Refrigerant leak diagnostic device, refrigerant leak diagnostic method, and learning model generation method

Abstract

It allows for more accurate detection of refrigerant leaks. [Solution] The present disclosure relates to a refrigerant leak diagnostic device 10 comprising a control unit 11, wherein the control unit 11 performs a first determination to estimate the amount of refrigerant in the high-pressure receiver 100 based on a refrigerant state prediction value obtained by inputting the operating data of the refrigeration and air conditioning equipment 1 during operation into a first prediction model generated based on the refrigerant state data in the high-pressure receiver 100 during normal operation of the refrigeration and air conditioning equipment 1 and the operating data of the refrigeration and air conditioning equipment 1, and determines whether there is a leak of refrigerant 101 in the refrigeration and air conditioning equipment 1 based on the output from the first prediction model.

Description

Technical Field

[0001] The present disclosure relates to a refrigerant leakage diagnosis device, a refrigerant leakage diagnosis method, and a learning model generation method.

Background Art

[0002] Leakage of the refrigerant enclosed in refrigeration and air conditioning equipment can be detected from changes in specific parameters in operation data such as the degree of subcooling of the refrigerant at the condenser outlet, the suction superheat degree of the compressor, and the expansion valve opening degree, which can be calculated based on the pressure and temperature of the refrigerant. However, in refrigeration and air conditioning equipment, since the required amount of refrigerant changes depending on the operating conditions, there may be a case where a container for temporarily storing surplus refrigerant is provided. Even if a refrigerant leakage occurs while the refrigerant is accumulated inside the container, since the shortage leaked from the refrigerant accumulated in the container is replenished, the state of the refrigeration cycle does not change and the specific parameters do not change either, so the leakage cannot be detected. Eventually, detection becomes possible after the refrigerant corresponding to the amount accumulated in the container has leaked. In particular, when the container is a high-pressure receiver located downstream of the condenser, the refrigerant tends to accumulate in the container. Therefore, a technique for detecting refrigerant leakage of refrigeration and air conditioning equipment based on changes in the liquid level height of the refrigerant accumulated inside the container provided in the refrigerant circuit has been proposed.

[0003] Patent Document 1 describes a technique for detecting the liquid level height of the refrigerant accumulated inside a container of refrigeration and air conditioning equipment with a level sensor, calculating the amount of refrigerant inside the container based on the detection values by a pressure sensor and a temperature sensor, and detecting the leakage of the refrigerant from the refrigerant circuit. Also, as the level sensor, a method for detecting the liquid level height by the temperature when the refrigerant passing through a plurality of branch pipes attached to the side surface of the container is depressurized and heated, and a method using a sensor that outputs a voltage corresponding to the position of a float floating on the refrigerant liquid surface inside the container are disclosed.

[0004] Patent Document 2 discloses a method for determining the liquid level height based on the capacitance detection value of a capacitance sensor composed of opposing electrode plates immersed in the refrigerant inside a liquid receiver. Further, a method for visually determining the liquid level height by attaching a sight glass with a liquid level scale to the side surface of the container is also disclosed. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2002-286333 [Patent Document 2] Patent No. 2997487 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] During operation of refrigeration and air conditioning equipment, the pressure and temperature of the refrigerant inside the equipment fluctuate in response to changes in operating conditions such as ambient temperature and refrigeration load, causing the refrigerant density to also fluctuate. Furthermore, changes in operating conditions cause fluctuations in the amount of refrigerant accumulated in the condenser and evaporator, resulting in fluctuations in the amount of refrigerant accumulated in the container. Due to these fluctuations in refrigerant density and distribution, the liquid level inside the container fluctuates even if there is no refrigerant leakage.

[0007] Furthermore, during operation, the refrigerant flows into the container at high speed, so a stable, stationary liquid level at the gas-liquid interface does not occur, and the liquid level fluctuates violently. Therefore, it is difficult to accurately detect the liquid level during operation, regardless of the level sensor used. In other words, in order to detect refrigerant leakage from changes in the liquid level inside the container, it is necessary to accurately detect the liquid level, correct for fluctuations in refrigerant density and distribution due to changes in operating conditions, and determine the amount of refrigerant inside the container under those conditions.

[0008] Therefore, in Patent Document 1, a pump-down operation is performed to recover the refrigerant in the condenser and container, and the stable liquid level is measured when the compressor is stopped after the pump-down operation is completed. This minimizes the influence of fluctuations in the refrigerant distribution, and the refrigerant mass in the container is calculated from the accurately detected liquid level height and the refrigerant density obtained from the detected values ​​of the pressure sensor and temperature sensor, thereby determining whether refrigerant leakage has occurred.

[0009] Patent Document 2 describes a method for controlling the compressor's suction pressure to a constant level by fixing the compressor rotation speed and controlling the expansion valve opening when detecting the liquid level. This stabilizes the refrigeration cycle and eliminates fluctuations in refrigerant density and distribution due to operating conditions, resulting in a stable liquid level and allowing for accurate determination of the refrigerant amount.

[0010] However, all of the above methods require special operation for leak detection, making continuous leak detection during normal operation impossible. In recent years, legal regulations to reduce refrigerant leakage have been strengthened in various countries, and the introduction of systems that continuously monitor refrigeration and air conditioning equipment and automatically detect leaks in their early stages is being recommended. Therefore, even for refrigeration and air conditioning equipment equipped with containers for storing excess refrigerant, achieving high-precision refrigerant leak detection by measuring liquid level during normal operation has become an urgent issue.

[0011] This disclosure provides a technology that enables high-precision detection of refrigerant leakage during normal operation in refrigeration and air conditioning equipment equipped with a container for storing excess refrigerant. [Means for solving the problem]

[0012] The first aspect of this disclosure is, A refrigerant leak diagnostic device comprising a control unit, The control unit, In a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally between a condenser and an expansion valve, a first determination is made to estimate the amount of refrigerant in the high-pressure receiver based on a refrigerant state prediction value obtained by inputting the operating data of the refrigeration and air conditioning system during operation into a first prediction model generated based on the refrigerant state data in the high-pressure receiver during normal operation of the refrigeration and air conditioning system and the operating data of the refrigeration and air conditioning system. Based on the output from the first prediction model, the leakage of refrigerant from the refrigeration and air conditioning equipment is determined.

[0013] According to a first aspect of this disclosure, in a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally, refrigerant leakage can be detected with higher precision.

[0014] A second aspect of this disclosure is a refrigerant leak diagnostic device as described in the first aspect, The control unit, During operation, the operating data is input into a second prediction model that predicts a refrigerant amount index corresponding to the amount of refrigerant in the refrigeration and air conditioning equipment, which is generated based on the normal operating data of the refrigeration and air conditioning equipment. Based on the predicted value of the refrigerant amount index obtained, a second determination is made to estimate the amount of refrigerant in the entire refrigerant circuit of the refrigeration and air conditioning equipment. Based on the first determination and the second determination, a leak of refrigerant from the refrigeration and air conditioning equipment is determined.

[0015] A third aspect of this disclosure is a refrigerant leak diagnostic device according to the first or second aspect, The refrigerant state data is a value based on the output signal of a liquid level sensor installed inside the high-pressure receiver.

[0016] A fourth aspect of this disclosure is a refrigerant leak diagnostic device according to the first or second aspect, The refrigerant status data is a value based on the output signal of a capacitance sensor installed inside the high-pressure receiver.

[0017] A fifth aspect of this disclosure is a refrigerant leak diagnostic device according to the first or second aspect, The aforementioned refrigerant state data is The side of the high-voltage receiver or the branch pipe connected to the high-voltage receiver has at least one observation hole, A light-transmitting member fitted into the observation hole, and a signal conversion unit that converts the refrigerant state observed through the observation hole into an electrical signal, This value is based on the output signal of a refrigerant state observation device.

[0018] A sixth aspect of this disclosure is a refrigerant leak diagnostic device as described in the fifth aspect, The signal conversion unit is an imaging device, The output signal is image data of the observation hole captured by the imaging device.

[0019] A seventh aspect of the present disclosure is the refrigerant leakage diagnostic apparatus according to the fifth aspect, wherein the signal conversion unit includes a light projector that projects specific light onto the observation hole, and a reflected light detector that detects reflected light of the specific light from the observation hole, and the output signal is an output signal of the reflected light detector.

[0020] An eighth aspect of the present disclosure is the refrigerant leakage diagnostic apparatus according to the fifth aspect, wherein the signal conversion unit includes a light projector that projects specific light onto the observation hole, and a transmitted light detector that detects transmitted light of the specific light that passes through the first observation hole, which is the observation hole, the refrigerant inside the high-pressure receiver or the refrigerant inside the branch pipe, and a second observation hole provided opposite to the first observation hole, and the output signal is an output signal of the transmitted light detector.

[0021] A ninth aspect of the present disclosure is the refrigerant leakage diagnostic apparatus according to the first aspect or the second aspect, wherein the refrigerant state data is a value based on an output signal of a temperature sensor, and the temperature sensor branches from the side surface of the high-pressure receiver, is connected to the compressor suction pipe via an expansion device, and is provided downstream of a heat exchanger for heating the refrigerant flowing out from the high-pressure receiver in a bypass pipe that bypasses the refrigerant in the high-pressure receiver to the compressor suction side and exchanges heat with another refrigerant pipe.

[0022] A tenth aspect of the present disclosure is a refrigerant leakage diagnostic apparatus including a control unit, wherein the control unit In a refrigeration and air conditioning system equipped with an accumulator that stores refrigerant internally between an evaporator and a compressor, a third determination is made to estimate the amount of refrigerant in the accumulator based on a refrigerant state prediction value obtained by inputting the operating data of the refrigeration and air conditioning system during operation into a third prediction model, which is a prediction model that predicts the amount of refrigerant in the accumulator based on the refrigerant state data in the accumulator during normal operation of the refrigeration and air conditioning system and the operating data of the refrigeration and air conditioning system. Based on the output from the third prediction model, the leakage of refrigerant from the refrigeration and air conditioning equipment is determined.

