refrigerant shortage is detected, and a data collection date, and a refrigerant charging date that is a date on which refrigerant is charged into the refrigerating and air-conditioning apparatus, the input unit 211 configured to receive, as information for refrigerant shortage prediction, input data including a prediction date, a prediction value of the outside air temperature on the prediction date, a target setting value of the operating frequency of the compressor 11 on the prediction date, and a target setting value of the evaporating temperature or the condensing temperature on the prediction date, the time period calculation unit 207 configured to calculate, on the basis of the data collection date, the prediction date, and the refrigerant charging date, an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the data collection date, using the data collection date as a reference, for each piece of operating data, and an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference, the probability calculation unit 208 configured to, by multivariate analysis on multiple pieces of past data including the operating frequency, the outside air temperature, and the evaporating temperature or the condensing temperature that are included in each piece of operating data, and the elapsed time period since refrigerant charging, classify the multiple pieces of past data into a plurality of groups, configured to identify, among the plurality of groups, a group to which prediction data including the input data and the estimated time period since refrigerant charging belongs, and configured to calculate, on the basis of the refrigerant shortage detection information of a piece of operating data corresponding to a piece of past data belonging to the identified group, a refrigerant shortage probability that is an index indicating whether a refrigerant shortage occurs on the prediction date, and the display unit 209 configured to output the refrigerant shortage probability calculated by the probability calculation unit 208.

[0051]

In Embodiment 1, the probability calculation unit 208, by multivariate analysis on pieces of data including operating data on an operating state of a refrigerating and air-conditioning apparatus, reduces disturbances included in the pieces of data, classifies the pieces of data into groups by the likelihood of a refrigerant shortage, and calculates a refrigerant shortage probability of a group to which a prediction date belongs, thus making it possible to predict the occurrence of a refrigerant shortage before the refrigerant shortage actually occurs in the refrigerating and air-conditioning apparatus.

[0052]

Furthermore, in Embodiment 1, the probability calculation unit 208 may perform principal component analysis on multiple pieces of past data to determine a twodimensional coordinate system with first and second principal components as axes, may perform clustering on the multiple pieces of past data plotted in the twodimensional coordinate system to classify the multiple pieces of past data into a plurality of clusters, may calculate, for each cluster, the proportion of the number of pieces of data of refrigerant shortage detection information including information that a refrigerant shortage has been detected to the number of all pieces of data in the cluster as a refrigerant shortage probability, and may cause the display unit 209 to output the refrigerant shortage probability of a cluster to which prediction data belongs.

Refrigerating and air-conditioning apparatuses are different in terms of the environment of the place where each refrigerating and air-conditioning apparatus is installed and a load condition, for example, and thus it is difficult to universally obtain parameters and disturbances for detecting a refrigerant shortage. On the other hand, in Embodiment 1, as described above, the probability calculation unit 208 performs principal component analysis to extract two principal components from pieces of past data including various parameter values related to an operating state of the refrigerating and air-conditioning apparatus and classifies the pieces of past data into clusters using a two-dimensional coordinate system with the two principal components as a reference. Thus, most appropriate principal components for detecting a refrigerant shortage are extracted corresponding to an operating state of the refrigerating and air-conditioning apparatus, and pieces of past data are appropriately classified into clusters on the basis of the extracted principal components.

[0053]

Furthermore, in Embodiment 1, the probability calculation unit 208 may cause the display unit 209 to output, in tabular form or graph form, multiple pieces of operating data including a temperature in a room to be subjected to air-conditioning by the refrigerating and air-conditioning apparatus, and diagnostic information including, as prediction results on a refrigerant shortage in the refrigerating and airconditioning apparatus, a plurality of refrigerant shortage probabilities on different prediction dates.

In this case, the multiple pieces of operating data and the plurality of prediction results are indicated chronologically in a table or a graph, thus making it easier for the manager to understand a history of an operating state and to identify a time when a refrigerant shortage occurs.

[0054]

Furthermore, in Embodiment 1, the reception unit may receive, from the centralized monitoring apparatus, multiple pieces of operating data and information on a refrigerant charging date of each of a plurality of refrigerating and airconditioning apparatuses.

