Abnormality detection method and system for refrigeration valve

Abstract

The invention relates to the technical field of data processing, in particular to an anomaly detection method and system for a refrigeration valve, and the method comprises the steps: obtaining the preprocessed operation data of the refrigeration valve, and for any data point in the operation data, calculating the sparse degree of the data point according to the distribution state of the data point in a local region; meanwhile, constructing a time sequence adjacent reference set of the data points, and obtaining an initial abnormal score of each data point in the adjacent reference set by using an LOF algorithm; constructing a trend anomaly degree according to the initial anomaly score of the data point in the time sequence adjacent reference set and the change of the data value; a preset initial abnormal threshold value is adjusted based on the trend abnormal degree and the sparse degree to obtain a dynamic threshold value, the dynamic threshold value is in negative correlation with the trend abnormal degree, and the operation data of the refrigeration valve are judged based on the dynamic threshold value. The method has the effect of improving the abnormality detection accuracy of the refrigeration valve.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and system for detecting abnormalities in refrigeration valves. Background Art

[0002] Refrigeration valves are an essential and common component in refrigeration systems, used to control and regulate fluid flow. Common refrigeration valves include shut-off valves, throttles, and safety valves. To ensure stable refrigeration system operation, refrigeration valve monitoring is often necessary. Traditional manual inspections have significant lag and are unable to detect system problems promptly. Currently, with the advancement of sensor and data processing technologies, machine learning-based methods are increasingly being used to monitor industrial refrigeration systems. For example, the LOF (Local Outlier Factor) algorithm compares the density of a data point with that of its neighbors to identify "outliers" where the local density deviates significantly. This method effectively mines hidden anomalies under unsupervised conditions and is suitable for multi-dimensional, nonlinear refrigeration system data modeling.

[0003] While traditional LOF algorithms have been effective in detecting anomalies in industrial data, they also suffer from some unavoidable flaws. For example, when the operating conditions of a refrigeration system change, industrial data can shift, creating areas with sparse data points. During LOF detection, these data points are typically identified as abnormal. However, these changes due to operating conditions are normal system performance and not true anomalies. Consequently, traditional LOF algorithms exhibit low accuracy. Summary of the Invention

[0004] In order to improve the accuracy of refrigeration valve operation data monitoring, the present application provides a method and system for detecting anomalies of refrigeration valves.

[0005] In a first aspect, the present application provides a method for detecting abnormalities in a refrigeration valve, which adopts the following technical solution: A method for detecting anomalies in refrigeration valves comprises: obtaining preprocessed operating data of the refrigeration valve, and for any data point in the operating data, calculating the sparsity of the data point based on the distribution state in the local area where the data point is located; simultaneously, constructing a temporal neighboring reference set of the data points, and using a LOF algorithm to obtain an initial anomaly score for each data point in the neighboring reference set; constructing a trend anomaly degree based on the initial anomaly scores of the data points in the temporal neighboring reference set and changes in data values; and adjusting a preset initial anomaly threshold based on the trend anomaly degree and the sparsity degree to obtain a dynamic threshold, wherein the dynamic threshold is negatively correlated with the trend anomaly degree, and judging the refrigeration valve operating data based on the dynamic threshold.

[0006] The beneficial effects are as follows: First, by using the spatial proximity reference set to calculate the sparsity, the density differences of data points in spatial distribution can be effectively distinguished, avoiding the misjudgment of sparse areas caused by changes in working conditions due to relying solely on local density, thereby reducing false alarms caused by working condition migration. Secondly, data trend analysis based on the temporal proximity reference set and the LOF initial anomaly score is introduced to calculate the trend anomaly degree, so that the detection not only considers the spatial distribution characteristics of a single moment, but also comprehensively reflects the changing trend of the data in the time dimension, and can identify abnormal patterns of slow offset or mutation. Furthermore, by constructing a dynamic threshold that is negatively correlated with the trend anomaly degree, the threshold is increased to suppress false alarms in the case of sparse but stable time series, and the threshold is lowered to provide early warning when the time series changes significantly, taking into account both missed alarms and false alarms. This dynamic threshold adjustment mechanism overcomes the poor adaptability problem caused by the static threshold of the traditional LOF algorithm, and realizes adaptive detection of anomalies under different operating conditions.

