Method for determining performance of air conditioning system based on remote monitoring

Abstract

The application relates to the field of airborne air conditioner performance monitoring, and discloses a remote monitoring air conditioner system performance judgment method, which comprises the following steps: S1, collecting A21 message and A24 message data through hardware FDIMU; S2, performing data decoding, regression calculation, mean variance calculation, left and right air conditioner data difference calculation, normalization processing on the A21 message and A24 message data, calculating trend analysis of each parameter, and marking + / ‑2sigma external data; S3, based on air conditioner steady-state core characteristic parameter data and an AHP (analytic hierarchy process) analysis method, establishing an important parameter evaluation model; and S4, applying a PNN (probabilistic neural network) to classify and identify different failure modes of the air conditioner according to a Bayesian criterion, and judging the failure sub-components of the air conditioner system according to characteristic samples. The application can remotely monitor the overall performance of the air conditioner system, assist in troubleshooting of the air conditioner system, and reduce unplanned replacement caused by sudden failure of the aircraft air conditioner and unnecessary cost increase caused by regular replacement based on reliability.

Description

Technical Field

[0001] This invention relates to the field of airborne air conditioning performance monitoring technology, specifically a method for determining the performance of air conditioning systems (including but not limited to heat exchangers, flow control valves, bypass valves, condensers, inlet damper actuators, and air compressor leaks) based on remote monitoring. Background Technology

[0002] Airborne air conditioning systems are complex and sophisticated systems whose performance is greatly affected by the external environment and operating conditions. Under different temperature, humidity, and flight altitude conditions, the heat exchange efficiency of the air conditioning system varies significantly, making it difficult to accurately assess the performance of the air conditioning radiator using traditional heat transfer analysis. For example, when an aircraft is operating in a high-temperature region, the air conditioning system needs to cope with large temperature differences and high humidity environments, while at high altitudes, it needs to cope with low temperatures and low pressure environments. This complex and variable operating environment makes the performance evaluation of the air conditioning system exceptionally difficult. Summary of the Invention

[0003] To address the issue that existing airborne air conditioning systems lack sufficient sensors, making it impossible to calculate the performance of sub-components using heat transfer theory, this invention provides a method for remotely monitoring and judging the performance of air conditioning systems based on statistics and PNN theory. The method utilizes FDIMU customer-customized programming to evaluate the air conditioning system's operating status under stable conditions in a fixed environment, primarily through data collected in message A21, supplemented by ground-based data collected in message A24.

[0004] To achieve the above objectives, the present invention provides a method for determining the performance of an air conditioning system based on remote monitoring, characterized by comprising the following steps: The air conditioning system's steady-state data in the air conditioning system (message A21) and the ground air conditioning operation data (message A24) are collected using the hardware FDIMU. Among them, the in-flight steady-state logic is as follows: within 100 seconds, the angle of the dual-engine turbine cooling valve is between 78% and 100%, and the fluctuation of the following parameters does not exceed the set value: The stability calculation formula is as follows:

[0005]

[0006] in, This represents the weight of each parameter in Table 1, with a default value of 1, which can be adjusted according to target requirements. As an independent individual within a set of parameters, The mean of a set of parameters. The standard deviation is denoted as .

[0007] in, =00 is the best, and 99 is the worst.

[0008] GS Aircraft ground speed ROLL ANGLE Roll angle TAT Total temperature EGT Exhaust temperature VACC Turbine cooling valve angle MN Mach number N1 Engine N1 speed PT2 Engine inlet pressure FF Fuel flow EPR (if any) Engine boost ratio Table 1 Data decoding, regression calculation, mean and variance calculation, left and right air conditioner data difference calculation, normalization processing are performed on the data of messages A21 and A24. Trend analysis of each parameter and + / -2σ external data calibration are also performed. Based on the steady-state core characterization parameter data of air conditioning and the AHP hierarchical analysis method, an evaluation model for important parameters is established; The PNN probabilistic neural network is applied to classify and identify the normal operation mode and different failure modes of the air conditioner based on Bayesian criteria, and the failure sub-components of the air conditioning system are determined based on feature samples.

[0009] Preferably, the A21 message data includes inlet damper angle RI, compressor outlet temperature COT, water separator outlet temperature TW, PACK flow rate PF, PACK outlet temperature TF, and bypass valve angle PBV.

