Industrial pollution discharge equipment operation data analysis method and system

Abstract

The invention relates to an industrial pollution discharge equipment operation data analysis method and system, and the method comprises the steps: carrying out the normalization of each index of collected industrial pollution discharge equipment operation data, calculating the distance matrix of each index, obtaining a minimum spanning tree through the distance matrixes, calculating the distance average value in the minimum spanning tree of each index, and obtaining a minimum spanning tree of each index; setting a weight for each index according to the distance average value; calculating the distance between the operation data according to the weight, obtaining the local density and the high-density minimum distance of the operation data according to the truncation distance and the distance, determining the number of nearest neighbors according to the minimum spanning tree of the index, calculating the operation data and the standard deviation of the local density of the nearest neighbors of the number of the nearest neighbors of the operation data, and obtaining the standard deviation of the local density of the nearest neighbors of the operation data. Determining a cluster center by using the local density, the high-density minimum distance and the standard deviation; and clustering the industrial data according to the cluster center, and displaying a clustering result as a data analysis result in a chart form.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and specifically to a method and system for analyzing the operation data of industrial sewage discharge equipment. Background Art

[0002] Industrial sewage discharge equipment is the core link of the pollution prevention and control system. The real-time monitoring and intelligent analysis of its operation status are important guarantees for realizing green production. With the increasingly strict global environmental protection policies, enterprises urgently need to reduce pollution emissions and energy consumption through refined operation and maintenance. Modern industrial sewage discharge equipment is usually equipped with multi-source sensors, such as sensors for flow rate, pH value, COD concentration, equipment vibration, etc., which can generate high-dimensional and strongly time-series correlated operation data in real time. However, limited by challenges such as large data noise, frequent dynamic switching of working conditions, and scarcity of abnormal samples, traditional threshold warning or statistical analysis methods are difficult to accurately capture the potential deterioration trend or inefficient operation mode of the equipment. For example, for the intermittent idling fault of a sewage pump in a chemical enterprise, its vibration signal is only abnormal under specific loads, and normal working condition data is likely to mask the fault with conventional methods.

[0003] The Density Peaks Clustering (DPC) algorithm based on density peaks has been widely used in industrial data analysis due to its advantages of not requiring the preset number of clusters and being able to adapt to the distribution of clusters of any shape. The classic DPC algorithm constructs a decision graph by calculating local density and relative distance, and manually selects density peak points as clustering centers. However, industrial sewage discharge data has characteristics such as high-dimensional heterogeneity and non-stationary time series, resulting in significant limitations for traditional DPC. Existing DPC requires manual selection of cluster centers, and when small clusters appear, due to the nature of the DPC algorithm itself, it is easy to confuse the centers of small clusters with noise, and small clusters are crucial for early detection of abnormal operation of industrial sewage discharge equipment. Summary of the Invention

[0004] Aiming at the problems that when using the existing DPC to analyze the operation data of industrial sewage discharge equipment, it is necessary to manually specify the cluster center and the cluster center of small clusters is easy to be confused with noise.

[0005] In a first aspect, the present invention provides a method for analyzing the operation data of industrial sewage discharge equipment, the method comprising: Normalize each index of the collected operation data of industrial sewage discharge equipment respectively, calculate the distance matrix of each index, obtain the minimum spanning tree using the distance matrix, calculate the average distance in the minimum spanning tree of each index, and set weights for each index according to the average distance; Calculate the distance between the operation data according to the weights, obtain the local density and the minimum distance of high density of the operation data based on the truncation distance and the distance, determine the number of nearest neighbors according to the minimum spanning tree of the metrics, calculate the standard deviation of the local density of the operation data and the nearest neighbors of the number of nearest neighbors of the operation data, and determine the cluster center by using the local density, the minimum distance of high density and the standard deviation; Cluster the industrial data according to the cluster center, and display the clustering result in the form of a chart as the data analysis result.

