An abnormal monitoring and early warning system for mining water pumps

Abstract

The present application relates to the field of data processing, and provides a mining water pump abnormality monitoring and early warning system, including: a data acquisition module, used to collect pressure data, temperature data and current data of the mining water pump at each moment; a coefficient calculation module, which calculates the distribution determination coefficient of the representative point of the pressure data at the corresponding moment based on the temperature data and current data at the same moment; a confidence calculation module, which determines the confidence of the representative point of the pressure data at the corresponding moment based on the distribution determination coefficient of the representative point of the pressure data at each moment; a data monitoring module, which is used to select the final representative point based on the confidence of the representative point of the pressure data at each moment, and cluster based on the representative point to monitor the mining water pump abnormality. It can improve the accuracy of data division, improve the accuracy of abnormal monitoring of mining water pumps and achieve timely early warning.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to an abnormal monitoring and early warning system for a mining water pump. Background Art

[0002] A mining water pump is a water pump used in mining, metallurgy, chemical industry and other industries. It is mainly used to transport water or other liquids containing particles or other media. Submersible electric pumps are commonly used in the coal mining industry, and the performance of electric pumps directly affects the safety of mining. Nowadays, with the rapid development of science and technology, the abnormal detection of mining water pumps during use is constantly improving. Due to the complexity of the application scenarios, the water pump may experience excessive temperature, excessive current, and abnormal flow. The occurrence of these faults will endanger the safety of on-site construction personnel.

[0003] At present, through the development of computer and other related technologies, an online monitoring system is used to monitor the abnormal warning of mining water pumps. Through this system, the abnormalities of mining water pumps can be monitored in real time, reducing the risk of damage caused by abnormal pumps. However, when using the traditional CURE clustering algorithm to monitor the abnormality of mining water pumps, the process of selecting representative points before data clustering only considers the distance between data, so the selection of representative points cannot better represent the shape of the cluster, resulting in inaccurate data division and low accuracy of data abnormality detection, causing delays in abnormal warning of mining water pumps. Summary of the invention

[0004] The present invention provides a mining water pump abnormality monitoring and early warning system, which can improve the accuracy of data division, improve the accuracy of mining water pump abnormality monitoring and achieve timely early warning.

[0005] In a first aspect, the present application provides a mining water pump abnormality monitoring and early warning system, comprising:

[0006] Data acquisition module, used to collect pressure data, temperature data and current data of mining water pumps at every moment;

[0007] A coefficient calculation module, which calculates the distribution determination coefficient of the representative point of the pressure data at the corresponding moment based on the temperature data and the current data at the same moment;

[0008] A confidence calculation module, which determines the confidence of the representative point of the pressure data at the corresponding moment based on the distribution determination coefficient of the representative point of the pressure data at each moment;

[0009] The data monitoring module is used to select the final representative point based on the confidence of the representative point of the pressure data at each moment, and cluster the representative points to monitor the abnormality of the mining water pump.

[0010] In an optional embodiment, the coefficient calculation module includes:

[0011] A change determination module, used to calculate the change of the nth temperature data and the nth current data;

[0012] An information entropy calculation module, used to calculate the information entropy of a K-neighborhood pressure data sequence of each pressure data in the pressure data sequence;

[0013] The distribution determination coefficient calculation module is used to calculate the distribution determination coefficient of the representative point of the pressure data at the corresponding moment based on the change of the nth temperature data and the nth current data and the information entropy of the K-neighborhood pressure data sequence of each pressure data.

[0014] In an optional embodiment, the change determination module is used to:

[0015] Calculating a first similarity between a K-neighborhood temperature data sequence of an n-th temperature data and a K-neighborhood temperature data sequence of an n+1-th temperature data in the temperature data sequence, and calculating a second similarity between a K-neighborhood temperature data sequence of an n-th temperature data and a K-neighborhood temperature data sequence of an n-1-th temperature data in the temperature data sequence; calculating a first difference between the first similarity and the second similarity, the first difference being used to characterize a change in the n-th temperature data;

[0016] Calculate the third similarity between the K-neighborhood pressure data sequence of the n-th current data in the pressure data sequence and the K-neighborhood pressure data sequence of the n+1-th current data, and calculate the fourth similarity between the K-neighborhood pressure data sequence of the n-th current data in the pressure data sequence and the K-neighborhood pressure data sequence of the n-1-th current data; calculate the second difference between the third similarity and the fourth similarity, and the second difference is used to characterize the change of the n-th current data.