[0010] Preferably, the A24 message data includes TW and COT data for evaluation reference.

[0011] Preferably, the evaluation model for the important parameters adopts a linear weighted summation method, and has different weight coefficients depending on the model.

[0012] Preferably, the feature samples include samples from cruising air and ground samples.

[0013] Preferably, the samples taken during the cruise flight include samples with a high RI (40 degrees), a COT temperature below 120 degrees, and a TW above 15 degrees, while the samples taken on the ground include samples with an RI of 40 degrees, a COT close to 200 degrees, and a TW exceeding 30 degrees.

[0014] Preferably, it also includes: in a stable cruise environment, setting parameter thresholds and evaluating the data in message A21 for reference, and triggering an alarm when parameters exceed the threshold.

[0015] Preferably, the alarm is triggered when the parameter exceeds the threshold twice consecutively.

[0016] Preferably, it also includes: improving the model's recognition ability by relying on human-assisted enhanced classification learning.

[0017] Preferably, the artificially assisted reinforcement classification learning includes using expert experience judgment and historical maintenance data analysis to correct and optimize the weight and threshold parameters of the PNN probabilistic neural network.

[0018] This invention provides a method for determining the performance of an air conditioning system based on remote monitoring. It has the following beneficial effects: This invention enables remote monitoring of the overall performance of the air conditioning system and the performance of its sub-components, and assists in troubleshooting. It also reduces unnecessary cost increases caused by unplanned replacements due to sudden aircraft air conditioning malfunctions and by periodic, reliability-based replacements. This method introduces a performance-based, planned air conditioning maintenance concept, combining aircraft operation, manpower, and resource allocation to optimize the cost of air conditioning maintenance, thereby minimizing aircraft maintenance costs while meeting airworthiness requirements. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the method logic processing of the present invention; Figure 2 This is a schematic diagram of the logic processing in one embodiment of the method of the present invention; Figure 3 This is a comparison chart of COT parameters before and after replacing the heat sink according to the present invention; Figure 4 This is a comparison chart of RI parameters before and after replacing the heat sink in this invention; Figure 5 This is a comparison chart of TW parameters before and after replacing the heat sink in this invention; Figure 6 This is a comparison image of the air leakage elimination process before and after the invention. Figure 7 This is a comparison image of the condenser before and after replacement according to the present invention; Figure 8 This is a comparison image of the imported damper actuator before and after replacement according to the present invention; Figure 9 This is a schematic diagram of the A21 message data of the present invention; Figure 10 This is a schematic diagram of the A24 message data of the present invention. Detailed Implementation

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see the appendix Figure 1 This invention provides a method for determining the performance of an air conditioning system based on remote monitoring, comprising the following steps: S1. Data Acquisition: The FDIMU (Flight Data Interface Management Unit) hardware is used to collect A21 message data from the air conditioning system during steady-state flight and A24 message data during ground operation. Figure 9 and 10 Examples of A21 and A24 message data are shown respectively; S2, Data Processing: The collected data from messages A21 and A24 are processed as follows: Data decoding: Converting raw data into an analyzable format.

[0022] Regression calculation: Determine the correlation between parameters through regression analysis.

[0023] Mean and variance calculation: Calculate the statistical properties of data, such as mean and variance.

[0024] Calculation of data difference between left and right air conditioning systems: Compare the data differences between the left and right air conditioning systems.

[0025] Normalization: Converting data to the same unit of measurement to facilitate comparison and analysis.

[0026] Trend analysis: Analyze the changing trends of each parameter over time.

[0027] + / -2σ outlier calibration: Marks outliers that fall outside the two standard deviations.

[0028] S3. Evaluation Model Establishment: The Analytic Hierarchy Process (AHP) was used to determine the core characterization parameters of the air conditioning system during steady-state operation, and an evaluation model for these important parameters was established based on these parameters. The AHP decomposes the problem by constructing a hierarchical model, then compares, ranks, and comprehensively evaluates the elements at each level to ultimately obtain the weights of each parameter. S4. Perform classification and identification: The Probabilistic Neural Network (PNN) is applied for classification and recognition. A PNN is a type of neural network based on Bayesian principles that can classify the operating status of an air conditioning system into normal mode or different failure modes based on feature samples. The PNN network learns the characteristics of normal operation and various failure states of the air conditioning system through training samples, and then uses statistical probability theory to classify and recognize new data samples.