[0006] Preferably, the calculating the distance between the operation data according to the weights is specifically as follows: For any two operation data whose distances are to be calculated, obtain the distances of the same metrics of the two operation data whose distances are to be calculated, weight the distances of the metrics by the weights of the metrics, and take the sum of all the weighted distances as the distance between the two operation data whose distances are to be calculated.

[0007] Preferably, the determining the number of nearest neighbors according to the minimum spanning tree of the metrics is specifically as follows: Divide the minimum spanning tree of each metric to obtain a plurality of subgraphs, obtain the minimum subgraph of the minimum spanning tree of each metric, and obtain the number of nearest neighbors according to the minimum subgraph.

[0008] Preferably, the obtaining the number of nearest neighbors according to the minimum subgraph is specifically as follows: Calculate the average value of the number of operation data in the minimum subgraphs of all metrics, and take the average value as the number of nearest neighbors.

[0009] Preferably, the determining the cluster center by using the local density, the minimum distance of high density and the standard deviation is specifically as follows: Sort the operation data in descending order according to the minimum distance of high density, and select the top k operation data, where k is a positive integer greater than 1; Calculate the average value of the local density of the top k operation data or the average value of the local density of the operation data in the minimum subgraphs of all metrics, and take the operation data whose local density of the top k operation data is less than the average value of the local density as candidate noise points; Obtain the average value or the maximum value of the standard deviation of the operation data in the minimum subgraphs of all metrics. If the standard deviation of the candidate noise points is greater than the average value or the maximum value of the standard deviation, then take the candidate noise points as noise; Take the non-noise operation data among the top k operation data as the cluster center.

[0010] In a second aspect, the present invention provides an operation data analysis system for industrial sewage treatment equipment, characterized in that the system includes: A data acquisition module, which is used to normalize each index of the operation data of industrial pollution discharge equipment collected respectively, calculate the distance matrix of each index, obtain the minimum spanning tree by using the distance matrix, calculate the average distance in the minimum spanning tree of each index, and set weights for each index according to the average distance; A parameter calculation module, which is used to calculate the distance between operation data according to the weights, obtain the local density and the minimum distance of high density of the operation data according to the truncation distance and the distance, determine the number of nearest neighbors according to the minimum spanning tree of the index, calculate the standard deviation of the local density of the operation data and the nearest neighbors of the number of nearest neighbors of the operation data, and determine the cluster center by using the local density, the minimum distance of high density and the standard deviation; A clustering module, which is used to cluster industrial data according to the cluster center and display the clustering result in the form of a chart as the data analysis result.

[0011] Preferably, the calculating the distance between operation data according to the weights is specifically as follows: For any two operation data whose distances are to be calculated, obtain the distances of the same index of the two operation data whose distances are to be calculated, weight the distances of the index by using the weights of the index, and take the sum of all weighted distances as the distance between the two operation data whose distances are to be calculated.

[0012] Preferably, the determining the number of nearest neighbors according to the minimum spanning tree of the index is specifically as follows: Split the minimum spanning tree of each index to obtain multiple subgraphs, obtain the minimum subgraph of the minimum spanning tree of each index, and obtain the number of nearest neighbors according to the minimum subgraph.

[0013] Preferably, the obtaining the number of nearest neighbors according to the minimum subgraph is specifically as follows: Calculate the average value of the number of operation data in the minimum subgraphs of all indexes, and take the average value as the number of nearest neighbors.

[0014] Preferably, the determining the cluster center by using the local density, the minimum distance of high density and the standard deviation is specifically as follows: Sort the operation data in descending order according to the minimum distance of high density, and select the first k operation data, where k is a positive integer greater than 1; Calculate the average value of the local density of the first k operation data or the average value of the local density of the operation data in the minimum subgraphs of all indexes, and take the operation data whose local density of the first k operation data is less than the average value of the local density as candidate noise points; Obtain the average value or the maximum value of the standard deviation of the operation data in the minimum subgraphs of all indexes. If the standard deviation of the candidate noise points is greater than the average value or the maximum value of the standard deviation, take the candidate noise points as noise; Take the non-noise running data among the first k running data as the cluster centers.

