Learning device, monitoring device, and air conditioning system

Abstract

The operating data (DA) of an air conditioner (40) includes a first data group (D1A1) and a second data group (D2A1) that does not overlap with the first data group (D1A1). The learning device (4) comprises: a first calculation unit (110A) configured to calculate a first feature quantity (F1A1) from a first data group (D1A1) of the air conditioner during a learning period; and a learning unit (120) configured to perform supervised learning by using the second data group (D2A1), by using the first feature quantity (F1A1) obtained through calculation by the first calculation unit (110A) as correct data, and generate a first inference model (M1) that infers a first normal range of the first feature quantity (F1A1) from a second data group (D2A1).

Description

Learning devices, monitoring devices and air conditioning systems

[0001] The present disclosure relates to a learning device, a monitoring device, and an air conditioning system.

[0002] Conventionally, methods have been considered for detecting refrigerant leaks in cooling systems based on a value that serves as an indicator of the amount of refrigerant filled in the system. One example of a refrigerant amount determination device that makes it easy to determine the amount of refrigerant is disclosed in Japanese Patent No. 6,791,429.

[0003] This refrigerant quantity determination device includes a calculation unit that calculates a refrigerant quantity index value from operating data of the air conditioning system, an inference unit that infers information regarding the correction of the refrigerant quantity index value using at least one of the operating data and the calculated refrigerant quantity index value, and a correction model, and a determination unit that determines the refrigerant quantity of the air conditioning system based on the information regarding the correction of the refrigerant quantity index value.

[0004] Patent No. 6791429

[0005] The refrigerant quantity determination device disclosed in Patent No. 6791429 had the problem of being unable to accurately grasp the impact that a decrease in refrigerant quantity has on air conditioning capacity, making it difficult to determine when maintenance should be performed. For example, if the refrigerant leakage rate is as low as 10%, the refrigerant quantity indicator may indicate that refrigerant is leaking, but the air conditioning capacity does not decrease significantly, making it difficult to determine whether to urgently investigate the leak location or maintain the status quo and wait for a while to see how it goes.

[0006] The present disclosure has been made to solve the above-mentioned problems, and its purpose is to provide a learning device that can obtain an inference model that is useful for improving the accuracy of anomaly detection in air conditioning systems.

[0007] The present disclosure relates to a learning device that learns the state of an air conditioning device in which a refrigerant circulates. The operating data of the air conditioning device includes a first data group and a second data group that does not include the same data elements as the first data group. The learning device includes a first calculation unit configured to calculate a first feature value from the first data group of the air conditioning device during a learning period, and a learning unit configured to perform supervised learning using the second data group with the first feature value as ground truth data to generate a first inference model that infers a first normal range of the first feature value from the second data group.

[0008] According to the learning device disclosed herein, a model is generated that infers the normal range of features from the behavior of driving data other than data that can be directly used to calculate features, thereby increasing the likelihood of early and accurate abnormality detection.

[0009] 5 is a block diagram showing the configuration of an air conditioning system according to embodiment 1. It is a functional block diagram showing the configuration of the air conditioning apparatus 40 of FIG. 1. It is a diagram showing an example of operating data reflecting the state of the air conditioning apparatus 40 of FIG. 1. It is a block diagram showing the configuration of the anomaly detection system 1 of FIG. 1. It is a flowchart for explaining the content of the learning process executed by the anomaly detection system 1. It is a diagram showing an example of a neural network Nw1 included in the inference model M1 constructed in step S3 of FIG. 5. It is a flowchart for explaining the content of the malfunction detection process executed by the anomaly detection system 1. It is a diagram showing an example of changes over time in the inferred values ​​and calculated values ​​of the heat exchange capacity Qo and the degree of subcooling SC of the refrigerant at the outlet of the condenser. It is a graph showing the relationship between the amount of refrigerant change and the amount of change in the degree of subcooling. It is a graph showing the relationship between the decrease in refrigerant amount, the performance degradation operation rate, and the performance degradation rate. It is a diagram showing an example of a trend graph displayed to a user.

[0010] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are designated by the same reference numerals, and their description will not be repeated in principle.

[0011] Embodiment 1. Fig. 1 is a block diagram showing the configuration of an air conditioning system according to embodiment 1. The air conditioning system 1000 includes an anomaly detection system 1 and an air conditioning device 40 whose state is monitored by the anomaly detection system 1. As shown in Fig. 1, the anomaly detection system 1 is connected to the air conditioning device 40 via a network 900.

[0012] The anomaly detection system 1 is configured to include a CPU (Central Processing Unit) 2, memory 3 (ROM (Read Only Memory) and RAM (Random Access Memory)), an input / output buffer (not shown), etc. The CPU 2 deploys a program stored in the ROM into the RAM, etc. and executes it. The program stored in the ROM is a program in which the processing procedures of the anomaly detection system 1 are written. The anomaly detection system 1 monitors each device in the air conditioning device 40 in accordance with these programs. This control is not limited to processing by software, and can also be processed by dedicated hardware (electronic circuits). The anomaly detection system 1 may be built on a server connected to the network 900.

[0013] The air conditioning apparatus 40 includes a plurality of indoor units 20, an outdoor unit 10, and a control device 30. Each of the plurality of indoor units 20 is disposed in an indoor space and connected to the outdoor unit 10 by liquid piping and gas piping through which a refrigerant passes. The outdoor unit 10 is disposed in a space outside the indoor space (outdoor space). Note that the number of indoor units 20 included in the air conditioning apparatus 40 may be one or two.

[0014] The outdoor unit 10 includes a compressor, an outdoor heat exchanger, and an outdoor fan. Each of the indoor units 20 includes an expansion valve and an indoor heat exchanger. A refrigerant is supplied to each of the indoor units 20 from the compressor included in the outdoor unit 10. The refrigerant circulates between the outdoor unit 10 and each of the indoor units 20.

[0015] The control device 30 includes a thermostat and performs integrated control of the air conditioning device 40. The control device 30 is connected to the anomaly detection system 1 via a network 900. The network 900 includes the Internet and a cloud system. The network 900 may be a LAN (Local Area Network).