[0029] This invention enables remote monitoring of the overall performance of the air conditioning system and the performance of its sub-components, and assists in troubleshooting. It also reduces unnecessary cost increases caused by unplanned replacements due to sudden aircraft air conditioning malfunctions and by periodic, reliability-based replacements. This method introduces a performance-based, planned air conditioning maintenance concept, combining aircraft operation, manpower, and resource allocation to optimize the cost of air conditioning maintenance, thereby minimizing aircraft maintenance costs while meeting airworthiness requirements.

[0030] Please see the appendix Figure 2 As one embodiment of the present invention, the method further includes: S5, improving the model's recognition ability based on artificially assisted reinforcement classification learning.

[0031] In some embodiments, the artificially assisted reinforcement classification learning includes using expert experience judgment and historical maintenance data analysis to correct and optimize the weights and threshold parameters of the PNN probabilistic neural network, combining the experience of human experts with the computational power of machine learning algorithms to achieve higher accuracy and efficiency.

[0032] Specifically, expert experience determines: An expert system is a knowledge processing system that simulates the decision-making process of human experts. It typically contains a large amount of domain-specific knowledge and judgment rules. In this invention, an expert system can be constructed that incorporates extensive experience in the operation and maintenance of air conditioning systems, as well as expertise in fault diagnosis. This system can be used to perform preliminary evaluations of the output of a Probabilistic Neural Network (PNN), or to provide decision support in ambiguous or difficult-to-determine situations.

[0033] Historical maintenance data analysis: Historical maintenance data is a valuable resource for diagnosing air conditioning system faults, containing all known failure cases and corresponding maintenance measures. Analyzing this data can reveal correlations between failure modes and specific data features, thereby allowing for the correction and optimization of the weights and threshold parameters of the PNN model. This analysis can be achieved through data mining techniques such as association rule learning and cluster analysis.

[0034] Weight and threshold parameter optimization: In probabilistic neural networks (PNNs), weights and threshold parameters are crucial in determining network performance. Through artificially assisted reinforcement learning, these parameters can be fine-tuned based on expert system judgments and historical maintenance data analysis. This may involve adjusting the connection strength (weights) and activation conditions (thresholds) of individual neurons in the neural network to improve the network's ability to identify specific fault modes.

[0035] By combining artificial intelligence with the experience of human experts, the overall performance of fault diagnosis systems can be significantly improved, thus providing a more powerful and reliable support tool for the operation and maintenance of air conditioning systems.

[0036] As one embodiment of the present invention, the A21 message data includes inlet damper angle RI, compressor outlet temperature COT, water separator outlet temperature TW, PACK flow rate PF, PACK outlet temperature TF, and bypass valve angle PBV. Monitoring and analyzing these parameters helps to assess the performance status of the air conditioning system in real time and to detect and prevent potential faults in a timely manner.

[0037] Specifically, the inlet damper angle RI: The inlet damper angle RI is a parameter used in the air conditioning system to control the position of the inlet damper, reflecting the entry of outside air. By collecting and recording the inlet damper angle RI data through the FDIMU, the entry of outside air can be monitored in real time and combined with other parameters for comprehensive analysis.

[0038] Compressor outlet temperature (COT): COT refers to the air temperature at the compressor outlet in an air conditioning system. Its changes reflect the compressor's operating status. Normalizing and trend analyzing COT data can help detect compressor malfunctions in a timely manner, thus aiding in the prevention of compressor failures.

[0039] Water separator outlet temperature TW: The water separator outlet temperature TW represents the temperature at the water separator outlet, which is an important operating parameter in the air conditioning system. Monitoring changes in TW data can promptly detect abnormal operating conditions of the water separator, ensuring the normal operation of the air conditioning system.

[0040] PACK flow rate (PF): PACK flow rate (PF) refers to the flow parameter in an air conditioning system, reflecting the flow of air or liquid. Normalizing the PACK flow rate (PF) and calibrating it with + / -2σ external data can eliminate outliers and ensure the accuracy of the evaluation model.

[0041] PACK Outlet Temperature TF: PACK outlet temperature TF represents the air temperature at the PACK outlet in the air conditioning system, and it is one of the important indicators of the air conditioning system's operating status. By monitoring changes in the PACK outlet temperature TF, it is possible to determine whether the PACK is operating normally.