[0015] The present invention has the following advantages: 1. By calculating different weights for the running data, the influence of the indexes with relatively dispersed data distribution on calculating the distance of the running data is eliminated; 2. Only the approximate range of the number of clusters needs to be specified, without manually specifying the specific number of clusters; 3. For the running data with both large clusters and small clusters, the present invention can identify the small clusters and eliminate the influence of noise. Description of the Drawings

[0016] Figure 1 Is the flowchart of the first embodiment; Figure 2 Is the schematic diagram of calculating the local density by the cut-off distance; Figure 3 Is the schematic diagram of subgraph partitioning of the minimum spanning tree; Figure 4 Is the flowchart of determining the cluster centers; Figure 5 Is the effect diagram of clustering the running data; Figure 6 Is the effect diagram of clustering the running data including small clusters. Detailed Embodiments

[0017] In order to enable those skilled in the art of the present technology to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0018] It should be noted that the terms "first", "second", etc. in the specification, claims and drawings of the present application are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data used may be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0019] The first embodiment provides an industrial sewage discharge equipment operation data analysis method, as Figure 1 shown, the method includes: Step 1: Normalize each indicator of the collected operation data of industrial pollution discharge equipment respectively, calculate the distance matrix of each indicator, obtain the minimum spanning tree using the distance matrix, calculate the average distance in the minimum spanning tree of each indicator, and set weights for each indicator according to the average distance. Industrial pollution discharge equipment is used to treat pollutants such as wastewater, waste gas, and solid waste to make the emissions meet the standards. Various data generated during the operation of the equipment, which reflect the operation status, treatment effect, and potential fault information of the equipment. The operation data includes equipment operation data, pollutant discharge data, equipment status data, etc.; among them, the equipment operation data includes but is not limited to the influent flow rate, effluent flow rate, pressure, temperature, pH value, dissolved oxygen, etc. of wastewater treatment equipment, or the air volume, air pressure, temperature, humidity, pollutant concentration, etc. of waste gas treatment equipment; the pollutant discharge data includes but is not limited to chemical oxygen demand (COD), ammonia nitrogen, total phosphorus, total nitrogen, etc. in wastewater, or sulfur dioxide, nitrogen oxides, particulate matter, volatile organic compounds (VOCs), etc. in waste gas; the equipment status data includes but is not limited to the motor speed, current, voltage, vibration, temperature, etc., the flow rate, pressure, operation time, etc. of the pump, and the opening degree, operation status, etc. of the valve. Each parameter in the above operation data is an indicator. For example, the influent flow rate is an indicator, and the effluent flow rate is another indicator. One piece of operation data includes the values of multiple indicators.

[0020] Since the units and magnitudes of different indicators are different, to eliminate this influence, each indicator of the operation data is normalized respectively. For example, a temperature of 50°C is normalized to 0.5, and a pressure of 5 MPa is normalized to 0.2, etc. The normalization methods include but are not limited to Z-score standardization or maximum-minimum normalization, etc.

[0021] Calculate the distance matrix for each indicator. When there are multiple pieces of operation data, each indicator will have multiple values. For example, if there are 10 pieces of operation data, the temperature indicator has 10 values, and each value is the corresponding value of the temperature in one piece of operation data. Calculate the distance matrix of the values of the same indicator under all operation data, and each indicator obtains a distance matrix. After obtaining the distance matrix, use the Prim algorithm or Kruskal algorithm to obtain the minimum spanning tree of the indicator. Similarly, each indicator will have a minimum spanning tree. Calculate the average distance of the minimum spanning tree of each indicator. If the average distance is relatively small, the data distribution of this indicator is relatively dense or concentrated, and vice versa.

[0022] If the distribution of the metrics is relatively dispersed, when calculating the distance between running data subsequently, the distance is mainly determined by these dispersed metrics. To eliminate this influence, weights are set for each metric according to the average distance. The larger the average distance, the smaller the weight. In one embodiment, setting weights for each metric according to the average distance specifically involves taking the reciprocal of the average distance, normalizing all the reciprocals so that the sum of all weights equals 1. In another embodiment, the exponential decay method is used to calculate the weights.