[0016] The control device 30 is configured to include a CPU 32, memory 33 (ROM and RAM), an input / output buffer (not shown), etc. The CPU 32 loads a program stored in the ROM into the RAM, etc., and executes it. The program stored in the ROM is a program in which the processing procedures of the control device 30 are written. The control device 30 controls each device in the air conditioning device 40 in accordance with these programs. This control is not limited to processing by software, but can also be processed by dedicated hardware (electronic circuitry).

[0017] Fig. 2 is a functional block diagram showing the configuration of the air conditioning apparatus 40 of Fig. 1. As shown in Fig. 2, the outdoor unit 10 includes a compressor 11, an outdoor heat exchanger 12, a four-way valve 13, an outdoor fan 14, an accumulator 15, temperature sensors 54 to 56, pressure sensors 61 and 63, and a humidity sensor 57.

[0018] Each of the indoor units 20 includes an expansion valve 21, an indoor heat exchanger 22, an indoor fan 23, and temperature sensors 51 to 53. The expansion valve 21 includes, for example, an LEV (Linear Expansion Valve).

[0019] The operating modes of the air conditioner 40 include a heating mode, a cooling mode, and a defrost mode. In the heating mode, the four-way valve 13 connects the discharge port of the compressor 11 to the indoor heat exchanger 22, and also connects the outdoor heat exchanger 12 to a refrigerant inlet of the accumulator 15. In the heating mode, the refrigerant circulates through the compressor 11, the four-way valve 13, the indoor heat exchanger 22, the expansion valve 21, and the outdoor heat exchanger 12, in that order. In the cooling mode and the defrost mode, the four-way valve 13 connects the discharge port of the compressor 11 to the outdoor heat exchanger 12, and also connects the indoor heat exchanger 22 to a refrigerant inlet of the accumulator 15. In the cooling mode and the defrost mode, the refrigerant circulates through the compressor 11, the four-way valve 13, the outdoor heat exchanger 12, the expansion valve 21, and the indoor heat exchanger 22, in that order.

[0020] The temperature sensor 51 measures the temperature of the air drawn into the indoor heat exchanger 22 (room temperature TH1ic) and outputs the measured temperature to the control device 30. The temperature sensors 52 and 53 measure the temperatures of the refrigerant before and after passing through the indoor heat exchanger 22 (indoor liquid temperature TH2ic, indoor gas temperature TH3ic), respectively, and output the measured temperatures to the control device 30.

[0021] The temperature sensor 54 measures the temperature (discharge temperature TH4) of the refrigerant discharged from the compressor 11 and outputs the discharge temperature to the control device 30. The temperature sensor 55 measures the temperature (suction temperature TH5) of the refrigerant sucked into the compressor 11 via the accumulator 15 and outputs the suction temperature to the control device 30. The temperature sensor 56 measures the temperature TH3 of the liquid refrigerant in the pipe connecting the outdoor heat exchanger 12 and the liquid pipe 41 and outputs the temperature to the control device 30.

[0022] The pressure sensor 61 measures the pressure of the refrigerant discharged from the compressor 11 (discharge pressure HS1) and outputs the discharge pressure HS1 to the control device 30. The pressure sensor 63 measures the pressure of the refrigerant sucked into the compressor 11 (suction pressure LS) and outputs the suction pressure LS to the control device 30.

[0023] Below, control during cooling will be described as a representative example. The control device 30 controls the operating frequency fCOMP of the compressor 11 so that the intake saturated gas temperature reaches a target temperature, thereby controlling the amount of refrigerant discharged per unit time from the compressor 11. The control device 30 controls the opening degree Li of the expansion valve 21 so that the degree of superheat SH (= TH3ic - TH2ic) of the refrigerant at the outlet of the indoor heat exchanger 22 reaches a target value. The control device 30 controls the four-way valve 13 to follow the flow path indicated by the solid line, thereby switching the refrigerant circulation direction. The control device 30 controls the rotational frequency fFANo of the outdoor fan 14 to control the fan's airflow per unit time so that the discharge saturated gas temperature reaches a target value. The control device 30 controls the rotational frequency fFANi of the indoor fan 23 to achieve the airflow volume set by the user. The control device 30 associates operating data reflecting the state of the air conditioning system with the measurement time and transmits it to the anomaly detection system 1.

[0024] Fig. 3 is a diagram showing an example of operating data reflecting the state of the air conditioning apparatus 40 of Fig. 1. As shown in Fig. 3, the operating data includes, for example, the outdoor air temperature, the discharge temperature (TH4), the evaporation temperature (TH2ic), the condensation temperature, the suction temperature (TH1ic), the discharge temperature, the high pressure (HS1), the low pressure (LS), the operating frequency (fCOMP) of the compressor 11, the opening of the expansion valve 21, the operating mode, the operating state (operating, stopped, or standby), the rotation speeds (fFANo, fFANi) of the outdoor fan 14 and the indoor fan 23, the indoor space temperature (set temperature) set by the user, the current value of the inverter of the compressor 11, the voltage value of the inverter, the temperature of the heat sink included in the outdoor unit 10, and the temperature (liquid pipe temperature TH3) of the liquid pipe (piping through which liquid refrigerant flows) connecting the outdoor unit 10 and the indoor unit 20. The operating frequency of the compressor 11, the opening of the expansion valve 21, and the rotation speed of the outdoor fan 14 are basic manipulated variables in VRF (Variable Refrigerant Flow) control.

[0025] The environment in which the air conditioning apparatus 40 operates may have characteristics specific to that environment (for example, the length of the refrigerant piping, the type of indoor units 20, the number of indoor units 20, and the difference in elevation between the indoor units 20 and the outdoor unit 10). For this reason, the judgment criteria (for example, threshold values) for detecting an abnormality in the air conditioning apparatus 40 may differ depending on the environment in which the air conditioning apparatus 40 operates. Therefore, if common judgment criteria are used regardless of the environment in which the air conditioning apparatus 40 operates, the accuracy of abnormality detection in the air conditioning apparatus 40 may decrease.

