A medium leakage early warning method and system for refrigerant filling equipment

Abstract

The present invention relates to the technical field of refrigerant leakage, and specifically to a method and system for early warning of medium leakage of refrigerant filling equipment; the method comprises: matching the current vibration data sequence with the historical vibration data sequence by using a DTW algorithm to analyze the slope change of the matching point between the current vibration data sequence and the historical vibration data sequence, establishing an abnormal value of the current vibration data sequence, and dynamically setting a threshold value by using the distribution of the abnormal value and the abnormal value of the historical vibration data sequence. If the abnormal value exceeds the set threshold, an abnormal warning is triggered, indicating the presence of a refrigerant leakage.

Description

Technical Field

[0001] The present invention relates to the technical field of refrigerant leakage, and in particular to a medium leakage early warning method and system for refrigerant filling equipment. Background Art

[0002] Refrigerant leakage is a frequent and significant fault in the operation of refrigeration equipment, especially air conditioning systems. As a critical medium in the refrigeration cycle, the integrity of the refrigerant directly determines the system's cooling efficiency and operational stability. Traditional methods for detecting refrigerant leaks in air conditioners rely primarily on monitoring and analyzing operating parameters such as the temperature and power of the air conditioning system. However, this approach has significant limitations. Because changes in parameters such as temperature and power are not only affected by refrigerant leakage but can also be caused by a variety of factors, including environmental changes, equipment aging, and load fluctuations, it is difficult to accurately determine whether a refrigerant leak has occurred based solely on simple monitoring of these parameters, and the detection results are subject to a high degree of lag. This not only prevents the system from issuing timely warnings of refrigerant leaks, but can also delay subsequent treatment measures, affecting the normal operation of the equipment, accelerating equipment aging, and shortening its service life.

[0003] To overcome the shortcomings of traditional methods, the industry has begun exploring the use of vibration data monitoring technology to detect refrigerant leaks in air conditioners. The basic principle of this technology is to monitor the vibration signals of the air conditioning system during operation, analyze the changes in the vibration data, and compare it with a preset threshold to determine whether a refrigerant leak exists. While this method has improved detection sensitivity and accuracy to a certain extent, the fixed preset threshold has become a new problem. Currently, most systems use a manual and subjective threshold selection method, which cannot adapt to the complex changes in air conditioning system operating conditions. Specifically, if the preset threshold is set too low, the system may generate false positives due to normal load fluctuations, resulting in unnecessary alarms and wasted resources. Conversely, if the preset threshold is set too high, a true refrigerant leak may be missed, missing the optimal repair opportunity and exacerbating equipment damage.

[0004] Therefore, it is necessary to propose a solution that can solve the problem that the existing technology has a fixed preset threshold when judging refrigerant leakage, resulting in low accuracy and high hysteresis in detecting refrigerant leakage, and cannot provide timely and accurate warning and subsequent processing of refrigerant leakage. Summary of the Invention

[0005] The purpose of the present invention is to propose a method for warning of medium leakage of refrigerant filling equipment, so as to solve the problem that the existing technology fixes the preset threshold when judging refrigerant leakage, resulting in low accuracy and high hysteresis in detecting refrigerant leakage, and cannot timely and accurately warn and subsequently handle refrigerant leakage; to this end, the present invention provides solutions in the following two aspects.

[0006] In a first aspect, the present invention provides a method for early warning of medium leakage in refrigerant filling equipment, comprising:

[0007] Obtain the current vibration data sequence of the refrigeration equipment and multiple historical vibration data sequences under different working conditions;

[0008] Taking the current moment as the starting point, sequence segments of different lengths are intercepted from the current vibration data sequence. The DTW algorithm is used to calculate the path distance between the sequence segments of each length and any historical vibration data sequence, and the sequence segment corresponding to the shortest path distance is taken as the optimal sequence segment.