[0017] In an optional embodiment, the distribution determination coefficient calculation module is used to calculate the distribution determination coefficient of the representative point of the pressure data at the corresponding moment using the following formula:

[0018] ;

[0019] in, The distribution determination coefficient of the representative point of the nth pressure data is represented by The information entropy of the K-neighborhood pressure data sequence of each pressure data, Indicates the change of the nth current data, Indicates the change of the nth temperature data, () represents exponential function.

[0020] In an optional embodiment, the confidence calculation module is further used to:

[0021] Calculating the mean value of the Euclidean distance between the nth pressure data in the pressure data set and each initial representative point in the initial representative point set of the pressure data, wherein the initial representative point set is determined from the collected pressure data, and the pressure data set includes the remaining pressure data of the collected pressure data except the data in the initial representative point set;

[0022] The confidence of the representative point of the pressure data at the corresponding moment is calculated based on each mean value and the distribution determination coefficient in the confidence determination set; wherein the confidence determination set is composed of the initial representative point of the pressure data and the first representative point in the pressure data set. The distribution determination coefficient of each pressure data is composed of

[0023] In an optional embodiment, the confidence calculation module is further used to calculate the confidence of the representative point of the pressure data at the corresponding moment using the following formula:

[0024] ;

[0025] in, represents the mean value of the Euclidean distance between the nth pressure data in the pressure data set and the initial representative point of the pressure data, Represents the confidence judgment set Middle The distribution determination coefficient corresponding to the value is Represents the confidence judgment set The mean of the distribution determination coefficient corresponding to the value in , Represents the confidence judgment set The number of values ​​in Indicates the confidence level of the representative point of the nth pressure data.

[0026] In an optional embodiment, the data monitoring module is further used to:

[0027] The point with the highest confidence in the initial representative point set is taken as the final representative point, and the final representative point is merged with the initial representative point as a new initial representative point set, and the next final representative point is calculated until the number of final representative points reaches the preset requirement.

[0028] In an optional embodiment, the data monitoring module is further used to:

[0029] The pressure data is clustered based on the final representative points using the CURE algorithm to divide the pressure data into different clusters.

[0030] In an optional embodiment, the data monitoring module is further used to:

[0031] Calculate the mean of the pressure data in each cluster;

[0032] Determine the water pump pressure judgment parameter corresponding to each cluster based on the mean value of the pressure data;

[0033] The water pump is monitored for abnormality based on the judgment parameters.

[0034] In an optional embodiment, if the judgment parameter is greater than a threshold, the data monitoring module determines that the water pump is abnormal.

[0035] The beneficial effects of the present application are different from the prior art. The present application is a mining water pump abnormality monitoring and early warning system, including: a data acquisition module, used to collect pressure data, temperature data and current data of the mining water pump at each moment; a coefficient calculation module, which calculates the distribution determination coefficient of the representative point of the pressure data at the corresponding moment based on the temperature data and current data at the same moment; a confidence calculation module, which determines the confidence of the representative point of the pressure data at the corresponding moment based on the distribution determination coefficient of the representative point of the pressure data at each moment; a data monitoring module, which is used to select the final representative point based on the confidence of the representative point of the pressure data at each moment, and cluster based on the representative point to monitor the mining water pump abnormality. It can improve the accuracy of data division, improve the accuracy of mining water pump abnormality monitoring and achieve timely early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a structural schematic diagram of an embodiment of a mining water pump abnormality monitoring and early warning system of the present invention;

[0037] Figure 2 for Figure 1 A schematic structural diagram of an embodiment of a coefficient calculation module. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0039] When the mining water pump is in abnormal operation, it causes high cost consumption and poses a safety hazard. Therefore, it is necessary to monitor the operating status of the mining water pump. The traditional abnormal data detection algorithm based on CURE is used. Due to the different data distribution characteristics, the selection of representative points may be concentrated on both sides. In this case, the selected representative points cannot reflect the data set and cannot reflect the properties of the representative points. Therefore, in the process of selecting representative points, the density of each sample point in the data set is considered in the present application, and the representative points are weighted according to the size of the density. The information carried by the representative points can be reflected according to the density. Therefore, the selection of representative points can be more accurate, thereby making the division of the data set more accurate, improving the accuracy of abnormal data detection, and issuing early warnings for abnormal operation of mining water pumps in a timely manner. The present application is described in detail below in conjunction with the accompanying drawings and embodiments.