[0023] Step 2: Calculate the distance between running data according to the weights, obtain the local density and the minimum high-density distance of the running data based on the truncation distance and the distance, determine the number of nearest neighbors according to the minimum spanning tree of the metrics, calculate the standard deviation of the local density of the running data and the nearest neighbors of the number of nearest neighbors of the running data, and determine the cluster center using the local density, the minimum high-density distance, and the standard deviation; After obtaining the weights of the metrics, according to the weights of the metrics, the distances of each metric are weighted and summed to obtain the distance between running data. Specifically, for any two running data whose distances are to be calculated, obtain the distances of the same metrics of the two running data whose distances are to be calculated, weight the distances of the metrics using the weights of the metrics, and take the sum of all weighted distances as the distance between the two running data whose distances are to be calculated. For two running data whose distances are to be calculated, calculate the differences between them on each same metric through the Euclidean distance, etc., multiply the differences or distances on each metric by the weight of the metric, and add up all the weighted metric distances to obtain the final distance between the two running data. For example, for two running data, each running data has two metrics, temperature and pressure, and the weights of temperature and pressure are 0.7 and 0.3 respectively. If the temperature difference between these two running data is 10 and the pressure difference is 2, then the distance between these two running data is 7.6.

[0024] After obtaining the distance between running data, the DPC clustering algorithm is used to obtain the local density and the minimum high-density distance of each running data. Among them, the local density is the number of running data whose distance from the running data is less than the truncation distance. As Figure 2 shown, Figure 2 the local density of the black dot in is 6; the minimum high-density distance is the minimum distance from a data point to all data points with a higher local density than it.

[0025] In the DPC clustering algorithm, the truncation distance generally makes the number of neighbors of each data point account for 1% to 2% of the total number of data points on average. However, due to different data distribution situations, the above setting of the truncation distance is not a general method, and the truncation distance directly affects the subsequent clustering effect. If the truncation distance is too large, small clusters cannot be independently identified. If the truncation distance is too small, small clusters may be misjudged as noise points. In one embodiment, the truncation distance is still determined by the proportion of the average number of neighbors, for example, the truncation distance is determined by the way that the number of neighbors of each data point accounts for 1% to 5% of the total number of data points. After obtaining the local density and the minimum distance of the high density using the truncation distance, the number of nearest neighbors is further determined according to the minimum spanning tree of the index, and then the standard deviation of the local density of the running data and the nearest neighbors of the number of nearest neighbors of the running data is calculated, and the cluster center is determined using the local density, the minimum distance of the high density, and the standard deviation.

[0026] Among them, the standard deviation is centered on a running data A, and the number of running data closest to the running data A is obtained, and the standard deviation of the local density of the running data A and the number of running data closest to the running data A is calculated. In one embodiment, after obtaining the number of running data closest to the running data A, it further includes removing the maximum and minimum values of the local density of the running data A and the number of running data closest to the running data A, and then calculating the standard deviation.

[0027] The smaller the standard deviation, the denser the periphery of the running data A; the larger the standard deviation, the greater the change in the periphery of the running data A, especially when the running data A is in a small cluster.

[0028] In one embodiment, the determining the number of nearest neighbors according to the minimum spanning tree of the index is specifically: The minimum spanning tree of each index is segmented to obtain a plurality of subgraphs, the minimum subgraph of the minimum spanning tree of each index is obtained, and the number of nearest neighbors is obtained according to the minimum subgraph.

[0029] For the minimum spanning tree of each index, the minimum spanning tree is segmented into a plurality of subgraphs by the size of the edge or the specified number of subgraphs, etc., such as Figure 3As shown, each metric will obtain a series of subgraphs or subtrees. The minimum subgraph of each metric is obtained, and the number of nearest neighbors is generated based on the minimum subgraphs of all metrics. In a preferred embodiment, the average number of running data in the minimum subgraphs of all metrics is calculated, and the average value is used as the number of nearest neighbors. The number of nearest neighbors actually reflects the number of minimum clusters of running data. In one embodiment, it further includes filtering of the subgraphs. If the number of running data in the segmented subgraph is less than a set value such as 5 or a preset percentage of the total number of running data, the subgraph is deleted from all the segmented subgraphs, which can avoid the influence of noise on the segmentation of the minimum spanning tree. For example, a certain noise point is segmented into a subgraph.