[0009] Calculating the matching degree between the current vibration data sequence and each working condition, wherein the matching degree is inversely correlated with the shortest path distance of each historical vibration data sequence under the corresponding working condition;

[0010] Obtaining a matching pair sequence of any historical vibration data sequence and a corresponding optimal sequence segment of the working condition corresponding to the maximum matching degree, wherein the matching pair is composed of vibration data in any historical vibration data sequence and vibration data in the optimal sequence segment;

[0011] Calculating an outlier value of the current vibration data sequence, wherein the outlier value is positively correlated with the coefficient of variation of each matching pair sequence and the shortest path distance; the coefficient of variation is obtained based on the slope of the matching pair in each matching pair sequence; the slope is the slope of the line connecting the two data in the matching pair in the coordinate system formed by the time and vibration data;

[0012] If the abnormal value exceeds the set threshold, an abnormal warning is triggered, indicating that there is a refrigerant leak.

[0013] The above scheme uses the DTW algorithm to match the optimal sequence segment with the historical vibration data sequence in order to analyze the slope change of the matching point between the optimal sequence segment and the historical vibration data sequence, establish the abnormal value of the current vibration data sequence, and use the distribution of the abnormal value and the abnormal value of the historical vibration data sequence to set a threshold. If the abnormal value exceeds the set threshold, an abnormal warning is triggered, indicating the presence of a refrigerant leak.

[0014] Optionally, the slope calculation process is specifically as follows:

[0015] stretching or compressing the optimal sequence segment and the historical vibration data sequence so as to align the start time points of the current vibration data sequence and the optimal sequence segment;

[0016] Aligning the start time points of the optimal sequence segment and the historical vibration data sequence and mapping them in a two-dimensional coordinate system, wherein the abscissa is the acquisition interval and the ordinate is the vibration intensity;

[0017] Obtain corresponding vectors of the path between the optimal sequence segment and the matching point of the historical vibration data sequence to calculate the cosine value of the angle between each corresponding vector and the abscissa axis as the slope of the path between the optimal sequence segment and the matching point of the historical vibration data sequence.

[0018] The above scheme obtains the corresponding vectors of the path between the matching points of the optimal sequence segment and the historical vibration data sequence, and calculates the cosine value of the angle between each corresponding vector and the horizontal axis as the slope of the path between the optimal sequence segment and the matching points of the historical vibration data sequence, so as to obtain an accurate outlier value of the current vibration data sequence, thereby improving the accuracy of the outlier value calculation.

[0019] Optionally, the abnormal value is specifically:

[0020] ;

[0021] Among them, S is the abnormal value of the current vibration data sequence, F ia is the coefficient of variation of the path slope between matching points in the matching pair sequence consisting of the ath historical vibration data sequence of the i-th working condition and the corresponding optimal sequence segment, B ia is the shortest path distance between the optimal sequence segment corresponding to the a-th historical vibration data sequence and the a-th historical vibration data of the i-th working condition, a is the a-th vibration data of the i-th working condition in the historical vibration data sequence, q is the number of historical vibration data sequences obtained under the i-th working condition, is a normalization function, and exp( ) is an exponential function with the natural constant e as the base.

[0022] The above solution establishes the outlier value of the current vibration data sequence by performing weighted calculation on the shortest path distance using the coefficient of variation of the path slope, thereby improving the accuracy and reliability of outlier value calculation.

[0023] Optionally, the coefficient of variation of the path slope is specifically:

[0024] :

[0025] Among them, F ia is the coefficient of variation of the path slope between matching points in the matching pair sequence consisting of the a-th historical vibration data sequence of the i-th working condition and the corresponding optimal sequence segment; is the mean value of the path slope between the nth data point in the optimal sequence segment and the mth data point in the ath historical vibration data sequence of the i-th working condition; E mn is the path slope between the nth data point in the optimal sequence segment and the mth data point in the ath historical vibration data sequence of the ith working condition; N is the total number of matching points between the optimal sequence segment and the historical vibration data sequence.