[0042] Bypass valve angle (PBV): The bypass valve angle (PBV) controls the position of the bypass valve, affecting the airflow in the air conditioning system. Monitoring the PBV data allows for timely detection of valve position anomalies, facilitating adjustments to airflow and ensuring the normal operation of the air conditioning system.

[0043] By real-time monitoring and data analysis of these key parameters, combined with PNN probabilistic neural networks and artificially assisted reinforcement classification learning, this invention can effectively identify and predict the performance status of air conditioning systems, providing scientific decision support for maintenance work. This not only improves the operating efficiency and reliability of air conditioning systems but also saves maintenance personnel a significant amount of time and resources when diagnosing and handling faults.

[0044] In one embodiment of the present invention, the A24 message data includes TW and COT data for evaluation reference.

[0045] Specifically, the TW (water separator outlet temperature) data used for evaluation: TW is the water separator outlet temperature data used as a reference for evaluation. The water separator outlet temperature is one of the important parameters in an air conditioning system, reflecting the working status and efficiency of the water separator. By monitoring and recording the TW data, it can serve as one of the reference indicators for evaluating radiator performance and help determine the working status of the air conditioning system.

[0046] COT (Compressor Outlet Temperature) data used for evaluation: COT is the compressor outlet temperature data used as a reference for evaluation. Compressor outlet temperature is another important operating parameter in an air conditioning system, reflecting the compressor's operating status and efficiency. Normalizing and trend analyzing COT data can serve as one of the reference indicators for evaluating radiator performance and help determine the operating status of the air conditioning system.

[0047] By collecting and recording TW and COT data for evaluation reference, and combining them with other parameter data in message A21, an evaluation model for radiator performance can be established. This evaluation model can comprehensively consider the changes in various parameters and utilize the PNN neural network failure mode intelligent identification method to determine and evaluate the working status of the radiator, thereby improving the accuracy and effectiveness of radiator performance assessment.

[0048] As one embodiment of the present invention, the important parameter evaluation model adopts a linear weighted summation method, and has different weight coefficients depending on the model.

[0049] For the A321 model, the performance evaluation formula is as follows: PDI_AC = 0.1789 * (RI - 7) / 23 + 0.3938 * (200 - COT) / 90 + 0.4779* ((PBV - 23) / 27) + 0.0493 * (PF_MAX - PF) / 0.25 For the A320 model, the performance evaluation formula is as follows: PDI_AC = 0.1789 * (RI - 7) / 23 + 0.3938 * (200 - COT) / 100 + 0.4779* ((PBV - 21) / 25) + 0.0493 * (PF_MAX - PF) / 0.25 For the A319 model, the performance evaluation formula is as follows: PDI_AC = 0.1789 * (RI - 7) / 23 + 0.3938 * (190 - COT) / 100 + 0.4779* ((PBV - 20) / 25) + 0.0493 * (PF_MAX - PF) / 0.25 Wherein, PF_MAX is the maximum PF value, which varies depending on the model: A321=0.7, A320=0.65, and A319=0.55.

[0050] The calculation method for the performance evaluation index PDI_AC is as follows: Each parameter is standardized by dividing the deviation between the measured value and the set benchmark value by a range value to normalize it, so that the contribution of the parameter is limited to between 0 and 1.

[0051] Each normalized parameter value is multiplied by its corresponding weighting coefficient.

[0052] Sum all the weighted parameter values ​​to obtain the performance evaluation index PDI_AC.

[0053] This method yields a quantifiable performance evaluation index, PDI_AC, which provides a clear indication of the air conditioning system's operational status and whether maintenance or adjustments are necessary. This evaluation model helps maintain the air conditioning system at its optimal operating level, improving aircraft operational efficiency and passenger comfort.

[0054] As one embodiment of the present invention, the feature samples include samples from cruising in the air and samples from the ground.

[0055] Specifically, the samples taken during cruising included RI at a high setting (40 degrees), COT below 120 degrees, and TW above 15 degrees. The samples taken on the ground included RI at 40 degrees, COT close to 200 degrees, and TW above 30 degrees.

[0056] By analyzing these feature samples, the PNN neural network model can be trained and optimized, thereby improving the ability to identify the performance status of air conditioning systems. These samples reflect the extreme conditions that air conditioning systems may encounter in actual operation, ensuring the generalization ability and practicality of the evaluation model.

