Radiator working state monitoring method and system based on data analysis

Abstract

The invention relates to the technical field of electric digital data processing, in particular to a radiator working state monitoring method and system based on data analysis, and the method comprises the steps: collecting multi-dimensional parameter data of a radiator at different moments, and taking the multi-dimensional parameter data of the radiator at the same moment as a data point; clustering all the data points to obtain a plurality of clusters; the abnormal degree of any cluster is calculated, the abnormal degree is in inverse correlation with the number of the data points in each cluster and is in positive correlation with the standard deviation of the data points in each cluster, the data points in the cluster with the abnormal degree larger than a set threshold value are responded to be abnormal data points, and it is judged that the radiator is abnormal at the moment corresponding to the abnormal data points. The problem that the monitoring efficiency and accuracy are not high in the process of monitoring the abnormality of the radiator is solved.

Description

Technical Field

[0001] This invention relates to the field of electronic digital data processing technology. More specifically, this invention relates to a method and system for monitoring the operating status of radiators based on data analysis. Background Technology

[0002] In industrial production and equipment operation, radiators, as critical heat dissipation components, directly affect equipment performance and lifespan. However, traditional radiator condition monitoring methods mainly rely on periodic manual inspections or simple temperature sensor monitoring, which suffers from problems such as untimely monitoring, insufficient accuracy, and difficulty in achieving intelligent analysis, making it difficult to meet the demands of modern industry for efficient and intelligent operation and maintenance.

[0003] With the development of data analysis and intelligent monitoring technologies, real-time monitoring and intelligent diagnosis of radiator operating status can be achieved through multi-dimensional data acquisition, analysis, and modeling. By combining multi-dimensional data such as temperature distribution, heat transfer efficiency, environmental parameters, and radiator workload, and utilizing algorithms such as machine learning, anomaly detection, and trend prediction, the operating status of radiators can be effectively identified, potential faults can be predicted, and thus the operational reliability and maintenance efficiency of the equipment can be improved. Therefore, there is an urgent need for a data analysis-based method and system for monitoring the operating status of radiators to achieve intelligent and precise monitoring of radiators, thereby improving equipment safety and operational efficiency.

[0004] Existing methods for monitoring the operating status of radiators mainly rely on manual inspections, temperature sensor monitoring, or alarm systems based on fixed thresholds. However, these methods suffer from limitations such as limited monitoring capabilities, insufficient data dimensions, delayed response, difficulty in accurately identifying the causes of anomalies, lack of intelligent analysis, and high maintenance costs. Because they only focus on surface temperature or local temperature distribution, existing methods fail to comprehensively consider multi-dimensional factors such as heat dissipation efficiency, heat conduction paths, and ambient temperature and humidity, resulting in limited monitoring accuracy. Furthermore, they do not fully account for the impact of complex environments on monitoring results, leading to low monitoring efficiency and accuracy in practical applications. Summary of the Invention

[0005] To address the problems of low monitoring efficiency and accuracy in actual monitoring as mentioned in the background art, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a data analysis-based method for monitoring the operating status of a radiator, comprising: collecting multi-dimensional parameter data of the radiator at different times, the multi-dimensional parameter data including temperature data, voltage data and current data of the radiator, and taking the multi-dimensional parameter data of the radiator at the same time as a data point; clustering all data points to obtain multiple clusters; calculating the anomaly degree of any cluster, the anomaly degree being inversely correlated with the number of data points in the cluster and positively correlated with the standard deviation of the data points in the cluster; and, in response to data points in clusters with an anomaly degree greater than a set threshold being identified as anomalous data points, determining that the radiator is malfunctioning at the time corresponding to the anomalous data point.

[0007] The aforementioned technical solution clusters multi-dimensional parameter data of the radiator at different times, and combines this with a comprehensive calculation of the number of data points and standard deviation to accurately assess the degree of anomaly in each cluster. When the degree of anomaly exceeds a set threshold, it can promptly identify the possibility of radiator malfunction. This solution effectively addresses data fluctuations and environmental changes in radiator monitoring, automatically identifies potential fault moments, and provides accurate early warning capabilities for radiator health management, thereby solving the problems of low efficiency and accuracy in radiator monitoring.

