Refrigerant Leakage Fault Prediction Method, System, Device and Medium for Refrigeration Equipment

Abstract

The present invention discloses a method, system, device and medium for predicting refrigerant leakage faults of a refrigeration device. The method includes: collecting the operating parameters of the refrigeration device under various working conditions, and obtaining a training set after feature screening of the operating parameters; training a first SVM classifier with normal steady-state data and leakage steady-state data, and training a second SVM classifier with interference fault steady-state data; constructing an LSTM time series prediction model according to transient leakage data; inputting the real-time operating parameters of the refrigeration device into the LSTM time series prediction model, outputting a predicted value, obtaining a predicted parameter set according to the predicted value, and inputting the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis. The present invention uses a time series prediction model to obtain a predicted parameter set, and performs pattern judgment on the parameter set to realize fault early warning, solves the problem that traditional methods do not have prediction ability, and can be widely applied to the field of leakage fault diagnosis and prediction.

Description

Technical Field

[0001] The present invention relates to the field of leakage fault diagnosis and prediction, and particularly to a method, system, device and medium for predicting refrigerant leakage faults in refrigeration equipment. Background Art

[0002] After long-term operation, refrigeration equipment may have refrigerant leakage faults due to aging of connecting components or pipeline damage caused by vibration and corrosion. Moreover, since the system is in a high-pressure state, it will lead to a decline in the performance of the refrigeration equipment and cause energy waste. More seriously, if the refrigerant is flammable such as R290, a large amount of leakage indoors will pose an explosion hazard. Therefore, it is particularly important to detect refrigerant leakage in the refrigeration system as early as possible.

[0003] For refrigerant leakage faults, the existing fault diagnosis methods mainly include the following aspects: one is to analyze the variation laws of various parameters caused by the leakage of the refrigeration system and use this as the judgment basis; the other is to establish a fault diagnosis model through data mining methods and complete fault identification by judging the test data. However, currently, the fault diagnosis methods can basically only complete the identification of faults after the faults occur and do not have the ability of early warning. Summary of the Invention

[0004] To at least partly solve one of the technical problems existing in the prior art, an object of the present invention is to provide a method, system, device and medium for predicting refrigerant leakage faults in refrigeration equipment.

[0005] The technical solution adopted by the present invention is as follows:

[0006] A method for predicting refrigerant leakage faults in refrigeration equipment includes the following steps:

[0007] Collect the operating parameters of the refrigeration equipment under various working conditions. After feature screening of the operating parameters, a training set is obtained; wherein, the training set includes normal steady-state data, leakage steady-state data, interference fault steady-state data and transient leakage data;

[0008] Train a first SVM classifier with the normal steady-state data and the leakage steady-state data, and train a second SVM classifier with the interference fault steady-state data;

[0009] Construct an LSTM time series prediction model according to the transient leakage data;

[0010] Input the real-time operating parameters of the refrigeration equipment into the LSTM time series prediction model, output the predicted value, obtain a predicted parameter set according to the predicted value, and input the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis.

[0011] Further, after collecting the operating parameters of the refrigeration equipment under various working conditions and performing feature screening on the operating parameters, a training set is obtained, including:

[0012] Control the refrigeration equipment to operate under normal conditions, insufficient refrigerant conditions, interference fault conditions, and refrigerant leakage conditions respectively, and collect the operating parameters under different conditions;

[0013] Label the collected operating parameters,

[0014] Perform feature screening on the labeled operating parameters to obtain a training set.

[0015] Further, the feature screening of the labeled operating parameters includes:

[0016] Adopt the Pearson correlation coefficient method to calculate the correlation between each operating parameter and the correlation between the operating parameter and the label, and select an operating parameter with the largest correlation with the label as the representative feature to complete the feature screening.