[0026] Optionally, the matching value is specifically:

[0027] ;

[0028] Among them, C i is the matching degree between the current vibration data sequence and the historical vibration data sequence of the i-th working condition, B ia is the shortest path distance to the ath historical vibration data sequence of the i-th working condition, a is the ath vibration data of the i-th working condition in the historical vibration data sequence, q is the number of historical vibration data sequences obtained under the i-th working condition, and norm() is the normalization function.

[0029] The above solution improves the accuracy of matching between the historical vibration data sequence and the current vibration data sequence by calculating the matching degree value.

[0030] Optionally, the current vibration data sequence and the plurality of historical vibration data sequences under different working conditions are collected by sensors installed in designated areas of the refrigeration equipment.

[0031] Optionally, the set threshold of the abnormal value is 1.

[0032] Optionally, after triggering the abnormal warning, the steps include:

[0033] Record abnormal data and update the historical database for use in working condition matching and abnormal value calculation in subsequent inspections.

[0034] Optionally, after collecting the current vibration data sequence and the historical vibration data sequence, the process includes:

[0035] The current vibration data sequence and the historical vibration data sequence are preprocessed, outliers in the historical vibration data sequence are screened out using a box plot method, and missing data are filled using mean interpolation.

[0036] In a second aspect, the present invention provides a refrigerant filling equipment medium leakage warning system, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned refrigerant filling equipment medium leakage warning method when executing the computer program.

[0037] The beneficial effects of the present invention are:

[0038] The solution of the present invention can respectively calculate the shortest path distance between the current vibration data sequence and the historical vibration data sequence by using the DTW algorithm, match the current vibration data sequence to the corresponding working condition, solve the problem that the length of the current vibration data sequence cannot be determined under different working conditions, and match the current vibration data sequence to the corresponding working condition, thereby reducing the misjudgment rate in subsequent comparisons; and establish outliers in combination with the path slope change between DTW matching point pairs, taking into account the problem of inaccurate similarity calculation caused by the stretching of the time axis during the DTW matching process, thereby improving the accuracy and reliability of outlier calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flowchart schematically illustrates a method for early warning of medium leakage in a refrigerant filling device according to this embodiment;

[0040] Figure 2 The structural block diagram of a medium leakage warning system for refrigerant filling equipment in this embodiment is schematically shown. DETAILED DESCRIPTION

[0041] The present invention is directed to a medium leakage early warning scheme for refrigerant filling equipment. The medium leakage early warning scheme for refrigerant filling equipment needs to use the DTW algorithm to calculate the shortest path distance between the current vibration data sequence and the historical vibration data sequence respectively, and match the current vibration data sequence with the corresponding working condition and sequence length; combine the slope change of the path between the matching point pairs of the current vibration data sequence and the historical vibration data sequence obtained by the DTW algorithm to establish the abnormal value of the current vibration data sequence; set a threshold for the abnormal value, detect abnormal conditions and issue an early warning.

[0042] Therefore, based on the above content, the present invention provides a method and system for early warning of medium leakage of refrigerant filling equipment, so as to accurately provide early warning of refrigerant leakage.

[0043] Specifically, if Figure 1 As shown, a method for early warning of medium leakage of refrigerant filling equipment in this embodiment includes the following steps:

[0044] Step S1: obtaining a current vibration data sequence of the refrigeration equipment and a plurality of historical vibration data sequences under different working conditions.

[0045] Specifically, the vibration data sequence in the present invention includes a current vibration data sequence and a historical vibration data sequence. By installing a vibration sensor at the compressor, the vibration data during the operation of the equipment is pre-collected at a fixed collection interval to obtain the current vibration data sequence and the historical vibration data sequences of multiple working conditions.

[0046] The current vibration data sequence is preprocessed with multiple historical vibration data sequences of multiple working conditions, wherein the preprocessing is to use a box plot method to filter out abnormal values ​​in the historical vibration data sequence and fill in missing data with mean interpolation.