[0026] Therefore, in the anomaly detection system 1, if a normal state is confirmed during a trial run after installation of the air conditioning system, the system is considered to be in a normal state for a certain period thereafter. A trained model is then generated that learns the relationship between the operating data of the air conditioning device 40 during that certain period and the normal values ​​of the state index values ​​(specific parameters) that represent the state of the air conditioning device 40 and correspond to that operating data. By using this trained model, it becomes possible to detect anomalies in the air conditioning device 40 using judgment criteria that are suited to the environment in which the air conditioning device 40 operates. As a result, the accuracy of anomaly detection in the air conditioning system can be improved.

[0027] Fig. 4 is a block diagram showing the configuration of the anomaly detection system 1 of Fig. 1. The anomaly detection system 1 acquires driving data for each of a plurality of consecutive driving periods (for example, one day, one week, or one month), and constructs trained inference models M1 and M2 using training data including the driving data. The consecutive driving periods may be the same period (a predetermined period) or different periods.

[0028] 4 , the anomaly detection system 1 includes a learning device 4 and a monitoring device 5. The learning device 4 includes a calculation unit 110A and a learning unit 120. The monitoring device 5 includes a calculation unit 110B, an inference unit 220, a determination unit 310, and a display unit 320.

[0029] The calculation unit 110A and the calculation unit 110B basically perform the same calculations, the only difference being whether they are used during learning or monitoring. Therefore, the calculation unit 110A may also be used as the calculation unit 110B.

[0030] The calculation unit 110A calculates a feature value F1A1 from normal operating data D1A1. The learning unit 120 constructs an inference model M1 through machine learning of the feature value F1A1 of the element equipment and normal operating data D2A1. The inference unit 220 infers a normal range (upper and lower limit values) F2B1 of the feature value by inputting operating data D2B1 at the time of judgment into the trained inference model M1. The calculation unit 110A also calculates a feature value F1A2 from normal operating data D1A2. The learning unit 120 constructs an inference model M2 through machine learning of the feature value F1A2 of the element equipment and normal operating data D2A2. The inference unit 220 constructs an inference model M2 by inputting operating data D2B2 at the time of judgment into the trained inference model M2.

[0031] The calculation unit 110B calculates a feature quantity F1B1 from the operating data D1B1 at the time of judgment. The calculation unit 110B calculates a feature quantity F1B2 from the operating data D1B2 at the time of judgment. The determination unit 310 includes a data processing unit 311 and a data processing unit 312. The data processing unit 311 counts the number of data items in which the feature quantity F1B1 (the degree of subcooling SC of the refrigerant at the condenser outlet) calculated by the calculation unit 110 from the operating data D1B1 at the time of judgment exceeds or falls below the normal range F2B1 inferred by the trained inference model M1 from the operating data during the malfunction (data items outside the normal range). The data processing unit 312 counts the number of data items in which the feature quantity F1B2 (the heat exchange capacity Qo of the heat exchanger) calculated from the operating data D1B2 at the time of judgment falls below a predetermined capacity. The determination unit 310 calculates a ratio by dividing the number of data items in which the feature value F1B1 (degree of subcooling SC) falls outside the normal range (is outside the normal range) and the feature value F1B2 (heat exchange capacity Qo) has decreased by a predetermined amount (is outside the normal range) at the same time by the number of determined operating data items. This ratio indicates the proportion of time per day during which operation in which the heat exchange capacity Qo is lower than normal due to malfunction of the element equipment occurs, and is referred to as the "degraded capacity operation rate" in this specification. The determination unit 310 determines whether the degraded capacity operation rate is equal to or greater than a threshold value.

[0032] The display unit 320 displays trend data showing the change in the amount of refrigerant and the change in the performance degradation operation rate per day for each component device (outdoor heat exchanger 12, indoor heat exchanger 22), and if the performance degradation operation rate is above a threshold value, it displays a message that maintenance is required.

[0033] 1 operates as calculation units 110A and 110B, a learning unit 120, an inference unit 220, and a judgment unit 310 in accordance with corresponding programs. In addition, the memory 3 stores an operating data set of the air conditioning device 40 and inference models M1 and M2.

[0034] Each of the inference models M1 and M2 is a regression model including a neural network that infers a normal value of the condition index value of the air conditioning device 40 from the operating data of the air conditioning device 40. Each of the inference models M1 and M2 may be a classification model that infers a stage (classification) of the condition index value. The normal value of the condition index value may be the maximum and minimum values ​​of the confidence interval that the condition index value can take when the air conditioning device 40 is normal. Furthermore, the confidence interval may be a range with the normal value as the median (for example, a range of ±10% of the normal value).

[0035] It should be noted that general AI (Artificial Intelligence) technology can be applied to clustering and weighting of parameters included in the driving data.

[0036] The learning unit 120 performs machine learning on the inference models M1 and M2 to construct trained inference models M1 and M2. The machine learning algorithm used by the learning unit 120 may be a known algorithm such as supervised learning, semi-supervised learning, unsupervised learning, or reinforcement learning. Furthermore, the machine learning algorithm may be deep learning, which learns to extract features themselves, or other known methods such as neural networks, genetic programming, functional logic programming, or support vector machines. The following describes the case where supervised learning is applied to a neural network.

[0037] The calculation unit 110A calculates a state index value (SC) for the first operating period from the operating data D1A1 included in the operating data DA. The calculation unit 110A calculates a state index value (Qo) for the first operating period from the operating data D1A2 included in the operating data DA. The learning unit 120 performs supervised learning on the inference model M1 using learning data in which the state index value for the first operating period is used as the correct answer data (teacher data) for the first operating data. The learning unit 120 constructs trained inference models M1 and M2 and stores them in the memory 3. Note that the first operating period is the period immediately after the air conditioning device 40 is installed at the installation location and its operation is confirmed through a trial run, so there is a low possibility that there is an abnormality in the air conditioning device 40. Since it can be assumed that the air conditioning device 40 is normal during the first operating period, the state index value for the first operating period is used as the correct answer data for the first operating data. In the following example, the first operating period is set to 365 days from immediately after installation.

