Energy efficiency optimization and fault diagnosis method and system for multi-split air conditioning system

Abstract

The invention provides an energy efficiency optimization and fault diagnosis method and system for a multi-split air conditioning system, and belongs to the technical field of air conditioners. The method comprises the steps that first data are obtained, and the first data comprise energy efficiency data, environment working condition data and equipment operation parameters; calculating the energy efficiency data to obtain an energy efficiency index; correcting the energy efficiency index based on the environment working condition data to obtain a target energy efficiency index, and comparing the target energy efficiency index with a preset first threshold value; if the target energy efficiency index is smaller than a first threshold value, it is judged that the multi-split air conditioning system is in an energy efficiency abnormal state, first data corresponding to the energy efficiency abnormal state is input into a pre-trained abnormal causal model, and a failure mechanism of the energy efficiency abnormal state is obtained; determining a fault source based on the fault mechanism; performing fault type identification and severity evaluation on the fault source based on a preset fault diagnosis rule base, and generating a fault diagnosis report; and determining and executing an energy efficiency optimization control strategy based on the fault diagnosis report. According to the invention, integrated management and control of energy efficiency and faults are realized.

Description

Technical Field

[0001] This application relates to the field of air conditioning technology, and in particular to a method and system for energy efficiency optimization and fault diagnosis of multi-split air conditioning systems. Background Technology

[0002] With the upgrading of building energy conservation requirements and the development of intelligent manufacturing, multi-split air conditioning systems are widely used in commercial and residential buildings due to their flexible control advantages. However, their operation involves complex mechanisms such as multi-compressor coordination and variable capacity regulation, making their energy efficiency susceptible to environmental conditions such as outdoor temperature and humidity and indoor load fluctuations, and potential faults are more hidden. The industry currently faces a dual challenge: first, systems often fail to intervene in time due to energy efficiency degradation, resulting in energy waste, which contradicts the dual-carbon goals; second, existing fault diagnosis relies on manual troubleshooting, which suffers from response lag and is prone to misjudging the source and severity of faults.

[0003] Therefore, there is an urgent need to develop a method and system for energy efficiency optimization and fault diagnosis of multi-split air conditioning systems. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides a method and system for energy efficiency optimization and fault diagnosis of multi-split air conditioning systems.

[0005] A first aspect of this application provides a method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system, including: Acquire first data, which includes energy efficiency data, environmental condition data, and equipment operating parameters; The energy efficiency data is calculated to obtain an energy efficiency index; the energy efficiency index is then corrected based on the environmental operating condition data to obtain a target energy efficiency index. The target energy efficiency index is compared with a preset first threshold; if the target energy efficiency index is less than the first threshold, the multi-split air conditioning system is determined to be in an abnormal energy efficiency state. The first data corresponding to the abnormal energy efficiency state is input into the pre-trained abnormal causal model to obtain the fault mechanism that leads to the abnormal energy efficiency state. Based on the aforementioned fault mechanism, the fault source is determined; based on a preset fault diagnosis rule base, the fault source is identified in terms of fault type and severity, and a fault diagnosis report is generated. Based on the fault diagnosis report, an energy efficiency optimization control strategy is determined and implemented.

[0006] A second aspect of this application provides an energy efficiency optimization and fault diagnosis system for multi-split air conditioning systems, comprising: The data acquisition module is used to acquire first data, which includes energy efficiency data, environmental condition data, and equipment operating parameters. The target energy efficiency determination module is used to calculate the energy efficiency data to obtain energy efficiency indicators; and to correct the energy efficiency indicators based on the environmental operating condition data to obtain the target energy efficiency indicators. An energy efficiency anomaly determination module is used to compare the target energy efficiency index with a preset first threshold; if the target energy efficiency index is less than the first threshold, the multi-split air conditioning system is determined to be in an energy efficiency anomaly state. The fault mechanism generation module is used to input the first data corresponding to the abnormal energy efficiency state into a pre-trained abnormal causal model to obtain the fault mechanism that leads to the abnormal energy efficiency state. The fault diagnosis module is used to determine the fault source based on the fault mechanism; to identify the fault type and assess the severity of the fault source based on a preset fault diagnosis rule base, and to generate a fault diagnosis report. The optimization strategy determination module is used to determine and execute the corresponding energy efficiency optimization control strategy based on the fault diagnosis report.

[0007] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system.

[0008] In a fourth aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system.

[0009] The beneficial effects of the energy efficiency optimization and fault diagnosis method and system for multi-split air conditioning systems provided in this application are as follows: This application effectively overcomes the dual challenges of current multi-split air conditioning system operation and maintenance. Regarding the problem of delayed intervention in energy efficiency degradation, by integrating multi-dimensional data and correcting energy efficiency indicators, the accuracy of identifying energy efficiency anomalies is improved, thereby enabling the multi-split air conditioning system to return to efficient operation in a timely manner, reducing energy waste, and contributing to the achievement of dual-carbon goals. Facing the problem of inefficient and misjudgment in traditional fault diagnosis, based on a pre-trained anomaly causal model and fault diagnosis rule base, the fault mechanism, source, and severity of the multi-split air conditioning system are located, reducing downtime losses. Compared to the fragmented limitations of existing technologies, this application achieves integrated management and control of energy efficiency and faults. Attached Figure Description

[0010] Figure 1 A flowchart illustrating a method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system provided in an embodiment of this application; Figure 2This is a structural block diagram of a multi-split air conditioning system energy efficiency optimization and fault diagnosis system provided in an embodiment of this application; Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0011] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0012] To make the purpose, technical solution, and advantages of this application clearer, the following will be described in conjunction with the appendix. Figure 1-3 The following is an explanation using specific examples.

[0013] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system according to an embodiment of this application. The method includes: S101: Obtain first data, which includes energy efficiency data, environmental condition data, and equipment operating parameters.

[0014] In this embodiment, the first data is collected through a distributed sensor network built into the multi-split air conditioning system. This sensor network is categorized into three types based on function: energy efficiency sensors, environmental sensors, and equipment operation sensors. Energy efficiency data is acquired through energy efficiency sensors, including: cooling (heating) capacity, input power, COP value, and cumulative power consumption. Environmental condition data is acquired through environmental sensors, including outdoor temperature and humidity, indoor temperature and humidity, and atmospheric pressure. Equipment operation parameters are acquired through equipment operation sensors, including: compressor exhaust / suction temperature and pressure, operating frequency, electronic expansion valve opening, indoor and outdoor fan speeds, and heat exchanger inlet and outlet temperatures.

[0015] The first data acquisition frequency is based on the adaptive adjustment mode of the operating conditions. This adaptive adjustment mode dynamically matches the acquisition frequency according to the operating status of the multi-split air conditioning system. For example, when the multi-split air conditioning system is in steady state operation, the acquisition cycle is set to 5-10 seconds; when the multi-split air conditioning system is in the start-up and shutdown phase, the load changes suddenly (multiple indoor units start at the same time), or the operating parameters fluctuate abnormally, the acquisition cycle is automatically shortened to 1-2 seconds. Among them, the steady state operation of the multi-split air conditioning system is defined as indoor temperature fluctuation less than or equal to ±0.5℃ and compressor frequency stable for more than 5 minutes.

[0016] S102: Calculate the energy efficiency data to obtain the energy efficiency index; correct the energy efficiency index based on the environmental operating condition data to obtain the target energy efficiency index.