[0007] Optionally, the step of calculating the sparsity of the data point based on the distribution state of the data point in the local area where the data point is located includes: constructing a spatial adjacent reference set; obtaining the data difference between each data point in the spatial adjacent reference set and the current data point, and taking the average of the data differences corresponding to multiple data points as the sparsity of the pixel point.

[0008] The beneficial effect is: constructing a spatially adjacent reference set, then calculating the data difference between the current data point and taking the mean as the sparsity level, which can accurately reflect the local density characteristics of the data points at the spatial distribution level.

[0009] Optionally, the step of obtaining the data difference between each data point in the spatially adjacent reference set and the current data point includes: calculating the value difference based on the difference between each data point in the spatially adjacent reference set and each data point; calculating the direction difference based on the cosine angle between each data point in the spatially adjacent reference set and the current data point; and taking the product of the value difference and the direction difference as the data difference.

[0010] The beneficial effects are: the value difference directly measures the degree of deviation in amplitude between the current data point and the data points in the spatially adjacent reference set in various physical quantities (such as pressure, temperature, flow, etc.), and can reflect the numerical changes of operating parameters; secondly, the direction difference measures the distribution direction consistency of data points in the multidimensional feature space through the cosine angle. When the cosine angle is large, it indicates that the data distribution trend difference is significant, which can reveal structural changes that cannot be discovered from the numerical difference alone.

[0011] Optionally, the step of constructing a spatially adjacent reference set includes: for any data point, obtaining a preset number of data points that are adjacent to the data point in spatial distribution as preliminary data points, and taking the average of the Euclidean distances between the preliminary data points and the current data point as the adjacent distance; and taking the set of data points whose Euclidean distances to the current data point are less than the adjacent distance as the spatially adjacent reference set.

[0012] The beneficial effects are: selecting a preset number of preliminary data points that are spatially close to the current data point, ensuring that these candidate points have a high degree of similarity in the multidimensional feature space, providing a reasonable comparison basis for subsequent sparsity and data discrepancy calculations. Secondly, using the mean of the Euclidean distance between the preliminary data points and the current data point as the proximity distance threshold, and further filtering out data points with excessive distances, can eliminate outliers in the preliminary set caused by occasional noise, extreme fluctuations, or sudden changes in operating conditions, thereby reducing the bias of local density estimates.

[0013] Optionally, the step of calculating the value difference based on the difference between the data of the data points in the spatially adjacent reference set and the data of the current data point includes: for any data of any data point, calculating the absolute value of the difference between the current data point and the data of the data point; taking the normalized result of the absolute value as the standard difference, and taking the mean of the standard differences of each data as the value difference.

[0014] Optionally, the step of constructing the degree of trend anomaly based on the initial anomaly scores of data points in the time series neighboring reference set and the changes in data values ​​includes: any two adjacent data points in the time series neighboring reference set constitute a time series reference point pair, and a local score difference is constructed based on the difference in the initial anomaly scores in the time series reference point pair; a local data change is constructed based on the difference in the data of the two data points in the time series neighboring reference set; and a trend anomaly degree is constructed based on the local data change and the local score difference.

[0015] Construct multiple time series reference point pairs, compare the initial anomaly score difference and local data change between two adjacent data points, and then combine multiple local score differences and local data changes to reflect the overall trend anomaly degree.

[0016] Optionally, the sum of local data changes is taken as the overall data change, the sum of multiple local score differences is taken as the overall score difference, the product of the overall score difference and the overall data change is taken as the anomaly index, and the normalized result of the anomaly index is taken as the trend anomaly degree.

[0017] Optionally, the step of obtaining a dynamic threshold by adjusting the preset initial abnormality threshold based on the trend abnormality degree and the sparsity degree includes: taking the ratio of the sparsity degree of the data point to the trend abnormality degree as the first control index, and taking the average of the first control index corresponding to each data point as the second control index; taking the ratio of the normalized result of the first control index to the second control index as the adjustment factor, and taking the product of the adjustment factor and the initial abnormality threshold as the dynamic threshold.

[0018] In a second aspect, the present application provides an abnormality detection system for a refrigeration valve, which adopts the following technical solution: A system for detecting an abnormality of a refrigeration valve comprises a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the system implements the above-mentioned method for detecting an abnormality of a refrigeration valve.