[0023] According to a tenth aspect of this disclosure, in a refrigeration and air conditioning system equipped with an accumulator that stores refrigerant internally, refrigerant leakage can be detected with higher precision.

[0024] An eleventh aspect of this disclosure is a refrigerant leak diagnostic device as described in the tenth aspect, The control unit, During operation, the operating data is input into a fourth prediction model that predicts a refrigerant amount index corresponding to the refrigerant amount of the refrigeration and air conditioning equipment based on the normal operating data of the refrigeration and air conditioning equipment. Based on the predicted value of the refrigerant amount index obtained, a fourth determination is made to estimate the total amount of refrigerant in the refrigeration and air conditioning equipment's refrigerant circuit. Based on the third and fourth determinations, a leak of refrigerant from the refrigeration and air conditioning equipment is determined.

[0025] A twelfth aspect of this disclosure is a refrigerant leak diagnostic device according to the tenth or eleventh aspect, The refrigerant state data is a value based on the output signal of a liquid level sensor installed inside the accumulator.

[0026] A thirteenth aspect of this disclosure is a refrigerant leak diagnostic device according to the tenth or eleventh aspect, The refrigerant state data is a value based on the output signal of a capacitance sensor installed inside the accumulator.

[0027] A fourteenth aspect of this disclosure is a refrigerant leak diagnostic device according to the tenth or eleventh aspect, At least one observation hole provided on the side of the accumulator or in the branch pipe connected to the accumulator, A light-transmitting member fitted into the observation hole, and a signal conversion unit that converts the refrigerant state observed through the observation hole into an electrical signal, This value is based on the output signal of a refrigerant state observation device.

[0028] A 15th aspect of this disclosure is a refrigerant leak diagnostic device as described in the 14th aspect, The signal conversion unit is an imaging device, The output signal is image data of the observation hole captured by the imaging device.

[0029] A sixteenth aspect of this disclosure is a refrigerant leak diagnostic device as described in the fourteenth aspect, The signal conversion unit is A light source that projects specific light into the observation hole, A reflected light detector for detecting the reflected light of the specific light from the observation hole, Equipped with, The output signal is the output signal of the reflected light detector.

[0030] A 17th aspect of this disclosure is a refrigerant leak diagnostic device as described in the 14th aspect, The signal conversion unit is A light source that projects specific light into the observation hole, A transmitted light detector for detecting transmitted light of the specific light passing through a first observation hole, which is the observation hole, the refrigerant inside the accumulator or the refrigerant inside the branch pipe, and a second observation hole provided opposite the first observation hole, Equipped with, The output signal is the output signal of the transmitted light detector.

[0031] The eighteenth aspect of this disclosure is: In a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally between a condenser and an expansion valve, the first step is to perform a first determination to estimate the amount of refrigerant in the high-pressure receiver based on a predicted refrigerant state value obtained by inputting the operating data of the refrigeration and air conditioning system during operation into a first prediction model generated based on the refrigerant state data in the high-pressure receiver during normal operation of the refrigeration and air conditioning system and the operating data of the refrigeration and air conditioning system. A step of determining whether there is a refrigerant leak in the refrigeration and air conditioning equipment based on the output from the first prediction model, This is a refrigerant leak diagnosis method that includes [specific component / method].

[0032] According to the 18th aspect of this disclosure, in a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally, refrigerant leakage can be detected with higher accuracy during normal operation.

[0033] The 19th aspect of this disclosure is: In a refrigeration and air conditioning system equipped with an accumulator that stores refrigerant internally between an evaporator and a compressor, a third determination is made by inputting operating data during the operation of the refrigeration and air conditioning system into a third prediction model, which is a prediction model that predicts the amount of refrigerant in the accumulator based on the refrigerant state prediction value obtained from the refrigerant state prediction value obtained from the refrigerant state prediction value, and The steps include determining whether there is a refrigerant leak in the refrigeration and air conditioning equipment based on the output from the third prediction model, This is a refrigerant leak diagnosis method that includes [specific component / method].

[0034] According to a 19th aspect of this disclosure, in a refrigeration and air conditioning system equipped with an accumulator for storing refrigerant internally, refrigerant leakage can be detected with higher accuracy during normal operation.

[0035] The 20th aspect of this disclosure is: In a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally between a condenser and an expansion valve, a predictive model is generated to predict the amount of refrigerant in the high-pressure receiver by machine learning using refrigerant state data in the high-pressure receiver and operating data of the refrigeration and air conditioning system. This is a method for generating a learning model.

[0036] According to a 20th aspect of this disclosure, in a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally, refrigerant leakage can be detected with higher accuracy during normal operation.

[0037] The 21st aspect of this disclosure is: In a refrigeration and air conditioning system equipped with an accumulator that stores refrigerant internally between an evaporator and a compressor, a predictive model is generated to predict the amount of refrigerant in the accumulator by machine learning using refrigerant state data in the accumulator and operating data of the refrigeration and air conditioning system. This is a method for generating a learning model.

[0038] According to a 21st aspect of this disclosure, in a refrigeration and air conditioning system equipped with an accumulator for storing refrigerant internally, refrigerant leakage can be detected with higher accuracy during normal operation. [Brief explanation of the drawing]

[0039] [Figure 1] This is a diagram showing the configuration of a refrigeration and air conditioning system according to the first embodiment of this disclosure. [Figure 2] This is a hardware configuration diagram of a refrigerant leak diagnostic device according to the first embodiment of this disclosure. [Figure 3] This is a configuration diagram of a high-voltage receiver according to the first embodiment of this disclosure. [Figure 4] This figure illustrates a method used by a refrigerant leak diagnostic device according to the first embodiment of this disclosure to determine the state of a refrigerant based on changes in the refrigerant's liquid level. [Figure 5]This figure illustrates a method used by a refrigerant leak diagnostic device according to the first embodiment of this disclosure to determine the state of a refrigerant based on changes in the refrigerant's liquid level. [Figure 6] This figure illustrates how a refrigerant leak diagnostic device according to the first embodiment of this disclosure determines the refrigerant state based on the output signal of a refrigerant state observation device. [Figure 7] This figure shows a modified example of a refrigerant state observation device that outputs a signal to a refrigerant leak diagnostic device according to the first embodiment of this disclosure. [Figure 8] This figure illustrates a method used by a refrigerant leak diagnostic device according to the first embodiment of this disclosure to determine the state of a refrigerant based on the output signal of a temperature sensor. [Figure 9] This figure illustrates the relationship between the refrigeration and air conditioning equipment, the refrigerant leak diagnostic device, and the learning device according to the first embodiment of this disclosure. [Figure 10] This is a functional block diagram of a learning device according to the first embodiment of this disclosure. [Figure 11] This is a flowchart of the refrigerant leak detection process based on a first determination according to the first embodiment of this disclosure. [Figure 12] This is a flowchart of the refrigerant leak detection process based on a second determination, according to the first embodiment of this disclosure. [Figure 13] This is a configuration diagram of a refrigeration and air conditioning system according to a second embodiment of the present disclosure. [Figure 14] This is a diagram showing the configuration of an accumulator according to a second embodiment of this disclosure. [Modes for carrying out the invention]

[0040] The embodiments for carrying out the invention will be described below with reference to the drawings. In each drawing, the same reference numerals are used for identical components, and redundant explanations may be omitted.

[0041] [First Embodiment] <Configuration of Refrigeration and Air Conditioning Equipment 1> The hardware configuration of the refrigeration and air conditioning equipment 1 according to the first embodiment of this disclosure will be described below with reference to Figure 1. Figure 1 is a configuration diagram of the refrigeration and air conditioning equipment 1 according to the first embodiment of this disclosure. The refrigeration and air conditioning equipment 1 has an outdoor unit 200 and one or more indoor units 300. The multiple indoor units 300 may include indoor units 300 with different performance, indoor units 300 with the same performance, or indoor units 300 that are stopped.

[0042] The refrigeration and air conditioning equipment 1 is connected to a refrigerant leak diagnostic device 10. The refrigerant leak diagnostic device 10 includes a control unit 11 and determines the state of the refrigerant based on changes in the liquid level of the refrigerant in the high-pressure receiver 100. The state of the refrigerant is whether or not there is a refrigerant leak in the refrigeration and air conditioning equipment 1. The refrigerant leak diagnostic device 10 may be included in the refrigeration and air conditioning equipment 1.

[0043] In the illustrated example, the outdoor heat exchanger 201, high-pressure receiver 100, outdoor unit main expansion valve 205, subcooling heat exchanger 203, indoor heat exchanger expansion valve 302, indoor heat exchanger 301, four-way switching valve 206, and compressor 202 are connected by refrigerant piping to form the main refrigerant circuit. In this example, since the cooling operation mode is active, the flow path of the four-way switching valve 206 is set to supply the discharge gas from the compressor 202 to the outdoor heat exchanger 201. The outdoor heat exchanger 201 is a heat exchanger that functions as a condenser for the high-pressure refrigerant in the refrigeration cycle. The indoor heat exchanger 301 is a heat exchanger that functions as an evaporator for the low-pressure refrigerant in the refrigeration cycle.

[0044] Furthermore, a subcooled heat exchanger expansion valve 204 is provided in a bypass pipe connected from the piping between the outdoor heat exchanger 201 and the subcooled heat exchanger 203 to the suction side piping of the compressor 202. The subcooled heat exchanger 203 is a heat exchanger that exchanges heat between the refrigerant that has passed through the subcooled heat exchanger expansion valve 204, which is provided in the bypass pipe connected from between the outdoor heat exchanger 201 and the subcooled heat exchanger 203 to the suction side piping of the compressor 202, and the refrigerant in the main refrigerant circuit. Note that the illustrated bypass example is just one example and is not limited to this.