[0017] In this embodiment, an evaluation system of indicators and auxiliary indicators is constructed based on the collected energy efficiency data, namely the COP and EER evaluation system. Among them, COP (Coefficient of Performance for Cooling / Heating) is selected as the indicator, and the calculation of energy efficiency indicators follows the provisions of the national standard GB / T18837-2015 "Multi-split Air Conditioning (Heat Pump) Units".

[0018] For example, the energy efficiency data is first filtered to remove abnormal data segments with cooling (heating) capacity fluctuations greater than ±5% and input power jumps greater than 10%, and data that has been running continuously and stably for more than 3 minutes is selected as the calculation sample. For cooling operation, COP = cooling capacity / input power; for heating operation, COP = heating capacity / input power. Cooling (heating) capacity is directly measured by a calorimeter, while input power is collected in real time by a power sensor, and the calculation result is rounded to two decimal places. At the same time, to avoid the limitations of a single indicator, EER (Energy Efficiency Ratio) is calculated as an auxiliary indicator. When the trends of COP and EER deviate, a second data verification is triggered to eliminate calculation errors caused by sensor malfunctions.

[0019] In this embodiment, the heat exchange efficiency of the multi-split air conditioner is strongly correlated with ambient temperature and humidity. For example, when the outdoor temperature rises from 30°C to 35°C, the condenser heat exchange temperature difference decreases, the compressor power consumption increases, and the COP decreases by 3%-5%. When the indoor humidity rises from 40% to 60%, condensation on the evaporator surface thickens, the heat exchange coefficient decreases, and the COP further decreases by 1%-2%. If the calculated COP is used directly for evaluation, energy efficiency fluctuations caused by environmental changes may be misjudged as abnormalities in the multi-split air conditioner system. The correction rule is based on the standard operating conditions specified in GB / T7725-2022 (cooling: outdoor 35°C / indoor 27°C, heating: outdoor 7°C / indoor 19°C), and the indicators and auxiliary indicators are corrected based on environmental correction coefficients. The correction coefficients are calculated using a linear regression algorithm, and the regression equation is obtained through training with historical data. K = a × T + b × RH + c × T + d × RH + e Where K is the correction coefficient, T and RH are the outdoor temperature and humidity, respectively, and a, b, c, d, and e are regression coefficients, which were obtained by fitting over 5000 sets of measured data under different working conditions, with a goodness of fit of R0. 2 ≥0.95.

[0020] In this embodiment, the target energy efficiency index = COP × K. The target energy efficiency index eliminates the interference of environmental factors and can represent the energy efficiency performance of the multi-split air conditioning system itself.

[0021] S103: Compare the target energy efficiency index with the preset first threshold; if the target energy efficiency index is less than the first threshold, the multi-split air conditioning system is determined to be in an abnormal energy efficiency state.

[0022] In this embodiment, the comparison between the target energy efficiency index and the first threshold is the trigger point for energy efficiency anomaly determination. The determination of the first threshold is based on a dynamic threshold system constructed from the full life cycle data of the multi-split air conditioning system and industry standards. Its setting also takes into account three dimensions: equipment characteristics, operating history, and energy efficiency standards. The setting of the first threshold is divided into three steps: Step 1: The rated COP value in the design document of the multi-split air conditioning system is used as the base benchmark. For example, the rated cooling COP of a 5-horsepower multi-split unit is 3.6. This value is used as the reference for setting the first threshold. Step 2: Based on the historical operating data of the same type of multi-split air conditioning system, the optimal COP range of the same model and usage scenario in the past 3 years is extracted through statistical analysis, and the lower limit of the range is taken as the historical reference threshold. Step 3: According to national energy efficiency standards, such as GB21454-2021 "Energy Efficiency Limits and Energy Efficiency Grades of Multi-Split Air Conditioning (Heat Pump) Units", the COP requirement for Grade 1 energy efficiency is ≥3.4, and the standard constraint threshold is determined. The initial first threshold is obtained by weighting and fusing the base benchmark with a weight of 0.4, the historical reference threshold with a weight of 0.3, and the standard constraint value with a weight of 0.3. The initial first threshold is obtained by reserving a fluctuation margin of 5%-10%.

[0023] During the comparison of the target energy efficiency index with the first threshold, a continuous multi-cycle verification mechanism is constructed. For example, firstly, the target energy efficiency index is compared with the first threshold every 10 seconds. When the target energy efficiency index is lower than the first threshold for the first time, the early warning monitoring mode is triggered. Subsequently, in the next 5 minutes, the data collection frequency is increased to 1 second / time, and 300 sets of target energy efficiency index data are collected continuously. If more than 80% of the values ​​are lower than the first threshold and the data fluctuation range is less than or equal to ±3%, the multi-split air conditioning system is determined to be in an abnormal energy efficiency state. If only a single or sporadic data is lower than the first threshold, and the target energy efficiency index subsequently recovers to above the first threshold, it is determined to be a momentary interference.

[0024] S104: Input the first data corresponding to the abnormal energy efficiency state into the pre-trained abnormal causal model to obtain the fault mechanism that leads to the abnormal energy efficiency state.

[0025] In this embodiment, the first data corresponding to the abnormal energy efficiency state of the multi-split air conditioning system is subjected to feature enhancement processing. For example, COP fluctuation features in energy efficiency data, steady-state / transient environmental parameters in environmental condition data, and abnormal component response features in equipment operating parameters are extracted, such as compressor frequency mutation and abnormal fluctuation of electronic expansion valve opening. These features are then reorganized into a structured feature matrix that can be identified by the abnormal causal model based on the time series-parameter type dimension. Next, the feature matrix is ​​input into the abnormal causal model, and the causal inference engine built into the abnormal causal model is used to mine the intrinsic relationship between data features and fault causes, and output the fault mechanism. For example, evaporator frost leads to a reduction in heat exchange area, which reduces the cooling capacity of the multi-split air conditioning system, thereby causing a decrease in COP, insufficient lubricating oil in the inverter compressor leading to increased friction loss, increased input power, and abnormal energy efficiency indicators, etc.

[0026] The input to the abnormal causal model in this embodiment is the first data corresponding to the abnormal energy efficiency state, i.e., the feature matrix, which has undergone feature enhancement and structured recombination. The feature matrix is ​​a fusion feature matrix of energy efficiency data, environmental operating condition data, and equipment operating parameters. The output is a description of the fault mechanism, including the logic of the fault and the key impact links. The abnormal causal model is hierarchically divided into a data preprocessing layer, a feature fusion layer, a causal inference layer, and a result output layer. Among them, the data preprocessing layer is responsible for feature extraction and recombination; the feature fusion layer uses an attention mechanism for weighted fusion to obtain multi-dimensional features; the causal inference layer constructs a causal relationship graph based on a Bayesian network; and the result output layer is responsible for converting the inference results into a natural language description of the fault mechanism and outputting it.

[0027] The training process of this abnormal causal model is based on historical data throughout the entire life cycle. First, a training set is constructed, including 5,000 normal operation samples and 3,000 typical failure samples. The failure samples include failure types such as compressor failure, heat exchanger failure, and electronic expansion valve failure. Each sample is labeled with complete operating data and corresponding failure mechanism labels. During the training of the abnormal causal model, the sample data is first cleaned and features are extracted through a data preprocessing layer. Then, the weights of key failure features are strengthened by a feature fusion layer. Subsequently, the EM algorithm is used in the causal inference layer to optimize the node probability parameters of the Bayesian network. Through iterative adjustment, the failure mechanism output by the abnormal causal model is optimized.