[0038] Fig. 5 is a flowchart illustrating the contents of the learning process executed by the anomaly detection system 1. When the learning process starts, the anomaly detection system 1 shown in Fig. 4 acquires air conditioning operation data DA for 365 days immediately after installation in step S1. The air conditioning operation data DA includes a first data group D1A1 and a second data group D2A1 that does not include the same data elements as the first data group D1A1. The air conditioning operation data DA also includes a third data group D1A2 that is different from the first data group D1A1 and a fourth data group D2A2 that does not include the same data elements as the third data group D1A2.

[0039] Next, in step S2, the calculation unit 110A calculates a feature quantity F1A1 from the first data group D1A1 and calculates a feature quantity F1A2 from the third data group D1A2. An example of the feature quantity F1A1 is the degree of subcooling SC of the refrigerant at the outlet of the condenser. An example of the feature quantity F1A2 is the heat exchange capacity Qo of the outdoor heat exchanger 12.

[0040] For example, the degree of subcooling SC at the outlet of the condenser can be calculated using the following formula (1): SC = Tc - TH3 (1) where Tc is the discharge saturated gas temperature, which is determined by the discharge pressure HS1, and TH3 is the liquid refrigerant temperature.

[0041] Furthermore, for example, the heat exchange capacity Qo can be calculated by the following formula (2): Qo=Gr×(Hd−Hco) (2) where Gr indicates the refrigerant circulation amount and is a value determined by the operating frequency fCOMP of the compressor 11, Hd is the specific enthalpy at the inlet of the outdoor heat exchanger 12 and is a value determined by the pressure HS1 and temperature TH4, and Hco is the specific enthalpy at the outlet of the outdoor heat exchanger 12 and is a value determined by the pressure HS1 and temperature TH3.

[0042] The upper part of Table 1 below shows the operating data DA used when training the inference model M1 for determining refrigerant leakage, the operating data D2B1 used as input to the trained inference model M1, and the operating data D1B1 used by the calculation unit 110B during the determination.

[0043] In addition, the lower part of Table 1 shows the operating data DA used when learning the inference model M2 for determining insufficient heat exchange capacity of the outdoor heat exchanger 12, the operating data D2B2 used as input to the learned inference model M2, and the operating data D1B2 used by the calculation unit 110B during the determination.

[0044]

[0045] 5, the learning unit 120 constructs an inference model (regression model) M1 of the objective variable (feature amount F1A1) corresponding to the explanatory variable (driving data group D2A1). Similarly, the learning unit 120 constructs an inference model (regression model) M2 of the objective variable (feature amount F1A2) corresponding to the explanatory variable (driving data group D2A2). At this time, the learning unit 120 constructs an inference model of the number of features multiplied by 3 (upper limit, median, lower limit). For example, the upper limit may be 97.5% of the confidence interval, and the lower limit may be 2.5% of the confidence interval (equivalent to 2σ).

[0046] Figure 6 is a diagram showing an example of a neural network Nw1 included in the inference model M1 constructed in step S3 of Figure 5. As shown in Figure 6, the neural network Nw1 includes an input layer X10, an intermediate layer (hidden layer) Y10, and an output layer Z10. The input layer X10 includes neurons X11, X12, and X13. The intermediate layer Y10 includes neurons Y11 and Y12. The output layer Z10 includes a neuron Z11. The input layer X10 and the intermediate layer Y10 are fully connected to each other. The intermediate layer Y10 and the output layer Z10 are fully connected to each other. The neural network Nw1 may include two or more intermediate layers.

[0047] When multiple inputs are input to neurons X11 to X13 in the input layer X10, the values ​​are multiplied by weights w11, w12, w13, w14, w15, and w16, and the result is input to neurons Y11 and Y12 in the intermediate layer Y10. The outputs from neurons Y11 and Y12 are multiplied by weights w21 and w22, and the result is output from neuron Z11 in the output layer Z10. The output result from the output layer Z10 varies depending on the values ​​of weights w11 to w16, w21, and w22. The weights and biases of the neural network Nw1 are updated by backpropagation of the error between driving data input to the input layer and the correct data so that the result output from the output layer approaches the correct data.

[0048] In this way, the learning of the inference model M1 is completed. The same applies to the inference model M2, so the explanation will not be repeated. Next, the contents of the malfunction detection process after the learning is completed will be explained.

[0049] FIG. 7 is a flowchart for explaining the malfunction detection process executed by the abnormality detection system 1.

[0050] First, in step S11, the abnormality detection system 1 shown in FIG. 4 acquires the operation data DB for one day that is the target of the malfunction judgment.

[0051] 4 calculates a feature quantity F1B1 from the operating data D1B1 in the acquired operating data DB, and calculates a feature quantity F1B2 from the operating data D1B2. The feature quantity F1B1 is, for example, the degree of subcooling SC of the refrigerant at the outlet of the condenser, and the feature quantity F1B2 is the heat exchange capacity Qo of the outdoor heat exchanger 12. The calculation process for these has already been explained using the above-mentioned formulas (1) and (2), so the explanation will not be repeated. In parallel with this, in step S12, the inference unit 220 of Figure 4 uses the learned model M1 to infer the degree of subcooling SCi (upper limit value, lower limit value, median value SCAI) at the condenser outlet from the operating data D2B1 as the inferred value F2B1, and the inference unit 220 uses the learned model M2 to infer the heat exchange capacity QoAI (median inferred value) of the outdoor heat exchanger 12 from the operating data D2B2 as the inferred value F2B2.

[0052] Next, in step S13, the determination unit 310 extracts operating data in which the heat exchange capacity Qo of the outdoor heat exchanger 12 calculated by the calculation unit 110B using the operating data acquired in step S11 has decreased by a predetermined amount (e.g., 20% or more) relative to the central inference value QoAI based on the inference model M2. The data is extracted, for example, by setting a flag for a set of operating data. Note that the predetermined amount includes at least the amount of decrease in the heat exchange capacity Qo that results in insufficient cooling (a decrease of -20% capacity).