[0045] On the outdoor unit 200 side, the outdoor heat exchanger 201, compressor 202, subcooled heat exchanger 203, subcooled heat exchanger expansion valve (bypass circuit) 204, and outdoor unit main expansion valve (main refrigerant circuit) 205 are connected to the piping. The outdoor unit 200 has various sensors (temperature sensors (e.g., thermistors) (1), (3), (4), (6), (7) and pressure sensors (2), (5), etc.).

[0046] On the indoor unit 300 side, the indoor heat exchanger 301 and the indoor heat exchanger expansion valve 302 are connected to the piping. The indoor unit 300 has various sensors (temperature sensors (e.g., thermistors) (8), (9), etc.).

[0047] The high-pressure receiver 100 is a refrigerant container installed between the condenser and the expansion valve, capable of storing excess refrigerant as liquid refrigerant. Excess refrigerant is the surplus of refrigerant filled in the refrigerant circuit that is generated in the refrigeration and air conditioning equipment 1 due to changes in the outside air temperature, indoor set temperature, changes in the number of indoor units 300 in operation, etc. The high-pressure receiver 100 is, for example, a vertical cylindrical container, and is a container for temporarily storing the refrigerant flowing through the main refrigerant circuit.

[0048] <Explanation of Terms> Herein, we will explain the terminology used in this disclosure. First, "operational data" refers to any data that can be acquired while the equipment is in operation and while it is stopped. For example, the operational data for equipment (e.g., refrigeration and air conditioning equipment 1) is as follows: This relates to the refrigerant state, including the temperature (condenser and evaporator inlet / outlet temperatures, compressor suction temperature, compressor discharge temperature, etc.), pressure (condensation pressure, evaporation pressure), and circulation rate of the refrigerant circulating inside the refrigeration and air conditioning equipment 1. • Values ​​calculated from refrigerant temperature and pressure data (e.g., condenser outlet supercooling degree, supercooled heat exchanger outlet supercooling degree, compressor suction superheating degree, compressor discharge superheating degree, etc.). • Operating amounts and control command values ​​of actuators that control the state of the refrigerant (opening degrees of each expansion valve in the outdoor unit 200 and indoor unit 300, compressor rotation speed, fan rotation speed, target value of compressor suction superheat, etc.). • Time information regarding the operating status of the equipment (compressor operating time, defrost operating time, etc.). • Environmental conditions of the location where the equipment is operating (outside temperature, room temperature, indoor relative humidity, etc.). • Input energy to the equipment, such as compressor current and fan current. • The refrigeration or air conditioning capacity, which is the output of the equipment. • Flag values ​​indicating the operating mode of the equipment (e.g., cooling operation, heating operation, defrost operation).

[0049] Furthermore, "operating status" refers to the state of the equipment (for example, an index value used to determine the operating status of the equipment) used to control the operation of the equipment and diagnose equipment malfunctions. For example, the operating status is the amount of refrigerant stored. If the amount of refrigerant stored is at the specified value, it is in a normal state; if the amount of refrigerant stored is below the specified value, it is in an abnormal state of gas shortage or leakage.

[0050] <Configuration of refrigerant leak diagnostic device 10> Figure 2 is a hardware configuration diagram of a refrigerant leak diagnostic device 10 according to the first embodiment of this disclosure. The refrigerant leak diagnostic device 10 includes an input device 51, a display device 52, an external I / F 53, RAM 54, ROM 55, CPU 56, a communication I / F 57, and an HDD 58, which are all connected by a bus 59. The input device 51 and the display device 52 may be connected in a configuration as needed.

[0051] The input device 51 includes a touch panel, operation keys and buttons, a keyboard and mouse, etc., used by the user to input various signals. The display device 52 consists of a display such as an LCD or organic EL that displays the screen, and a speaker that outputs sound data such as voice and music. The communication I / F 57 is an interface for the refrigerant leak diagnostic device 10 to communicate data via a network.

[0052] Furthermore, HDD58 is an example of a non-volatile storage device that stores programs and data. The programs and data stored include the OS, which is the basic software that controls the entire refrigerant leak diagnostic device 10, and applications that provide various functions on the OS. The refrigerant leak diagnostic device 10 may also use a drive device that uses flash memory as the storage medium instead of HDD58. For example, an SSD may be used as the drive device.

[0053] The external I / F 53 is an interface to an external device. An external device may be a recording medium 53a. This allows the refrigerant leak diagnostic device 10 to read from and write to the recording medium 53a via the external I / F 53. The recording medium 53a may include a flexible disk, CD, DVD, SD memory card, USB memory, etc.

[0054] RAM54 is an example of volatile semiconductor memory that temporarily holds programs and data. ROM55 is an example of non-volatile semiconductor memory that can retain programs and data even when the power is turned off. ROM55 stores programs and data such as the BIOS, OS settings, and network settings that are executed when the refrigerant leak diagnostic device 10 is started up.

[0055] The CPU 56 is a computing unit that controls and implements the overall functions of the refrigerant leak diagnostic device 10 by reading programs and data from storage devices such as the ROM 55 and HDD 58 onto the RAM 54 and executing processing.

[0056] <Configuration of high-voltage receiver 100> Figure 3 is a configuration diagram of a high-pressure receiver 100 according to the first embodiment of this disclosure. As shown in the figure, refrigerant 101 is stored in the high-pressure receiver 100. Various liquid level detection means connected to the refrigerant leak diagnostic device 10 can also be installed in the high-pressure receiver 100. One or more types of liquid level detection means may be installed.

[0057] More specifically, a liquid level sensor 110 and a voltage output unit 111 may be provided inside the high-pressure receiver 100 to measure the liquid level of the refrigerant 101. Alternatively, a capacitance sensor 121 may be provided inside the high-pressure receiver 100 to measure the liquid level of the refrigerant 101. Furthermore, a refrigerant state observation device 130 may be provided in the high-pressure receiver 100 to measure the liquid level of the refrigerant 101. The refrigerant state observation device 130 is connected to the high-pressure receiver 100 and includes a branch pipe 133 with an observation hole 131 and a signal conversion unit 132 that converts the refrigerant state observable from the observation hole 131 into an electrical signal.

[0058] Alternatively, a liquid level detection device 140 may be provided on the high-pressure receiver 100 to measure the liquid level of the refrigerant 101. The liquid level detection device 140 detects the liquid level at the bypass pipe connection of the high-pressure receiver 100 based on the temperature change of the refrigerant 101 flowing through the bypass pipe 144 connected to the suction side of the compressor 202 from the side of the high-pressure receiver 100. Note that multiple observation holes 131 and liquid level detection devices 140 may be provided depending on the number of liquid level heights to be detected.

[0059] <<Liquid level sensor 110>> The refrigerant status data is a value based on the output signal of a liquid level sensor 110 installed inside the high-pressure receiver 100. The liquid level sensor 110 may be, for example, a float. The liquid level sensor 110 is a level sensor that can linearly measure the height of the liquid level of the refrigerant 101 inside the high-pressure receiver 100. The voltage output unit 111 holds the liquid level sensor 110 so that it can slide up and down and outputs a signal corresponding to the height position of the liquid level sensor 110 to the refrigerant leak diagnostic device 10.

[0060] Figures 4 and 5 illustrate a method used by a refrigerant leak diagnostic device 10 according to the first embodiment of this disclosure to determine the state of the refrigerant 101 based on changes in the liquid level of the refrigerant 101. Figure 4 shows the change in the liquid level of the refrigerant 101 when the state of the refrigerant 101 is normal, and Figure 5 shows the change in the liquid level of the refrigerant 101 when there is an abnormal state of insufficient refrigerant 101.

[0061] Figure 4(a) is a graph showing the time-series changes in the measured and predicted liquid level of refrigerant 101, with the vertical axis representing the liquid level of refrigerant 101 and the horizontal axis representing time. Figure 4(b) is a graph showing the time-series changes in Δliquid level, defined as the difference between the measured and predicted liquid level of refrigerant 101, with the vertical axis representing time. Figure 4(c) is a graph of Figure 4(b) with moving average data and threshold th1 added.

[0062] The predicted liquid level is an output value obtained by inputting operating data into a prediction model, which is created by machine learning using liquid level data detected by various liquid level detection devices 140 installed in the high-pressure receiver 100 of the refrigeration and air conditioning equipment 1 when it is functioning normally, and operating data of the refrigeration and air conditioning equipment 1 as the explanatory variable. This value represents the liquid level that is expected to appear when the refrigerant 101 is in a normal state.

[0063] The predicted liquid level takes into account fluctuations in the liquid level due to variations in environmental and operating conditions. Therefore, the difference between the measured value and the predicted value under normal conditions, Δliquid level, ideally cancels out fluctuations due to factors other than refrigerant quantity fluctuations and changes only in response to fluctuations in the refrigerant quantity. Consequently, Δliquid level is always 0 under normal conditions when refrigerant 101 is at the appropriate level, becomes a positive value when refrigerant 101 is overfilled, and becomes a negative value when it is underfilled. Therefore, by setting an appropriate threshold for Δliquid level, it is possible to determine overfilling, underfilling, or leakage of refrigerant 101 based on the time-series data fluctuations of Δliquid level.

[0064] When the refrigeration and air conditioning equipment 1 is in operation, the refrigerant 101 flows into the high-pressure receiver 100 at high speed, so a stable, stationary liquid surface at the gas-liquid interface does not appear, and the liquid surface fluctuates violently. Therefore, as shown in Figure 4(a), the actual measured liquid surface height fluctuates violently in a short period of time.