[0017] Further, the leakage steady-state data includes 10% leakage data, 20% leakage data, and 30% leakage data;

[0018] The training of the first SVM classifier using the normal steady-state data and the leakage steady-state data, and the training of the second SVM classifier using the interference fault steady-state data includes:

[0019] Use the normal steady-state data, 10% leakage data, 20% leakage data, and 30% leakage data to preliminarily train the first SVM classifier, and output the prediction scores corresponding to 4 labels;

[0020] Input the interference fault steady-state data and the transient leakage data into the preliminarily trained first SVM classifier. After repeated verification on multiple data sets, obtain the law between the prediction scores and the interference fault and the transient leakage fault, and obtain the prediction score threshold θ. Distinguish various operating modes according to the prediction score threshold θ;

[0021] Use the interference fault steady-state data to train the second SVM classifier;

[0022] Among them, the first SVM classifier is used to identify the leakage fault and the interference fault, and the second SVM classifier makes a further judgment on the data that has been diagnosed as an interference fault to determine the type of the interference fault.

[0023] Further, the distinguishing of various operating modes according to the prediction score threshold θ includes:

[0024] When the maximum prediction score is greater than θ, directly judge that the refrigeration equipment is operating in the normal mode or the leakage mode at this time;

[0025] When the maximum predicted score is less than θ, and the sum of the maximum predicted score and the second - largest predicted score is less than θ, it is determined that the refrigeration equipment is operating in the interference fault mode at this time;

[0026] When the maximum predicted score is less than θ, and the sum of the maximum predicted score and the second - largest predicted score is greater than θ, it is determined that the refrigeration equipment is in the process of leakage between the labels corresponding to these two predicted scores at this time.

[0027] Furthermore, the construction of the LSTM time - series prediction model based on transient leakage data includes:

[0028] Determine the input step number n and output step number m of the time series;

[0029] According to the input step number n and output step number m, standardize the original transient leakage data and convert it into time - series format data;

[0030] Construct an LSTM time - series prediction model, and use the time - series format data to train the LSTM time - series prediction model.

[0031] Furthermore, inputting the real - time operating parameters of the refrigeration equipment into the LSTM time - series prediction model, outputting predicted values, obtaining a predicted parameter set according to the predicted values, and inputting the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis includes:

[0032] Collect n - step data to form a test data of a multivariate time series, and standardize the test data;

[0033] Input the standardized test data into the LSTM time - series prediction model, and output m - step predicted data;

[0034] Input the data at the t + m moment into the first SVM classifier to judge the working mode of the refrigeration equipment at this time;

[0035] If the output of the first SVM classifier is the normal mode, do nothing and continue with fault monitoring;

[0036] If the output of the first SVM classifier is the leakage fault mode, trigger a leakage alarm;

[0037] If the output of the first SVM classifier is the interference fault mode, trigger an interference fault alarm, and at the same time input the data into the second SVM classifier model to judge the specific type of the interference fault.

[0038] Another technical solution adopted by the present invention is:

[0039] A refrigerant leakage fault prediction system for a refrigeration equipment, comprising:

[0040] A data acquisition module, which is used to collect the operation parameters of the refrigeration equipment under various working conditions, and obtain a training set after feature screening of the operation parameters; among them, the training set includes normal steady-state data, leakage steady-state data, interference fault steady-state data, and transient leakage data.

[0041] A classifier training module, which is used to train a first SVM classifier with normal steady-state data and leakage steady-state data, and train a second SVM classifier with interference fault steady-state data.

[0042] A model construction module, which is used to construct an LSTM time series prediction model according to the transient leakage data.

[0043] A model coupling module, which is used to input the real-time operation parameters of the refrigeration equipment into the LSTM model, output a predicted value, obtain a predicted parameter set according to the predicted value, and input the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis.

[0044] Another technical solution adopted by the present invention is:

[0045] A refrigerant leakage fault prediction device for a refrigeration equipment, comprising:

[0046] At least one processor;

[0047] At least one memory, which is used to store at least one program;

[0048] When the at least one program is executed by the at least one processor, the at least one processor implements the refrigerant leakage fault prediction method for a refrigeration equipment.

[0049] Another technical solution adopted by the present invention is:

[0050] A computer-readable storage medium, in which a program executable by a processor is stored, and the program executable by the processor is used to execute the refrigerant leakage fault prediction method for a refrigeration equipment as described above when executed by the processor.