[0053] 8 shows an example of time-dependent changes in the inferred and calculated values ​​of the heat exchange capacity Qo and the degree of subcooling SC of the refrigerant at the outlet of the condenser. From time t0 to t1, the refrigerant amount decreases by less than -10%, then during the subsequent period TP1, the refrigerant amount decreases by -10%, and during the period TP2, the refrigerant amount decreases by -20%.

[0054] The greater the amount of refrigerant decrease shown in the upper part of FIG. 8, the greater the number of extracted data represented by white circles.

[0055] 7, step S14 is executed. In step S14, the determination unit 310 extracts data from the operating data acquired in step S11 in which the degree of subcooling SC of the refrigerant at the outlet of the condenser falls below a lower limit. This lower limit is a value inferred by the inference model M1.

[0056] Next, in step S15, the determination unit 310 calculates the ratio of data in which the heat exchange capacity Qo is lower than the lower limit and the degree of subcooling SC of the refrigerant at the outlet of the condenser is lower than the lower limit (degraded performance operation rate R). The ratio R is the value obtained by dividing the number of data extracted in step S13 and also in step S14 by the total number of operation data. In other words, the degraded performance operation rate R is (the number of observed data in which a flag is set in both S13 and S14) / (the total number of observations).

[0057] In this way, by setting the extraction condition to AND data in which the heat exchange capacity Qo has decreased and data in which each feature value falls outside the appropriate range, it is possible to identify performance degradation due to malfunction of each element. As a result, performance degradation due to poor heat transfer in the outdoor heat exchanger 12 caused by, for example, rust or dust adhesion is not detected. The data extracted in this way is indicated by circles in Figure 8.

[0058] For example, at time t2, the heat exchange capacity Qo of the outdoor heat exchanger 12 is smaller than the lower limit of the median value minus 20%, and the degree of subcooling SC of the refrigerant at the condenser outlet is equal to or less than the inferred lower limit, so data is extracted and used to calculate the reduced capacity operation rate in step S15. On the other hand, at time t7, although the degree of subcooling SC of the refrigerant at the condenser outlet is equal to or less than the inferred lower limit, the heat exchange capacity Qo is not equal to or less than the lower limit of the median value minus 20%, so no data is extracted in step S15.

[0059] The number of data extracted from the one-day's worth of operating data in this manner is divided by the total number of one-day's worth of operating data to calculate the performance degradation operating rate R.

[0060] Next, in step S16, the display unit 320 displays a trend graph showing the change in the feature amount over time, and if the feature amount exceeds a threshold, notifies the user (maintenance worker) of the malfunction.

[0061] Figure 9 is a graph showing the relationship between the amount of refrigerant change and the amount of change in the degree of subcooling. As shown in the lower part of Figure 8, the difference ΔSC between the estimated median SCAI and the calculated feature value of subcooling SC tends to increase as the amount of refrigerant decreases. As shown in Figure 9, a unique relationship can be found between the amount of refrigerant change (decrease) and ΔSC. For example, if the correlation between ΔSC and the amount of liquid refrigerant shown in Figure 9 is set for each model, the amount of refrigerant change can be estimated. ΔSC can also be calculated using the following formula, where the subscript ave indicates the average over a predetermined period: ΔSC = SCAI_ave - SC_ave. Figure 10 is a graph showing the relationship between the decrease in refrigerant amount, the degraded operation rate, and the degraded operation rate. In Figure 10, the horizontal axis represents the degraded operation rate (Qo / QoAI), and the vertical axis represents the degraded operation rate.

[0062] Fig. 11 is a diagram showing an example of a trend graph displayed to the user. Fig. 11 shows the change over time in the performance degradation operation rate when the performance degradation rate in Fig. 10 is -10%. For example, in the case of a slow refrigerant leak, a maintenance worker can confirm that the duration of performance degradation caused by the refrigerant leak is increasing day by day, and can determine the timing of maintenance.

[0063] If the refrigerant leakage progresses over time (-10% to -30%), the trend data of the capacity degradation rate rises and exceeds the threshold value R1. When the data exceeds the threshold value R1, the display unit 320 displays a malfunction on the screen so that the malfunction can be easily recognized.

[0064] Furthermore, if the amount of refrigerant deficiency corresponding to ΔSC derived from the operating data based on the relationship in FIG. 9 is also recorded, the maintenance worker can predict the amount of refrigerant to be additionally charged and prepare it.

[0065] The monitoring device 5 of the present embodiment described above organizes the percentage of time during which the degree of subcooling SC, a characteristic quantity indicative of a refrigerant leak, falls outside the normal range and the heat exchange capacity Qo is lower than normal (decreased capacity operation rate) as daily trend data, and notifies a maintenance worker of the timing of maintenance when the malfunction determination threshold is exceeded, thereby enabling the maintenance worker to determine when to perform maintenance work.

[0066] Furthermore, by displaying trend data instead of instantaneous values, it is possible to correctly determine malfunction even when the threshold value is instantaneously exceeded due to internal or external disturbances.

[0067] For example, even if a refrigerant shortage occurs, the same capacity Qo may be exhibited under the same environmental conditions as under normal conditions. This is because the capacity Qo is balanced with the air conditioning load through control. However, when there is a refrigerant shortage, the operating state of the air conditioning device 40 changes, causing the central inferred value of the capacity Qo inferred from the operating state to rise, and the calculated value of the capacity Qo falls below the inferred normal range, allowing a decrease in capacity to be detected. Therefore, since a refrigerant shortage can be detected in intermediate seasons such as spring and autumn, it is possible to prevent a capacity shortage when full-scale air conditioning operation is performed in the summer.

[0068] (Various Modifications) Note that the flowcharts in Figures 5 and 7 describe refrigerant leakage during cooling and are explained using equation (1), but for poor heat transfer in the outdoor heat exchanger 12 during cooling, poor heat transfer in the indoor heat exchanger 22, and a faulty indoor expansion valve, it is possible to change the feature values ​​as shown in Table 2 below and determine malfunctions using a similar process.