[0028] The feature extraction window for the abnormal causal model is 60 seconds, which is one complete system operation cycle. The initial value of the feature weights for the attention mechanism is based on the parameter importance allocation, for example, the weight of equipment operating parameters is 0.5, the weight of energy efficiency data is 0.3, and the weight of environmental operating condition data is 0.2. The number of iterations of the EM algorithm is set to 200 rounds, and the convergence threshold is set to 0.001. Training stops when the change in the loss function value of 10 consecutive iterations is less than the convergence threshold.

[0029] S105: Based on the fault mechanism, determine the fault source; based on the preset fault diagnosis rule base, identify the fault type and assess the severity of the fault source, and generate a fault diagnosis report.

[0030] In this embodiment, after obtaining the fault mechanism, candidate fault sources are initially screened by searching the association rules in the database using keyword retrieval. Next, the equipment operating parameters corresponding to each candidate fault source are extracted from the equipment operating parameters to obtain the target equipment operating parameters. These parameters are then compared with a preset fault mechanism feature template library to calculate the matching degree, resulting in a matching degree score. Based on the matching degree score, a candidate fault sequence is obtained. Finally, candidate fault sources in the candidate fault source sequence with a matching degree score greater than or equal to a preset second threshold are identified as fault sources; candidate fault sources in the candidate fault source sequence with a matching degree score less than the second threshold are identified as components to be monitored.

[0031] This embodiment's fault diagnosis rule base includes a mapping of fault characteristics, fault types, severity, and handling suggestions. For example, if the compressor exhaust temperature is greater than 110℃, the operating frequency fluctuation is greater than ±10Hz, and the COP decreases by more than 20%, the fault type is compressor overload, the severity is level two, and the handling suggestion is to check the lubricating oil level and cooling system. Each rule has been verified by actual fault cases, and the diagnostic accuracy is greater than or equal to 93%. During fault diagnosis, the characteristic parameters of the fault source and the fault mechanism description are first extracted as input conditions for rule matching. The optimal rule is matched through a forward reasoning algorithm to complete the fault type identification.

[0032] The severity assessment employs a multi-indicator weighted scoring mechanism, quantifying the severity across four dimensions: scope of impact, urgency of repair, degree of energy efficiency loss, and safety risk level. A total score of 0-100 corresponds to three severity levels: 80-100 points = Level 1, requiring immediate shutdown to prevent component damage; 50-79 points = Level 2, requiring action within 24 hours; 0-49 points = Level 3, requiring action within 72 hours. After type identification and severity assessment, a standardized fault diagnosis report is generated. In addition to core information such as fault source name, type, and severity, the report also includes associated operating parameter curves, fault development trend predictions, recommended maintenance procedures, and spare parts models.

[0033] S106: Based on the fault diagnosis report, determine and implement energy efficiency optimization control strategies.

[0034] In this embodiment, based on the fault diagnosis report, an objective function with the goal of optimizing system energy efficiency is constructed. Based on the equipment operating parameters and environmental condition data, the operating constraints of the multi-split air conditioning system are set. The objective function is then solved under the constraints using a genetic algorithm to obtain the optimal combination of equipment operating parameters. Based on the optimal combination of equipment operating parameters, an energy efficiency optimization control strategy is generated and then distributed to the corresponding execution components of the multi-split air conditioning system for execution.

[0035] Each item in the fault diagnosis report corresponds to an optimization logic: if the report determines that the problem is condenser dust accumulation, the optimization focus is to improve heat exchange efficiency by adjusting the fan speed to enhance heat dissipation; if the problem is compressor overload, the control strategy prioritizes reducing the compressor load while also taking into account indoor temperature control requirements, resulting in a fault-adaptive optimization solution.

[0036] As can be seen from the above, this application effectively overcomes the dual challenges of current multi-split air conditioning system operation and maintenance. Regarding the issue of delayed intervention in energy efficiency degradation, by integrating multi-dimensional data and correcting energy efficiency indicators, the accuracy of identifying energy efficiency anomalies is improved, enabling multi-split air conditioning systems to return to efficient operation in a timely manner, reducing energy waste, and contributing to the achievement of dual-carbon goals. Addressing the problem of inefficient and misjudgment in traditional fault diagnosis, based on a pre-trained anomaly causal model and fault diagnosis rule base, the fault mechanism, source, and severity of the multi-split air conditioning system are located, reducing downtime losses. Compared to the fragmented limitations of existing technologies, this application achieves integrated management and control of energy efficiency and faults.

[0037] In one embodiment of this application, determining the fault source based on the fault mechanism includes: Based on the aforementioned fault mechanism, a search is performed in a preset fault mechanism-component association database to obtain candidate fault sources; Extract the equipment operating parameters corresponding to each candidate fault source from the equipment operating parameters to obtain the target equipment operating parameters; The matching degree of the target equipment's operating parameters is calculated by comparing them with a preset fault mechanism feature template library, and a matching degree score is obtained. All candidate fault sources are sorted according to their matching scores to obtain a sequence of candidate fault sources; Candidate fault sources are determined and classified based on matching scores, including: Candidate fault sources in the candidate fault source sequence whose matching score is greater than or equal to a preset second threshold are identified as fault sources; Candidate fault sources with a matching score less than the second threshold in the candidate fault source sequence are designated as components to be monitored. The fault source is identified by verifying the fault in the monitored component.

[0038] In this embodiment, fault mechanisms are retrieved based on a pre-defined fault mechanism-component association database to obtain candidate fault sources. This database is built using MySQL and indexed by both fault mechanism keywords and component functional attributes. It includes the association relationships between typical faults, subdivided fault mechanisms, and various components. Each association rule includes an association confidence level and an explanation of the applicable operating conditions. For example, abnormal return gas superheat is associated with the electronic expansion valve and the evaporator, respectively, with association confidence levels of 92% and 85%, and the applicable operating condition is marked as cooling mode. During the retrieval, the description of the fault mechanism is first segmented into words. For example, the phrase "evaporator frosting leading to decreased heat exchange efficiency" is broken down into keywords such as "evaporator," "frosting," and "decreased heat exchange efficiency." Boolean logic is then used to search and match the association rules in the database, and candidate fault sources are output from high to low based on the association confidence level.

[0039] In this embodiment, firstly, the equipment operating parameters corresponding to each candidate fault source are extracted from the equipment operating parameters to obtain the target equipment operating parameters. During the extraction process, the collection timestamp and fluctuation characteristics of each equipment operating parameter are marked to obtain the target equipment operating parameters in the dimensions of candidate fault source-parameter type-parameter value-time characteristics.

[0040] Secondly, the matching degree between the target equipment's operating parameters and the fault mechanism feature template library is calculated to obtain a matching degree score. The fault mechanism feature template library is organized hierarchically based on fault mechanism-component-feature parameter, storing standard parameter templates for various fault states. For example, the template for evaporator frosting includes threshold ranges such as surface temperature less than or equal to 5℃, inlet and outlet air temperature difference greater than or equal to 8℃, and frosting sensor value greater than or equal to 0.8V, as well as the linear decreasing trend of temperature over time. During the matching degree score calculation, multi-scale feature extraction is performed on the target equipment's operating parameters to obtain a first multi-dimensional feature vector. Then, a second multi-dimensional feature vector template corresponding to the fault mechanism is retrieved from the template library, and the matching degree between the first and second multi-dimensional feature vectors is calculated.