[0057] In practical applications, these feature samples are used to train neural networks, enabling them to identify the performance status of air conditioning systems under different operating conditions. Furthermore, data collected during actual operation is compared with these feature samples to determine whether the system's current state is normal or if potential performance problems exist. Such diagnostic capabilities are crucial for aircraft maintenance and operation, helping to identify and resolve issues proactively and prevent potential system failures.

[0058] As one embodiment of the present invention, the method further includes: S6, in a stable cruise environment, setting parameter thresholds and evaluating the data of message A21 for reference, and triggering an alarm when parameters exceed the thresholds.

[0059] Specifically, this step is crucial because a stable environment during cruise flight is essential for accurately assessing radiator performance. Below are some implementation methods for this step: Parameter threshold settings under stable cruise conditions: During cruise flight, the system can monitor and record parameters related to radiator performance, such as temperature, flow rate, and damper angle. For each parameter, corresponding thresholds can be set according to the aircraft's design specifications and performance requirements. These threshold values ​​can be determined based on the specific requirements of the aircraft and actual test data to ensure a reasonable evaluation of radiator performance under stable cruise conditions.

[0060] Evaluation and reference of A21 message data: The system can evaluate and reference the real-time collected A21 message data by comparing it with preset parameter thresholds. If any parameter exceeds the preset threshold range, the system will trigger an alarm.

[0061] Alarm Trigger: Once a parameter exceeds a threshold range, the system can immediately trigger an alarm, notifying the pilot or ground control personnel that there may be an abnormality in the radiator performance. This allows for timely implementation of necessary measures to ensure the safe operation of the aircraft.

[0062] The system can monitor and evaluate radiator performance in real time under stable cruise conditions. Once an anomaly is detected, timely measures can be taken, which improves the monitoring capability of aircraft radiator performance and helps ensure the safe flight of the aircraft.

[0063] In one embodiment of the present invention, the alarm is triggered when the parameter exceeds the threshold twice consecutively.

[0064] Specifically, the system will only trigger an alarm when a parameter exceeds a preset threshold in two consecutive monitoring sessions. This setting takes into account potential noise or transient interference during data acquisition. By requiring two consecutive exceedances of the threshold, the possibility of false alarms can be reduced, thus improving the accuracy of the alarm. Example 1:

[0065] Because an air-borne air conditioning system is a complex system, the heat exchange efficiency of the entire system varies greatly under different external environments and operating conditions. It is difficult to accurately analyze the performance of the air conditioning radiator through traditional heat transfer analysis.

[0066] To further illustrate the application of the proposed remote monitoring-based air conditioning system performance assessment method in aircraft air conditioning system performance monitoring, a practical radiator performance monitoring case will be used as an example: The message indicates that the performance of air conditioning component PACK1# is poor, with the component outlet temperature exceeding 15°C. The main / auxiliary exchangers of radiator PACK1 need to be replaced. A comparison of COT, RI, and TW parameters before and after radiator replacement is attached. Figure 3-5 As shown.

[0067] Depend on Figure 2-4 Therefore, replacing the air conditioner radiator based on performance evaluation helps optimize spare parts reserves, reduce the occurrence of abnormal events, smooth out peak and trough maintenance work, reduce manpower load, and lower maintenance costs. Example 2:

[0068] Air leaks in the air conditioning system (damaged seals or plugs at the ACM) can lead to decreased ACM performance, severely reduced air conditioning efficiency, and excessively cold condenser inlet, causing condenser cracks. Therefore, early detection of ACM seal damage is crucial to prevent further damage and impact on the air conditioning system. The performance assessment method of this invention only requires adding ACM leakage characteristic samples to the sample library.

[0069] As one implementation method of this embodiment, the ACM leakage feature sample library is listed below: Cruise (RI > 15, COT < 120, COT drops rapidly with a small K value, TW rises).

[0070] Ground level: CTO < 140, TW > 25.