[0069]

[0070] In Table 2 above, the contents of each variable are as follows. Note that OC_ is a symbol representing the outdoor unit, and IC_ is a symbol representing the indoor unit. Also, the following f() indicates that it is defined as a function calculated using the value obtained from the refrigerant physical property table for the value in parentheses detected by the sensor. Tc: Compressor discharge saturated gas temperature f(HS1) [°C] hd: Compressor discharge enthalpy f(HS1,TH4) [kJ / kg] hco: Refrigerant enthalpy at the outlet of the outdoor heat exchanger 12 f(HS1,TH3) [kJ / kg] ρs: Compressor suction density f(LS,TH5) [kg / m 3 ] fCOMP: Compressor frequency [Hz] St: Displacement volume [cc] ηv: Volumetric efficiency 0.95 [-] hs: Compressor suction enthalpy f(LS,TH5) [kJ / kg] A: Coefficient (value according to pressure unit) Cv: Flow coefficient (value according to LEV specifications and opening) ρco: Density of refrigerant at outlet of outdoor heat exchanger 12 f(HS1,TH3) [kg / m 3 ] Furthermore, the tendency of the above feature values ​​to deviate from the normal range during cooling is as follows: Refrigerant leakage: protruding downwards Poor heat transfer to outdoor heat exchanger 12: protruding downwards Poor heat transfer to indoor heat exchanger 22: protruding downwards Poor indoor expansion valve: protruding upwards & downwards Furthermore, while a malfunction in the heat exchange capacity of the outdoor heat exchanger 12 during cooling is shown in equation (2), a malfunction in the heat exchange capacity of the indoor heat exchanger 22 can also be determined by changing the feature values ​​as shown in Table 3 below, using a similar process.

[0071]

[0072] In Table 3 above, the contents of each variable are as follows. Note that OC_ is a symbol indicating the outdoor unit, and IC_ is a symbol indicating the indoor unit. Also, the following f() indicates that it is defined as a function calculated using the value obtained from the refrigerant physical property table for the value in parentheses detected by the sensor. hd: Compressor discharge enthalpy f(HS1,TH4) [kJ / kg] hco: Enthalpy of refrigerant at the outlet of the outdoor heat exchanger 12 f(HS1,TH3) [kJ / kg] ρs: Compressor suction density f(LS,TH5) [kg / m 3] fCOMP: Compressor frequency [Hz] St: Displacement volume [cc] ηv: Volumetric efficiency 0.95 [-] hs: Compressor suction enthalpy f(LS,TH5) [kJ / kg] A: Coefficient (value according to pressure unit) Cv: Flow coefficient (value according to LEV specifications and opening) ρco: Density of refrigerant at outlet of outdoor heat exchanger 12 f(HS1,TH3) [kg / m 3 ] In addition, for refrigerant leakage during heating, poor heat transfer in the outdoor heat exchanger 12, poor heat transfer in the indoor heat exchanger 22, and a malfunction of the indoor expansion valve, the feature values ​​can be changed as shown in Table 4 below, and a similar process can be used to determine malfunctions.

[0073]

[0074] In Table 4 above, the contents of each variable are as follows. OC_ is a symbol representing the outdoor unit, and IC_ is a symbol representing the indoor unit. Also, f() below indicates that it is specified as a function calculated using the value obtained from the refrigerant property table for the numerical value in parentheses detected by the sensor. Te: Compressor suction saturated gas temperature f(LS) [°C] hs: Compressor suction enthalpy f(LS,TH5) [kJ / kg] hco_ave: Refrigerant enthalpy at the outlet of the indoor heat exchanger 22 f(HS1,TH2ic_ave) [kJ / kg] *TH2ic_ave is the average value of the outlet temperatures Th2ic of all indoor heat exchangers 22 Ia7: Outdoor air enthalpy f(TH7, relative humidity) [kJ / kg] Ie: Surface enthalpy of the outdoor heat exchanger 12 f(Te, relative humidity 100%) [kJ / kg] ρs: Compressor suction density f(LS,TH5) [kg / m 3 ] fCOMP: Compressor frequency [Hz] St: Displacement volume [cc] ηv: Volumetric efficiency 0.95 [-] hco: Enthalpy of refrigerant at outlet of indoor heat exchanger 22 f(HS1,TH2ic) [kJ / kg] A: Coefficient (value according to pressure unit) Cv: Flow coefficient (value according to LEV specifications and opening) ρco: Density of refrigerant at outlet of indoor heat exchanger 22 f(HS1,TH2ic) [kg / m 3] Furthermore, the tendency for the above feature values ​​to deviate from the normal range during heating is as follows: Refrigerant leakage: protruding upward Poor heat transfer to outdoor heat exchanger 12: protruding downward Poor heat transfer to indoor heat exchanger 22: protruding downward Malfunctioning indoor expansion valve: protruding upward & downward Furthermore, for malfunctions in the heat exchange capacity of the outdoor heat exchanger 12 and the indoor heat exchanger 22 during heating, the feature values ​​can be changed as shown in Table 5 below, and a similar process can be used to determine malfunctions.

[0075]

[0076] In Table 5 above, the contents of each variable are as follows. Note that OC_ is a symbol representing the outdoor unit, and IC_ is a symbol representing the indoor unit. Also, the following f() indicates that it is defined as a function calculated using the value obtained from the refrigerant property table for the value in parentheses detected by the sensor. hs: Compressor suction enthalpy f(LS,TH5) [kJ / kg] hco_ave: Refrigerant enthalpy at the outlet of the indoor heat exchanger 22 f(HS1,TH2ic_ave) [kJ / kg] *TH2ic_ave is the average value of the refrigerant temperature Th2ic at the outlet of all indoor heat exchangers 22 ρs: Compressor suction density f(LS,TH5) [kg / m 3 ] fCOMP: Compressor frequency [Hz] St: Displaced volume [cc] ηv: Volumetric efficiency 0.95 [-] hco: Enthalpy of refrigerant at outlet of indoor heat exchanger 22 f(HS1,TH2ic) [kJ / kg] A: Coefficient (value according to unit of pressure) Cv: Flow coefficient (value according to opening determined by the specifications of the expansion valve) ρco: Density of refrigerant at outlet of indoor heat exchanger 22 f(HS1,TH2ic) [kg / m3] From the above, a refrigerant shortage can be determined when a feature related to refrigerant leakage listed in Tables 2 and 4 falls outside the normal range and a drop in heat exchange capacity OC_Q from the normal range is detected simultaneously.