In this case, in the probability calculation unit, by multivariate analysis on multiple pieces of past data of each of the plurality of refrigerating and air-conditioning apparatuses, the accuracy of analysis increases as the number of pieces of data to be analyzed increases.

Reference Signs List [0055] centralized monitoring apparatus 2 heat source unit refrigerant pipe 8 refrigerating and air-conditioning apparatus

14, 18, 20 apparatus 3 network 4A to 4D to 6E signal line 7Ato7D to 10D exchanger 13, 19 fan expansion valve 16 controller operating state reception unit 104 detection information reception unit 10 108,201 reception unit 110,210 temperature acquisition unit 203 evaporating temperature input unit refrigerant shortage prediction

5Ato5D load-side unit 6A cold storage warehouse 10, 10A compressor 12, 17 heat temperature sensor 15 electronic

101,211 input unit 102,209 display unit 103 refrigerant charging date reception unit 105

106,206 memory 107 transmission unit controller 202 prediction outside air prediction frequency input unit 204 prediction 205 prediction date input unit 207 time

Claims (6)

1. CLAIMS [Claim 1]

   A refrigerant shortage prediction apparatus configured to be connected, via a network, to a centralized monitoring apparatus configured to collect operating data that is data on an operating state from a refrigerating and air-conditioning apparatus including a heat source unit including a compressor and a heat exchanger, the refrigerant shortage prediction apparatus comprising:

   a reception unit configured to receive, from the centralized monitoring apparatus, multiple pieces of the operating data including an operating frequency of the compressor, an outside air temperature that is a temperature of outside air with which the heat exchanger exchanges heat, an evaporating temperature or a condensing temperature in the heat exchanger, refrigerant shortage detection information that is information on whether a refrigerant shortage is detected, and a data collection date that is a date on which these pieces of data are collected, and a refrigerant charging date that is a date on which refrigerant is charged into the refrigerating and air-conditioning apparatus;

   an input unit configured to receive, as information for refrigerant shortage prediction, input data including a prediction date, a prediction value of the outside air temperature on the prediction date, a target setting value of the operating frequency of the compressor on the prediction date, and a target setting value of the evaporating temperature or the condensing temperature on the prediction date;

   a time period calculation unit configured to calculate, on a basis of the data collection date, the prediction date, and the refrigerant charging date, an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the data collection date, using the data collection date as a reference, for each piece of the operating data, and an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference;

   a probability calculation unit configured to, by multivariate analysis on multiple pieces of past data including the operating frequency, the outside air temperature, and the evaporating temperature or the condensing temperature that are included in each piece of the operating data, and the elapsed time period since refrigerant charging, classify the multiple pieces of past data into a plurality of groups, configured to identify, among the plurality of groups, a group to which prediction data including the input data and the estimated time period since refrigerant charging belongs, and configured to calculate, on a basis of the refrigerant shortage detection information of a piece of the operating data corresponding to a piece of the multiple pieces of past data belonging to the identified group, a refrigerant shortage probability that is an index indicating whether a refrigerant shortage occurs on the prediction date; and a display unit configured to output the refrigerant shortage probability calculated by the probability calculation unit.
2. [Claim 2]

   The refrigerant shortage prediction apparatus of claim 1, wherein the probability calculation unit is configured to perform principal component analysis on the multiple pieces of past data to determine a two-dimensional coordinate system with first and second principal components as axes, perform clustering on the multiple pieces of past data plotted in the two-dimensional coordinate system to classify the multiple pieces of past data into a plurality of clusters, calculate, for each cluster, a proportion of a number of pieces of data of the refrigerant shortage detection information including information that a refrigerant shortage is detected to a number of all pieces of data in the cluster as the refrigerant shortage probability, and cause the display unit to output the refrigerant shortage probability of a cluster to which the prediction data belongs.
3. [Claim 3]

   The refrigerant shortage prediction apparatus of claim 1 or 2, wherein the probability calculation unit is configured to cause the display unit to output, in tabular form or graph form, multiple pieces of the operating data including a temperature in a room to be subjected to air-conditioning by the refrigerating and air-conditioning apparatus, and diagnostic information including, as prediction results on a refrigerant shortage in the refrigerating and air-conditioning apparatus, a plurality of the refrigerant shortage probabilities on different prediction dates.
4. [Claim 4]