[0030] Cluster centers are usually located in high-density regions and are far from other high-density regions. However, when small clusters appear, the local density of small clusters may not be much different from the local density of running data that is not a cluster center, which may lead to the mixing of the centers of small clusters and other data, especially noise. In one embodiment, as Figure 4 shown, the cluster center is determined by using the local density, the minimum distance to high density, and the standard deviation, specifically as follows: Sort the running data in descending order according to the minimum distance to high density, and select the first k running data, where k is a positive integer greater than 1; Calculate the average local density of the first k running data or the average local density of the running data in the minimum subgraphs of all metrics, and use the running data whose local density of the first k running data is less than the average local density as candidate noise points; Obtain the average value or maximum value of the standard deviation of the running data in the minimum subgraphs of all metrics. If the standard deviation of the candidate noise points is greater than the average value or maximum value of the standard deviation, the candidate noise points are regarded as noise; Use the non-noise running data among the first k running data as the cluster center.

[0031] Sort the running data in descending order according to the minimum distance to high density to obtain a sorted data sequence, with data points having a higher potential to be cluster centers arranged in the front. Select the first k running data in the sorted data sequence as candidate cluster centers, where k is a positive integer greater than 1, and the value of k can be adjusted according to the scale of the data set and the expected number of clusters. If the running data is only divided into two categories: normal and abnormal, then k can be taken as 4. If the running data needs to be classified in more detail, then k can be set larger. If a running data is a cluster, especially the center of a large cluster, then both the local density and the minimum distance to high density are relatively large. If it is noise or the center of a small cluster, then the local density may be relatively small.

[0032] Take the running data whose local density of the first k running data is less than the average as candidate noise points, and further use the standard deviation to determine whether it is a cluster center. If it is a cluster center, a tight area is formed, and the standard deviation of the running data around it is also small. If it is noise, its distance from other clusters is far, the local density changes greatly, and the standard deviation is large. And if it is a small cluster, its standard deviation is less than that of the noise points. Further obtain the average value or maximum value of the standard deviation of the running data in the minimum subgraph of all indicators. If the standard deviation of the candidate noise point is greater than the average value or maximum value of the standard deviation, then take the candidate noise point as noise.

[0033] In another embodiment, use the local density, the minimum distance to the high density, and the standard deviation to determine the cluster center. Specifically: Set a local density threshold, a minimum distance threshold to the high density, and a standard deviation threshold. If the local density of a running data is greater than the local density threshold, and the minimum distance to the high density is greater than the minimum distance threshold to the high density, and the standard deviation is less than the standard deviation threshold, then take this running data as the cluster center; or, By setting weights for the local density, the minimum distance to the high density, and the standard deviation respectively, calculate the weighted sum of the local density and the minimum distance to the high density, and subtract the product of the standard deviation and the weight, and then obtain the score of each running data. Sort the running data in descending order of the score, and select the first n running data as the cluster center.

[0034] Step 3, cluster the industrial data according to the cluster center, and display the clustering result in the form of a chart as the data analysis result.

[0035] According to the determined cluster center, allocate each running data in the data set or the non-noise points in the data set to the cluster to which the cluster center with the closest distance belongs, and display the clustering result in the form of a chart. Figure 5 It is the effect diagram of clustering industrial data. Figure 6 It is the effect diagram of clustering the running data containing small clusters. It can be seen that the present invention can accurately identify small clusters and avoid taking small clusters as noise. The chart form includes but is not limited to two-dimensional scatter plots, three-dimensional scatter plots, heat maps, etc.