[0051] The beneficial effects of the present invention are: The present invention adopts an LSTM-SVM coupling model, uses a time series prediction model to obtain a predicted parameter set, and then performs pattern judgment on the parameter set, which can realize fault early warning to a certain extent, and solves the problem that the traditional method can only realize single fault diagnosis and does not have the prediction ability. Description of the Drawings

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following introduces the accompanying drawings of the related technical solutions in the embodiments of the present invention or the prior art. It should be understood that the accompanying drawings in the following introduction are only for conveniently and clearly presenting some embodiments of the technical solutions of the present invention. For those skilled in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0053] Figure 1 is a flowchart of a method for predicting refrigerant leakage faults in a refrigeration device according to an embodiment of the present invention;

[0054] Figure 2 is a data acquisition parameter curve graph according to an embodiment of the present invention;

[0055] Figure 3 is a heat map of the correlation between features before feature screening according to an embodiment of the present invention;

[0056] Figure 4 is a heat map of the correlation between features after feature screening according to an embodiment of the present invention;

[0057] Figure 5 is a prediction curve of the condenser outlet pressure according to an embodiment of the present invention;

[0058] Figure 6 is a schematic diagram of the fault prediction result according to an embodiment of the present invention;

[0059] Figure 7 is a step flowchart of a method for predicting refrigerant leakage faults in a refrigeration device according to an embodiment of the present invention. Detailed Embodiments

[0060] The following details the embodiments of the present invention. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation to the present invention. For the step numbers in the following embodiments, they are only set for the convenience of elaboration and explanation, and no limitation is imposed on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0061] In the description of the present invention, it should be understood that for the orientation description, such as the orientation or positional relationship indicated by up, down, front, back, left, right, etc., is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention.

[0062] In the description of the present invention, "several" means one or more, "multiple" means more than two, "greater than", "less than", "exceeding", etc. are understood not to include the recited number, and "above", "below", "within", etc. are understood to include the recited number. If "first" and "second" are described, they are only used for the purpose of distinguishing technical features and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features or implicitly specifying the sequence relationship of the indicated technical features.

[0063] In the description of the present invention, unless otherwise clearly defined, words such as "set", "installed", "connected", etc. should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above words in the present invention in combination with the specific content of the technical solution.

[0064] As Figure 7 shown, this embodiment provides a method for predicting refrigerant leakage faults of a refrigeration device, including the following steps:

[0065] S1. Collect the operating parameters of the refrigeration device under various working conditions. After feature screening of the operating parameters, a training set is obtained; wherein, the training set includes normal steady-state data, leakage steady-state data, interference fault steady-state data, and transient leakage data;

[0066] S2. Train the first SVM classifier with the normal steady-state data and the leakage steady-state data, and train the second SVM classifier with the interference fault steady-state data;

[0067] S3. Construct an LSTM time series prediction model according to the transient leakage data;

[0068] S4. Input the real-time operating parameters of the refrigeration device into the LSTM time series prediction model, output the predicted value, obtain the predicted parameter set according to the predicted value, and input the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis.

[0069] This embodiment provides a method for predicting refrigerant leakage faults of a refrigeration device that couples a pattern recognition algorithm (SVM) and a time series prediction algorithm (LSTM), thereby realizing early detection of refrigerant leakage in the refrigeration device, that is, early warning of large refrigerant leakage.

[0070] Embodiment 1

[0071] As Figure 1 shown, this embodiment provides a method for predicting refrigerant leakage faults of a refrigeration device based on LSTM-SVM. The steps of this method are as follows:

[0072] S101. Data collection and feature selection.

[0073] Design and conduct experiments on refrigerant leakage faults and interference faults. Under the conditions of normal system operation, insufficient refrigerant quantity operation, interference fault operation, and refrigerant leakage operation, collect the system operation parameters under the corresponding working conditions. Since there is a large correlation between the parameters, it is necessary to perform feature screening on the parameter set, and use the Pearson correlation coefficient method to obtain the screened parameter set. The specific steps are as shown in steps S1011 - S1013:

[0074] S1011. Conduct experiments and collect the operating parameters of the refrigeration system to form a training set. The training set includes normal, leakage, and interference fault steady-state data, as well as transient leakage data;

[0075] S1012. Label various types of steady-state data. For example, the label for normal data is 0, and the label for 10% leakage data is 1, and so on. There are a total of 8 modes.