   The refrigerant shortage prediction apparatus of any one of claims 1 to 3, wherein the reception unit is configured to receive, from the centralized monitoring apparatus, multiple pieces of the operating data and information on a refrigerant charging date of each of a plurality of the refrigerating and air-conditioning apparatuses.
5. [Claim 5]

   A refrigerant shortage prediction method performed by a refrigerant shortage prediction apparatus configured to be connected, via a network, to a centralized monitoring apparatus configured to collect operating data that is data on an operating state from a refrigerating and air-conditioning apparatus including a heat source unit including a compressor and a heat exchanger, the refrigerant shortage prediction method comprising:

   receiving, from the centralized monitoring apparatus, multiple pieces of the operating data including an operating frequency of the compressor, an outside air temperature that is a temperature of outside air with which the heat exchanger exchanges heat, an evaporating temperature or a condensing temperature in the heat exchanger, refrigerant shortage detection information that is information on whether a refrigerant shortage is detected, and a data collection date that is a date on which these pieces of data are collected, and a refrigerant charging date that is a date on which refrigerant is charged into the refrigerating and air-conditioning apparatus;

   in response to input of, as information for refrigerant shortage prediction, input data including a prediction date, a prediction value of the outside air temperature on the prediction date, a target setting value of the operating frequency of the compressor on the prediction date, and a target setting value of the evaporating temperature or the condensing temperature on the prediction date, calculating, on a basis of the data collection date, the prediction date, and the refrigerant charging date, an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the data collection date, using the data collection date as a reference, for each piece of the operating data, and an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference;

   by multivariate analysis on multiple pieces of past data including the operating frequency, the outside air temperature, and the evaporating temperature or the condensing temperature that are included in each piece of the operating data, and the elapsed time period since refrigerant charging, classifying the multiple pieces of past data into a plurality of groups, identifying, among the plurality of groups, a group to which prediction data including the input data and the estimated time period since refrigerant charging belongs, and calculating, on a basis of the refrigerant shortage detection information of a piece of the operating data corresponding to a piece of the multiple pieces of past data belonging to the identified group, a refrigerant shortage probability that is an index indicating whether a refrigerant shortage occurs on the prediction date; and outputting the refrigerant shortage probability.
6. [Claim 6]

   A program for causing a computer connected, via a network, to a centralized monitoring apparatus configured to collect operating data that is data on an operating state from a refrigerating and air-conditioning apparatus including a heat source unit including a compressor and a heat exchanger to execute steps comprising:

   receiving, from the centralized monitoring apparatus, multiple pieces of the operating data including an operating frequency of the compressor, an outside air temperature that is a temperature of outside air with which the heat exchanger exchanges heat, an evaporating temperature or a condensing temperature in the heat exchanger, refrigerant shortage detection information that is information on whether a refrigerant shortage is detected, and a data collection date that is a date on which these pieces of data are collected, and a refrigerant charging date that is a date on which refrigerant is charged into the refrigerating and air-conditioning apparatus;

   in response to input of, as information for refrigerant shortage prediction, input data including a prediction date, a prediction value of the outside air temperature on the prediction date, a target setting value of the operating frequency of the compressor on the prediction date, and a target setting value of the evaporating temperature or the condensing temperature on the prediction date, calculating, on a basis of the data collection date, the prediction date, and the refrigerant charging date, an elapsed time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the data collection date, using the data collection date as a reference, for each piece of the operating data, and an estimated time period since refrigerant charging that is a time period from a most recent refrigerant charging date to the prediction date, using the prediction date as a reference;

   by multivariate analysis on multiple pieces of past data including the operating frequency, the outside air temperature, and the evaporating temperature or the condensing temperature that are included in each piece of the operating data, and the elapsed time period since refrigerant charging, classifying the multiple pieces of past data into a plurality of groups, identifying, among the plurality of groups, a group to which prediction data including the input data and the estimated time period since refrigerant charging belongs, and calculating, on a basis of the refrigerant shortage detection information of a piece of the operating data corresponding to a piece of the multiple pieces of past data belonging to the identified group, a refrigerant shortage probability that is an index indicating whether a refrigerant shortage occurs on the prediction date; and