[0019] The beneficial effect is that the above-mentioned abnormality detection method for refrigeration valves is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a system based on the memory and the processor for easy use.

[0020] This application has the following technical effects: In this application, the sparsity degree is constructed based on the spatial distribution of data points, and the trend anomaly index is constructed based on the changes in the initial anomaly scores of data points in time series and the changes in data. The trend anomaly degree can reflect the degree of anomaly of data changes in time series, thereby reducing the situation of false alarms of data anomalies caused by working condition migration. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a method flow chart of a method for detecting abnormality of a refrigeration valve in an embodiment of the present application.

[0022] Figure 2 This is a method flow chart of step S2 in a method for detecting an abnormality of a refrigeration valve in an embodiment of the present application. DETAILED DESCRIPTION

[0023] The present application discloses an anomaly detection method for a refrigeration valve. The method constructs a sparsity index reflecting the density of the local area where the current data point is located based on the spatial position of the data point. Simultaneously, a temporal proximity reference set adjacent to the current data point is constructed. By analyzing the initial anomaly score and data changes in the temporal proximity reference set, a trend anomaly index reflecting the degree of anomaly of the current data point in the temporal sequence is constructed. Finally, the initial anomaly threshold is adjusted based on the two to obtain a dynamic threshold. Based on the dynamic threshold, the state of the current data point is determined, and the operating state of the refrigeration valve is then determined.

[0024] Reference Figure 1, a method for detecting an abnormality of a refrigeration valve, comprising steps S1 to S3.

[0025] S1: Acquire pre-processed operating data of the refrigeration valve.

[0026] In order to achieve comprehensive perception of the operating status of the refrigeration valve and accurate identification of abnormal behavior, it is necessary to collect multiple key operating parameters of the refrigeration valve in real time during the normal operation of the refrigeration system. Specific parameters collected may include: the pressure value at the front end of the refrigeration valve, the pressure value at the back end, the valve body temperature, the valve drive current, and the flow rate through the valve body. Among them, the pressure before the valve and the pressure after the valve are used to calculate the pressure difference to reflect the system load, the valve body temperature and drive current can reflect the operating intensity and heat dissipation status of the valve, and the flow rate directly reflects the refrigerant passage efficiency. These parameters are collected in real time through the deployed industrial-grade pressure sensors, temperature sensors, current transformers, and flow meters. The sensors have a measurement accuracy of more than 0.1 and are connected to the data acquisition module via RS485 or CAN (Controller Area Network) bus.

[0027] The frequency of data collection is determined by the speed of changes in the refrigeration valve operating conditions, typically set to every 10 seconds to cover common periodic regulation behaviors. For each data point generated during the collection process, it includes the aforementioned multidimensional data, resulting in raw data.

[0028] To improve the accuracy and robustness of subsequent anomaly detection and reduce noise interference, the raw data is preprocessed. This preprocessing includes missing value filling and normalization. These operations are conventional techniques in the field and will not be detailed here. The preprocessed raw data is used as the running data and provides input for subsequent steps.

[0029] S2: For any data point in the running data, the sparsity of the data point is calculated based on the distribution state of the local area where the data point is located.

[0030] For any data point, whether the local area where the data point is located belongs to a sparse area is determined based on its spatial position, so as to further judge whether the data point is truly abnormal data.

[0031] Reference Figure 2 , step S2 includes steps S21 to S23.

[0032] S21: Constructing a spatial proximity reference set.

[0033] In the process of constructing the spatially adjacent reference set, first, for any data point, a preset number of data points that are adjacent to the data point in spatial distribution are obtained as preliminary data points.

[0034] When constructing a proximity set for any data point, a preset number of data points should be obtained as preliminary data points based on the distance between the data point and other data points. In this embodiment, the preset number of preliminary data points is 30. In other embodiments, those skilled in the art may adjust the preset number based on experience and actual production conditions. For example, the preset number may be set to 40 or 50.

[0035] Then, the mean of the Euclidean distances between the preliminary data points and the current data point is taken as the proximity distance, and the set of data points whose Euclidean distances to the current data point are less than the proximity distance is taken as the spatial proximity reference set.