[0040] See also Figure 1 , Figure 1 The structure diagram of an embodiment of a mining water pump abnormality monitoring and early warning system of the present invention comprises: a data acquisition module 11, a coefficient calculation module 12, a confidence calculation module 13 and a data monitoring module 14.

[0041] The data acquisition module 11 is used to collect the pressure data, temperature data and current data of the mining water pump at every moment.

[0042] Specifically, the data acquisition module 11 collects multi-dimensional working parameters of the mining water pump during operation through sensors, obtains the time series data of the multi-dimensional monitoring parameters of the mining water pump, and uses it as the basic data for detecting abnormal status of the mining water pump. There are many monitoring parameters for mining water pumps. In the present invention, pressure, current, and temperature are selected, and pressure sensors, current sensors, and temperature sensors are used for data acquisition. The model of the sensor can be selected according to the actual situation. The sequence length of the pressure, current, and temperature data acquisition is n, where the data sequence length is Can be defined by oneself, in the present invention .

[0043] The data acquisition module 11 acquires the pressure, current, and temperature data of the mining water pump at the same time through sensors. The data sequence under three dimensions is analyzed, and the abnormal data under three dimensions are extracted to realize the abnormal analysis of the mining water pump. Based on the collected monitoring parameters of each mining water pump, the abnormality of the mining water pump can be analyzed according to the relationship between different parameters when the mining water pump is running at different times. Specifically, the data sequence under three dimensions can be used to construct an abnormality detection matrix for the mining water pump, which is:

[0044] ;

[0045] In the above formula Indicates the nth temperature data collected at the same time. Indicates the nth current data collected at the same time. represents the nth pressure data collected at the same time. That is, the above matrix It consists of pressure, current and temperature data collected at the same time.

[0046] The coefficient calculation module 11 calculates the distribution determination coefficient of the representative point of the pressure data at the corresponding time based on the temperature data and the current data at the same time.

[0047] When the flow rate of mining water pumps is abnormal, inaccurate monitoring will lead to reduced efficiency, equipment damage, and potential safety hazards. Based on the CURE clustering algorithm, the flow rate abnormality data of mining water pumps is monitored, and pressure sensors are selected to reflect the flow changes of mining water pumps. However, in the process of anomaly detection of pressure data, due to inappropriate selection of representative points, the division of the data set will be inaccurate, and ultimately the accuracy of anomaly detection will be low.

[0048] During the operation of the mining water pump, the value of the detected pressure data will decrease when the flow is insufficient. At the same time, due to the insufficient flow, the water flow in the mining water pump is small, the cooling effect of the water pump is reduced, and the overload of the water pump will cause the temperature of the water pump to be higher at this time. In the case of low flow, the load of the water pump is large, resulting in increased current, and because the temperature is difficult to dissipate, the temperature of the heat-prone coil in the pump increases, and the current may increase additionally. Based on the above characteristics, the relationship between the flow, temperature and current of the mining water pump can be analyzed, and the abnormal flow of the mining water pump can be further monitored.

[0049] Specifically, the collected temperature data forms a temperature data sequence, and the collected current data forms a current data sequence. These two types of data are affected by pressure anomalies to different degrees. Based on the relationship between pressure data and temperature and current data, a judgment model for the correlation between data can be constructed. The greater the correlation of information changes, the greater the contribution to flow data analysis.

[0050] In one embodiment, the coefficient calculation module includes: a change determination module 21, an information entropy calculation module 22 and a distribution determination coefficient calculation module 23. Figure 2 shown.

[0051] The change determination module 21 is used to calculate the change of the nth temperature data and the nth current data.

[0052] According to the above matrix K, the K-neighborhood temperature data sequence of the nth temperature data in the corresponding temperature data sequence can be obtained: And the K-neighborhood pressure data sequence of the nth current data in the current data sequence .

[0053] Specifically, the change determination module 21 is used to calculate the first similarity between the K-neighborhood temperature data sequence of the nth temperature data and the K-neighborhood temperature data sequence of the n+1th temperature data in the temperature data sequence. . First similarity The calculation method is: .