[0076] S1013. Use the Pearson correlation coefficient method to calculate the correlation between each feature parameter and the correlation between each parameter and the label. Select 1 feature with the largest correlation with the label among the features with strong correlation as the representative feature to complete the feature screening.

[0077] S102. Build a classifier model.

[0078] Based on the screened parameter set, use the steady-state data in the normal and leakage modes to train the SVM classifier svm1 (i.e., the first SVM classifier), and use the steady-state data in the interference fault mode to train the SVM classifier svm2 (i.e., the second SVM classifier). svm1 can accurately identify leakage faults and interference faults, avoiding misjudgment of the model, while svm2 can further determine which interference fault the data diagnosed as an interference fault is in. The specific steps are as shown in steps S1021 - S1024:

[0079] S1021. Use the steady-state parameter sets of normal, 10% leakage, 20% leakage, and 30% leakage to train the SVM classifier model svm1, and perform parameter tuning on the validation set to obtain the best model;

[0080] S1022. For the obtained classifier model svm1, output the prediction scores of each label, and initially obtain the rule between the prediction scores and the prediction labels;

[0081] S1023. Input the interference fault parameter set and the transient leakage parameter set into the svm1 model respectively. After repeated verification on multiple data sets, obtain the rule between the prediction scores and the interference faults and transient leakage faults, and based on this, obtain the prediction score threshold θ for distinguishing various modes. The judgment rule is as follows:

[0082] When the maximum predicted score of the label is greater than θ, it can be directly determined that it is in this label mode (normal or leakage level) at this time;

[0083] When the maximum predicted score of the label is less than θ and the sum of the maximum and the second - largest label scores is less than θ, it is determined that it is in the interference fault mode at this time;

[0084] When the maximum predicted score of the label is less than θ and the sum of the maximum and the second - largest label scores is greater than θ, it is determined that it is in the leakage process between these two labels at this time.

[0085] S1024. Use the interference fault parameter set to train the SVM classifier model svm2. This model further judges the data determined as interference faults by the model svm1 to identify which type of interference fault it is.

[0086] S103. Time - series prediction modeling.

[0087] Based on the transient leakage parameter set, first determine the number of prediction input steps n and the number of output steps m. Temporalize the data set, and then construct an LSTM time - series prediction model and conduct model training. Its specific steps are as described in steps S1031 - S1034:

[0088] S1031. Conduct a small - sample test experiment to determine the optimal number of prediction input steps n, which is determined according to the minimization of the mean square error of the model on the validation set;

[0089] S1032. Based on the data acquisition step size and the backward prediction time requirement, preliminarily determine the number of prediction output steps m. It can be iteratively improved by checking whether the error between the predicted value and the actual value meets the requirements during the subsequent test process;

[0090] S1033. Based on the selected parameters n and m, temporalize the parameter set. Use the data from steps t - n to t as input and the data from t + 1 to t + m as output to form a training set, and perform standardization processing on the training set;

[0091] S1034. Construct a basic LSTM model, use the training set for model training, and perform parameter tuning on the test set until the mean square error on the validation set is minimized.

[0092] S104. Coupled LSTM - SVM model and fault prediction.

[0093] Input the real - time operating parameters of the refrigeration equipment into the LSTM model to obtain the predicted values of each parameter. Then input the predicted parameter set into the svm1 model for refrigerant leakage fault diagnosis. If it is diagnosed as an interference fault, the data can be further input into the svm2 model to diagnose the type of interference fault. Its specific steps are as described in steps S1041 - S1043:

[0094] S1041. Collect n-step data using sensors to form a multi-variable time series test data, and standardize the test data using the standardization model of the training set.

[0095] S1042. Input the standardized test data into the LSTM model to output m-step prediction data.

[0096] S1043. Input the data at the t + m moment into the classifier model svm1 to determine the state of the system at this time, and perform different outputs or feedbacks for different states.