[0065] Furthermore, while the prediction model can predict long-period fluctuations such as ambient temperature and load changes, it cannot predict such non-stationary, short-period fluctuations in the liquid level. As a result, the predicted liquid level height under normal conditions will be an average time-series data that predicts only long-period fluctuations and does not include the short-period fluctuations of the actual measured value.

[0066] Therefore, as shown in Figure 4(b), the actual Δ liquid level height is obtained by removing only the long-period fluctuations from the measured liquid level height, leaving data with short-period fluctuations centered around 0. When determining the amount of refrigerant in this state, the threshold must be set to a value that is the sum of the acceptable refrigerant amount fluctuation and the maximum amplitude of the short-period oscillation in order to avoid misjudgment. As a result, the detection accuracy decreases.

[0067] Therefore, as shown in Figure 4(c), by using moving average data of the time-series data of Δ liquid level height, the influence of short-period oscillations is eliminated and detection accuracy is improved. With the threshold th1 setting shown in the figure, using the original Δ liquid level data may result in a false detection of leakage, but using the moving average data allows for a consistently normal detection.

[0068] Similar to Figure 4, Figure 5(a) is a graph showing the time-series changes in the measured liquid level of refrigerant 101 on the vertical axis and time on the horizontal axis. Figure 5(b) is a graph showing the time-series changes in Δliquid level, with Δliquid level on the vertical axis and time on the horizontal axis. Figure 5(c) is a graph that adds moving average data and threshold th1 to the graph in Figure 5(b), showing the moving average data of the time-series data of Δliquid level and the leak detection threshold.

[0069] As illustrated, in the case of an abnormal situation where refrigerant 101 is insufficient, as shown in Figure 5(a), the measured liquid level is a time-series data with long-term and short-term fluctuations overlapping, similar to the normal situation, but it has shifted downwards on the graph overall compared to the normal predicted value. As shown in Figure 5(b), the Δ liquid level is generally negative, with short-term fluctuations remaining. As shown in Figure 5(c), with the threshold th1 setting in the figure, using the original Δ liquid level data may result in a false positive for normal, but using the moving average data always results in a leak detection.

[0070] <<Capacitive Sensor 121>> The refrigerant state data is a value based on the output signal of a capacitance sensor 121 installed inside the high-pressure receiver 100. The capacitance sensor 121 includes opposing plates 121a and 121b. The capacitance sensor 121 is sealed inside the high-pressure receiver 100 such that the space between the opposing plates 121a and 121b is filled with refrigerant 101 in a liquid or gaseous state, and detects the liquid level of the refrigerant 101.

[0071] When the amount of refrigerant 101 changes and the liquid level changes, the ratio of the portion of the electrode plates 121a and 121b that is in contact with the liquid refrigerant to the portion that is in contact with the gaseous refrigerant changes. As a result, there is a difference of several times in the relative permittivity of the liquid refrigerant and the gaseous refrigerant, so the capacitance detected by the capacitance sensor 121 changes. Based on the capacitance measured by the capacitance sensor 121, the liquid level detection circuit 120 can calculate the height of the refrigerant 101 and the amount of refrigerant 101 at the time of measurement.

[0072] The liquid level detection circuit 120 measures the amount of refrigerant 101 stored in the high-pressure receiver 100, that is, the height of the refrigerant 101 in the high-pressure receiver 100, using a capacitance sensor 121. More specifically, the liquid level detection circuit 120 converts the capacitance detected by the capacitance sensor 121 according to the height of the refrigerant 101 liquid level between the electrodes 121a and 121b into an oscillation frequency or oscillation period. Then, the liquid level detection circuit 120 can detect the height of the refrigerant 101 liquid level in the high-pressure receiver 100 by converting the oscillation frequency due to the change in capacitance or the change in oscillation frequency into a DC voltage.

[0073] The time-series data of liquid level detected by the capacitance sensor 121 has the same characteristics as the time-series data of liquid level detected by the liquid level sensor 110. Therefore, it is possible to determine the amount of refrigerant using the time-series data of liquid level detected by the capacitance sensor 121 in the same manner as the method described using Figures 4 and 5.

[0074] <<Refrigerant state observation device 130>> The refrigerant state data is a value based on the output signal of the refrigerant state observation device 130. The refrigerant state observation device 130 includes an observation hole 131 and a signal conversion unit 132. The observation hole 131 is provided on the side of the high-pressure receiver 100 so that the liquid level of the refrigerant 101 in the high-pressure receiver 100 can be confirmed. The observation hole 131 may also be provided on the branch pipe 133 connected to the high-pressure receiver 100, or it may be provided on both the side of the high-pressure receiver 100 and the branch pipe 133. A light-transmitting member such as a glass member is fitted into the observation hole 131. The light-transmitting member transmits visible light.

[0075] The signal conversion unit 132 converts the refrigerant state observed through the observation hole 131 into an electrical signal. The signal conversion unit 132 is, for example, an imaging device such as a digital camera. At this time, the output signal output to the refrigerant leak diagnostic device 10 is image data of the observation hole 131 captured by the imaging device. The imaging device is positioned to face the light-transmitting member of the observation hole 131. The imaging device captures an image of the inside of the high-pressure receiver 100 or the inside of the branch pipe 133 through the light-transmitting member. The refrigerant leak diagnostic device 10 analyzes the image received from the refrigerant state observation device 130 to detect the state of the refrigerant 101 inside the high-pressure receiver 100.

[0076] More specifically, the state of the refrigerant 101 in the high-pressure receiver 100 fluctuates depending on the pressure, temperature, and circulation rate of the refrigerant 101. For example, when the observation hole 131 is located on the side of the high-pressure receiver 100, the liquid level of the refrigerant 101 under normal conditions fluctuates around the vicinity of the observation hole 131. When a leak of refrigerant 101 occurs and an abnormal state develops, the center of the liquid level fluctuation moves below the observation hole, and as the leak progresses further, the center of the liquid level fluctuation descends outside the range of the observation hole, so that a state in which droplets are scattered in the refrigerant gas can be observed from the observation hole 131. Even when the observation hole 131 is located in the branch pipe 133, the refrigerant state transitions similarly as the amount of refrigerant decreases, but a more stable state with less liquid level fluctuation can be observed than when it is located on the side of the high-pressure receiver 100.

[0077] Figure 6 is a diagram illustrating the determination of the refrigerant state based on the output signal of the refrigerant state observation device 130, performed by the refrigerant leak diagnostic device 10 according to the first embodiment of this disclosure. Figure 6(a) shows each pixel of the observation hole 131 captured by the imaging device, where p represents one pixel.

[0078] Figure 6(b) is a graph where the vertical axis represents the ratio of gray to white pixels in all pixels, and the horizontal axis represents time. It shows the time-series changes in the measured value of the ratio of gray to white pixels in all pixels and the time-series changes in the predicted value of the ratio of gray to white pixels in all pixels. Figure 6(c) is a graph where the vertical axis represents the difference between the measured value and the predicted value of the ratio of gray to white pixels in all pixels, and the horizontal axis represents time. It shows the time-series changes in the difference between the measured value and the predicted value of the ratio of gray to white pixels in all pixels.

[0079] The predicted value of the ratio of black and white pixels in all pixels is the output value obtained by inputting the operating data into a prediction model, which is obtained by machine learning using the operating data of the refrigeration and air conditioning equipment 1 as the explanatory variable. The actual value of the ratio of black and white pixels is calculated by binarizing the image data of the observation hole 131 acquired by the refrigerant state observation device 130 during normal operation of the refrigeration and air conditioning equipment 1 using the refrigerant leak diagnosis device 10, and the operating data of the refrigeration and air conditioning equipment 1 is used as the dependent variable.

[0080] The refrigerant leak diagnostic device 10 converts each pixel p of the image data from the observation hole 131 acquired from the refrigerant state observation device 130 into either white or black to create a binary image. Furthermore, the refrigerant leak diagnostic device 10 calculates the ratio of white to black pixels as a representative value for evaluating the created binary image, and evaluates the state of the refrigerant 101 observed from the observation hole 131.

[0081] In the example shown in Figure 6, the value of each pixel p in the image of the observation hole 131 is black and white, and the refrigerant leak diagnostic device 10 evaluates the overall state of the observation hole 131 using the ratio of black and white pixels in all pixels. The ratio of black and white pixels in all pixels is, for example, the ratio of black pixels to white pixels (number of black pixels / number of white pixels).

[0082] Referring to Figures 6(b) and (c), when the refrigerant 101 is in a normal state, the time-series change of the measured value of the ratio of black and white pixels in all pixels will generally follow the waveform of the time-series change of the predicted value of the ratio of black and white pixels in all pixels. When the refrigerant 101 is in an abnormal state, the time-series change of the measured value of the ratio of black and white pixels in all pixels will be a waveform that is more than a predetermined value away from the time-series change of the predicted value of the ratio of black and white pixels in all pixels. Therefore, the refrigerant leak diagnostic device 10 can determine that the refrigerant 101 is in an abnormal state if the difference between the measured value of the ratio of black and white pixels for each pixel p and the predicted value of the ratio of black and white pixels for each pixel p is greater than or equal to a predetermined threshold shown in th2 in Figure 6(c).

[0083] When determining the state of the refrigerant 101 based on the ratio of black and white pixels in all pixels, under normal conditions, the proportion of liquid is large in the field of view of the observation hole 131, and the color of the high-pressure receiver 100 and the metal surface inside the observation hole is reflected through the transparent liquid, resulting in a relatively large number of black pixels. Under abnormal conditions where the amount of refrigerant decreases, the proportion of liquid decreases, and the mixing of liquid and gas increases, leading to foaming and cloudiness, resulting in a relatively large number of white pixels. Therefore, the refrigerant leak diagnostic device 10 can determine that the state of the refrigerant 101 is abnormal if the measured value of the ratio of black and white pixels per pixel p is lower than or equal to a predetermined threshold th2 than the predicted value of the ratio of black and white pixels per pixel p.