[0008] Furthermore, the degree of abnormality , In the formula, For the first The number of data points in each cluster For the first The standard deviation of data points in each cluster It is an exponential function with the natural constant e as its base.

[0009] Furthermore, the density peak clustering algorithm is used to cluster all the data points to obtain multiple clusters, specifically: Any data point is taken as a cluster, and the neighboring clusters of any cluster are obtained. The neighboring clusters are all clusters whose local density is greater than that of any cluster. The neighboring clusters are sorted from smallest to largest based on the Euclidean distance between any cluster and its neighboring clusters to obtain a neighboring cluster sequence. If the Euclidean distance between each cluster in the neighboring cluster sequence and any cluster is less than the fluctuation of any cluster, the clusters are merged until all clusters are merged.

[0010] The aforementioned technical solution effectively identifies natural structures in the data and performs refined grouping. By dynamically comparing the local density of clusters, it prioritizes merging clusters with higher density and greater proximity, ensuring that data points with strong similarity are grouped together, while data points with significant density differences are correctly separated. Based on the assessment of fluctuation levels, it further avoids overly loose or unstable clustering results, improving the stability and accuracy of clustering through flexible merging strategies. This solution not only improves the accuracy of data clustering but also enhances sensitivity to outlier data points, effectively handling high-dimensional, complexly distributed radiator monitoring data, providing a more reliable and accurate foundation for subsequent anomaly monitoring.

[0011] Furthermore, hierarchical clustering algorithm is used to cluster all the data points to obtain multiple clusters. Specifically, any data point is taken as a cluster, and the inter-cluster distance between each cluster is calculated using the single connection method or the fully connected method. The two clusters with the smallest inter-cluster distance are merged into a new cluster. The above operation is iterated until the preset number of clusters is reached.

[0012] Furthermore, the local density of any data point is calculated as follows: Calculate the first Local density of data points , In the formula, For the first The data point of the th data point The values ​​of the dimension parameters, For the first Within a defined neighborhood radius centered on the data point, the first... The data point of the th data point The values ​​of the dimension parameters, The total number of dimensions, This represents the size of the neighborhood radius.

[0013] The aforementioned technical solution can effectively assess the density of data points within a local area, thereby reflecting the spatial distribution characteristics of the data. The calculation method for local density integrates the relative distance between a data point and its neighboring data points, sensitively capturing local structure and changes within the dataset. When data points are densely distributed within a local area, the local density is high; conversely, the density is low. Through this density assessment, the algorithm can more accurately identify anomalies in the areas where data points are located, especially those anomalous data points with low density.

[0014] Further, the fluctuation level of any data point is calculated, specifically: the Euclidean distance from each data point to other data points is calculated and sorted from smallest to largest to obtain the target sequence of each data point and the number of data points within the target sequence; the fluctuation level of the first data point is calculated. The local density difference of a data point represents the degree of anomaly of the data point; if the local density difference is greater than a set local threshold, the number of preset data points is incremented by 1 until the local density difference is less than or equal to the set local threshold, thereby obtaining the number of second preset data points, and all data points within the range of the second preset data point number are compared with the number of the first preset data points. The maximum Euclidean distance of the nth data point is used as the nth... The degree of fluctuation of each data point.

[0015] Further, calculate the first Local density differences of data points , In the formula, The preset number of data points, For the first Within the target sequence of the data points, the first... Local density of data points For the first Local density of data points For the first The local density mean of all data points within a target sequence of data points. This is the normalization function.

[0016] The aforementioned technical solution can accurately measure the relative anomaly of a data point within its cluster. Local density difference comprehensively considers the density difference between the data point and other data points within the target sequence, and removes the scale effect through a normalization function, ensuring consistency and comparability of density differences. This method can sensitively detect anomalous data whose local density differs significantly from surrounding data points, thereby enhancing anomaly detection capabilities.

[0017] Further, calculate the first Local density differences of data points , In the formula, The preset number of data points, For the first Within the target sequence of the data points, the first... Local density of data points For the first Local density of data points For the first The local density mean of all data points within a target sequence of data points. This is the normalization function.