[0036] This is primarily used to screen the initial data points to reduce the introduction of outliers in the spatially adjacent reference set, improving the accuracy and robustness of subsequent sparsity calculations. For example, suppose one or two of the initial data points have a Euclidean distance of 30 from the current data point, while the other data points have Euclidean distances between 10 and 15. This step can eliminate data points with Euclidean distances greater than 30, thereby improving the accuracy of extreme sparsity calculations.

[0037] S22: Obtain the data difference between each data point in the spatially adjacent reference set and the current data point, and use the average of the data differences corresponding to the multiple data points as the sparsity degree of the pixel point.

[0038] In one embodiment, the step of obtaining the data difference between each data point in the spatially adjacent reference set and the current data point includes: calculating the value difference based on the difference between each data point in the spatially adjacent reference set and each data point; calculating the direction difference based on the cosine angle between each data point in the spatially adjacent reference set and the current data point; and taking the product of the value difference and the direction difference as the data difference.

[0039] The step of calculating the value difference based on the difference between the data of the data points in the spatially adjacent reference set and the data of the current data point includes: for any data of any data point, calculating the absolute value of the difference between the current data point and the data of the data point; taking the normalized result of the absolute value as the standard difference, and taking the mean of the standard differences of each data as the value difference.

[0040] Directional differences are calculated based on the cosine angles of the data points of each data point in the spatially adjacent reference set and the data points of the current data point. The cosine angle between two data points indicates the directional consistency of the vectors corresponding to the two data points. A larger cosine angle indicates that the two data points are more spatially dispersed and irregular, indicating that the current data point may be in a sparse area.

[0041] The data difference between the current data point and the data points in the spatially adjacent reference set is calculated by combining the value difference and direction difference to improve the accuracy and robustness of the sparsity calculation.

[0042] Based on the above steps, the data difference between each pixel in the spatially adjacent reference set and the current data point is calculated, and the mean of the data differences is used as the sparsity of the current data point.

[0043] Specifically, the calculation formula of sparsity can be expressed as: Where, Represents a data point The sparsity of the neighborhood data points, Represents a data point The number of data points in the corresponding spatially adjacent reference set, Represents a data point No. Item data value, Represents a data point The first in the spatial proximity reference set The data point Item data value, Represents a data point The corresponding spatial proximity of all data points in the reference set The standard deviation of the data value, which is mainly used to standardize the numerator. Represents a data point The first reference set in the corresponding space The minimum cosine angle value of the data points; Indicates the dimension of the data contained in the data point. Taking this application as an example, the data points in this application include the pressure value at the front end of the refrigeration valve, the pressure value at the back end, the valve body temperature, the valve drive current, and the flow rate through the valve body. Then is 5.

[0044] This part reflects the difference between one of the data points in the current data point and one of the data points in the corresponding spatially adjacent reference set. The differences of multiple data points corresponding to the data point are summed and averaged to obtain the overall value difference of each data point between the current data point and the data point.

[0045] Indicates the minimum cosine angle between the current data point and the corresponding data point, representing the difference in the distribution directions of the two data points.

[0046] This section reflects the data difference between two data points, taking into account both the difference in data value and the difference in distribution direction. The data differences are summed and averaged to reflect the overall distance between the current data point and the data points in the spatially adjacent reference set. A larger distance indicates a greater sparsity at the current data point's location, and vice versa.

[0047] S3: Construct a temporal neighborhood reference set of data points and use the LOF algorithm to obtain the initial anomaly score of each data point in the neighborhood reference set; construct the trend anomaly degree based on the initial anomaly score of the data points in the temporal neighborhood reference set and the change of data values.

[0048] The sparsity can be used to obtain the density of data points distributed in space. Data points with a large sparsity may be true abnormal data points, but they may also be normal data migration caused by changes in working conditions.

[0049] Therefore, in this embodiment, a temporal neighboring reference set of data points is constructed.

[0050] In the step of constructing a temporally proximate reference set of data points, for any data point, a preset number of data points preceding the data point's collection time are obtained based on the data point's collection time to form the temporally proximate reference set. The data points in the temporally proximate reference set are close to the collection time of the current data point. In this embodiment, the preset collection number is 30. This number can be determined based on actual production conditions and the experience of those skilled in the art.