[0084] Furthermore, when the refrigerant leak diagnostic device 10 determines the state of the refrigerant 101 using the refrigerant state observation device 130, it can detect changes in the state, such as the inside of the observation hole 131 changing from transparent to cloudy. Therefore, the refrigerant leak diagnostic device 10 can detect changes in the state of the refrigerant 101 in real time. Note that the method for evaluating the image of the observation hole 131 is not limited to the method shown in Figure 6. A black and white image with 256 grayscale values ​​for each pixel p, or a color image with RGB values, may also be used for evaluation.

[0085] Figure 7 shows a modified example of a refrigerant state observation device 130 that outputs a signal to a refrigerant leak diagnostic device 10 according to the first embodiment of this disclosure. The signal conversion unit 132 shown in Figure 7(a) includes a light emitter 1321 that emits specific light L1 into an observation hole 131 and a reflected light detector 1322 that detects reflected light L2 of the specific light L1 from the observation hole 131. The reflected light detector 1322 transmits an output signal to the refrigerant leak diagnostic device 10.

[0086] In the example shown in Figure 7(a), the reflected light detector 1322 detects reflected light L2 that is reflected from the specific light L1 from the light emitter 1321 as it passes through the light-transmitting member of the observation hole 131 and the refrigerant 101 inside the high-pressure receiver 100 or branch pipe 133, reaching the inner surface of the high-pressure receiver 100 or branch pipe 133, at the interface where the air and the outer surface of the light-transmitting member are in contact, the interface where the inner surface of the light-transmitting member is in contact with the liquid or gaseous refrigerant 101, the gas-liquid interface present in the refrigerant 101, and the interface where the liquid or gaseous refrigerant 101 is in contact with the inner surface of the high-pressure receiver 100 or branch pipe 133.

[0087] In the example shown in Figure 7(a), the refrigerant leak diagnostic device 10 can determine the state of the refrigerant 101 based on the reflectance of light calculated based on the intensity of specific light L1 and the intensity of reflected light L2.

[0088] The signal conversion unit 132 shown in Figure 7(b) includes a light emitter 1321 that emits specific light L1 into the first observation hole 131a and a transmitted light detector 1323. The specific light L1 from the light emitter 1321 passes through the first observation hole 131a, the refrigerant 101 inside the high-pressure receiver 100 or branch pipe 133, and the second observation hole 131b which is provided opposite the first observation hole 131a. The transmitted light L3, which is the transmitted specific light L1, is detected by the transmitted light detector 1323. The transmitted light detector 1323 transmits an output signal to the refrigerant leak diagnostic device 10.

[0089] In the example shown in Figure 7(b), the refrigerant leak diagnostic device 10 can determine the state of the refrigerant 101 based on the light transmittance calculated based on the intensity of the specific light L1 and the intensity of the transmitted light L3.

[0090] <<Liquid level detection device 140>> The liquid level detection device 140 is connected to the side of the high-pressure receiver 100 and includes a bypass pipe 144 that bypasses the internal refrigerant 101 to the suction side of the compressor 202, a heat exchanger 141 that heats the refrigerant 101 flowing through the bypass pipe 144 by heat exchange with high-temperature refrigerant flowing through another refrigerant pipe (e.g., compressor outlet pipe or condenser outlet pipe), a temperature sensor 142 that measures the refrigerant temperature after passing through the heat exchanger 141, and an expansion device 143 that adjusts the flow rate of refrigerant flowing through the bypass pipe 144. The refrigerant state data is a value based on the output signal of the temperature sensor 142. The temperature sensor 142 is not limited to a thermistor, for example. The expansion device 143 is for example an expansion valve and a capillary tube.

[0091] The liquid level of the refrigerant 101 in the high-pressure receiver 100 is determined based on the value measured by the temperature sensor 142. If the liquid level of the refrigerant 101 in the high-pressure receiver 100 is above the connection point of the bypass pipe 144 connected to the side, liquid refrigerant flows out into the bypass pipe 144. If the liquid level is at the same position as the connection point, refrigerant 101 in a mixed state of liquid and gas flows out. If the liquid level is below the connection point, gaseous refrigerant flows out. The more gaseous refrigerant contained in the flowing refrigerant 101, the greater the temperature rise after being heated by the high-temperature refrigerant 101 as it flows through the heat exchanger 141. Therefore, the position of the liquid level of the refrigerant 101 relative to the connection point of the bypass pipe 144 can be determined based on the temperature detected by the temperature sensor 142.

[0092] Therefore, multiple liquid level detection devices 140 may be provided on the side of the high-pressure receiver 100, depending on the number of liquid level heights to be detected. By providing multiple liquid level detection devices 140 according to the liquid level height of the refrigerant 101 to be detected, it is possible to determine, for example, whether the liquid level of the refrigerant 101 is within a set range.

[0093] Figure 8 is a diagram illustrating a method by which a refrigerant leak diagnostic device 10 according to a first embodiment of the present disclosure determines the state of the refrigerant 101 based on the output signal of a temperature sensor 142. In Figure 8(a), the vertical axis represents the temperature measured by the temperature sensor 142, the horizontal axis represents time, and the graph shows the time-series change of the measured temperature and the time-series change of the predicted temperature.

[0094] Furthermore, in Figure 8(b), the vertical axis represents the difference between the temperature measured by the temperature sensor 142 and the predicted temperature, and the horizontal axis represents time. The graph shows the time-series change of the difference between the temperature and the predicted temperature. In the figure, th3 indicates a predetermined threshold. The predicted temperature is the value of the refrigerant temperature that is estimated to be detected by the liquid level detection device 140 when the refrigerant 101 is in a normal state. This value is obtained by machine learning with the refrigerant temperature detected by the liquid level detection device 140 connected to the side of the high-pressure receiver 100 of the refrigeration and air conditioning equipment 1 as the dependent variable, and the operating data of the refrigeration and air conditioning equipment 1 as the independent variable.

[0095] Referring to Figures 8(a) and (b), when the refrigerant 101 is in a normal state, the time-series change of the measured temperature will generally follow the waveform of the time-series change of the predicted value. When the refrigerant 101 is in an abnormal state, the time-series change of the measured temperature will deviate from the time-series change of the predicted value by a predetermined amount or more. Therefore, the refrigerant leak diagnostic device 10 can detect that the liquid level of the refrigerant 101 has fallen below the connection point of the liquid level detection device 140 when the difference between the measured temperature and the predicted temperature is greater than or equal to a predetermined threshold th3, and can determine that the state of the refrigerant 101 is abnormal.

[0096] <Learning Model> The refrigerant leak diagnostic device 10 according to this embodiment generates various predictive models using the learning device 500 and performs various judgments. More specifically, the refrigerant leak diagnostic device 10 generates a first predictive model that predicts the amount of refrigerant in the high-pressure receiver 100 by machine learning, using the refrigerant state data in the high-pressure receiver 100 under normal conditions as the target variable and the operating data of the refrigeration and air conditioning equipment 1 as the explanatory variable. The first predictive model can use a predictive model that predicts the normal detection values ​​of the liquid level sensor 110, capacitance sensor 121, refrigerant state observation device 130, and liquid level detection device 140 as described above.

[0097] The refrigerant leak diagnostic device 10 then performs a first determination to estimate the amount of refrigerant in the high-pressure receiver 100 based on a predicted refrigerant state value obtained by inputting the operating data of the refrigeration and air conditioning equipment 1 during operation into a first prediction model. Based on the output from the first prediction model, the refrigerant leak diagnostic device 10 determines whether there is a leak of refrigerant 101 from the refrigeration and air conditioning equipment 1.

[0098] Furthermore, the refrigerant leak diagnostic device 10 generates a second predictive model that predicts the refrigerant amount index corresponding to the refrigerant amount of the refrigeration and air conditioning equipment 1 by using machine learning with a refrigerant amount index that has a particularly strong correlation with the refrigerant amount of the refrigeration and air conditioning equipment 1, extracted from the normal operating data of the refrigeration and air conditioning equipment 1 or calculated values ​​from the normal operating data, as the dependent variable, and the normal operating data as the independent variable.

[0099] Furthermore, the refrigerant leak diagnostic device 10 performs a second determination to estimate the total amount of refrigerant in the refrigerant circuit of the refrigeration and air conditioning equipment 1 based on the predicted value of the refrigerant amount index obtained by inputting operating data into a second prediction model during operation. Since the refrigerant amount index value is an index of the balance change of the refrigeration cycle, even if a refrigerant leak occurs, the refrigerant 101 in the high-pressure receiver 100 compensates for the leak, so a change only occurs when the amount of excess refrigerant becomes small.

[0100] When the amount of excess refrigerant decreases, the liquid level becomes even more unstable, making it difficult to detect the liquid level using the liquid level detection means. Therefore, the refrigerant leak diagnostic device 10 can more reliably determine refrigerant leakage from the refrigeration and air conditioning equipment 1 by considering a second leak determination in addition to the first leak determination.

[0101] Figure 9 is a diagram illustrating the relationship between a refrigeration and air conditioning unit 1, a refrigerant leak diagnostic device 10, and a learning device 500 according to one embodiment of the present disclosure.

[0102] As shown in <Example 1> in the figure, the refrigerant leak diagnostic device 10 may be implemented on a computer installed in the same building as the refrigeration and air conditioning equipment 1. Alternatively, the learning device 500 may be implemented on a cloud server located away from the refrigeration and air conditioning equipment 1 and the refrigerant leak diagnostic device 10.