[0097] A1. If the svm1 model outputs the normal mode, do nothing and continue with fault monitoring.

[0098] A2. If the svm1 model outputs the leakage fault mode, trigger a leakage alarm.

[0099] A3. If the svm1 model outputs the interference fault mode, trigger an interference fault alarm, and at the same time input the data into the svm2 model to determine the specific type of interference fault.

[0100] Embodiment 2

[0101] In this embodiment, an R290 air conditioner system is used as the object for predicting refrigerant leakage faults. The system uses R290 refrigerant with a standard filling amount of 300 g. Both the condenser and the evaporator use fin-tube heat exchangers. Since the R290 refrigerant is flammable, its leakage indoors will cause relatively serious safety accidents. Therefore, it is particularly important to detect refrigerant leakage at an early stage.

[0102] The specific steps of applying the refrigerant leakage fault prediction method based on LSTM-SVM provided by the embodiment of the present invention in this system are as follows:

[0103] S201. Data collection and feature selection.

[0104] Among them, step S201 specifically includes steps S2011 - S2013:

[0105] S2011. For common environmental conditions, collect the steady-state operation data of the system under normal operation, 10%, 20%, 30% refrigerant leakage, and interference fault conditions. In addition, it is also necessary to collect the transient leakage data of the system when the refrigerant leaks from 100% to 70%. The collected parameters include the temperature, pressure, etc. of the refrigeration system. Some parameter curves are as Figure 2 shown.

[0106] S2012. Tag the collected steady-state data. The data categories are normal, 10% leakage, 20% leakage, 30% leakage, 1 / 3 condenser occlusion, 2 / 3 condenser occlusion, 1 / 3 evaporator occlusion, and 2 / 3 evaporator occlusion. Therefore, the corresponding tags are {0, 1, 2, 3, 4, 5, 6, 7}. Among them, 1 / 3 condenser occlusion, 2 / 3 condenser occlusion, 1 / 3 evaporator occlusion, and 2 / 3 evaporator occlusion are types of interference faults.

[0107] S2013. Use the Pearson correlation coefficient for feature screening. Calculate the correlation coefficients between features and the correlation coefficients between features and tags. Select one of the most representative features from the features with strong correlation. The correlation heatmaps before and after feature screening are as Figure 3 、 Figure 4 shown.

[0108] S202. Classifier modeling.

[0109] Step S202 specifically includes steps S2021 - S2024:

[0110] S2021. Construct a leakage diagnosis model training set with the steady-state parameter sets of normal, 10% leakage, 20% leakage, and 30% leakage. The sample size is 6,008. Train the SVM to obtain the classifier model svm1.

[0111] S2022. Input the test set into svm1, output the predicted scores of each tag, and initially obtain the rule between the predicted scores and the predicted tags, that is, the predicted scores of the model output tags are greater than 0.99.

[0112] S2023. Input the interference fault parameter set and the transient leakage parameter set into the svm1 model respectively. After repeated verification on multiple data sets, obtain the rules between the predicted scores and the interference faults and transient leakage faults:

[0113] B1. For interference faults, the predicted scores of the 4 tags of the model are all less than 0.99, and the sum of the highest score and the second highest score is still less than 0.99;

[0114] B2. For leakage faults, the predicted scores of the 4 tags of the model are all less than 0.99, but the sum of the highest score and the second highest score is greater than 0.99, and these two tags are the leakage intervals at this time;

[0115] B3. Therefore, the predicted score threshold is determined to be 0.99. When the maximum predicted score of the label is greater than 0.99, it can be directly determined that it is in the mode of this label (normal or leakage level) at this time; when the maximum predicted score of the label is less than 0.99 and the sum of the maximum and the second - largest label scores is less than 0.99, it is determined that it is in the interference fault mode at this time; when the maximum predicted score of the label is less than 0.99 and the sum of the maximum and the second - largest label scores is greater than 0.99, it is determined that it is in the leakage process between these two labels at this time.

[0116] S2024. The steady - state parameter sets collected under 4 interference fault modes are used to form an interference fault diagnosis training set with a sample size of 4056, and the SVM model is trained to obtain the classifier model svm2.