[0018] The aforementioned technical solution not only considers the density differences between individual data points and their neighboring data points but also introduces an overall mean comparison, thereby enabling more accurate identification of outlier data points. In this way, the algorithm can more precisely capture data points exhibiting significant deviations within local areas, particularly those regions showing marked fluctuations in density distribution. Normalization ensures that density differences at different scales can be compared uniformly, making the monitoring results more reliable and consistent.

[0019] Furthermore, it also includes data standardization and missing data completion processing of the multi-dimensional parameter data.

[0020] In a second aspect, the present invention provides a heat sink operating status monitoring system based on data analysis, including a memory and a processor. The memory stores computer program instructions, which, when executed by the processor, implement the heat sink operating status monitoring method based on data analysis described above.

[0021] The beneficial effects of this invention are as follows: This invention combines multiple clustering algorithms and local density analysis methods to enable precise monitoring and early warning of radiators. By collecting multi-dimensional parameter data from the radiator and performing clustering processing, abnormal data points can be effectively identified, reflecting abnormal conditions of the radiator at specific times. By calculating the degree of anomaly of the clusters and combining it with local density differences, it is possible to accurately determine which data points represent potential radiator anomalies. Furthermore, the introduction of density peak clustering and hierarchical clustering algorithms not only ensures high accuracy in data analysis but also enables the handling of noise and data fluctuations in complex environments. This method not only improves the sensitivity of radiator monitoring but also enhances the efficiency and accuracy of anomaly detection, helping to promptly identify potential safety hazards and thus providing strong support for radiator maintenance and management. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating a data analysis-based method for monitoring the operating status of a radiator according to an embodiment of the present invention; Figure 2 This is a schematic block diagram illustrating the structure of a data analysis-based radiator operating status monitoring system according to an embodiment of the present invention. Detailed Implementation

[0023] Implementation example of a data analysis-based method for monitoring the operating status of radiators.

[0024] like Figure 1 The flowchart of the data analysis-based radiator operating status monitoring method of the present invention is shown, including the following steps: S1: Collect multi-dimensional parameter data of the radiator at different times, and take the multi-dimensional parameter data of the radiator at the same time as a data point.

[0025] In one embodiment, the multi-dimensional parameter data includes temperature data, voltage data, and current data of the heat sink. These data are crucial for describing the performance changes of the heat sink under different load and environmental conditions. Temperature data reflects the heat sink's heat dissipation efficiency and thermal conductivity, voltage data is used to assess the heat sink's stability under different power supply states, and current data reveals the actual power consumption of the heat sink and its load matching.

[0026] Temperature data is typically acquired using distributed temperature sensors, such as thermocouples, PT100 platinum resistance thermometers, or infrared temperature sensors. These sensors can be mounted on the surface of the heat sink or key heat dissipation components to obtain real-time temperature distribution. Data deviations are corrected using ambient temperature sensors, thereby improving the accuracy of temperature monitoring. Voltage data is acquired by connecting a high-precision voltage sensor or voltage sampling module to the heat sink circuit, and then using a data acquisition card or microcontroller to monitor fluctuations in the supply voltage in real time. Current data can be acquired using Hall effect current sensors or shunt resistor measurements. The former enables non-contact current measurement, suitable for high-current scenarios, while the latter detects current changes using high-precision resistors and combines this with analog-to-digital conversion technology for data acquisition.

[0027] To better process the multi-dimensional parameter data of the radiator, this embodiment employs data standardization and missing data completion techniques. Data standardization transforms various types of data to a unified scale, facilitating effective comparisons in subsequent analyses. This process not only eliminates data differences caused by different sensor types, ranges, or measurement methods but also improves the stability and accuracy of data analysis algorithms. Missing data completion techniques address the unavoidable missing data during data acquisition, employing strategies based on interpolation, regression analysis, or multiple interpolation to ensure the continuity and reliability of analysis even with missing data. Through these data processing techniques, noise and errors caused by environmental fluctuations, equipment failures, or transmission delays can be effectively eliminated, resulting in a more complete and consistent dataset across dimensions and improving the accuracy of radiator health status assessments.

[0028] S2: Cluster all data points to obtain multiple clusters; calculate the anomaly degree of any cluster.