[0041] Subsequently, candidate fault sources are sorted and classified according to their matching scores. The candidate fault source sequence is obtained by sorting them from highest to lowest matching score. The second threshold is set based on historical fault case data, calculating the fault identification accuracy and false negative rate under different thresholds, ensuring that the fault identification accuracy is greater than or equal to 91% and the false negative rate is less than or equal to 3% at this threshold. Candidate fault sources with matching scores greater than or equal to the second threshold are marked as fault sources, and their key parameter deviation characteristics during the matching process are simultaneously associated; candidate components with matching scores less than the second threshold are included in the list of components to be monitored.

[0042] Fault verification of the monitored components is a secondary screening of candidate fault sources. Candidate components that pass the secondary screening are also considered as fault sources.

[0043] From the above, it can be concluded that this embodiment not only defines the scope of candidate fault sources through the fault mechanism-component association database, avoiding the inefficiency caused by blind investigation, but also constructs a quantitative evaluation standard through precise matching of target equipment operating parameters and feature template library, ensuring the objectivity and scientific nature of fault source determination. Furthermore, by using the processing mechanism of fault source locking and verification of monitored components, it not only ensures the rapid location of explicit faults, but also investigates the risk of implicit faults through extended cycle monitoring and performance degradation assessment, effectively reducing the probability of missed or false judgments. Ultimately, it improves the efficiency and accuracy of fault source location, while reducing the ineffective maintenance costs and system downtime losses caused by inaccurate fault location, further improving the stability and economy of multi-split air conditioning system operation.

[0044] In one embodiment of this application, the matching degree of the target device operating parameters with the fault mechanism feature template library is calculated to obtain a matching degree score, including: Multi-scale feature extraction is performed on the operating parameters of the target equipment to obtain the first multi-dimensional feature vector; From the fault mechanism feature template library, retrieve the second multidimensional feature vector template corresponding to the fault mechanism. The second multidimensional feature vector template is the standard multidimensional feature vector template of the fault mechanism. The difference between the first multidimensional feature vector and the second multidimensional feature vector template is calculated to obtain the feature deviation vector. Based on the feature deviation vector, the similarity between the first multidimensional feature vector and the second multidimensional feature vector template is calculated to obtain the matching score.

[0045] In this embodiment, the first multidimensional feature vector includes time-domain statistical features, frequency-domain energy features, and trend change features. Correspondingly, the second multidimensional feature vector includes standard time-domain statistical features, standard frequency-domain energy features, and standard trend change features.

[0046] While the target equipment operating parameters include raw information on the component operating status, direct use is susceptible to noise interference. Multi-scale feature extraction enhances features through a three-dimensional analysis of the time domain, frequency domain, and statistical domain: The time domain scale includes the parameter's variation characteristics over time, using a sliding window method to extract indicators such as mean, variance, peak value, valley value, and rate of change, for example, the maximum fluctuation amplitude of the electronic expansion valve opening within 5 minutes and the average response delay time; The frequency domain scale decomposes the time domain signal into different frequency components through wavelet transform, extracting low-frequency trend features and high-frequency abrupt change features, such as the fundamental frequency component amplitude of the compressor vibration frequency and the proportion of high-frequency interference energy; The statistical domain scale calculates indicators such as skewness, kurtosis, and quantiles based on 30 sets of continuous data to determine whether the parameter distribution deviates from the normal range, such as the 90th percentile of the evaporator surface temperature and the distribution skewness coefficient. Finally, the three-dimensional features are fused according to the proportions of time domain features, frequency domain features, and statistical domain features to obtain the first multi-dimensional feature vector.

[0047] The second multidimensional feature vector template is retrieved from the fault mechanism feature template library. The standardization of this library determines the reliability of the matching results. This library employs a hierarchical storage structure, with a three-level index: fault mechanism category, sub-fault type, and component name. Each index corresponds to a unique second multidimensional feature vector template, which is a two-dimensional vector of interval and trend constructed based on standard test data of similar faults. For example, in the template corresponding to evaporator frosting, the surface temperature feature dimension is in the range of [2℃, 5℃] with a linear decreasing trend; the inlet and outlet air temperature difference is in the range of [8℃, 12℃] with a stable trend; and the frosting sensor value is in the range of [0.8V, 1.2V] with a step-like increasing trend. When retrieving the second multidimensional feature vector, the index is first established using keywords in the fault mechanism description, and then the target template is locked based on the candidate fault source.

[0048] The calculation of the difference between the first and second multidimensional feature vector templates is the foundation for quantifying the actual value and the standard deviation. For example, since the first multidimensional feature vector is a numerical vector and the second multidimensional feature vector represents interval-trend information, the second multidimensional feature vector is first converted to a numerical value: the midpoint of the interval is used as the standard value for interval features, and the trend features are converted into trend slope values ​​through linear fitting. Then, the difference is calculated based on the one-to-one correspondence of the feature dimensions to obtain the feature deviation vector. The process involves calculating similarity based on feature deviation vectors to obtain a matching score, and achieving a comprehensive score based on the magnitude and direction of the feature deviation vectors. Specifically, firstly, deviation standardization is performed, converting the deviation values ​​of each dimension into standardized deviations within the range of [-1, 1] using the formula: deviation value ÷ standard value range. For example, a temperature deviation of -1.7℃ corresponds to a standard range of [2℃, 5℃], and the standardized deviation is -1.7 / 3 ≈ -0.57. Secondly, based on the standardized deviation vectors, the cosine value of the angle between the two vectors is calculated using the cosine theorem to obtain the cosine similarity. The closer the cosine similarity is to 1, the smaller the overall deviation and the higher the similarity.

[0049] As can be seen from the above, this embodiment comprehensively captures the static numerical characteristics and dynamic change patterns of the target equipment's operating parameters through three-dimensional feature extraction in the time domain, frequency domain, and statistical domain, avoiding the limitations and noise interference of single-scale analysis. Furthermore, by retrieving a standardized second multi-dimensional feature vector template, a comparison benchmark is established between the first multi-dimensional feature vector and the standard state of the fault mechanism. In addition, through feature deviation vector calculation and similarity quantification, the multi-dimensional parameter deviation is transformed into a matching degree score, which not only achieves an objective assessment of the correlation between candidate fault sources and fault mechanisms, but also avoids the subjectivity and randomness of manual judgment, further improving the scientificity and reliability of the matching results.

[0050] In one embodiment of this application, based on the feature deviation vector, the similarity between the first multidimensional feature vector and the second multidimensional feature vector is calculated to obtain a matching score, including: The deviation level of the feature deviation vector is determined based on the absolute value of the feature deviation vector and a preset deviation-level mapping table; the deviation level includes a first deviation level, a second deviation level and a third deviation level, and the first deviation level is greater than the second deviation level, and the second deviation level is greater than the third deviation level. The adjustment direction is determined based on the deviation direction of the characteristic deviation vector and the standard change direction of the fault mechanism; The baseline weights of the first and second multidimensional feature vectors are adjusted based on the deviation level and adjustment direction, including: If the direction of deviation is consistent with the standard change direction under the fault mechanism, then positive adjustment is performed: If the direction of deviation is inconsistent with the standard change direction under the fault mechanism, then negative adjustment is performed: If the deviation level of the feature deviation vector is the first deviation level, then its benchmark weight coefficient is adjusted using the first adjustment amplitude. If the deviation level of the feature deviation vector is the second deviation level, then the baseline weight coefficient is adjusted using the second adjustment amplitude. If the deviation level of the feature deviation vector is the third deviation level, then keep its initial weight coefficient unchanged; Among them, the first adjustment range is greater than the second adjustment range; Based on the adjusted weight coefficients, the similarity between the first multidimensional feature vector and the second multidimensional feature vector template is calculated using a weighted similarity fusion algorithm to obtain the matching score.