[0117] S203. Time - series prediction modeling.

[0118] Step S203 specifically includes steps S2031 - S2032:

[0119] S2031. Combine the LSTM model to explore the transient leakage data set, determine that the number of input steps of the time series n = 20, the number of output steps m = 5, and convert the original data set into a time - series format after standardization.

[0120] S2032. Build an LSTM time - series prediction model, train the model using the above - mentioned time - series data, use the minimization of the mean - square error as the optimization goal, and the fitting curve of the model on the test set is as Figure 5 shown, which is the condenser outlet pressure.

[0121] S204. Couple the LSTM - SVM model and fault prediction.

[0122] Step S204 specifically includes steps S2041 - S2042:

[0123] S2041. Take 20 - step time - series data, that is, X t-20 ~X t , after standardizing this sequence using the standardization model of the training set, input it into the LSTM prediction model to obtain the predicted values Y t+1 ~Y t+5 .

[0124] S2042. First, standardize and transform Y t+5 using the standardization model of the training set, and then use it as the input value of the leakage fault diagnosis model svm1. The output result of the model is a leakage fault, that is, a refrigerant leakage fault is occurring at this time. As Figure 6 shown, the system fault can be detected 3 steps in advance on the predicted parameter set.

[0125] In summary, compared with the corresponding technologies, the method of this embodiment has the following advantages and beneficial effects:

[0126] (1) The present invention designs a relatively complete experimental scheme for refrigerant leakage and interference faults, conducts experiments and collects the system operation parameters in each mode, ensuring the quality of the modeling data and effectively improving the performance of the model from the perspective of underlying data.

[0127] (2) Based on the traditional support vector machine modeling, the present invention discovers that the prediction scores of different modes have significant regularities, and thus determines the prediction score threshold. Using the prediction score threshold to distinguish normal, leakage, and interference modes, this method can reduce the misjudgment between leakage faults and interference faults to a certain extent and avoid the problem of difficult acquisition of interference condition data.

[0128] (3) The present invention adopts an LSTM-SVM coupling model, uses a time series prediction model to obtain the predicted parameter set, and then judges the mode of the parameter set, which can realize the early warning of faults to a certain extent, while the traditional method can only achieve single fault diagnosis without prediction ability.

[0129] This embodiment also provides a refrigerant leakage fault prediction system for a refrigeration device, including:

[0130] A data acquisition module, configured to collect the operation parameters of the refrigeration device under various working conditions, and obtain a training set after feature screening of the operation parameters; wherein, the training set includes normal steady-state data, leakage steady-state data, interference fault steady-state data, and transient leakage data;

[0131] A classifier training module, configured to train a first SVM classifier using normal steady-state data and leakage steady-state data, and train a second SVM classifier using interference fault steady-state data;

[0132] A model construction module, configured to construct an LSTM time series prediction model according to the transient leakage data;

[0133] A model coupling module, configured to input the real-time operation parameters of the refrigeration device into the LSTM model, output a prediction value, obtain a predicted parameter set according to the prediction value, and input the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis.

[0134] The refrigerant leakage fault prediction system for a refrigeration device of this embodiment can execute a refrigerant leakage fault prediction method provided by the method embodiment of the present invention, can execute any combination of the implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.

[0135] This embodiment also provides a refrigerant leakage fault prediction device for a refrigeration device, including:

[0136] At least one processor;

[0137] At least one memory for storing at least one program;

[0138] When the at least one program is executed by the at least one processor, the at least one processor implements the Figure 7 method shown.

[0139] A refrigerant leakage fault prediction system for a refrigeration device according to this embodiment can execute a refrigerant leakage fault prediction method provided by an embodiment of the method of the present invention, can execute any combination of implementation steps of the method embodiment, and has corresponding functions and beneficial effects of the method.

[0140] This application embodiment also discloses a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes Figure 7 the method shown.