[0036] Embodiment 2 provides an operating data analysis system for industrial sewage treatment equipment. The system includes: A data acquisition module, which is used to normalize each index of the collected operating data of the industrial sewage treatment equipment, calculate the distance matrix of each index, obtain the minimum spanning tree using the distance matrix, calculate the average value of the distances in the minimum spanning tree of each index, and set weights for each index according to the average value of the distances. A parameter calculation module, configured to calculate the distance between running data according to the weight, obtain the local density and the minimum distance of high density of the running data based on the truncation distance and the distance, determine the number of nearest neighbors according to the minimum spanning tree of the index, calculate the standard deviation of the local density of the running data and the nearest neighbors of the number of nearest neighbors of the running data, and determine the cluster center by using the local density, the minimum distance of high density and the standard deviation; A clustering module, configured to cluster the industrial data according to the cluster center and display the clustering result in the form of a chart as the data analysis result.

[0037] Preferably, the calculating the distance between running data according to the weight is specifically as follows: For any two running data whose distances are to be calculated, obtain the distances of the same index of the two running data whose distances are to be calculated, weight the distances of the index by the weight of the index, and take the sum of all weighted distances as the distance between the two running data whose distances are to be calculated.

[0038] Preferably, the determining the number of nearest neighbors according to the minimum spanning tree of the index is specifically as follows: Split the minimum spanning tree of each index to obtain multiple subgraphs, obtain the minimum subgraph of the minimum spanning tree of each index, and obtain the number of nearest neighbors according to the minimum subgraph.

[0039] Preferably, the obtaining the number of nearest neighbors according to the minimum subgraph is specifically as follows: Calculate the average value of the number of running data in the minimum subgraphs of all indexes, and take the average value as the number of nearest neighbors.

[0040] Preferably, the determining the cluster center by using the local density, the minimum distance of high density and the standard deviation is specifically as follows: Sort the running data in descending order according to the minimum distance of high density, and select the first k running data, where k is a positive integer greater than 1; Calculate the average value of the local density of the first k running data or the average value of the local density of the running data in the minimum subgraphs of all indexes, and take the running data whose local density of the first k running data is less than the average value of the local density as candidate noise points; Obtain the average value or the maximum value of the standard deviation of the running data in the minimum subgraphs of all indexes. If the standard deviation of the candidate noise points is greater than the average value or the maximum value of the standard deviation, take the candidate noise points as noise; Take the non-noise running data among the first k running data as the cluster center.

[0041] Embodiment 3 provides a computer-readable storage medium. Optionally, in this embodiment, the above computer program product can be used to save the program code executed by the data analysis method provided in the above Embodiment 1.

[0042] Embodiment 4 provides a computer program product. Optionally, in this embodiment, the above computer program product can be used to store the program code executed by the data analysis method provided in Embodiment 1 above.

[0043] Optionally, in this embodiment, the above computer program product can be located in any one of the computer terminals in the computer terminal group in the computer network, or in any one of the mobile terminals in the mobile terminal group.

[0044] In the above embodiments of the present application, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0045] In the several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of the units or modules can be in an electrical or other form.

[0046] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0047] In addition, the functional units in the various embodiments of the present application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0048] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical discs.

[0049] The above are only the preferred embodiments of this application. It should be noted that for those of ordinary skill in the art, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A method for analyzing the operation data of industrial sewage discharge equipment, characterized in that: The method comprises: Normalize each indicator of the collected industrial sewage discharge equipment operation data, calculate the distance matrix of each indicator, use the distance matrix to obtain the minimum spanning tree, calculate the average distance in the minimum spanning tree of each indicator, and set a weight for each indicator according to the average distance; Calculate the distance between the running data according to the weight, obtain the local density and high-density minimum distance of the running data according to the cutoff distance and the distance, determine the number of nearest neighbors according to the minimum spanning tree of the indicator, calculate the standard deviation of the local density of the nearest neighbors of the running data and the number of nearest neighbors of the running data, and determine the cluster center using the local density and high-density minimum distance and standard deviation; The industrial data is clustered according to the cluster centers, and the clustering results are displayed in a chart form as data analysis results.