[0051] In this embodiment, the deviation level is first determined based on a preset deviation-level mapping table. This mapping table is constructed based on statistical analysis of a large amount of historical fault data and represents the correspondence between the absolute value of the deviation and the deviation level under different feature dimensions. For example, in the compressor frequency feature dimension, an absolute deviation value greater than or equal to 10Hz is the first deviation level, an absolute deviation value between 5-10Hz is the second deviation level, and an absolute deviation value less than 5Hz is the third deviation level. In the electronic expansion valve opening feature dimension, an absolute deviation value greater than or equal to 15% is the first deviation level, an absolute deviation value between 5-15% is the second deviation level, and an absolute deviation value less than 5% is the third deviation level. When classifying the deviation level, the mapping table is used to determine the level for each dimension in the feature deviation vector.

[0052] Secondly, the adjustment direction is determined based on the consistency between the deviation direction and the standard change direction of the fault mechanism. The standard change direction of the fault mechanism is predetermined based on the physical laws governing the occurrence of the fault. For example, the standard change direction corresponding to the condenser ash accumulation fault is a decrease in heat transfer coefficient and an increase in condensing pressure. If the deviation direction of the condensing pressure dimension in the characteristic deviation vector is upward, it is determined that the direction is consistent, and positive adjustment is performed; if the deviation direction is downward, it is determined that the direction is inconsistent, and negative adjustment is performed.

[0053] Based on this, weight adjustment follows a dual rule: the deviation level determines the magnitude, and the adjustment direction determines the trend. The baseline weight coefficient is pre-assigned based on the importance of the feature dimension to fault diagnosis. For example, the baseline weight for features such as compressor exhaust temperature and electronic expansion valve opening is 0.2, while the baseline weight for auxiliary features such as fan speed is 0.05. During weight adjustment, if the deviation level is the first deviation level, the baseline weight is adjusted using the first adjustment magnitude; if the deviation level is the second deviation level, the baseline weight is adjusted using the second adjustment magnitude, and the first adjustment magnitude is greater than the second adjustment magnitude; if the deviation level is the third deviation level, the baseline weight remains unchanged. Simultaneously, the weight change trend is determined according to the adjustment direction: a positive adjustment increases the weight, and a negative adjustment decreases the weight. For example, if a core feature dimension is at the first deviation level and the direction is consistent, its baseline weight of 0.2 will be increased by 20% to 0.24; if the direction is inconsistent, it will be decreased by 20% to 0.16. The baseline weight coefficient is pre-set based on the importance of different feature dimensions to fault diagnosis.

[0054] Finally, based on the adjusted weight coefficients, a weighted similarity fusion algorithm is used to calculate the final matching score. This algorithm first calculates the cosine similarity of each dimension between the first multidimensional feature vector and the second multidimensional feature vector template, then multiplies each dimension's similarity with the adjusted weight coefficients, and finally sums and normalizes the weighted similarities over all dimensions to obtain a matching score ranging from 0 to 100. This weighted fusion method highlights the contribution of key feature dimensions such as the first deviation level and direction consistency, while also taking into account the auxiliary role of other feature dimensions, enabling the matching score to reflect the degree of correlation between the candidate fault source and the fault mechanism.

[0055] The determination of the first and second adjustment ranges includes: First, retrieving past cases from the historical database that are similar to the current fault mechanism and environmental conditions to obtain the diagnostic accuracy assessment results corresponding to different adjustment ranges in the past cases. Second, selecting the adjustment range used in the case with the best diagnostic accuracy assessment result from the assessment results as the benchmark value of the first or second adjustment range. Based on the data of the current fault mechanism and environmental conditions, the benchmark values ​​of the first and second adjustment ranges are adjusted to obtain the first and second adjustment ranges used to adjust the benchmark weight coefficients of each feature vector.

[0056] From the above, it can be concluded that this embodiment, through the dual rules of adjusting the magnitude of the deviation level and the targeted trend adjustment of the deviation direction, not only highlights the contribution of the first deviation level and the key feature dimension consistent with the change direction of the fault mechanism standard to the matching result, but also weakens the interference of the third deviation level and the reverse deviation dimension. This effectively solves the problem of insufficient matching accuracy caused by the equal weight of each feature dimension in similarity calculation. At the same time, based on the weighted similarity fusion algorithm, the deviation information of multi-dimensional features is transformed into an intuitive and quantitative matching score, which improves the distinguishability and scientificity of the score. This provides a reliable quantitative basis for the ranking and classification of candidate fault sources, thereby ensuring the efficiency and accuracy of fault source localization.

[0057] In one embodiment of this application, fault verification is performed on the component to be monitored to obtain the fault source, including: Generate a monitoring task for each candidate component among the components to be monitored; The second data, collected based on the monitoring task, is input into the pre-trained performance degradation evaluation model to obtain the health index and degradation rate of each candidate component; the second data is data collected based on the monitoring task and is of the same data type as the first data. If the health index is less than the preset third threshold, or the degradation rate is greater than the preset fourth threshold, then the corresponding candidate component is identified as a source of failure.

[0058] In this embodiment, the monitoring task includes: key detection parameters, detection frequency and duration, wherein the key monitoring parameters are characteristic parameter data that are strongly correlated with the performance of candidate components; and the monitoring frequency and duration are adjusted according to the magnitude of the target energy efficiency index. The second set of data collected based on the monitoring task is input into the pre-trained performance degradation assessment model, which calculates the health index and degradation rate of each candidate component. The health index comprehensively considers the deviation of the component's operating parameters and the stability of the parameter change trend; the degradation rate is calculated by fitting the performance decay curve of continuous monitoring data, which can represent the real-time degradation status of the candidate component.

[0059] If the health index is less than a preset third threshold, or the degradation rate is greater than a preset fourth threshold, the corresponding candidate component is identified as a fault source. The third and fourth thresholds are determined based on the candidate component's safe operating standards and historical fault data. The determination uses an OR logic; if either condition is met, the candidate component is identified as a fault source. Candidate components with a health index greater than or equal to the third threshold and / or a degradation rate less than or equal to the fourth threshold are continuously monitored. The monitoring task can be cancelled when the target energy efficiency index exceeds the first threshold, and the component's health index and degradation rate tend to stabilize over multiple consecutive monitoring cycles. The third and fourth thresholds are predetermined based on the candidate component's safe operating standards and historical fault data.

[0060] The performance degradation assessment model is an LSTM-based time-series data analysis model. The input is the second data, which specifically includes time-series data of characteristic parameters that are strongly correlated with the performance of candidate components, as well as auxiliary interference parameters such as environmental conditions and system load that are recorded simultaneously. The output consists of two indicators: a health index that represents the current performance status of the candidate component and a degradation rate that represents the speed of performance degradation of the component. This embodiment's performance degradation assessment model consists of three layers: a data preprocessing layer, an LSTM feature extraction layer, and an index output layer. The data preprocessing layer is responsible for cleaning, normalizing, segmenting, and denoising the input time-series data. The LSTM feature extraction layer uses multiple LSTM units to capture the dynamic degradation characteristics of candidate component performance-related feature parameters over time, uncovering potential patterns of performance decline. The index output layer uses a fully connected network combined with a regression algorithm to map the extracted time-series features into specific values ​​for health index and degradation rate.