[0141] This embodiment also provides a storage medium storing instructions or a program that can execute a refrigerant leakage fault prediction method provided by an embodiment of the method of the present invention. When the instructions or the program are run, any combination of implementation steps of the method embodiment can be executed, and corresponding functions and beneficial effects of the method are provided.

[0142] In some alternative embodiments, the functions / operations mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the functions / operations involved, two consecutive blocks shown may actually be executed substantially simultaneously or the blocks can sometimes be executed in the reverse order. In addition, the embodiments presented and described in the flowcharts of the present invention are provided by way of example for the purpose of providing a more comprehensive understanding of the technology. The disclosed method is not limited to the operations and logical flows presented herein. Alternative embodiments are contemplated, in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0143] In addition, although the present invention has been described in the context of functional modules, it should be understood that, unless otherwise stated to the contrary, one or more of the described functions and / or features may be integrated in a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It should also be understood that a detailed discussion of the actual implementation of each module is not necessary for understanding the present invention. Rather, considering the attributes, functions, and internal relationships of the various functional modules in the devices disclosed herein, the actual implementation of the modules will be understood within the ordinary skills of an engineer. Thus, those skilled in the art can implement the present invention as set forth in the claims without undue experimentation. It should also be understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0144] If the described functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0145] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in conjunction with such instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0146] More specific examples (nonexhaustive list) of computer-readable media include the following: an electrical connection (electronic device) having one or more wirings, a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable media can even be paper or other suitable media on which the program can be printed, as the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0147] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, the multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having suitable combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0148] In the foregoing description of the present specification, the description with reference to the terms "one embodiment / example", "another embodiment / example", or "certain embodiments / examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0149] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the claims and their equivalents.

[0150] The above has specifically described the preferred embodiments of the present invention, but the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent deformations or substitutions without departing from the spirit of the present invention, and these equivalent deformations or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A refrigerant leakage fault prediction method for a refrigeration device, characterized in that, It includes the following steps: Collect the operating parameters of the refrigeration equipment under various working conditions, and obtain a training set after feature screening of the operating parameters; among them, The training set includes normal steady-state data, leakage steady-state data, interference fault steady-state data, and transient leakage data; Use the normal steady-state data and leakage steady-state data to train the first SVM classifier, and use the interference fault steady-state data to train the second SVM classifier; Construct an LSTM time series prediction model according to the transient leakage data; Input the real-time operating parameters of the refrigeration equipment into the LSTM time series prediction model, output the predicted values, obtain the predicted parameter set according to the predicted values, and input the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis; The leakage steady-state data includes 10% leakage data, 20% leakage data, and 30% leakage data; The step of using the normal steady-state data and leakage steady-state data to train the first SVM classifier, and using the interference fault steady-state data to train the second SVM classifier includes: Use the normal steady-state data, 10% leakage data, 20% leakage data, and 30% leakage data to preliminarily train the first SVM classifier, and output the predicted scores corresponding to 4 labels; Input the interference fault steady-state data and transient leakage data into the preliminarily trained first SVM classifier. After repeated verification on multiple data sets, obtain the rules between the predicted scores and interference faults, transient leakage faults, and obtain the predicted score threshold θ. Distinguish various operating modes according to the predicted score threshold θ; Use the interference fault steady-state data to train the second SVM classifier; Among them, the first SVM classifier is used to identify leakage faults and interference faults, and the second SVM classifier makes further judgments on the data that has been diagnosed as interference faults to determine the type of interference faults.

2. The refrigerant leakage fault prediction method of a refrigeration device according to claim 1, characterized in that The step of collecting the operating parameters of the refrigeration equipment under various working conditions, and obtaining a training set after feature screening of the operating parameters includes: Control the refrigeration equipment to operate under normal conditions, insufficient refrigerant conditions, interference fault conditions, and refrigerant leakage conditions respectively, and collect the operating parameters under different conditions; Label the collected operating parameters, Perform feature screening on the labeled operating parameters to obtain a training set.

3. A refrigerant leakage fault prediction method for a refrigeration device according to claim 2, characterized in that, The step of performing feature screening on the labeled operating parameters includes: Adopt the Pearson correlation coefficient method to calculate the correlation between each operating parameter and the correlation between the operating parameter and the label, and select an operating parameter with the largest correlation with the label as the representative feature to complete the feature screening.