2. The method according to claim 1, characterized in that The distance between the running data is calculated according to the weight, specifically: For any two running data whose distance is to be calculated, the distance of the same indicator of the two running data whose distance is to be calculated is obtained, the distance of the indicator is weighted by the weight of the indicator, and the sum of all weighted distances is taken as the distance between the two running data whose distance is to be calculated.

3. The method according to claim 1, characterized in that The method of determining the number of nearest neighbors according to the minimum spanning tree of the indicator is specifically as follows: The minimum spanning tree of each indicator is divided into multiple subgraphs, the minimum subgraph of the minimum spanning tree of each indicator is obtained, and the number of nearest neighbors is obtained according to the minimum subgraph.

4. The method according to claim 3, characterized in that The number of nearest neighbors is obtained according to the minimum subgraph, specifically: The average number of running data items in the minimum subgraph of all indicators is calculated, and the average value is used as the number of nearest neighbors.

5. The method according to claim 3 or 4, characterized in that: The cluster center is determined by using the local density and the minimum distance and standard deviation of the high density, specifically: Sort the running data in descending order according to the high-density minimum distance, and select the first k running data, among which, k is a positive integer greater than 1; Calculate the local density average of the first k running data or the local density average of the running data in the minimum subgraph of all indicators, and take the running data whose local density of the first k running data is less than the local density average as the candidate noise point; Obtain the average or maximum value of the standard deviation of the running data in the minimum subgraph of all indicators. If the standard deviation of the candidate noise point is greater than the average or maximum value of the standard deviation, the candidate noise point is regarded as noise; The non-noise running data among the first k running data are taken as the cluster centers.

6. An industrial sewage equipment operation data analysis system, characterized in that: The system comprises: The data collection module is used to normalize each indicator of the collected industrial sewage equipment operation data, calculate the distance matrix of each indicator, obtain the minimum spanning tree using the distance matrix, calculate the average distance in the minimum spanning tree of each indicator, and set a weight for each indicator according to the average distance; A parameter calculation module, used to calculate the distance between the running data according to the weight, obtain the local density and high-density minimum distance of the running data according to the cutoff distance and the distance, determine the number of nearest neighbors according to the minimum spanning tree of the indicator, calculate the standard deviation of the local density of the nearest neighbors of the running data and the number of nearest neighbors of the running data, and determine the cluster center using the local density and high-density minimum distance and standard deviation; The clustering module is used to cluster the industrial data according to the cluster centers and display the clustering results in the form of charts as data analysis results.

7. The system according to claim 6, characterized in that The distance between the running data is calculated according to the weight, specifically: For any two running data whose distance is to be calculated, the distance of the same indicator of the two running data whose distance is to be calculated is obtained, the distance of the indicator is weighted by the weight of the indicator, and the sum of all weighted distances is taken as the distance between the two running data whose distance is to be calculated.

8. The system according to claim 6, characterized in that The method of determining the number of nearest neighbors according to the minimum spanning tree of the indicator is specifically as follows: The minimum spanning tree of each indicator is divided into multiple subgraphs, the minimum subgraph of the minimum spanning tree of each indicator is obtained, and the number of nearest neighbors is obtained according to the minimum subgraph.

9. The system according to claim 8, characterized in that The number of nearest neighbors is obtained according to the minimum subgraph, specifically: The average number of running data items in the minimum subgraph of all indicators is calculated, and the average value is used as the number of nearest neighbors.

10. The system according to claim 8 or 9, characterized in that The cluster center is determined by using the local density and the minimum distance and standard deviation of the high density, specifically: Sort the running data in descending order according to the high-density minimum distance, and select the first k running data, among which, k is a positive integer greater than 1; Calculate the local density average of the first k running data or the local density average of the running data in the minimum subgraph of all indicators, and take the running data whose local density of the first k running data is less than the local density average as the candidate noise point; Obtain the average or maximum value of the standard deviation of the running data in the minimum subgraph of all indicators. If the standard deviation of the candidate noise point is greater than the average or maximum value of the standard deviation, the candidate noise point is regarded as noise; The non-noise running data among the first k running data are taken as the cluster centers.

Images