[0103] As shown in <Example 2> in the figure, the refrigerant leak diagnostic device 10 may be implemented as part of the refrigeration and air conditioning equipment 1 (for example, installed inside the outdoor unit 200 or indoor unit 300). Alternatively, the learning device 500 may be implemented on a cloud server separate from the refrigeration and air conditioning equipment 1 and the refrigerant leak diagnostic device 10.

[0104] As shown in <Example 3> in the figure, the refrigerant leak diagnostic device 10 and the learning device 500 may be implemented on a cloud server located away from the refrigeration and air conditioning equipment 1.

[0105] As shown in <Example 4> in the figure, the refrigerant leak diagnostic device 10 and the learning device 500 may be implemented as part of the refrigeration and air conditioning equipment 1 (for example, installed inside the outdoor unit 200 or the indoor unit 300).

[0106] Figure 10 is a functional block diagram of a learning device 500 according to the first embodiment of this disclosure. The learning device 500 may include a teacher data acquisition unit 501, a teacher data storage unit 502, and a learning unit 503. Furthermore, the learning device 500 can function as the teacher data acquisition unit 501 and the learning unit 503 by executing a program.

[0107] The training data acquisition unit 501 acquires training data. The training data acquisition unit 501 stores the acquired training data in the training data storage unit 502. The training data consists of operating data and refrigerant status data under normal conditions (i.e., when the refrigerant 101 of the refrigeration and air conditioning equipment 1 is at the appropriate amount (hereinafter also referred to as the appropriate amount of refrigerant)). The training data storage unit 502 stores the training data.

[0108] The learning unit 503 extracts data from the normal operating data of the refrigeration and air conditioning equipment 1, which is in a state where the refrigerant charge is appropriate and there are no refrigerant leaks or other malfunctions, and uses only the data of items that have a particularly strong correlation with the refrigerant status data as training data. It then uses each item as an explanatory variable and the refrigerant status data as the objective variable for machine learning. Examples of operating data items that have a strong correlation with the refrigerant amount index value include the outside temperature, the rotational speed of the compressor 202, the condensation temperature, the evaporation temperature, the opening degree of the subcooling heat exchanger expansion valve 204, and the current value of the compressor 202.

[0109] As a result of training using the training data, a first prediction model is generated as a trained model. When another test data consisting of the same items as the training data is input to the first prediction model, the predicted refrigerant state value is output to the refrigerant leak diagnostic device 10.

[0110] Furthermore, a second prediction model is generated that predicts a refrigerant quantity index corresponding to the refrigerant quantity of the refrigeration and air conditioning equipment 1, as a result of learning using training data, which is the normal operating data of the refrigeration and air conditioning equipment 1. When another test data consisting of the same items as the training data is input to the second prediction model, the predicted value of the refrigerant quantity index is output to the refrigerant leak diagnostic device 10.

[0111] Furthermore, the training data does not necessarily have to be extracted from the normal operating data of the refrigeration and air conditioning equipment 1 whose refrigerant quantity index value we want to predict. The training data may be extracted from the normal operating data of a different refrigeration and air conditioning equipment 1, or from the normal operating data of multiple refrigeration and air conditioning equipment 1. In addition, machine learning algorithms such as random forests and support vector machines can be used to create the trained model.

[0112] <Method for diagnosing refrigerant leaks> Figure 11 is a flowchart of a refrigerant leak diagnosis process based on a first determination according to the first embodiment of the present disclosure. The training data acquisition unit 501 acquires operating data and refrigerant status data as training data (S61). The training data acquisition unit 501 stores the acquired training data in the training data storage unit 502.

[0113] The learning unit 503 uses normal operating data as an explanatory variable and refrigerant state data as an objective variable, and performs machine learning by relating the two (S62). As a result of learning by relating the operating data and refrigerant state data, a first predictive model is generated (S63).

[0114] The refrigerant leak diagnostic device 10 performs a first determination to estimate the amount of refrigerant in the high-pressure receiver 100 based on a predicted refrigerant state value obtained by inputting operating data from the refrigeration and air conditioning equipment 1 during operation into a first prediction model (S64). Based on the first determination, the refrigerant leak diagnostic device 10 determines whether there is a refrigerant leak in the refrigeration and air conditioning equipment 1 (S65).

[0115] Figure 12 is a flowchart of the refrigerant leak diagnosis process based on the second determination according to the first embodiment of the present disclosure. The training data acquisition unit 501 acquires normal operating data as training data (S71). The training data acquisition unit 501 stores the acquired training data in the training data storage unit 502.

[0116] The learning unit 503 extracts a refrigerant quantity index that has a strong correlation with the total refrigerant quantity of the refrigerant circuit of the refrigeration and air conditioning equipment 1, based on normal operating data or calculated values ​​thereof, as the target variable, and extracts appropriate explanatory variables from operating data other than the refrigerant quantity index to perform machine learning (S72). As a result of learning from normal operating data, a second predictive model is generated (S73).

[0117] For example, refrigerant quantity indices include the degree of subcooling at the condenser outlet, the degree of subcooling at the subcooling heat exchanger outlet, the degree of compressor suction superheating, and the degree of compressor discharge superheating. Appropriate explanatory variables include ambient temperature, the rotational speed of the compressor 202, the condensation temperature, the evaporation temperature, the opening degree of the subcooling heat exchanger expansion valve 204, and the current value of the compressor 202.

[0118] The refrigerant leak diagnostic device 10 performs a second determination to estimate the total amount of refrigerant in the refrigerant circuit of the refrigeration and air conditioning equipment 1 based on the predicted value of the refrigerant amount index obtained by inputting operating data into the second prediction model during operation (S74). Based on the first determination and the second determination, the refrigerant leak diagnostic device 10 determines whether there is a leak of refrigerant 101 in the refrigeration and air conditioning equipment 1 (S75).

[0119] These steps carry out a refrigerant leak diagnosis method and a learning model generation method according to one aspect of the present invention. However, the refrigerant leak diagnosis method and learning model generation method according to one aspect of the present invention may include other steps as appropriate, depending on the measurement conditions, measurement environment, etc.

[0120] <Main effects of the first embodiment> In a refrigeration and air conditioning system 1 that stores excess refrigerant in a high-pressure receiver 100, the amount of refrigerant accumulated inside the high-pressure receiver 100 fluctuates according to changes in operating conditions such as ambient temperature and load. Because high-speed refrigerant flows into a confined space, the liquid level of the refrigerant 101 inside fluctuates violently, and at times a gas-liquid mixture state without a liquid level appears. Therefore, even with detection means such as a liquid level sensor 110, it has been difficult to accurately detect the liquid level during normal operation.

[0121] According to the refrigerant leak diagnostic device 10 of this embodiment, the liquid level of the refrigerant 101 is determined by comparing the difference between the predicted liquid level value obtained by a first prediction model created by machine learning the refrigerant state data and operating data, which are liquid level information inside the high-pressure receiver 100, and the actual measured liquid level value obtained by the detection means.

[0122] The predicted values ​​do not reflect nonlinear, high-speed liquid level fluctuations, but they do reflect fluctuations in liquid level height due to operating conditions such as ambient temperature and load. Therefore, only the high-speed fluctuation components remain in the difference values. As a result, the refrigerant leak diagnostic device 10 can accurately detect fluctuations in the liquid level height of the refrigerant 101 in the high-pressure receiver 100 by averaging the time-series data of the difference values. This enables high-precision detection of leaks in the refrigeration and air conditioning equipment 1 equipped with the high-pressure receiver 100.

[0123] Furthermore, the refrigerant leak diagnostic device 10 can detect refrigerant leaks during normal operation of the refrigeration and air conditioning equipment 1 by using a second prediction model created with machine learning to analyze changes in the amount of refrigerant in the entire refrigerant circuit. The second prediction model cannot detect leaks while there is a large amount of excess refrigerant, but it becomes capable of detecting leaks when the amount of excess refrigerant decreases and the detection accuracy of the first prediction model declines. Therefore, if there is little excess refrigerant from the time of installation, using both the first and second prediction models together can more reliably detect refrigerant leaks in the refrigeration and air conditioning equipment 1.

[0124] [Second Embodiment] <Configuration of Refrigeration and Air Conditioning Equipment 1> The hardware configuration of the refrigeration and air conditioning equipment 1 according to the second embodiment of this disclosure will be described below with reference to Figure 13. Figure 13 is a configuration diagram of the refrigeration and air conditioning equipment 1 according to the second embodiment of this disclosure. The refrigeration and air conditioning equipment 1 has an outdoor unit 200 and one or more indoor units 300. Unlike the first embodiment, the refrigeration and air conditioning equipment 1 according to this embodiment has an accumulator 400 provided between the outdoor heat exchanger 201 and the compressor 202. The refrigerant leak diagnostic device 10 includes a control unit 11 that determines the state of the refrigerant 101 based on the change in the liquid level of the refrigerant 101 in the accumulator 400.

[0125] In the illustrated example, the outdoor heat exchanger 201, the outdoor unit main expansion valve 205, the subcooling heat exchanger 203, the indoor heat exchanger expansion valve 302, the indoor heat exchanger 301, the four-way switching valve 206, the accumulator 400, and the compressor 202 are connected by refrigerant piping to form the main refrigerant circuit. The four-way switching valve 206 is configured to supply the discharge gas from the compressor 202 to the outdoor heat exchanger 201. Note that components identical to those already described are denoted by the same reference numerals, and redundant explanations are omitted.

[0126] <Configuration of Accumulator 400> Figure 14 is a configuration diagram of an accumulator 400 according to a second embodiment of the present disclosure. As shown in the figure, the accumulator 400 can be equipped with various liquid level detection means connected to the refrigerant leak diagnostic device 10. One or more types of liquid level detection means may be installed.