[0071] To further illustrate the application of the proposed remote monitoring-based air conditioning system performance assessment method in aircraft air conditioning system performance monitoring, a real-world case study of air conditioning system leakage monitoring will be used: A leak was found in the left pack, specifically in the square seal at the ACM compressor inlet and the seal on the piping connecting the ACM to the main heat exchanger. Replacement resolved the issue. A comparison before and after leak repair is attached. Figure 6 As shown. Example 3:

[0072] A ruptured condenser in an air conditioning system causes hot and cold gases to mix at the condenser, severely reducing air conditioning efficiency and significantly decreasing cooling performance. However, the method of this invention only requires adding condenser characteristic samples to the sample library for determination.

[0073] As one implementation method of this embodiment, the following condenser feature samples are listed: Cruise: (RI > 20, COT < 110, TW < 10, TP > 15 degrees).

[0074] Ground COT is normal, TW > 25 degrees, TP > 15 degrees. Comparing the two packs, the TP temperature is higher on the side with the condenser rupture.

[0075] To further illustrate the application of the proposed remote monitoring-based air conditioning system performance assessment method in aircraft air conditioning system performance monitoring, a real-world case study of condenser rupture assessment in an air conditioning system will be used: Performance monitoring revealed a low COT temperature during cruise control and a decrease in overall PACK performance, prompting an inspection of the air conditioning compartment for leaks, with a suspected condenser malfunction. Inspection confirmed a condenser leak. A comparison of the condenser before and after replacement is attached. Figure 7 As shown. Example 4:

[0076] An abnormal return position of the inlet damper actuator cylinder in the air conditioning system can lead to an abnormal inlet damper angle. However, when using the method of this invention to determine this, it is only necessary to add the abnormal feature samples of the inlet damper actuator cylinder to the sample library.

[0077] As one implementation method of this embodiment, the following are examples of abnormal characteristics of the inlet damper actuator: Cruise: (RI angle is unstable, RI o is relatively large, other parameters are stable.)

[0078] To further illustrate the application of the proposed remote monitoring-based air conditioning system performance assessment method in aircraft air conditioning system performance monitoring, a real-world case study of inlet damper actuator performance assessment and monitoring will be used: RI actuator malfunction. Comparison of the imported damper actuator before and after replacement is attached. Figure 8 As shown.

[0079] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for determining the performance of an air conditioning system based on remote monitoring, characterized in that, Includes the following steps: The air conditioning system's steady-state data in the air (message A21) and ground air conditioning operation data in the air (message A24) are collected via hardware FDIMU. The steady-state data in the air conditioning system in message A21 includes inlet damper angle RI, compressor outlet temperature COT, water separator outlet temperature TW, PACK flow rate PF, PACK outlet temperature TF, and bypass valve angle PBV. The ground air conditioning operation data in message A24 includes TW and COT data for evaluation reference. In a stable cruise environment, set parameter thresholds and evaluate the steady-state data of the air conditioning system in the air in message A21. Trigger an alarm when parameters exceed the threshold. Data decoding, regression calculation, mean and variance calculation, left and right air conditioning data difference calculation, and normalization processing are performed on the air conditioning system steady-state data in message A21 and the ground air conditioning operation data in message A24. Trend analysis of each parameter and + / -2σ external data calibration are also performed. Based on the steady-state core characterization parameter data of air conditioners and the AHP hierarchical analysis method, an evaluation model for important parameters is established. The evaluation model for important parameters adopts a linear weighted summation method and has different weight coefficients according to different models. By applying a probabilistic neural network (PNN), based on feature samples, the normal operating mode and different failure modes of the air conditioner are classified and identified according to Bayesian criteria to determine the failed sub-components of the air conditioning system. The feature samples include samples from the air and samples from the ground. The samples from the air include RI at 40 degrees, COT temperature below 120 degrees, and TW above 15 degrees. The samples from the ground include RI at 40 degrees, COT close to 200 degrees, and TW exceeding 30 degrees.

2. The method for determining the performance of an air conditioning system based on remote monitoring according to claim 1, characterized in that, The alarm is triggered when a parameter exceeds the threshold twice consecutively.

3. The method for determining the performance of an air conditioning system based on remote monitoring according to claim 1, characterized in that, Also includes: Enhance the model's recognition capabilities by relying on human-assisted reinforcement classification learning.

4. The method for determining the performance of an air conditioning system based on remote monitoring according to claim 3, characterized in that, The artificially assisted reinforcement classification learning includes using expert experience judgment and historical maintenance data analysis to correct and optimize the weight and threshold parameters of the PNN probabilistic neural network.

Images