[0051] After the time series adjacent reference set is constructed, the LOF algorithm is used to obtain the initial anomaly score of each data point in the adjacent reference set, and the trend anomaly degree is constructed based on the initial anomaly score of the data point in the time series adjacent reference set and the change of data values.

[0052] In the process of constructing the trend anomaly degree, any two adjacent data points in the time series neighborhood reference set form a time series reference point pair. The local score difference is constructed based on the difference in the initial anomaly scores of the time series reference point pair. The local data variation is constructed based on the difference in the data items of the two data points in the time series neighborhood reference set. The trend anomaly degree is constructed based on the local data variation and the local score difference. The sum of the local data variation is used as the overall data variation, the sum of multiple local score differences is used as the overall score difference, the product of the overall score difference and the overall data variation is used as the anomaly index, and the normalized result of the anomaly index is used as the trend anomaly degree.

[0053] Specifically, the calculation formula for the degree of trend anomaly can be expressed as: Where, Represents pixel points The degree of abnormality of the trend; Represents the number of data points in the temporal neighborhood reference set; Represents a data point The corresponding time sequence is close to the reference set The initial anomaly score of the data point; Represents a data point The corresponding time sequence is close to the reference set The initial anomaly score of the data point; Represents a data point The corresponding time sequence is close to the reference set The first pixel Item data; Represents a data point The corresponding time sequence is close to the reference set The first pixel Item data; Indicates the first of all data points in the time series close to the reference set The mean of the data items is mainly used to standardize the molecular part; Indicates the number of data points in the time series adjacent to the reference set; Indicates the dimension in which the data point contains the data.

[0054] The first data points and The data points are two adjacent data points, forming a time series reference point pair. Represents the local score difference between two data points in a temporal reference point pair; This section calculates the total difference in anomaly scores for all time series reference points, that is, the overall score difference. The larger the value, the greater the anomaly score of each data point in the time series, and thus the more likely the current data point is a true anomaly.

[0055] This represents the difference between a particular data item between the later and earlier data points in a time series reference point pair. The local data delta corresponding to each data item is calculated by averaging the differences. The overall data delta is then calculated by summing the local deltas for each time series reference point pair.

[0056] This part represents the abnormality index; This part is mainly used to limit the value range of the abnormal index to avoid the calculated result of the abnormal index being too large.

[0057] S4: adjusting a preset initial abnormality threshold based on the trend abnormality degree and the sparsity degree to obtain a dynamic threshold, wherein the dynamic threshold is negatively correlated with the trend abnormality degree, and judging the refrigeration valve operation data based on the dynamic threshold.

[0058] In the process of obtaining the dynamic threshold, the ratio of the sparsity of the data points to the trend anomaly degree is used as the first control index, and the mean of the first control index corresponding to each data point is used as the second control index; the ratio of the normalized result of the first control index to the second control index is used as the adjustment factor, and the product of the adjustment factor and the initial anomaly threshold is used as the dynamic threshold.

[0059] Specifically, the calculation formula of the dynamic threshold can be expressed as: Where, Represents a data point The corresponding regulatory factors; Represents a data point The sparsity of Represents a data point The degree of abnormality of the trend; It represents the mean of the ratio of the sparsity degree of all data points in the entire running data to the trend anomaly degree.

[0060] In this formula, when the sparsity is large but the trend anomaly is small, it means that although the current data point is in an area with sparse data density, there is no major anomaly in the time series, which may be due to data sparsity caused by working condition migration; therefore, for this type of situation, the adjustment factor is greater than 1, which causes the dynamic threshold to become larger to avoid false alarms.

[0061] At the same time, if the trend of a data point is highly abnormal, it may indicate a true data anomaly. This situation will result in a smaller numerator and an adjustment factor less than 1, which in turn will cause the dynamic threshold to become smaller, allowing the system to issue an early alarm and reduce missed reports.

[0062] In addition, during the normal operation of the refrigeration valve, the sparsity is relatively small. Once the data shows a slow abnormality in the time series, it will still cause the dynamic threshold to become smaller, and an alarm will be issued in advance, reducing the situation where the traditional LOF algorithm cannot detect due to slow data offset, thereby improving the accuracy and robustness of system detection.