[0127] More specifically, the accumulator 400 may be equipped with a liquid level sensor 110 and a voltage output unit 111 to measure the liquid level of the refrigerant 101. Alternatively, a capacitance sensor 121 may be equipped inside the accumulator 400 to measure the liquid level of the refrigerant 101. Furthermore, the accumulator 400 may be equipped with a refrigerant state observation device 130 to measure the liquid level of the refrigerant 101. The refrigerant state observation device 130 is connected to the high-pressure receiver 100 and includes a branch pipe 133 equipped with an observation hole 131 and a signal conversion unit 132 that converts the refrigerant state observable from the observation hole 131 into an electrical signal.

[0128] The manner in which the refrigerant leak diagnostic device 10 according to this embodiment acquires refrigerant state data using various sensors is the same as in the first embodiment. The refrigerant state data acquired by the refrigerant leak diagnostic device 10 is a value based on the output signal of the liquid level sensor 110 installed inside the accumulator 400. In addition, the refrigerant state data acquired by the refrigerant leak diagnostic device 10 is a value based on the output signal of the capacitance sensor 121 installed inside the accumulator 400.

[0129] Furthermore, the refrigerant state data acquired by the refrigerant leak diagnostic device 10 is a value based on the output signal of the refrigerant state observation device 130. The refrigerant state observation device 130 includes, as in the first embodiment, at least one observation hole 131 provided on the side of the accumulator 400 or in the branch pipe 133 connected to the accumulator 400.

[0130] Furthermore, the refrigerant state observation device 130 includes a light-transmitting member attached to the observation hole 131 and a signal conversion unit 132 that converts the refrigerant state observed through the observation hole 131 into an electrical signal. Note that since the accumulator 400 is installed on the low-pressure side of the refrigeration cycle, the liquid level detection device 140 that utilizes the high-low pressure difference of the refrigeration cycle cannot be used.

[0131] <Learning Model> The refrigerant leak diagnostic device 10 according to this embodiment generates various prediction models using the learning device 500, similar to the first embodiment, and performs various judgments. More specifically, the refrigerant leak diagnostic device 10 generates a third prediction model that predicts the amount of refrigerant in the accumulator 400 by machine learning the refrigerant state data in the accumulator 400 when the refrigeration and air conditioning equipment 1 is functioning normally, and the operating data of the refrigeration and air conditioning equipment 1.

[0132] The refrigerant leak diagnostic device 10 then performs a third determination to estimate the amount of refrigerant in the accumulator 400 based on a predicted refrigerant state value obtained by inputting the operating data of the refrigeration and air conditioning equipment 1 during operation into a third prediction model. Based on the output from the third prediction model, the refrigerant leak diagnostic device 10 determines whether there is a leak of refrigerant 101 from the refrigeration and air conditioning equipment 1.

[0133] Furthermore, similar to the first embodiment, the refrigerant leak diagnostic device 10 generates a fourth prediction model that predicts a refrigerant amount index corresponding to the amount of refrigerant in the refrigeration and air conditioning equipment 1 by machine learning the normal operating data of the refrigeration and air conditioning equipment 1. In addition, the refrigerant leak diagnostic device 10 performs a fourth determination to estimate the amount of refrigerant in the entire refrigerant circuit of the refrigeration and air conditioning equipment 1 based on the predicted value of the refrigerant amount index obtained by inputting operating data during operation into the fourth prediction model. Based on the third determination and the fourth determination, the refrigerant leak diagnostic device 10 determines whether there is a leak of refrigerant 101 in the refrigeration and air conditioning equipment 1.

[0134] The relationship between the refrigeration and air conditioning equipment 1, the refrigerant leak diagnostic device 10, and the learning device 500 according to this embodiment, as well as the functional blocks of the learning device 500, are the same as in the first embodiment, so their explanation will be omitted.

[0135] In this embodiment, a third prediction model is generated as a trained model as a result of training using the training data. When another test data consisting of the same items as the training data is input to the third prediction model, the refrigerant state prediction value is output to the refrigerant leak diagnostic device 10.

[0136] Furthermore, similar to the first embodiment, a fourth prediction model is generated that predicts a refrigerant quantity index corresponding to the refrigerant quantity of the refrigeration and air conditioning equipment 1, as a result of learning using training data, which is the normal operating data of the refrigeration and air conditioning equipment 1. When another test data consisting of the same items as the training data is input to the fourth prediction model, the predicted value of the refrigerant quantity index is output to the refrigerant leak diagnostic device 10.

[0137] <Main effects of the second embodiment> Similar to the case of the high-pressure receiver 100, it is difficult to detect the liquid level of the refrigerant 101 accumulating inside the accumulator 400 during normal operation. According to the refrigerant leak diagnostic device 10 of this embodiment, in a refrigeration and air conditioning equipment 1 that stores excess refrigerant in the accumulator 400, the amount of refrigerant in the accumulator 400, which fluctuates violently, can be accurately estimated by using a third prediction model created by machine learning, and a leak of refrigerant 101 can be detected during the normal operation of the refrigeration and air conditioning equipment 1.

[0138] Furthermore, the refrigerant leak diagnostic device 10 can detect refrigerant leaks during normal operation of the refrigeration and air conditioning equipment 1 by using a fourth prediction model created with machine learning to analyze changes in the amount of refrigerant in the entire refrigerant circuit. Similar to the case of the high-pressure receiver 100, when the amount of excess refrigerant inside the accumulator 400 becomes small, prediction becomes difficult with the third prediction model, and detection becomes possible with the fourth prediction model. Therefore, by using both the third and fourth prediction models in combination, refrigerant leaks in the refrigeration and air conditioning equipment 1 can be detected more reliably.

[0139] <Other application examples> Although embodiments have been described above, it should be understood that various modifications to the form and details are possible without departing from the spirit and scope of the claims.

[0140] The apparatus described in the examples represents only one of several computing environments for carrying out the embodiments disclosed herein. In one embodiment, the refrigerant leak diagnostic apparatus 10 includes several computing devices, such as a server cluster. The several computing devices are configured to communicate with each other via any type of communication link, including a network or shared memory, and to perform the processes disclosed herein.

[0141] Each of the functions of this disclosure described above can be implemented not only by software processing through program execution, but also by one or more processing circuits. Here, "processing circuit" as used herein includes processors programmed to execute each function by software, such as processors implemented by electronic circuits, as well as devices such as ASICs (Application Specific Integrated Circuits), DSPs (Digital Signal Processors), FPGAs (Field Programmable Gate Arrays), and conventional circuit modules designed to execute each of the functions described above.

[0142] <Reasons why the effect occurs> A first aspect of this disclosure is a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally between a condenser and an expansion valve. The system inputs operating data from the refrigeration and air conditioning system into a first prediction model generated based on the refrigerant state prediction value obtained from the refrigerant state data in the high-pressure receiver during normal operation and the operating data of the refrigeration and air conditioning system. Based on this prediction value, a first determination is made to estimate the amount of refrigerant in the high-pressure receiver, and based on the output from the first prediction model, refrigerant leakage from the refrigeration and air conditioning system is determined. Since the state of the refrigerant 101 is determined after predicting the amount of refrigerant by machine learning using the refrigerant state data and operating data, refrigerant leakage can be detected in real time and with higher accuracy based on fluctuations in the liquid level and gas-liquid mixing conditions of the refrigerant 101 in the high-pressure receiver 100.

[0143] A second aspect of this disclosure involves "inputting operating data during operation into a second prediction model that predicts a refrigerant amount index corresponding to the amount of refrigerant in the refrigeration and air conditioning equipment, which is generated based on normal operating data of the refrigeration and air conditioning equipment, and performing a second determination to estimate the amount of refrigerant in the entire refrigerant circuit of the refrigeration and air conditioning equipment based on the predicted value of the refrigerant amount index obtained, and determining refrigerant leakage from the refrigeration and air conditioning equipment based on the first determination and the second determination," thereby enabling more reliable detection of refrigerant leakage 101 from the refrigeration and air conditioning equipment 1.

[0144] A third aspect of this disclosure is that, since "the refrigerant state data is a value based on the output signal of a liquid level sensor installed inside the high-pressure receiver," the liquid level height of the refrigerant 101 can be continuously estimated with a simple configuration, and the amount of change in the refrigerant 101 can be determined with high accuracy by combining it with machine learning.

[0145] A fourth aspect of this disclosure is that "the refrigerant state data is a value based on the output signal of a capacitance sensor installed inside the high-pressure receiver," so the state of the refrigerant 101 can be continuously determined by the capacitance value, and by combining this with machine learning, the amount of change in the refrigerant 101 can be determined with high accuracy.

[0146] A fifth aspect of this disclosure is that "the refrigerant state data is a value based on the output signal of a refrigerant state observation device having at least one observation hole on the side of the high-pressure receiver or on a branch pipe connected to the high-pressure receiver, a light-transmitting member mounted on the observation hole, and a signal conversion unit that converts the refrigerant state observed through the observation hole into an electrical signal." Therefore, changes in the state of the refrigerant 101 inside the high-pressure receiver 100 or the branch pipe 133 can be determined by appearance, and by combining this with machine learning, the amount of change in the refrigerant 101 can be determined with high accuracy.

[0147] A sixth aspect of this disclosure is that "the signal conversion unit is an imaging device, and the output signal is image data of the observation hole captured by the imaging device," so that the state of the refrigerant 101 inside the high-pressure receiver 100 or the branch pipe 133 can be determined using the imaging device, and the amount of change in the refrigerant 101 can be determined with high accuracy by combining it with machine learning.