4. A refrigerant leakage fault prediction method for a refrigeration device according to claim 1, characterized in that, The step of distinguishing various operating modes according to the predicted score threshold θ includes: When the maximum predicted score is greater than θ, directly judge that the refrigeration equipment is working in the normal mode or leakage mode at this time; When the maximum predicted score is less than θ, and the sum of the maximum predicted score and the second largest predicted score is less than θ, judge that the refrigeration equipment is working in the interference fault mode at this time; When the maximum predicted score is less than θ, and the sum of the maximum predicted score and the second largest predicted score is greater than θ, judge that the refrigeration equipment is working in the leakage process between the labels corresponding to these two predicted scores at this time.

5. A refrigerant leakage fault prediction method for a refrigeration device according to claim 1, characterized in that, Constructing an LSTM time series prediction model based on transient leakage data includes: Determining the input step number n and output step number m of the time series; According to the input step number n and output step number m, standardizing the original transient leakage data and converting it into time series format data; Constructing an LSTM time series prediction model and training the LSTM time series prediction model using the time series format data.

6. A refrigerant leakage fault prediction method for a refrigeration device according to claim 5, characterized in that, Inputting the real-time operating parameters of the refrigeration equipment into the LSTM time series prediction model, outputting predicted values, obtaining a predicted parameter set according to the predicted values, and inputting the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis, including: Collecting n-step data to form a test data of a multivariate time series, and standardizing the test data; Inputting the standardized test data into the LSTM time series prediction model to output m-step predicted data; Inputting the data at the t + m moment into the first SVM classifier to judge the working mode of the refrigeration equipment at this time; If the first SVM classifier outputs a normal mode, no processing is performed and fault monitoring continues; If the first SVM classifier outputs a leakage fault mode, a leakage alarm is triggered; If the first SVM classifier outputs an interference fault mode, an interference fault alarm is triggered, and at the same time, the data is input into the second SVM classifier model to judge the specific type of the interference fault.

7. A refrigerant leakage fault prediction system for a refrigeration device, characterized in that, Including: A data acquisition module for collecting the operating parameters of the refrigeration equipment under various working conditions, performing feature screening on the operating parameters, and obtaining a training set; wherein, the training set includes normal steady-state data, leakage steady-state data, interference fault steady-state data, and transient leakage data; A classifier training module for training the first SVM classifier using normal steady-state data and leakage steady-state data, and training the second SVM classifier using interference fault steady-state data; A model construction module for constructing an LSTM time series prediction model based on transient leakage data; A model coupling module for inputting the real-time operating parameters of the refrigeration equipment into the LSTM model, outputting predicted values, obtaining a predicted parameter set according to the predicted values, and inputting the predicted parameter set into the first SVM classifier and the second SVM classifier for refrigerant leakage fault diagnosis; The leakage steady-state data includes 10% leakage data, 20% leakage data, and 30% leakage data; Training the first SVM classifier using normal steady-state data and leakage steady-state data, and training the second SVM classifier using interference fault steady-state data, including: Preliminarily training the first SVM classifier using normal steady-state data, 10% leakage data, 20% leakage data, and 30% leakage data, and outputting the predicted scores corresponding to 4 labels; Inputting the interference fault steady-state data and transient leakage data into the preliminarily trained first SVM classifier, obtaining the law between the predicted scores and interference faults, transient leakage faults through repeated verification on multiple data sets, and obtaining a predicted score threshold θ, and distinguishing various operating modes according to the predicted score threshold θ; Training the second SVM classifier using interference fault steady-state data; Among them, the first SVM classifier is used to identify leakage faults and interference faults, and the second SVM classifier further judges the data that has been diagnosed as interference faults to determine the type of interference faults.

8. A refrigerant leakage fault prediction device for a refrigeration equipment, characterized in that, Including: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1-6.

9. A computer-readable storage medium storing a program executable by a processor, characterized in that, The program executable by the processor is used to execute the method according to any one of claims 1-6 when executed by the processor.

Images