[0054] The change determination module 21 is used to calculate the second similarity between the K-neighborhood temperature data sequence of the nth temperature data and the K-neighborhood temperature data sequence of the n-1th temperature data in the temperature data sequence. The second similarity is calculated as: The change determination module 21 is further configured to calculate a first difference between the first similarity and the second similarity. ,Right now , the first difference is used to characterize the change of the nth temperature data. Indicates that ED edit distance is used to calculate similarity.

[0055] Specifically, the change determination module 21 is further used to calculate the third similarity between the K-neighborhood pressure data sequence of the nth current data and the K-neighborhood pressure data sequence of the n+1th current data in the pressure data sequence. , and calculating the fourth similarity between the K-neighborhood pressure data sequence of the nth current data in the pressure data sequence and the K-neighborhood pressure data sequence of the n-1th current data ; Calculate the second difference between the third similarity and the fourth similarity ,Right now , the second difference is used to characterize the change of the nth current data. Among them, the third similarity Similarity to the fourth The calculation method of is the same as the calculation method of the first similarity and the second similarity mentioned above, which will not be repeated here.

[0056] First Difference and the second difference Respectively represent The temperature data and The information changes of the current data.

[0057] In this embodiment, the information entropy calculation module 22 is used to calculate the information entropy of the K-neighborhood pressure data sequence of each pressure data in the pressure data sequence, wherein the pressure data sequence includes a plurality of the pressure data. Specifically, when the data is densely distributed, the information carried by the pressure data is richer, and the points in this part are selected as representative points. The k-nearest neighbor algorithm (k takes an empirical value of 10) is used, and the input is the data of the row where the pressure parameter is located in the matrix K, and the K-neighborhood pressure data sequence of each pressure data is obtained. The information entropy calculation module 22 calculates the K-neighborhood pressure data sequence Information entropy.

[0058] In addition to the increase in current and temperature caused by abnormal flow of mining water pumps, the increase in temperature may cause an additional increase in current, so abnormal changes in current are more sensitive. Based on the changes in temperature and current obtained at the same time, the distribution determination coefficient of the representative point of the pressure data can be constructed.

[0059] Specifically, the distribution determination coefficient calculation module 23 is used to calculate the distribution determination coefficient of the representative point of the pressure data at the corresponding moment based on the change of the nth temperature data and the nth current data and the information entropy of the pressure data sequence. The distribution determination coefficient calculation module is used to calculate the distribution determination coefficient of the representative point of the pressure data at the corresponding moment using the following formula:

[0060] ;

[0061] in, The distribution determination coefficient of the representative point of the nth pressure data is represented by The information entropy of the K-neighborhood pressure data sequence of each pressure data, Indicates the change of the nth current data, Indicates the change of the nth temperature data, () represents exponential function.

[0062] In the above formula, the information carried by the pressure data is calculated based on the information entropy of the data in the K neighborhood of the pressure data and the corresponding changes in temperature and current caused by the abnormal flow of the water pump, that is, the distribution determination coefficient of the representative point of the pressure data When the data distribution is concentrated and the pump is working normally at this moment, the calculated The smaller the value of is, the more consistent the changes in current and temperature are, that is, The smaller the value of The smaller the value of is. When the data distribution is discrete and the water pump is working abnormally at this moment, The larger the value of The larger the value of The larger the value is. At this time, each pressure data can get the distribution determination coefficient of the representative point. When selecting the representative point, the information characteristics carried by the data are considered. Therefore, some of the pressure data can be randomly selected as the initial representative points. The initial representative point set is .

[0063] The confidence calculation module 13 determines the confidence of the representative point of the pressure data at each moment based on the distribution determination coefficient of the representative point of the pressure data at the corresponding moment.

[0064] In one embodiment, the confidence calculation module 13 is further used to calculate the mean of the Euclidean distance between the nth pressure data in the pressure data set and each initial representative point in the initial representative point set of the pressure data, wherein the initial representative point set is determined from the collected pressure data, and the pressure data set includes the remaining pressure data of the collected pressure data except the data in the initial representative point set. Specifically, some of the pressure data are randomly selected from the collected pressure data as initial representative points, and the initial representative point set composed of the initial representative points is The pressure data collected, except for the data in the initial representative point set, constitute the pressure data set. Calculate the relationship between each initial representative point in the initial representative point set and the first point in the pressure data set. The Euclidean distance of the pressure data set is calculated based on the The mean value is obtained by calculating the average of the Euclidean distances between each pressure data and all initial representative points. .