[0047] It should be noted that each working condition corresponds to multiple historical vibration data sequences.

[0048] In this embodiment, the current vibration data sequence and the historical vibration data sequence are collected and preprocessed so as to calculate the abnormal value using the preprocessed data.

[0049] Step S2: Taking the current moment as the starting point, extract sequence segments of different lengths from the current vibration data sequence; use the DTW algorithm to calculate the path distance between the sequence segments of each length and any historical vibration data sequence, and take the sequence segment corresponding to the shortest path distance as the optimal sequence segment. Calculate the matching degree between the current vibration data sequence and each working condition, and select the working condition corresponding to the maximum matching degree as the working condition to which the current vibration data sequence belongs.

[0050] Specifically, in order for the threshold selection to be able to more accurately judge abnormal situations, it is necessary to first match the current vibration data sequence when the equipment is running at the current moment with the corresponding working conditions. Under different working conditions, the length of the current vibration data sequence generated when the equipment is running is different, and the characteristics of the current vibration data sequence are also different under different vibration data sequence lengths. Therefore, it is necessary to divide the length of the collected current vibration data sequence according to the working condition matching results.

[0051] Specifically, the vibration data sequence of the compressor under different working conditions closest to the historical vibration data sequence is intercepted from the current vibration data sequence. Three vibration data sequences are intercepted for each working condition. There are a total of I working conditions, where the ath vibration data sequence of the i-th working condition is recorded as , assuming that the sequence length of the working condition to which the current vibration data sequence belongs is k, and the value range of k is 1 hour to 24 hours, it can be determined according to the actual situation, and the DTW algorithm is used to calculate the similarity between the historical vibration data sequence and the current vibration data sequence, traverse the value range of the historical vibration data sequence, select the best sequence segment and the shortest path distance from the best sequence segment to the historical vibration data sequence. Use the shortest path distance to calculate the matching degree of the historical vibration data sequence corresponding to the best sequence segment, and select the value with the smallest shortest path distance as the matching degree value. After obtaining the matching degree between the current vibration data sequence and the historical vibration data sequence of each working condition, select the working condition corresponding to the maximum matching degree as the working condition to which the current vibration data sequence belongs.

[0052] The matching degree is calculated as follows:

[0053] ;

[0054] Among them, C i is the matching degree between the current vibration data sequence and the historical vibration data sequence of the i-th working condition, B ia is the shortest path distance to the ath historical vibration data sequence of the i-th working condition, a is the ath vibration data of the i-th working condition in the historical vibration data sequence, q is the number of historical vibration data sequences obtained under the i-th working condition, and norm() is the normalization function.

[0055] In this embodiment, the DTW algorithm is used to match the current vibration data sequence with the historical vibration data sequence to match the most suitable working condition of the optimal sequence segment.

[0056] Step S3: establishing an abnormal value of the current vibration data sequence by analyzing the slope change of the matching point between the current vibration data sequence and the historical vibration data sequence.

[0057] Specifically, after determining that the current vibration data sequence belongs to the i-th working condition, it is necessary to analyze the abnormal values ​​of the current vibration data sequence by combining the matching point information of the three historical vibration data sequences with the current vibration data sequence.

[0058] In this embodiment, three historical vibration data sequences of the i-th working condition are obtained and paired with the matching points of the best sequence segment, wherein the sequence of matching points of the a-th historical vibration data sequence of the i-th working condition and the best sequence segment is , where C iam is the ath historical vibration data sequence A of the i-th working condition ia The mth data point in D kn is the nth data point in the optimal sequence segment of current length k. At this time, the optimal sequence segment length k is k ia In the DTW matching results, these two data points are matched with each other. Due to the different lengths of the two data sequences, the matching results of the DTW algorithm may have one-to-many or many-to-one situations, and there may also be repeated single data points in adjacent point pairs in the matching point sequence.