[0077] In addition, poor heat transfer in the outdoor heat exchanger 12 can be determined when the feature values ​​related to poor heat transfer in the outdoor heat exchanger 12 listed in Tables 2 and 4 deviate from the normal range and the heat exchange capacity OC_Q listed in Tables 3 and 5 decrease from the normal range are detected simultaneously.

[0078] In addition, poor heat transfer of the indoor heat exchanger 22 can be determined when the feature values ​​related to poor heat transfer of the indoor heat exchanger listed in Tables 2 and 4 deviate from the normal range and the heat exchange capacity IC_Q listed in Tables 3 and 5 decrease from the normal range are detected simultaneously.

[0079] Furthermore, a malfunction of the indoor expansion valve can be determined when the deviation rate of the feature quantity IC_LEV related to the indoor expansion valve and a decrease in the heat exchange capacity IC_Q from the normal range are detected simultaneously.

[0080] <Summary> Referring again to the drawings, the learning device, monitoring device, and air conditioning system of the present disclosure will be summarized.

[0081] (Section 1) The present disclosure relates to a learning device 4 that learns the state of an air conditioning device 40 in which a refrigerant circulates. As shown in Fig. 4, operating data DA of the air conditioning device 40 includes a first data group D1A1 and a second data group D2A1 that does not include the same data elements as the first data group D1A1. The learning device 4 includes a first calculation unit 110A configured to calculate a first feature amount F1A1 from the first data group D1A1 of the air conditioning device during a learning period, and a learning unit 120 configured to perform supervised learning using the second data group D2A1 with the first feature amount F1A1 as correct answer data, and to generate a first inference model M1 that infers a first normal range of the first feature amount F1A1 from the second data group D2A1.

[0082] (Section 2) In the learning device described in Section 1, as shown in Fig. 4, the driving data DA includes a third data group D1A2 different from the first data group D1A1 and a fourth data group D2A2 that does not include the same data elements as the third data group D1A2. The first calculation unit 110A is configured to calculate a second feature amount F1A2 from the third data group D1A2 during a learning period. The learning unit 120 is configured to perform supervised learning using the fourth data group D2A2, with the second feature amount F1A2 calculated by the first calculation unit 110A as correct answer data, and to generate a second inference model M2 that infers a second normal range of the second feature amount F1A2 from the fourth data group D2A2.

[0083] (Section 3) In another aspect, the present disclosure relates to a monitoring device 5 for an air conditioning device 40 that uses a trained first inference model M1 and a second inference model M2 generated by the learning device 4 described in Section 2. As shown in Fig. 4, the monitoring device 5 includes a second calculation unit 110B configured to calculate a first feature amount F1B1 from a first data group D1B1 of the operating data DB of the air conditioning device 40 during a monitoring period, and to calculate a second feature amount F1B2 from a third data group D1B2, and a second calculation unit 110B configured to use the trained first inference model M1 to output a first normal range F2B1 from a second data group D2B1 of the operating data DB of the air conditioning device during the monitoring period, and to calculate a second normal range F2B1 from the operating data of the air conditioning device during the monitoring period. The data processing system includes an inference unit (220) configured to output a second normal range (F2B2) from the fourth data group (D2B2) of the first data, and a judgment unit (310) that judges whether the state of the air conditioning apparatus is normal or not based on the frequency with which a first malfunction state in which the first feature value (F1B1) calculated by the second calculation unit (110B) falls outside the first normal range (F2B1) inferred by the inference unit (220) and a second malfunction state in which the second feature value (F1B2) calculated by the second calculation unit (110B) falls outside the second normal range (F2B2) inferred by the inference unit (220) are simultaneously detected.

[0084] (4) In the monitoring device 5 described in 3, the air conditioning apparatus 40 includes a refrigerant circuit C1 through which a refrigerant circulates. As shown in Fig. 2, the refrigerant circuit C1 includes a compressor 11, a condenser (outdoor heat exchanger 12), an expansion valve 21, and an evaporator (indoor heat exchanger 22). The first feature amount is the degree of subcooling SC of the refrigerant at the outlet of the condenser, and the second feature amount is the heat exchange capacity Qo of the condenser.

[0085] (Item 5) In the monitoring device 5 described in item 4, as shown in Table 1, the first data group D1B1 includes at least the pressure HS1 of the refrigerant discharged from the compressor and the refrigerant temperature TH3 at the outlet of the condenser, the second data group D2B1 includes at least the temperature TH1ic of the air-conditioned space, the operating frequency fCOMP of the compressor 11, and the opening Li of the expansion valve 21. The third data group D1B2 includes at least the pressure HS1 of the refrigerant discharged from the compressor 11, the refrigerant temperature TH3 at the outlet of the condenser, and the operating frequency fCOMP of the compressor 11, and the fourth data group D2B2 includes at least the temperature TH1ic of the air-conditioned space and the opening Li of the expansion valve 21.

[0086] (Item 6) The monitoring device 5 described in item 3 further includes a display unit 320 that displays changes in frequency over a fixed period of time and indicates that maintenance is required when the frequency exceeds a judgment threshold.

[0087] (Section 7) In yet another aspect, the present disclosure relates to a monitoring device 5 for an air conditioning device 40 that uses a trained first inference model M1 generated by the learning device 4 described in Section 1. As shown in Fig. 4, the monitoring device 5 includes a second calculation unit 110B configured to calculate a first feature amount F1B1 from a first data group D1B1 in an operating data DB of the air conditioning device 40 during a monitoring period, an inference unit 220 configured to output a first normal range F2B1 from a second data group D2B1 in the operating data DB of the air conditioning device 40 during the monitoring period using the trained first inference model M1, and a determination unit 310 that compares the first feature amount F1B1 calculated by the second calculation unit 110B with the first normal range F2B1 inferred by the inference unit 220 to determine whether the state of the air conditioning device 40 is normal.