[0148] A seventh aspect of this disclosure is that "the signal conversion unit comprises a light projector that projects specific light into the observation hole, and a reflected light detector that detects reflected light of the specific light from the observation hole, and the output signal is the output signal of the reflected light detector." Therefore, the state of the refrigerant 101 can be determined by the reflectance of light based on the intensity of the specific light and the intensity of the reflected light, and by combining this with machine learning, the amount of change in the refrigerant 101 can be determined with high accuracy.

[0149] An eighth aspect of this disclosure is that "the signal conversion unit comprises a light emitter that emits specific light into the observation hole, a first observation hole which is the observation hole, the refrigerant inside the high-pressure receiver or the refrigerant inside the branch pipe, a second observation hole provided opposite the first observation hole, and a transmitted light detector that detects the transmitted light of the specific light passing through the first observation hole, and the output signal is the output signal of the transmitted light detector." Therefore, the state of the refrigerant 101 can be determined by the transmittance of light based on the intensity of the specific light and the intensity of the transmitted light, and by combining this with machine learning, the amount of change in the refrigerant 101 can be determined with high accuracy.

[0150] A ninth aspect of this disclosure is that, since "the refrigerant state data is a value based on the output signal of the temperature sensor," changes in the state of the refrigerant 101 can be easily detected based on temperature changes, and the amount of change in the refrigerant 101 can be determined with even greater accuracy by combining it with machine learning. [Explanation of symbols]

[0151] 1 Refrigeration and air conditioning equipment 10 Refrigerant Leakage Diagnostic Device 11 Control Unit 100 High-voltage receiver 101 Refrigerant 110 Liquid level sensor 121 Capacitive Sensor 130 Refrigerant State Observation Device 131 Observation hole 131a First observation hole 131b Second observation hole 132 Signal Conversion Section 1321 Floodlight 1322 Reflected light detector 1323 Transmitted light detector 133 Branch pipe 140 Liquid level detection device 141 Heat exchanger 142 Temperature Sensor 143 Expansion device 144 Bypass piping 400 Accumulator

Claims

1. A refrigerant leak diagnostic device comprising a control unit, The control unit, In a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally between a condenser and an expansion valve, a first determination is made to estimate the amount of refrigerant in the high-pressure receiver based on a refrigerant state prediction value obtained by inputting the operating data of the refrigeration and air conditioning system during operation into a first prediction model generated based on the refrigerant state data in the high-pressure receiver during normal operation of the refrigeration and air conditioning system and the operating data of the refrigeration and air conditioning system. Based on the output from the first prediction model, the leakage of refrigerant from the refrigeration and air conditioning equipment is determined. Refrigerant leak diagnostic device.

2. The control unit, During operation, the operating data is input into a second prediction model that predicts a refrigerant amount index corresponding to the amount of refrigerant in the refrigeration and air conditioning equipment, which is generated based on the normal operating data of the refrigeration and air conditioning equipment. Based on the predicted value of the refrigerant amount index obtained, a second determination is made to estimate the amount of refrigerant in the entire refrigerant circuit of the refrigeration and air conditioning equipment. Based on the first determination and the second determination, a leak of refrigerant from the refrigeration and air conditioning equipment is determined. The refrigerant leak diagnostic device according to claim 1.

3. The refrigerant state data is a value based on the output signal of a liquid level sensor installed inside the high-pressure receiver. A refrigerant leak diagnostic device according to claim 1 or 2.

4. The refrigerant state data is a value based on the output signal of a capacitance sensor installed inside the high-pressure receiver. A refrigerant leak diagnostic device according to claim 1 or 2.

5. The aforementioned refrigerant state data is The side of the high-voltage receiver or the branch pipe connected to the high-voltage receiver has at least one observation hole, A light-transmitting member fitted into the observation hole, and a signal conversion unit that converts the refrigerant state observed through the observation hole into an electrical signal, This value is based on the output signal of a refrigerant state observation device. A refrigerant leak diagnostic device according to claim 1 or 2.

6. The signal conversion unit is an imaging device, The output signal is image data of the observation hole captured by the imaging device. The refrigerant leak diagnostic device according to claim 5.

7. The signal conversion unit is A light source that projects specific light into the observation hole, A reflected light detector for detecting the reflected light of the specific light from the observation hole, Equipped with, The output signal is the output signal of the reflected light detector. The refrigerant leak diagnostic device according to claim 5.

8. The signal conversion unit is A light source that projects specific light into the observation hole, A transmitted light detector for detecting transmitted light of the specific light passing through a first observation hole, which is the observation hole, the refrigerant inside the high-pressure receiver or the refrigerant inside the branch pipe, and a second observation hole provided opposite the first observation hole, Equipped with, The output signal is the output signal of the transmitted light detector. The refrigerant leak diagnostic device according to claim 5.

9. The aforementioned refrigerant state data is a value based on the output signal of the temperature sensor. The aforementioned temperature sensor is In the bypass piping that branches off from the side of the high-pressure receiver and is connected to the compressor suction piping via an expansion device, and bypasses the refrigerant in the high-pressure receiver to the compressor suction side, a heat exchanger is provided upstream of the expansion device and downstream of the heat exchanger for heating the refrigerant flowing out of the high-pressure receiver by heat exchange with other refrigerant piping. A refrigerant leak diagnostic device according to claim 1 or 2.

10. A refrigerant leak diagnostic device comprising a control unit, The control unit, In a refrigeration and air conditioning system equipped with an accumulator that stores refrigerant internally between an evaporator and a compressor, a third determination is made to estimate the amount of refrigerant in the accumulator based on a refrigerant state prediction value obtained by inputting the operating data of the refrigeration and air conditioning system during operation into a third prediction model, which is a prediction model that predicts the amount of refrigerant in the accumulator based on the refrigerant state data in the accumulator during normal operation of the refrigeration and air conditioning system and the operating data of the refrigeration and air conditioning system. Based on the output from the third prediction model, the leakage of refrigerant from the refrigeration and air conditioning equipment is determined. Refrigerant leak diagnostic device.

11. The control unit, During operation, the operating data is input into a fourth prediction model that predicts a refrigerant amount index corresponding to the refrigerant amount of the refrigeration and air conditioning equipment based on the normal operating data of the refrigeration and air conditioning equipment. Based on the predicted value of the refrigerant amount index obtained, a fourth determination is made to estimate the total amount of refrigerant in the refrigeration and air conditioning equipment's refrigerant circuit. Based on the third and fourth determinations, a leak of refrigerant from the refrigeration and air conditioning equipment is determined. The refrigerant leak diagnostic device according to claim 10.

12. The refrigerant state data is a value based on the output signal of a liquid level sensor installed inside the accumulator. The refrigerant leak diagnostic device according to claim 10 or 11.

13. The refrigerant state data is a value based on the output signal of a capacitance sensor installed inside the accumulator. The refrigerant leak diagnostic device according to claim 10 or 11.

14. The aforementioned refrigerant state data is At least one observation hole provided on the side of the accumulator or in the branch pipe connected to the accumulator, A light-transmitting member fitted into the observation hole, and a signal conversion unit that converts the refrigerant state observed through the observation hole into an electrical signal, This value is based on the output signal of a refrigerant state observation device. The refrigerant leak diagnostic device according to claim 10 or 11.

15. The signal conversion unit is an imaging device, The output signal is image data of the observation hole captured by the imaging device. The refrigerant leak diagnostic device according to claim 14.

16. The signal conversion unit is A light source that projects specific light into the observation hole, A reflected light detector for detecting the reflected light of the specific light from the observation hole, Equipped with, The output signal is the output signal of the reflected light detector. The refrigerant leak diagnostic device according to claim 14.

17. The signal conversion unit is A light source that projects specific light into the observation hole, A transmitted light detector for detecting transmitted light of the specific light passing through a first observation hole, which is the observation hole, the refrigerant inside the accumulator or the refrigerant inside the branch pipe, and a second observation hole provided opposite the first observation hole, Equipped with, The output signal is the output signal of the transmitted light detector. The refrigerant leak diagnostic device according to claim 14.

18. In a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally between a condenser and an expansion valve, the first step is to perform a first determination to estimate the amount of refrigerant in the high-pressure receiver based on a refrigerant state prediction value obtained by inputting the operating data of the refrigeration and air conditioning system during operation into a first prediction model generated based on the refrigerant state data in the high-pressure receiver during normal operation of the refrigeration and air conditioning system and the operating data of the refrigeration and air conditioning system. A step of determining whether there is a refrigerant leak in the refrigeration and air conditioning equipment based on the output from the first prediction model, A refrigerant leak diagnostic method that includes [specific component / feature].

19. In a refrigeration and air conditioning system equipped with an accumulator that stores refrigerant internally between an evaporator and a compressor, a third determination is made by inputting operating data during the operation of the refrigeration and air conditioning system into a third prediction model, which is a prediction model that predicts the amount of refrigerant in the accumulator based on the refrigerant state prediction value obtained from the refrigerant state prediction value obtained from the refrigerant state prediction value, and The steps include determining whether there is a refrigerant leak in the refrigeration and air conditioning equipment based on the output from the third prediction model, A refrigerant leak diagnostic method that includes [specific component / feature].

20. In a refrigeration and air conditioning system equipped with a high-pressure receiver that stores refrigerant internally between a condenser and an expansion valve, a predictive model is generated to predict the amount of refrigerant in the high-pressure receiver by machine learning using refrigerant state data in the high-pressure receiver and operating data of the refrigeration and air conditioning system. Method for generating a learning model.

21. In a refrigeration and air conditioning system equipped with an accumulator that stores refrigerant internally between an evaporator and a compressor, a predictive model is generated to predict the amount of refrigerant in the accumulator by machine learning using refrigerant state data in the accumulator and operating data of the refrigeration and air conditioning system. Method for generating a learning model.

Images