[0065] The confidence calculation module 13 determines the set based on each mean and confidence The confidence of the representative point of the pressure data at the corresponding moment is calculated using the distribution determination coefficient in; wherein the confidence determination set is composed of the initial representative point of the pressure data and the first representative point in the pressure data set. The distribution determination coefficient of each pressure data is composed of

[0066] In one embodiment, the confidence calculation module 13 is further used to calculate the confidence of the representative point of the pressure data at the corresponding moment using the following formula:

[0067] ;

[0068] in, represents the mean value of the Euclidean distance between the nth pressure data in the pressure data set and the initial representative point of the pressure data, Represents the confidence judgment set Middle The distribution determination coefficient corresponding to the value is Represents the confidence judgment set The mean of the distribution determination coefficient corresponding to the value in , Represents the confidence judgment set The number of values ​​in Indicates the confidence level of the representative point of the nth pressure data.

[0069] Since the selection of representative points directly affects the quality of clustering effect, in this embodiment, when selecting representative points, the distribution determination coefficient and distance of the data are used as the confidence of representative point selection, which not only considers the distribution characteristics of the sample points, but also reduces the impact of abnormal points.

[0070] The data monitoring module 14 is used to select a final representative point based on the confidence of the representative point of the pressure data at each moment, and to perform clustering based on the representative point to perform abnormal monitoring on the mining water pump.

[0071] Specifically, when The farther the pressure data is from all the initial representative points, the The smaller the difference between the distribution determination coefficient of each pressure data and the initial representative point, that is, The smaller the value of The larger the value of The confidence level of the representative point of the pressure data The larger the value is, the more suitable the data is for selection as a representative point. Based on this, the data monitoring module 14 selects the point with the highest confidence in the initial representative point set as the final representative point.

[0072] The selected final representative point and the initial representative point are combined as a new set of initial representative points, and the above steps are repeated: the mean of the Euclidean distances between the nth pressure data in the pressure data set and each initial representative point in the initial representative point set of the pressure data is calculated, so as to screen the next final representative point until the number of final representative points obtained reaches the preset requirement.

[0073] In one embodiment, assuming that the required number of final representative points is 10, after selecting 10 final representative points, the number of clusters is set to r, the shrinkage factor is 0.5, and the CURE algorithm is used to cluster the pressure data based on the final representative points, and the pressure data is divided into different clusters. According to the pre-set number of clusters, r different clusters are finally obtained.

[0074] The data monitoring module 14 further calculates the mean of the pressure data in each cluster and obtains [ ], represents the mean value of the pressure data of the rth cluster. The water pump pressure judgment parameter corresponding to each cluster is determined based on the mean value of the pressure data. Specifically, when the mine water pump is in normal operation, its pressure data is , thus the difference between the mean value of pressure data and the normal value of pressure data in each cluster of data division can be calculated Based on this difference and pressure data Calculate the water pump pressure judgment parameter. The specific calculation method is:

[0075] ;

[0076] In the above formula Indicates the pressure data divided into The mean of the values ​​within the cluster, is the pressure data value when the mining water pump is operating normally, calculated The pressure data is divided into The water pump pressure judgment parameters within a cluster.

[0077] Based on the judgment parameter, the water pump is monitored for abnormality. Specifically, a threshold value is set, for example, 0.2. When the pressure judgment parameter When it is greater than the threshold, the mine water pump is judged to be operating abnormally, and an abnormal mine water pump operation warning is issued at the control terminal.

[0078] The present invention analyzes the abnormal flow of the water pump by collecting the pressure, current and temperature parameters of the mining water pump at the same time. The representative point confidence is constructed according to the corresponding relationship between the pressure change and the current and temperature parameter changes when the mining water pump has abnormal flow. The representative point confidence takes into account the relationship between the change of pressure data caused by abnormal flow and the change of current and temperature, and can more accurately select data with more information in the pressure data as representative points. Based on the representative point confidence, representative points in the pressure data are selected, and more appropriate representative points can be selected to cluster the data, thereby improving the accuracy of data division, improving the accuracy of abnormal monitoring of mining water pumps, and achieving timely warning.