[0029] In one embodiment, the density peak clustering algorithm is used to cluster all the data points to obtain multiple clusters. Specifically, any data point is taken as a cluster, and the neighboring clusters of any cluster are obtained. The neighboring clusters are all clusters whose local density is greater than the local density of any cluster. The neighboring clusters are sorted from smallest to largest based on the Euclidean distance between any cluster and the neighboring clusters to obtain a neighboring cluster sequence. If the Euclidean distance between each cluster in the neighboring cluster sequence and any cluster is less than the fluctuation degree of any cluster, the clusters are merged until the merging of all clusters is completed.

[0030] Clustering data points using density peak clustering can effectively identify the inherent structure and density distribution differences in the data, thereby achieving efficient grouping and outlier detection. By initially setting any data point as a cluster and sorting and merging neighboring clusters based on local density and Euclidean distance, the algorithm can dynamically adjust the cluster density and structure, ensuring that data points with strong similarity are grouped into the same cluster, while clusters in different density regions are reasonably divided.

[0031] Specifically, calculating the local density of any data point involves: calculating the... Local density of data points , In the formula, For the first The data point of the th data point The values ​​of the dimension parameters, For the first Within a defined neighborhood radius centered on the data point, the first... The data point of the th data point The values ​​of the dimension parameters, The total number of dimensions, This represents the size of the neighborhood radius.

[0032] By calculating the local density of each data point, particularly using the relative distance within a defined neighborhood radius, the distribution density of data points in multidimensional space can be more accurately assessed. This local density-based metric allows the clustering process to reflect the similarity between data points more precisely, especially in high-dimensional spaces, avoiding the biases that may arise from simple global distance metrics. Through the calculation of local density, high-density regions in the data—the core regions of clusters—can be identified, along with low-density noise points or outliers, thereby improving the accuracy and robustness of clustering.

[0033] To calculate the volatility of any data point, specifically: calculate the Euclidean distance from each data point to other data points and sort them in ascending order to obtain the target sequence of each data point and the number of data points within the target sequence; calculate the... The local density difference of a data point represents the degree of anomaly of the data point; if the local density difference is greater than a set local threshold, the number of preset data points is incremented by 1 until the local density difference is less than or equal to the set local threshold, thereby obtaining the number of second preset data points, and all data points within the range of the second preset data point number are compared with the number of the first preset data points. The maximum Euclidean distance of the nth data point is used as the nth... The degree of fluctuation of each data point.

[0034] By calculating the volatility and local density differences of data points, this method can effectively identify and quantify anomalous behavior of data points within local regions. By analyzing the Euclidean distance from each data point to other data points and combining this with the calculation of local density differences, it not only reflects the density of data points but also reveals potential anomalies or changes within specific regions. The introduction of volatility makes the clustering algorithm more flexible, allowing it to adjust to dynamic changes in data points, thereby effectively capturing nonlinear changes and potential anomalous patterns. This volatility analysis improves clustering accuracy, avoids oversimplification of stable regions in traditional methods, and enhances the algorithm's robustness to noisy data and outliers, ensuring effective grouping and more accurate clustering results in complex datasets.

[0035] Calculate the first Local density differences of data points , In the formula, The preset number of data points, For the first Within the target sequence of the data points, the first... Local density of data points For the first Local density of data points For the first The local density mean of all data points within a target sequence of data points. This is the normalization function.

[0036] By calculating the local density difference of data points, their density fluctuations relative to surrounding data points can be quantified, thereby revealing potential anomalies. Local density difference considers the local density variations between a data point and its neighbors, and normalization eliminates the influence of different scales, allowing density differences to be compared under a unified standard. This metric not only effectively captures the density variations of data points within local regions but also identifies data points exhibiting different clustering characteristics due to density differences. Through the calculation of local density differences, clustering algorithms can more acutely detect anomalies in the local structure of data points, avoiding the neglect of unevenly distributed density data by traditional methods, and further improving the fine-grained identification and anomaly detection capabilities of clustering.

[0037] In another embodiment, the calculation of the first Local density differences of data points , In the formula, The preset number of data points, For the first Within the target sequence of the data points, the first... Local density of data points For the first Local density of data points For the first The local density mean of all data points within a target sequence of data points. This is the normalization function.