[0059] Therefore, it is necessary to calculate the slope of the path between the matching points of two vibration sequences to quantify the degree of stretching of the time axis during the matching process.

[0060] The calculation process of the slope of the matching point between the optimal sequence segment and the historical vibration data sequence is specifically as follows:

[0061] The optimal sequence segment and the historical vibration data sequence are stretched or compressed so that the start time points of the current vibration data sequence and the optimal sequence segment are aligned.

[0062] The optimal sequence segment and the starting time point of the historical vibration data sequence are aligned and then mapped into a two-dimensional coordinate system.

[0063] Obtain the path corresponding vectors between the optimal sequence segment and the matching points of the historical vibration data sequence, and calculate the cosine value of the angle between each corresponding vector and the horizontal axis as the slope of the path between the optimal sequence segment and the matching points of the historical vibration data sequence, wherein the matching points The slope of the path between them is E mn .

[0064] If there are anomalies in the optimal sequence segment, the path slope between matching points will change unevenly during the DTW matching process. The greater the slope change, the more time axis is distorted between the optimal sequence segment and the historical vibration data sequence during the matching process, which affects the calculation of the shortest path distance.

[0065] Therefore, the abnormal value of the current vibration data sequence can be quantified by using the change in the path slope between the matching points of the optimal sequence segment and the historical vibration data sequence in combination with the shortest path distance calculated in step S2.

[0066] Calculate the current vibration data sequence and the ath historical vibration data sequence of the i-th working condition The coefficient of variation of the path slope between matching points .

[0067] in, is the standard deviation of the path slope between the matching points of the two vibration data sequences, F ia is the coefficient of variation of the path slope between matching points in the matching pair sequence consisting of the a-th historical vibration data sequence of the i-th working condition and the corresponding optimal sequence segment; is the mean value of the path slope between the nth data point in the optimal sequence segment and the mth data point in the ath historical vibration data sequence of the i-th working condition; E mn is the path slope between the nth data point in the optimal sequence segment and the mth data point in the ath historical vibration data sequence of the ith working condition; N is the total number of matching points between the optimal sequence segment and the historical vibration data sequence.

[0068] The coefficient of variation of the path slope between matching points reflects the consistency of the degree of time axis stretching of the two vibration data sequences during the matching process. If the coefficient of variation is small, it indicates that the time axis is not over-stretched during the matching process. In addition to the sequence length, the vibration information contained in the two data sequences is highly similar, and the current vibration data sequence is likely to be free of anomalies. Conversely, if the coefficient of variation is large, it indicates that the degree of time axis stretching between different matching points is different during the matching process, the difference between the two data sequences is large, and the current vibration data sequence may be anomaly.

[0069] The coefficient of variation of the path slope of the matching point is used to perform weighted calculation on the shortest path distance calculated in step S2 to obtain the abnormal value of the current vibration data sequence:

[0070] ;

[0071] Among them, S is the abnormal value of the current vibration data sequence, F ia is the coefficient of variation of the path slope between matching points in the matching pair sequence consisting of the ath historical vibration data sequence of the i-th working condition and the corresponding optimal sequence segment, B ia is the shortest path distance between the optimal sequence segment corresponding to the a-th historical vibration data sequence and the a-th historical vibration data of the i-th working condition, a is the a-th vibration data of the i-th working condition in the historical vibration data sequence, is a normalization function, and exp( ) is an exponential function with the natural constant e as the base.

[0072] B ia Reflects the similarity between the best sequence segment and the historical vibration data sequence, B ia The smaller the value, the higher the similarity. Using the norm function to normalize it can map the outliers to a smaller range, which is convenient for setting the threshold later.

[0073] In this embodiment, the shortest path distance is weightedly calculated by using the coefficient of variation of the path slope to establish the abnormal value of the current vibration data sequence, so as to improve the accuracy and reliability of the dynamic early warning.