[0079] The above are only implementation methods of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A mining water pump abnormality monitoring and early warning system, characterized in that: include: Data acquisition module, used to collect pressure data, temperature data and current data of mining water pumps at every moment; A coefficient calculation module, which calculates the distribution determination coefficient of the representative point of the pressure data at the corresponding moment based on the temperature data and the current data at the same moment; A confidence calculation module, which determines the confidence of the representative point of the pressure data at the corresponding moment based on the distribution determination coefficient of the representative point of the pressure data at each moment; A data monitoring module, used to select a final representative point based on the confidence of the representative point of the pressure data at each moment, and cluster the representative points to perform abnormal monitoring on the mining water pump; The coefficient calculation module includes: A change determination module, used to calculate the change of the nth temperature data and the nth current data; An information entropy calculation module, used to calculate the information entropy of a K-neighborhood pressure data sequence of each pressure data in the pressure data sequence; A distribution determination coefficient calculation module, used to calculate the distribution determination coefficient of the representative point of the pressure data at the corresponding moment based on the change of the nth temperature data and the nth current data and the information entropy of the K-neighborhood pressure data sequence of each pressure data; The change determination module is used for: Calculating a first similarity between a K-neighborhood temperature data sequence of an n-th temperature data and a K-neighborhood temperature data sequence of an n+1-th temperature data in the temperature data sequence, and calculating a second similarity between a K-neighborhood temperature data sequence of an n-th temperature data and a K-neighborhood temperature data sequence of an n-1-th temperature data in the temperature data sequence; calculating a first difference between the first similarity and the second similarity, the first difference being used to characterize a change in the n-th temperature data; Calculating a third similarity between a K-neighborhood pressure data sequence of an n-th current data and a K-neighborhood pressure data sequence of an n+1-th current data in the pressure data sequence, and calculating a fourth similarity between a K-neighborhood pressure data sequence of an n-th current data and a K-neighborhood pressure data sequence of an n-1-th current data in the pressure data sequence; calculating a second difference between the third similarity and the fourth similarity, wherein the second difference is used to characterize a change in the n-th current data; The distribution determination coefficient calculation module is used to calculate the distribution determination coefficient of the representative point of the pressure data at the corresponding moment using the following formula: ； in, The distribution determination coefficient of the representative point of the nth pressure data is represented by The information entropy of the K-neighborhood pressure data sequence of each pressure data, Indicates the change of the nth current data, Indicates the change of the nth temperature data, () represents exponential function; The confidence calculation module is also used for: Calculating the mean value of the Euclidean distance between the nth pressure data in the pressure data set and each initial representative point in the initial representative point set of the pressure data, wherein the initial representative point set is determined from the collected pressure data, and the pressure data set includes the remaining pressure data of the collected pressure data except the data in the initial representative point set; The confidence of the representative point of the pressure data at the corresponding moment is calculated based on each mean value and the distribution determination coefficient in the confidence determination set; wherein the confidence determination set is composed of the initial representative point of the pressure data and the first representative point in the pressure data set. The distribution determination coefficient of each pressure data is composed of; The confidence calculation module is also used to calculate the confidence of the representative point of the pressure data at the corresponding moment using the following formula: ； in, represents the mean value of the Euclidean distance between the nth pressure data in the pressure data set and the initial representative point of the pressure data, Represents the confidence judgment set Middle The distribution determination coefficient corresponding to the value is Represents the confidence judgment set The mean of the distribution determination coefficient corresponding to the value in , Represents the confidence judgment set The number of values ​​in Indicates the confidence level of the representative point of the nth pressure data.

2. The abnormal monitoring and early warning system for mining water pumps according to claim 1 is characterized in that: The data monitoring module is also used for: The point with the highest confidence in the initial representative point set is taken as the final representative point, and the final representative point is merged with the initial representative point as a new initial representative point set, and the next final representative point is calculated until the number of final representative points reaches the preset requirement.

3. The abnormal monitoring and early warning system for mining water pumps according to claim 2 is characterized in that: The data monitoring module is also used for: The pressure data is clustered based on the final representative points using the CURE algorithm to divide the pressure data into different clusters.

4. The abnormal monitoring and early warning system for mining water pumps according to claim 3 is characterized in that: The data monitoring module is also used for: Calculate the mean of the pressure data in each cluster; Determine the water pump pressure judgment parameter corresponding to each cluster based on the mean value of the pressure data; The water pump is monitored for abnormality based on the judgment parameters.

5. The abnormal monitoring and early warning system for mining water pumps according to claim 4 is characterized in that: If the judgment parameter is greater than the threshold, the data monitoring module determines that the water pump is abnormal.

Images