[0038] Employing a metric based on the absolute difference and mean allows for a more sensitive reflection of local density variations among data points. This calculation method not only considers the differences in local density between data points and other data points within the target sequence but also, by introducing the relationship between absolute difference and mean, enables the clustering algorithm to capture subtle density fluctuations between data points, especially in areas with complex or highly variable density distributions. Normalization further eliminates the influence of scale differences, making density differences between different data points comparable and enhancing the stability and consistency of clustering. This method helps improve the accuracy of the clustering process, particularly in complex datasets, and better identifies areas with large density fluctuations. This provides more targeted and effective support for anomaly detection and fine-grained classification, optimizing data grouping results.

[0039] In another embodiment, a hierarchical clustering algorithm is used to cluster all the data points to obtain multiple clusters, specifically: Treat any data point as a cluster, calculate the inter-cluster distance between each cluster using the single-connection or full-connection method, merge the two clusters with the smallest inter-cluster distance into a new cluster, and iterate the above operation until the preset number of clusters is reached.

[0040] By using hierarchical clustering algorithms, combined with single-connection or fully-connection methods to calculate inter-cluster distances and progressively merge clusters, layer-by-layer grouping and refined clustering of data points can be achieved. This method effectively identifies the hierarchical structure of the dataset by dynamically adjusting inter-cluster distances, thus avoiding the over-reliance on data distribution assumptions in traditional clustering methods. During the clustering process, the progressive merging of clusters allows the algorithm to adaptively adjust the clustering scale, making it particularly suitable for handling multi-level relationships and regions of varying densities in the dataset. The choice between single-connection and fully-connection methods allows for flexible adjustment of cluster density, making the clustering results more consistent with the structural characteristics of the actual data, effectively improving the accuracy and interpretability of clustering.

[0041] S3: In response to data points in clusters where the anomaly level is greater than a set threshold, the data points are identified as anomalous data points, and the corresponding time point of the anomalous data point is determined to be a radiator malfunction.

[0042] The threshold value set above can be 0.8, but it can also be determined according to the actual situation.

[0043] In one embodiment, anomalies in the radiator at a given moment can be effectively identified by comparing the anomaly level within clusters with a set threshold. When the anomaly level of a cluster exceeds the set threshold, the data points within that cluster are marked as anomalous, indicating a potential radiator malfunction at that moment. To further optimize this process, dynamic threshold settings can be employed, adjusting the threshold based on the radiator's historical health data and real-time monitoring results for more accurate anomaly identification. Time series analysis techniques can be introduced to further reduce the probability of misjudgments by analyzing the changing trends of data points over different time periods. For example, if the anomaly level of a cluster exceeds the threshold at multiple consecutive moments, this moment can be further confirmed as a critical moment for the radiator.

[0044] This optimized method can more accurately capture structural changes in radiators under different loads and environmental conditions, provide timely warnings of potential safety hazards, improve the ability to monitor the long-term health status of radiators, reduce false alarms caused by environmental factors or short-term fluctuations, and further enhance the intelligence and adaptability of monitoring.

[0045] This invention effectively monitors radiator anomalies through the collection and clustering analysis of multi-dimensional parameter data. By classifying data points and calculating the degree of anomaly using clustering algorithms, radiator malfunctions can be sensitively identified. The introduction of multiple clustering methods, such as density peak clustering and hierarchical clustering, combined with the calculation of local density and fluctuation levels, further enhances the monitoring capability for abnormal data points, especially in the face of complex data and noise, enabling precise differentiation between normal and abnormal regions. Utilizing the calculation of local density differences, the degree of anomaly for each data point can be effectively assessed, thereby improving monitoring accuracy.

[0046] Example of a data analysis-based radiator operating status monitoring system: like Figure 2 As shown in the figure, the structural block diagram of the heat sink operating status monitoring system based on data analysis according to an embodiment of the present invention includes a processor and a memory.

[0047] This invention also provides a heat sink operating status monitoring system based on data analysis. For example... Figure 2As shown, the system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the data analysis-based radiator operating status monitoring method according to the present invention.

[0048] The data analysis-based radiator operating status monitoring system also includes other components well known to those skilled in the art, such as communication interfaces. Their settings and functions are known in the art and will not be described in detail here.