[0063] An embodiment of the present application also discloses an abnormality detection system for a refrigeration valve, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, an abnormality detection method for a refrigeration valve according to the present application is implemented.

[0064] The above system also 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 will not be described in detail here.

[0065] The above are all preferred embodiments of the present application, and are not intended to limit the scope of protection of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A method for detecting abnormality of a refrigeration valve, characterized in that: Obtaining pre-processed operating data of the refrigeration valve, and for any data point in the operating data, calculating the sparsity of the data point based on the distribution state of the data point in the local area; At the same time, a temporal neighboring reference set of data points is constructed, and the LOF algorithm is used to obtain the initial anomaly score of each data point in the neighboring reference set; Construct the trend anomaly degree based on the initial anomaly scores of the data points in the time series neighboring reference set and the changes in the data values; A dynamic threshold is obtained by adjusting a preset initial abnormality threshold based on the trend abnormality degree and the sparseness degree, wherein the dynamic threshold is negatively correlated with the trend abnormality degree, and the refrigeration valve operation data is judged based on the dynamic threshold.

2. The abnormality detection method for a refrigeration valve according to claim 1, characterized in that: The steps of calculating the sparsity of a data point based on the distribution state of the data point in the local area where the data point is located include: constructing a spatial adjacent reference set; obtaining the data difference between each data point in the spatial adjacent reference set and the current data point, and taking the average of the data differences corresponding to multiple data points as the sparsity of the pixel point.

3. The abnormality detection method for a refrigeration valve according to claim 2, characterized in that: The steps of obtaining the data difference between each data point in the spatially adjacent reference set and the current data point include: calculating the value difference based on the difference between each data point in the spatially adjacent reference set and each data point; calculating the direction difference based on the cosine angle between each data point in the spatially adjacent reference set and the current data point; and taking the product of the value difference and the direction difference as the data difference.

4. The abnormality detection method for a refrigeration valve according to claim 2, characterized in that: The steps of constructing a spatial proximity reference set include: for any data point, obtaining a preset number of data points that are close to the data point in spatial distribution as preliminary data points, and taking the average of the Euclidean distances between the preliminary data points and the current data point as the proximity distance; and taking the set of data points whose Euclidean distances to the current data point are less than the proximity distance as the spatial proximity reference set.

5. The abnormality detection method for a refrigeration valve according to claim 1, characterized in that: The step of calculating the value difference based on the difference between the data of the data points in the spatially adjacent reference set and the data of the current data point includes: for any data of any data point, calculating the absolute value of the difference between the current data point and the data of the data point; taking the normalized result of the absolute value as the standard difference, and taking the mean of the standard differences of each data as the value difference.

6. The abnormality detection method for a refrigeration valve according to claim 1, characterized in that: The steps of constructing the degree of trend anomaly based on the initial anomaly scores of data points in the time series neighboring reference set and the changes in data values ​​include: any two adjacent data points in the time series neighboring reference set constitute a time series reference point pair, and a local score difference is constructed based on the difference in the initial anomaly scores in the time series reference point pair; a local data change is constructed based on the difference in the data of the two data points in the time series neighboring reference set; and a trend anomaly degree is constructed based on the local data change and the local score difference.

7. The abnormality detection method for a refrigeration valve according to claim 1, characterized in that: The sum of local data changes is taken as the overall data change, the sum of multiple local score differences is taken as the overall score difference, the product of the overall score difference and the overall data change is taken as the anomaly index, and the normalized result of the anomaly index is taken as the trend anomaly degree.

8. The abnormality detection method for a refrigeration valve according to claim 1, characterized in that: The steps of adjusting the preset initial abnormality threshold based on the trend abnormality degree and the sparsity degree to obtain the dynamic threshold include: taking the ratio of the sparsity degree of the data point to the trend abnormality degree as the first control index, and taking the average of the first control index corresponding to each data point as the second control index; taking the ratio of the normalized result of the first control index to the second control index as the adjustment factor, and taking the product of the adjustment factor and the initial abnormality threshold as the dynamic threshold.

9. An abnormality detection system for a refrigeration valve, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an abnormality detection method for a refrigeration valve according to any one of claims 1 to 8 is implemented.

Images