[0074] Step S4: dynamically set a threshold value using the distribution of the abnormal value and the abnormal value of the historical vibration data sequence. If the abnormal value exceeds the set threshold value, an abnormal warning is triggered, indicating that there is a refrigerant leak.

[0075] Specifically, by analyzing the slope change of the matching point between the optimal sequence segment and the historical vibration data sequence, the abnormal value of the current vibration data sequence is established, and a threshold is set for the current vibration data sequence. If the abnormal value of the current vibration data sequence exceeds the preset threshold, an abnormal refrigerant leakage occurs during the operation of the current equipment, and an early warning needs to be issued in time so that the staff can carry out subsequent processing to avoid greater losses.

[0076] In this embodiment, by utilizing the distribution of abnormal values, a threshold is dynamically set. If the abnormal value exceeds the set threshold, an abnormal warning is triggered, indicating that there is a refrigerant leak, thereby improving the accuracy of the refrigerant leak warning.

[0077] Furthermore, when leak tests are performed at different locations, when a leak is detected, a corresponding light indication alarm is given to indicate the presence of a leak.

[0078] In the present invention, by obtaining the current vibration data sequence of the refrigeration equipment and multiple historical vibration data sequences under different working conditions, taking the current moment as the starting point, sequence segments of different lengths are intercepted from the current vibration data sequence, and the path distance between the sequence segments of each length and any historical vibration data sequence is calculated using the DTW algorithm. The sequence segment corresponding to the shortest path distance is taken as the optimal sequence segment, and the matching degree between the current vibration data sequence and each working condition is calculated. A matching pair sequence of any historical vibration data sequence of the corresponding working condition when the matching degree is maximum and the corresponding optimal sequence segment is obtained. The matching pair is composed of vibration data in any historical vibration data sequence and vibration data in the optimal sequence segment. The abnormal value of the current vibration data sequence is calculated. The abnormal value is positively correlated with the coefficient of variation of each matching pair sequence and the shortest path distance. The coefficient of variation is obtained based on the slope of the matching pair in each matching pair sequence. The slope is the slope of the line connecting the two data in the matching pair in the coordinate system composed of time and vibration data. If the abnormal value exceeds a set threshold, an abnormal warning is triggered, indicating the presence of a refrigerant leak. The present invention dynamically sets a threshold value. If the abnormal value exceeds the set threshold value, an abnormal warning is triggered to indicate the presence of a refrigerant leak, thereby reducing the false positive rate and the missed positive rate.

[0079] The present invention also provides a refrigerant filling equipment medium leakage warning system. Figure 2 As shown, the system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a method for warning of medium leakage of a refrigerant filling device according to the present invention is implemented.

[0080] The system further includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and thus will not be described in detail here.

[0081] In the present invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, the computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium can be part of, accessible to, or connectable to a device. Any application or module described in the present invention can be implemented by computer-readable / executable instructions stored or otherwise retained by such a computer-readable medium.

Claims

1. A method for early warning of medium leakage in refrigerant filling equipment, characterized in that: include: Obtain the current vibration data sequence of the refrigeration equipment and multiple historical vibration data sequences under different working conditions; Taking the current moment as the starting point, sequence segments of different lengths are intercepted from the current vibration data sequence. The DTW algorithm is used to calculate the path distance between the sequence segments of each length and any historical vibration data sequence, and the sequence segment corresponding to the shortest path distance is taken as the optimal sequence segment. Calculating the matching degree between the current vibration data sequence and each working condition, wherein the matching degree is inversely correlated with the shortest path distance of each historical vibration data sequence under the corresponding working condition; Obtaining a matching pair sequence of any historical vibration data sequence and a corresponding optimal sequence segment of the working condition corresponding to the maximum matching degree, wherein the matching pair is composed of vibration data in any historical vibration data sequence and vibration data in the optimal sequence segment; Calculate the abnormal value of the current vibration data sequence, specifically: , S is the abnormal value of the current vibration data sequence, F ia is the coefficient of variation of the path slope between matching pairs in the matching pair sequence composed of the a-th historical vibration data sequence of the i-th working condition and the corresponding optimal sequence segment, B ia is the shortest path distance between the optimal sequence segment corresponding to the a-th historical vibration data sequence and the a-th historical vibration data of the i-th working condition, a is the a-th vibration data of the i-th working condition in the historical vibration data sequence, q is the number of historical vibration data sequences obtained under the i-th working condition, norm() is a normalization function, and exp() is an exponential function with the natural constant e as the base; the coefficient of variation of the outlier and each matching pair sequence is positively correlated with the shortest path distance; the coefficient of variation is obtained based on the slope of the matching pair in each matching pair sequence; the slope is the slope of the line connecting the two data in the matching pair in the coordinate system formed by time and vibration data; If the abnormal value exceeds the set threshold, an abnormal warning is triggered, indicating that there is a refrigerant leak.