[0049] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a 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 desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented by computer-readable / executable instructions stored or otherwise maintained on such a computer-readable medium.

[0050] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise explicitly specified.

[0051] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for monitoring the operating status of a radiator based on data analysis, characterized in that, include: Collect multi-dimensional parameter data of the radiator at different times. The multi-dimensional parameter data includes the temperature data, voltage data and current data of the radiator. Collect the multi-dimensional parameter data of the radiator at the same time as a data point. All data points are clustered to obtain multiple clusters; the anomaly degree of any cluster is calculated, wherein the anomaly degree is inversely correlated with the number of data points in any cluster and positively correlated with the standard deviation of the data points in any cluster; Data points in clusters with an anomaly level greater than a set threshold are identified as anomalous data points, and the corresponding time point of the anomalous data point is determined to be an radiator malfunction.

2. The method for monitoring the working status of a radiator based on data analysis according to claim 1, characterized in that, The degree of abnormality , In the formula, For the first The number of data points in each cluster For the first The standard deviation of data points in each cluster It is an exponential function with the natural constant e as its base.

3. The method for monitoring the working status of a radiator based on data analysis according to claim 1, characterized in that, The density peak clustering algorithm is used to cluster all the data points to obtain multiple clusters, specifically: Any data point is taken as a cluster, and the neighboring clusters of any cluster are obtained. The neighboring clusters are all clusters whose local density is greater than that of any cluster. The neighboring clusters are sorted from smallest to largest based on the Euclidean distance between any cluster and its neighboring clusters to obtain a neighboring cluster sequence. If the Euclidean distance between each cluster in the neighboring cluster sequence and any cluster is less than the fluctuation of any cluster, the clusters are merged until all clusters are merged.

4. The method for monitoring the working status of a radiator based on data analysis according to claim 1, characterized in that, Hierarchical clustering algorithm is used to cluster all the data points to obtain multiple clusters, specifically: Treat any data point as a cluster, calculate the inter-cluster distance between each cluster using the single-connection or full-connection method, merge the two clusters with the smallest inter-cluster distance into a new cluster, and iterate the above operation until the preset number of clusters is reached.

5. The radiator operating status monitoring method based on data analysis according to claim 3, characterized in that, The local density of any data point is calculated as follows: Calculate the first Local density of data points , In the formula, For the first The data point of the th data point The values ​​of the dimension parameters, For the first Within a defined neighborhood radius centered on the data point, the first... The data point of the th data point The values ​​of the dimension parameters, The total number of dimensions, This represents the size of the neighborhood radius.

6. The method for monitoring the working status of a radiator based on data analysis according to claim 3, characterized in that, To calculate the fluctuation level of any data point, specifically: calculate the Euclidean distance from each data point to other data points and sort them from smallest to largest to obtain the target sequence of each data point and the number of data points in the target sequence; Calculate the first The local density difference of a data point represents the degree of anomaly of the data point; if the local density difference is greater than a set local threshold, the number of preset data points is incremented by 1 until the local density difference is less than or equal to the set local threshold, thereby obtaining the number of second preset data points, and all data points within the range of the second preset data point number are compared with the number of the first preset data points. The maximum Euclidean distance of the nth data point is used as the nth... The degree of fluctuation of each data point.

7. The radiator operating status monitoring method based on data analysis according to claim 6, characterized in that, Calculate the first Local density differences of data points , In the formula, The preset number of data points, For the first Within the target sequence of the data points, the first... Local density of data points For the first Local density of data points For the first The local density mean of all data points within a target sequence of data points. This is the normalization function.

8. The method for monitoring the working status of a radiator based on data analysis according to claim 6, characterized in that, Calculate the first Local density differences of data points , In the formula, The preset number of data points, For the first Within the target sequence of the data points, the first... Local density of data points For the first Local density of data points For the first The local density mean of all data points within a target sequence of data points. This is the normalization function.

9. The method for monitoring the working status of a radiator based on data analysis according to claim 1, characterized in that, It also includes data standardization and missing data completion processing of the multi-dimensional parameter data.

10. A radiator operating status monitoring system based on data analysis, characterized in that, It includes a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the heat sink operating status monitoring method based on data analysis as described in any one of claims 1 to 9 is implemented.

Images