[0088] (Item 8) In the monitoring device 5 described in item 7, the air conditioning apparatus 40 includes a refrigerant circuit C1 through which a refrigerant circulates. As shown in Fig. 2, the refrigerant circuit C1 includes a compressor 11, a condenser (outdoor heat exchanger 12), an expansion valve 21, and an evaporator (indoor heat exchanger 22). The first feature quantity F1B1 is the degree of subcooling SC of the refrigerant at the outlet of the condenser.

[0089] (Item 9) In the monitoring device described in Item 8, as shown in Table 1, the first data group D1B1 includes the pressure HS1 of the refrigerant discharged from the compressor 11 and the refrigerant temperature TH3 at the outlet of the condenser, and the second data group D2B1 includes at least the temperature TH1ic of the space to be air-conditioned and the operating frequency fCOMP of the compressor 11.

[0090] (10th Clause) In still another aspect, the present disclosure relates to the air conditioning system 1000 shown in FIG. 1 , which includes the monitoring device 5 according to any one of the third to ninth clauses and an air conditioning device 40 .

[0091] The embodiments disclosed herein are intended to be combined as appropriate within the scope of compatibility. The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The scope of the present disclosure is defined by the claims, not the above description, and is intended to include all modifications within the meaning and scope of the claims.

[0092] 1 Anomaly detection system, 2, 32 CPU, 3, 33 Memory, 4 Learning device, 5 Monitoring device, 10 Outdoor unit, 11 Compressor, 12, 22 Heat exchanger, 13 Four-way valve, 14 Outdoor fan, 15 Accumulator, 20 Indoor unit, 21 Expansion valve, 23 Indoor fan, 30 Control device, 40 Air conditioning device, 51, 52, 53, 54, 55, 56 Temperature sensor, 57 Humidity sensor, 61, 63 Pressure sensor, 110, 110A, 110B Calculation unit, 120 Learning unit, 220 Inference unit, 310 Determination unit, 311, 312 Data processing unit, 320 Display unit, 900 Network, 1000 Air conditioning system, C1 Refrigerant circuit, M1, M2 Inference model, Nw1 Neural network.

Claims

1. A learning device that learns a state of an air conditioning device in which a refrigerant circulates, The operating data of the air conditioning device includes a first data group and a second data group that does not include the same data elements as the first data group, The learning device includes: A first calculation unit configured to calculate a first feature amount from the first data group of the air conditioning device during a learning period; a learning unit configured to perform supervised learning using the second data group with the first feature as correct answer data, and to generate a first inference model that infers a first normal range of the first feature from the second data group.

2. the driving data includes a third data group different from the first data group, and a fourth data group that does not include the same data elements as the third data group; the first calculation unit is configured to calculate a second feature amount from the third data group during the learning period; The learning device of claim 1, wherein the learning unit is configured to perform supervised learning using the fourth data group, with the second feature obtained by calculation by the first calculation unit as correct answer data, and to generate a second inference model that infers a second normal range of the second feature from the fourth data group.

3. 3. A monitoring device for an air-conditioning device that uses the first inference model and the second inference model that have been trained and generated by the learning device according to claim 2, a second calculation unit configured to calculate the first feature amount from the first data group of the operation data of the air conditioning device during a monitoring period and to calculate the second feature amount from the third data group; an inference unit configured to output the first normal range from the second data group of the operation data of the air conditioning device during the monitoring period using the first inference model that has been trained, and to output the second normal range from the fourth data group of the operation data of the air conditioning device during the monitoring period; a judgment unit that judges whether the state of the air conditioning apparatus is normal or not based on the frequency with which a first malfunction state in which the first feature amount calculated by the second calculation unit falls outside the first normal range inferred by the inference unit and a second malfunction state in which the second feature amount calculated by the second calculation unit falls outside the second normal range inferred by the inference unit are simultaneously detected.

4. The air conditioning device includes a refrigerant circuit through which a refrigerant circulates, The refrigerant circuit includes a compressor, a condenser, an expansion valve, and an evaporator. the first characteristic amount is a degree of subcooling of the refrigerant at an outlet of the condenser, The monitoring device according to claim 3 , wherein the second characteristic amount is a heat exchange capacity of the condenser.

5. The first data group includes at least a pressure of a refrigerant discharged from the compressor and a refrigerant temperature at an outlet of the condenser, The second data group includes at least a temperature of a space to be air-conditioned, an operating frequency of the compressor, and an opening degree of the expansion valve, The third data group includes at least a pressure of a refrigerant discharged from the compressor, a refrigerant temperature at an outlet of the condenser, and an operating frequency of the compressor; The monitoring device according to claim 4 , wherein the fourth data group includes at least a temperature of a space to be air conditioned and an opening degree of the expansion valve.

6. 4. The monitoring device according to claim 3, further comprising a display unit that displays a change in the frequency over a fixed period of time and displays a message that maintenance is required when the frequency exceeds a judgment threshold value.

7. 2. A monitoring device for an air conditioning device that uses the first inference model that has been trained by the learning device according to claim 1, a second calculation unit configured to calculate the first feature amount from the first data group of operation data of the air conditioning device during a monitoring period; an inference unit configured to output the first normal range from the second data group of operation data of the air conditioning device during the monitoring period using the trained first inference model; A monitoring device comprising: a judgment unit that compares the first characteristic calculated by the second calculation unit with the first normal range inferred by the inference unit to determine whether or not the state of the air conditioning device is normal.

8. The air conditioning device includes a refrigerant circuit through which a refrigerant circulates, The refrigerant circuit includes a compressor, a condenser, an expansion valve, and an evaporator. The monitoring device according to claim 7 , wherein the first characteristic amount is a degree of subcooling of the refrigerant at an outlet of the condenser.

9. The first data set includes a pressure of a refrigerant discharged from the compressor and a refrigerant temperature at an outlet of the condenser, The monitoring device according to claim 8 , wherein the second data group includes at least a temperature of a space to be air conditioned and an operating frequency of the compressor.

10. A monitoring device according to claim 3 or claim 7; An air conditioning system comprising the air conditioner.