[0061] During the training process of the performance degradation evaluation model, a training dataset is first constructed, including time-series operational data of candidate components with different degrees of degradation. Each data set is labeled with the true health index and degradation rate under the corresponding operating conditions. The training dataset is then divided into training, validation, and test sets in a 7:2:1 ratio. The parameter settings include: setting the LSTM feature extraction layer to 3 hidden layers with 64 hidden units per layer; setting the input sequence length to 300; setting the batch size to 32; setting the initial learning rate to 0.001 and adaptively decreasing it with training iterations; setting the number of training iterations to 200 epochs; and stopping training when the validation set loss value decreases by less than 0.0001 for 10 consecutive epochs, thus obtaining the final performance degradation evaluation model.

[0062] From the above, it can be concluded that this embodiment generates a monitoring task for each candidate component and collects second data that is strongly related to the component's performance in a targeted manner, thus avoiding the loss of required data. Furthermore, based on the pre-trained performance degradation assessment model, it outputs a quantitative health index and degradation rate, thereby realizing the identification of the component's latent degradation state and breaking through the limitations of qualitative judgment in the prior art. At the same time, the use of logical judgment rules based on the health index or degradation rate effectively reduces the probability of missing latent faults.

[0063] In one embodiment of this application, an energy efficiency optimization control strategy is determined and executed based on a fault diagnosis report, including: Based on the fault diagnosis report, an objective function with the goal of optimizing system energy efficiency is constructed, and the operating constraints of the multi-split air conditioning system are set based on equipment operating parameters and environmental condition data. A genetic algorithm is used to solve the objective function under constraints to obtain the optimal combination of equipment operating parameters; Generate an energy efficiency optimization control strategy based on the optimal combination of equipment operating parameters; The energy efficiency optimization control strategy is distributed to the corresponding execution components of the multi-split air conditioning system.

[0064] In this embodiment, the objective function is constructed based on the fault type, severity, and component health status in the fault diagnosis report. The expression of the objective function is as follows: F = w1 × COP + w2 × (1 − Failure Risk Coefficient) + w3 × (1 − Temperature Control Deviation Rate) Among them, the temperature control deviation rate is a relative indicator that measures the degree of deviation between the actual temperature control effect and the target temperature control effect of the air conditioning system. It is calculated as: |Actual Indoor Temperature - Target Set Temperature| ÷ Target Set Temperature, with a value range of 0-1. A smaller value indicates higher system temperature control accuracy and better comfort. The weighting coefficient is adjusted according to the severity of the fault. Specifically, for a Level 1 fault, w2=0.5, prioritizing safe equipment operation; for a Level 3 fault, w1=0.5, focusing on energy efficiency improvement; and for a Level 2 fault, w1=w2. The constraint settings are divided into hard constraints and soft constraints. Hard constraints are derived from fault diagnosis reports and component safety operation standards; soft constraints are based on real-time environmental operating condition data. The fault risk coefficient is obtained by weighting and summing the fault type, severity, and health index in the fault diagnosis report based on their importance. Its value range is 0-1; a larger value indicates a higher fault risk in the multi-split air conditioning system, while a smaller value indicates a more stable system operation. The calculation formula is: R... R is the fault risk coefficient; n is the number of fault types currently existing in the system; The type weight of the i-th type of fault is set (the weight is set based on the severity of the fault type); is the severity of the i-th type of fault; is the deterioration coefficient of the component corresponding to the i-th type of fault (the lower the health index, the higher the deterioration coefficient).

[0065] After obtaining the fault risk coefficient, based on the first-level fault: R≥0.7; the second-level fault: 0.3≤R<0.7; the third-level fault: 0<R<0.3, the fault levels are divided to obtain the first-level fault, the second-level fault and the third-level fault.

[0066] In this embodiment, the genetic algorithm is used to solve the objective function under the constraint conditions to obtain the optimal combination of device operation parameters. Specifically, the execution process of the genetic algorithm is customized based on the operation characteristics of the air-conditioning system. For example, first, the core operation parameters such as the compressor frequency, the opening degree of the electronic expansion valve, and the fan speed are encoded into binary chromosomes with a length of 64 bits, and a parameter combination group with an initial population size of 100 is constructed; secondly, the objective function value is used as the fitness function, and the appropriate chromosomes are screened through the selection operator to retain the high-quality parameter combinations; then, the gene recombination and random mutation of the parameter combinations are realized through the crossover operator and the mutation operator; based on the constraint condition judgment mechanism, the chromosomes that violate the hard constraints are directly eliminated. The number of iterations of this genetic algorithm is set to 200 rounds, and the iteration stops when the change amount of the optimal fitness value in 10 consecutive rounds is less than 0.001, and the parameter combination corresponding to the optimal chromosome is output, which is the optimal combination of device operation parameters.

[0067] In this embodiment, the optimal combination of device operation parameters is analyzed to obtain the target control parameters of each execution component in the multi-connected air-conditioning system. Among them, the execution components include variable-frequency compressors, indoor fans and electronic expansion valves. The target control parameters include: the target operating frequency of the variable-frequency compressor, the target speed of the indoor fan, and the target opening degree of the electronic expansion valve. Based on the target operating frequency, the target speed and the target opening degree, the corresponding variable-frequency control instructions, speed control instructions and opening degree adjustment instructions are generated respectively, and an energy efficiency optimization control strategy is constructed.

[0068] After obtaining the energy efficiency optimization control strategy, it is sent to the corresponding execution components of the multi-connected air-conditioning system. Among them, the transmission process adopts the CAN bus and 5G dual-mode communication method. The CAN bus is responsible for issuing millisecond-level instructions to the local execution components to ensure the rapid response of core components such as compressors and electronic expansion valves; 5G communication is used for the instruction synchronization and status feedback of the remote monitoring platform. After the instruction is issued, the feedback data of the execution component is collected, compared with the optimal parameter combination, and the deviation is calculated. If the deviation is greater than the fourth threshold, the PID adjustment mechanism is started to correct the corresponding adjustment instruction.

[0069] From the above, it can be concluded that this embodiment constructs a multi-objective optimization objective function based on the fault diagnosis report and sets constraints based on equipment operating parameters and environmental condition data. This avoids blind optimization out of the fault state and takes into account the multiple needs of energy efficiency, equipment safety, and user comfort of the multi-split air conditioning system. By using the global search capability of the genetic algorithm to solve for the optimal parameter combination, the drawback of optimization algorithms being prone to getting trapped in local optima is effectively avoided, so that the optimal operating parameter combination maximizes energy efficiency under the premise of satisfying the constraints. Then, the optimal parameters are transformed into control strategies and distributed to the corresponding components, which can not only improve the energy efficiency level of the multi-split air conditioning system under fault conditions, but also reduce the deterioration rate of faulty components and reduce ineffective operation and maintenance costs.

[0070] In one embodiment of this application, an energy efficiency optimization control strategy is generated based on the optimal combination of equipment operating parameters, including: Analyze the optimal combination of equipment operating parameters to determine the target control parameters for each actuator in the multi-split air conditioning system; the actuators include the inverter compressor, indoor fan, and electronic expansion valve. The target control parameters include: the target operating frequency of the variable frequency compressor, the target speed of the indoor fan, and the target opening degree of the electronic expansion valve; Based on the target operating frequency, target speed, and target opening degree, corresponding frequency conversion control commands, speed regulation commands, and opening degree adjustment commands are generated respectively. An energy efficiency optimization control strategy is formed based on frequency conversion control commands, speed regulation control commands, and opening degree adjustment commands.