2. A refrigerant filling equipment medium leakage early warning method according to claim 1, characterized in that: The calculation process of the slope is specifically as follows: stretching or compressing the optimal sequence segment and the historical vibration data sequence so as to align the start time points of the historical vibration data sequence and the optimal sequence segment; Aligning the start time points of the optimal sequence segment and the historical vibration data sequence and mapping them in a two-dimensional coordinate system, wherein the abscissa is the acquisition interval and the ordinate is the vibration intensity; Obtain corresponding vectors of the path between the matching pair of the optimal sequence segment and the historical vibration data sequence, and calculate the cosine value of the angle between each corresponding vector and the abscissa axis as the slope of the path between the matching pair of the optimal sequence segment and the historical vibration data sequence.

3. The method for early warning of medium leakage of refrigerant filling equipment according to claim 1, characterized in that: The coefficient of variation of the path slope is specifically: ； Among them, F ia is the coefficient of variation of the path slope between matching pairs in the matching pair sequence consisting of the a-th historical vibration data sequence of the i-th working condition and the corresponding optimal sequence segment; is the mean value of the path slope between the nth data point in the optimal sequence segment and the mth data point in the ath historical vibration data sequence of the i-th working condition; E mn is the path slope between the nth data point in the optimal sequence segment and the mth data point in the ath historical vibration data sequence of the i-th working condition; N is the total number of matching pairs of the optimal sequence segment and the historical vibration data sequence.

4. The method for early warning of medium leakage of refrigerant filling equipment according to claim 1, characterized in that: The matching degree is specifically: ； Among them, C i is the matching degree between the current vibration data sequence and the historical vibration data sequence of the i-th working condition, B ia is the shortest path distance to the ath historical vibration data sequence of the i-th working condition, a is the ath vibration data of the i-th working condition in the historical vibration data sequence, q is the number of historical vibration data sequences obtained under the i-th working condition, and norm() is the normalization function.

5. The method for early warning of medium leakage of refrigerant filling equipment according to claim 1, characterized in that: The current vibration data sequence and the plurality of historical vibration data sequences under different working conditions are collected by sensors installed in designated areas of the refrigeration equipment.

6. A refrigerant filling equipment medium leakage early warning method according to claim 1, characterized in that: The threshold value of the abnormal value is set to 1.

7. The method for early warning of medium leakage of refrigerant filling equipment according to claim 1, characterized in that: After a refrigerant leak is reported, the following actions are performed: Record abnormal data and update the historical database for use in working condition matching and abnormal value calculation in subsequent inspections.

8. The method for early warning of medium leakage of refrigerant filling equipment according to claim 1, characterized in that: After obtaining the current vibration data sequence and the historical vibration data sequence, including: The current vibration data sequence and the historical vibration data sequence are preprocessed, outliers in the historical vibration data sequence are screened out using a box plot method, and missing data are filled using mean interpolation.

9. A refrigerant filling equipment medium leakage early warning system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, a medium leakage early warning method for refrigerant filling equipment according to any one of claims 1 to 8 is implemented.

Images