[0071] In this embodiment, the actuators include a variable frequency compressor, an indoor fan, and an electronic expansion valve, which determine the system's cooling / heating efficiency and operational stability. The optimal combination of equipment operating parameters is analyzed to obtain the target control parameters, including the target operating frequency of the compressor, the target speed of the indoor fan, and the target opening degree of the electronic expansion valve.

[0072] Based on the target operating frequency, target speed, and target opening degree, corresponding control commands are generated following the principles of parameter precision and instruction standardization. For frequency converter control commands, the precision extends to the specific value of the compressor's target operating frequency and the slope of frequency adjustment. For speed control commands, based on the indoor ambient temperature deviation, the target speed is converted into multi-level control signals, such as 1500 r / min for high fan speed and 1000 r / min for medium fan speed, while setting the allowable range for speed fluctuations to balance comfort and energy efficiency. For opening degree adjustment commands, the target opening percentage, single adjustment step size, and response delay time of the electronic expansion valve need to be calibrated. In addition, all commands include fault protection thresholds; for example, when the compressor frequency is ≥60Hz, real-time monitoring of the exhaust temperature is triggered.

[0073] An energy efficiency optimization control strategy is constructed based on frequency conversion control commands, speed regulation control commands, and opening degree adjustment commands. Specifically, the construction of the energy efficiency optimization control strategy prioritizes commands according to the principle of safety first and energy efficiency second: the first priority is fault protection commands (such as compressor frequency reduction protection commands), the second priority is core parameter adjustment commands (such as electronic expansion valve opening commands), and the third priority is comfort optimization commands (such as indoor fan speed regulation commands). Simultaneously, command coordination logic is set for different fault levels. For example, when the compressor is overloaded (level 1 fault), the energy efficiency optimization control strategy simultaneously issues coordinated commands to reduce the compressor frequency to 30Hz, increase the electronic expansion valve opening to 40%, and adjust the indoor fan to medium speed to reduce the compressor load; when the filter is slightly clogged (level 3 fault), the energy efficiency optimization control strategy can issue only a single command to increase the fan speed by 10%, ultimately forming the energy efficiency optimization control strategy.

[0074] This embodiment also includes adjusting the parameter configuration of the genetic algorithm. Specifically, the current operating status parameters of the multi-split air conditioning system are obtained, and the parameter configuration of the genetic algorithm is determined based on the current operating status parameters. The parameter configuration includes population size, crossover probability, and mutation probability. During the iteration process of the genetic algorithm, after each generation iteration, the diversity index and the optimal solution improvement rate of the current population are evaluated to obtain the evaluation results. Based on the evaluation results, the crossover probability and mutation probability of the next generation iteration are adjusted. If the diversity index is greater than a preset diversity threshold, the mutation probability is increased; if the optimal solution improvement rate is less than a preset improvement rate threshold, the crossover probability is increased. The next generation iteration is executed according to the adjusted parameters, and the above iteration and parameter adjustment process is repeated until the iteration meets the convergence condition and the optimal combination of equipment operating parameters is output.

[0075] Population size refers to the number of individuals included in each generation of a genetic algorithm. Each individual corresponds to a set of operating parameters of a multi-split air conditioning system to be optimized. The population size determines the search range of a single iteration of the algorithm. Crossover probability is the probability that two parent individuals in a genetic algorithm will exchange gene segments to generate offspring individuals. Crossover is the core means for the algorithm to generate new solutions and explore the optimal solution space. The probability affects the evolution speed of the population. Mutation probability refers to the probability that an individual's gene segments will be randomly changed. Mutation can break the limitation of local optima in the population and maintain population diversity. The population diversity index is a quantitative indicator used to represent the degree of difference in the parameter combinations of individuals in a certain generation of the population. If the diversity is high, it means that the population includes a wide range of solution spaces. If the diversity is low, it means that the individuals in the population tend to be homogeneous and are prone to getting trapped in local optima. The optimal solution improvement rate refers to the objective function corresponding to the best individual in the population in two adjacent iterations. The increase in the value indicates the optimization efficiency of the genetic algorithm at the current stage; the diversity threshold is a pre-set benchmark value for judging whether the degree of differentiation of the current population needs to be adjusted for the mutation probability. It is based on historical optimization data of multi-split air conditioning system parameters and on the experience value of avoiding local optima in genetic algorithms; the improvement rate threshold is a pre-set benchmark value for judging whether the optimization efficiency of the current algorithm needs to be adjusted for the optimal solution improvement rate. It is set according to the minimum effective improvement rate requirement of the target optimization effect of the multi-split air conditioning system; the convergence condition is the basis for judging whether the genetic algorithm stops iterating. It includes the number of iterations reaching a preset value, the optimal solution improvement rate being less than the preset improvement rate threshold for several consecutive generations, and the population individuals tending to stabilize. It is set by referring to the iterative experience of genetic algorithms in similar multi-split air conditioning system parameter optimization scenarios and based on the target optimization accuracy requirements.

[0076] From the above, it can be concluded that this embodiment obtains the target control parameters of the variable frequency compressor, indoor fan, and electronic expansion valve actuators by decomposing the optimal combination of equipment operating parameters, thus avoiding the disconnect between the target control parameters and the control requirements of the components. Secondly, it generates standardized variable frequency, speed regulation, and opening degree adjustment commands based on the target parameters, improving the executability of the commands. Finally, it forms an energy efficiency optimization control strategy based on the commands, realizing the coordinated unification of control commands for multiple components. This not only avoids operational conflicts caused by independent adjustment of each component, but also improves the accuracy of the equipment operating parameter combination by adjusting the parameters in the genetic algorithm, further improving the precision of the energy efficiency optimization control strategy.

[0077] Corresponding to the multi-split air conditioning system energy efficiency optimization and fault diagnosis method in the above embodiments, Figure 2 This is a structural block diagram of a multi-split air conditioning system energy efficiency optimization and fault diagnosis system provided in one embodiment of this application. For ease of explanation, only the parts relevant to the embodiment of this application are shown. References Figure 2The multi-split air conditioning system energy efficiency optimization and fault diagnosis system 20 includes: a data acquisition module 21, a target energy efficiency determination module 22, an energy efficiency anomaly judgment module 23, a fault mechanism generation module 24, a fault diagnosis module 25, and an optimization strategy determination module 26.

[0078] Among them, the data acquisition module 21 is used to acquire the first data, which includes energy efficiency data, environmental condition data and equipment operating parameters; The target energy efficiency determination module 22 is used to calculate energy efficiency data to obtain energy efficiency indicators; and to correct the energy efficiency indicators based on environmental operating condition data to obtain the target energy efficiency indicators. The energy efficiency anomaly determination module 23 is used to compare the target energy efficiency index with a preset first threshold; if the target energy efficiency index is less than the first threshold, the multi-split air conditioning system is determined to be in an energy efficiency anomaly state. The fault mechanism generation module 24 is used to input the first data corresponding to the abnormal energy efficiency state into the pre-trained abnormal causal model to obtain the fault mechanism that leads to the abnormal energy efficiency state. The fault diagnosis module 25 is used to determine the fault source based on the fault mechanism; to identify the fault type and assess the severity of the fault source based on the preset fault diagnosis rule base, and to generate a fault diagnosis report. The optimization strategy determination module 26 is used to determine and execute energy efficiency optimization control strategies based on fault diagnosis reports.

[0079] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules in the aforementioned device embodiments, for example... Figure 2 The functions of the data acquisition module 21, target energy efficiency determination module 22, energy efficiency anomaly judgment module 23, fault mechanism generation module 24, fault diagnosis module 25, and optimization strategy determination module 26 are shown.

[0080] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0081] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0082] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store device type information.

[0083] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation methods described in any embodiment of the multi-split air conditioning system energy efficiency optimization and fault diagnosis method provided in the embodiments of this application, or they can execute the implementation methods of the electronic devices described in the embodiments of this application, which will not be repeated here.

[0084] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0085] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0086] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0087] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0088] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or units, or it may be an electrical, mechanical, or other form of connection.

[0089] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0090] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0091] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system, characterized in that, include: Acquire first data, which includes energy efficiency data, environmental condition data, and equipment operating parameters; The energy efficiency data is calculated to obtain energy efficiency indicators; The energy efficiency index is corrected based on the environmental operating condition data to obtain the target energy efficiency index. The target energy efficiency index is compared with a preset first threshold; if the target energy efficiency index is less than the first threshold, the multi-split air conditioning system is determined to be in an abnormal energy efficiency state. The first data corresponding to the abnormal energy efficiency state is input into the pre-trained abnormal causal model to obtain the fault mechanism that leads to the abnormal energy efficiency state. Based on the aforementioned fault mechanism, the fault source is determined; Based on a preset fault diagnosis rule base, the fault source is identified in terms of fault type and severity is assessed, and a fault diagnosis report is generated. Based on the fault diagnosis report, an energy efficiency optimization control strategy is determined and implemented.

2. The method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system according to claim 1, characterized in that, The process of determining the fault source based on the fault mechanism includes: Based on the aforementioned fault mechanism, a search is performed in a preset fault mechanism-component association database to obtain candidate fault sources; Extract the equipment operating parameters corresponding to each candidate fault source from the equipment operating parameters to obtain the target equipment operating parameters; The matching degree of the target device's operating parameters is calculated by comparing them with a preset fault mechanism feature template library to obtain a matching degree score; All candidate fault sources are sorted according to the matching score to obtain a candidate fault source sequence; Candidate fault sources in the candidate fault source sequence whose matching score is greater than or equal to a preset second threshold are identified as the fault sources. Candidate fault sources in the candidate fault source sequence whose matching score is less than the second threshold are designated as components to be monitored. The monitored component is subjected to fault verification, and the monitored component that passes the fault verification is taken as the fault source.

3. The method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system according to claim 2, characterized in that, The step of calculating the matching degree between the target device operating parameters and the fault mechanism feature template library to obtain a matching degree score includes: Multi-scale feature extraction is performed on the operating parameters of the target device to obtain a first multi-dimensional feature vector; From the fault mechanism feature template library, retrieve the second multidimensional feature vector template corresponding to the fault mechanism. The second multidimensional feature vector template is the standard multidimensional feature vector template of the fault mechanism. The difference between the first multidimensional feature vector and the second multidimensional feature vector template is calculated to obtain the feature deviation vector. Based on the feature deviation vector, the similarity between the first multidimensional feature vector and the second multidimensional feature vector template is calculated to obtain the matching score.

4. The method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system according to claim 3, characterized in that, The step of calculating the similarity between the first multidimensional feature vector and the second multidimensional feature vector based on the feature deviation vector to obtain the matching score includes: The deviation level of the feature deviation vector is determined based on the absolute value of the feature deviation vector and a preset deviation-level mapping table; the deviation level includes a first deviation level, a second deviation level and a third deviation level, and the first deviation level is greater than the second deviation level, and the second deviation level is greater than the third deviation level. The adjustment direction is determined based on the deviation direction of the characteristic deviation vector and the standard change direction of the fault mechanism; Adjusting the baseline weights of the first multidimensional feature vector and the second multidimensional feature vector based on the deviation level and the adjustment direction includes: If the direction of the deviation is consistent with the standard change direction under the fault mechanism, then positive adjustment is performed: If the direction of the deviation is inconsistent with the standard change direction under the fault mechanism, then negative adjustment is performed: If the deviation level of the feature deviation vector is the first deviation level, then its benchmark weight coefficient is adjusted using the first adjustment amplitude. If the deviation level of the feature deviation vector is the second deviation level, then its baseline weight coefficient is adjusted using the second adjustment amplitude. If the deviation level of the feature deviation vector is the third deviation level, then its initial weight coefficient remains unchanged; Wherein, the first adjustment range is greater than the second adjustment range; Based on the adjusted weight coefficients, the similarity between the first multidimensional feature vector and the second multidimensional feature vector template is calculated using a weighted similarity fusion algorithm to obtain the matching score.

5. The method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system according to claim 2, characterized in that, The step of fault verification of the monitored component, wherein the monitored component that passes the fault verification is taken as the fault source, includes: Generate a monitoring task for each candidate component among the components to be monitored; The second data, collected based on the monitoring task, is input into the pre-trained performance degradation evaluation model to obtain the health index and degradation rate of each candidate component; the second data is data of the same type as the first data, collected based on the monitoring task. If the health index is less than a preset third threshold, or the degradation rate is greater than a preset fourth threshold, then the corresponding monitored component is taken as the source of the fault.

6. The method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system according to claim 1, characterized in that, The step of determining and implementing an energy efficiency optimization control strategy based on the fault diagnosis report includes: Based on the fault diagnosis report, an objective function with the goal of optimizing system energy efficiency is constructed, and the operating constraints of the multi-split air conditioning system are set based on the equipment operating parameters and the environmental condition data. The objective function is solved using a genetic algorithm under the constraints to obtain the optimal combination of equipment operating parameters; Based on the optimal combination of equipment operating parameters, an energy efficiency optimization control strategy is generated. The energy efficiency optimization control strategy is sent to the corresponding execution component of the multi-split air conditioning system.

7. The method for energy efficiency optimization and fault diagnosis of a multi-split air conditioning system according to claim 6, characterized in that, The step of generating an energy efficiency optimization control strategy based on the optimal combination of equipment operating parameters includes: The optimal combination of equipment operating parameters is analyzed to determine the target control parameters for each actuator in the multi-split air conditioning system; wherein, the actuator includes a variable frequency compressor, an indoor fan, and an electronic expansion valve; The target control parameters include: the target operating frequency of the variable frequency compressor, the target speed of the indoor fan, and the target opening degree of the electronic expansion valve; Based on the target operating frequency, target speed and target opening degree, corresponding frequency conversion control command, speed regulation control command and opening degree adjustment command are generated respectively. The energy efficiency optimization control strategy is based on the frequency conversion control command, speed regulation control command, and opening degree adjustment command.

8. A multi-split air conditioning system energy efficiency optimization and fault diagnosis system, characterized in that, include: The data acquisition module is used to acquire first data, which includes energy efficiency data, environmental condition data, and equipment operating parameters. The target energy efficiency determination module is used to calculate the energy efficiency data to obtain energy efficiency indicators; and to correct the energy efficiency indicators based on the environmental operating condition data to obtain the target energy efficiency indicators. An energy efficiency anomaly determination module is used to compare the target energy efficiency index with a preset first threshold; if the target energy efficiency index is less than the first threshold, the multi-split air conditioning system is determined to be in an energy efficiency anomaly state. The fault mechanism generation module is used to input the first data corresponding to the abnormal energy efficiency state into a pre-trained abnormal causal model to obtain the fault mechanism that leads to the abnormal energy efficiency state. The fault diagnosis module is used to determine the fault source based on the fault mechanism. Based on a preset fault diagnosis rule base, the fault source is identified in terms of fault type and severity is assessed, and a fault diagnosis report is generated. The optimization strategy determination module is used to determine and execute energy efficiency optimization control strategies based on the fault diagnosis report.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images