Detection method and device for mining middle-high voltage direct current power supply
By constructing a comprehensive impact index and power supply reliability coefficient, combining the clustering algorithm to extract the unreliable moment of power supply, and only calculate the load current LOF value at that moment, it solves the problem of slow speed of the traditional LOF abnormality detection algorithm and improves the high-voltage DC power supply fault detection efficiency.
Patent Information
- Application Number
- CN202411533690.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-31
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-10-31
AI Technical Summary
The traditional LOF abnormality detection algorithm runs slowly in high-voltage DC power supply fault detection, resulting in low abnormality fault detection efficiency for mining high-voltage DC power supply.
By collecting the load current and voltage data of the high-voltage DC power supply, a comprehensive impact index and power supply reliability coefficient are constructed, combined with the clustering algorithm to extract the unreliable moment of power supply, and only calculate the load current LOF value at that moment for abnormal detection.
It improves the efficiency of abnormal fault detection of high-voltage DC power supply, avoids the need for LOF abnormal detection algorithm to calculate LOF values at all times, and significantly improves the detection speed.
Smart Images

Figure CN119044816A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of fault detection, and in particular to a method and device for detecting a medium- and high-voltage direct current power supply for mining. Background Art
[0002] With the development of industrial and mining enterprises and the simultaneous exploitation of mining resources, the safety of high-voltage power supplies in mines has attracted increasing attention. High-voltage DC power supplies (HVDCs) are powerful power supplies used for both voltage and current regulation. Due to their high stability and energy storage, HVDCs are widely used in high-voltage power supply systems in mines. However, mining operations often require a large amount of high-power production equipment, consuming significant amounts of electricity. This can easily damage HVDC power supplies due to prolonged high-load operation, leading to abnormal failures.
[0003] With the continuous development of machine learning technology, real-time monitoring of the load current and load voltage of high-voltage DC power supplies and the use of anomaly detection technology can accurately identify abnormal faults during the use of high-voltage DC power supplies. For example, the common LOF (Local Outlier Factor) anomaly detection algorithm can accurately identify abnormal faults during the use of high-voltage DC power supplies. However, traditional LOF anomaly detection algorithms require traversing the entire data set to calculate the LOF value for each sample point, making the LOF anomaly detection algorithm slow and resulting in low efficiency in detecting abnormal faults in mining high-voltage DC power supplies. Summary of the Invention
[0004] In order to solve the above technical problems, the purpose of this application is to provide a detection method and device for a medium- and high-voltage DC power supply for mining. The technical solutions adopted are as follows: In a first aspect, an embodiment of the present application provides a method for detecting a medium- and high-voltage DC power supply for a mine, the method comprising the following steps: Collect the load current and load voltage at each moment during the operation of the high-voltage DC power supply, and build a load current monitoring sequence and a load voltage monitoring sequence; Construct a comprehensive impact index at each acquisition moment based on the data changes in the load current monitoring sequence and the load voltage monitoring sequence; Based on the difference between the comprehensive impact indexes at different collection times, the harmonic component action intensity at each collection time is calculated, and the power supply reliability coefficient at each collection time is determined; The power supply unreliable moment is obtained based on the power supply reliability coefficient combined with the clustering algorithm, and the high-voltage DC power supply anomaly detection is performed based on the abnormal degree of load current at the moment of unreliable power supply.
[0005] In one embodiment, the process of obtaining the comprehensive impact index is as follows: Based on the load current time series, a long-term characteristic sequence of each load current is constructed; based on the load voltage monitoring sequence, the long-term characteristic sequence of each load voltage is obtained using the same acquisition method as the long-term characteristic sequence of the current; The long-term high-load characteristic value of the load current at each acquisition moment is calculated based on the data dispersion and overall value in the long-term characteristic sequence of current; the load severity characteristic value of the load voltage at each acquisition moment is calculated based on the data changes in the long-term characteristic sequence of voltage; The fusion value of the long-term high load characteristic value of the load current at each acquisition moment and the load severity characteristic value of the load voltage is used as the comprehensive impact index at each acquisition moment.
[0006] In one embodiment, the process of acquiring the long-term characteristic sequence of current is as follows: A sliding window of each load current is set with each load current in the load current monitoring sequence as the center; and a sequence composed of all load currents in the sliding window is used as a long-term characteristic sequence of each load current.
[0007] In one embodiment, the process of obtaining the long-term high load characteristic value is as follows: Calculate the mean and dispersion of all elements in the long-term characteristic sequence of the load current at each acquisition moment, and record them as the first current mean and the first dispersion respectively; The long-term high-load characteristic value of the load current at each acquisition moment is positively correlated with the first current mean value and negatively correlated with the first dispersion.
[0008] In one embodiment, the process of obtaining the load severity characteristic value is as follows: Calculate the range value within the long-term voltage characteristic sequence of the load voltage at each acquisition moment; calculate the mean of all elements in the long-term voltage characteristic sequence of the load voltage at each acquisition moment, and record it as the first voltage mean; the load severity characteristic value of the load voltage at each acquisition moment is positively correlated with the range value and negatively correlated with the first voltage mean.
[0009] In one embodiment, the process of obtaining the harmonic component intensity is as follows: A sequence consisting of comprehensive influence indexes of a preset first number of collection moments closest to any collection moment is used as a parallel influence feature sequence of the any collection moment; The difference between any two adjacent elements in the parallel influence feature sequence at each acquisition moment is calculated and recorded as the first difference. The mean of all elements in the parallel influence feature sequence at each acquisition moment is calculated and recorded as the second mean. The intensity of the harmonic component at each acquisition moment is positively correlated with the second mean and negatively correlated with the first difference.
[0010] In one embodiment, the power supply reliability coefficient is a calculation result of an exponential function with a natural constant as the base and a negative number of the harmonic component action intensity at each acquisition moment as the exponent.
[0011] In one embodiment, the process of obtaining the time when power supply is unreliable is as follows: The power supply reliability coefficients at all acquisition moments are used as input to the clustering algorithm, and each cluster is output. The mean of all elements in each cluster is calculated and recorded as the third mean. The acquisition moments corresponding to all elements in the cluster with the smallest third mean are taken as the power supply unreliable moments.
[0012] In one embodiment, the high-voltage DC power supply abnormality detection based on the abnormality of the load current at the time of unreliable power supply is specifically performed as follows: The load current at all moments of unreliable power supply is used as the input of the anomaly detection algorithm, and the output is the anomaly value of each abnormal load current; the average of the anomaly values of all abnormal load currents is used as the abnormal fault metric value. If the abnormal fault metric value is greater than the abnormal fault judgment threshold, then the high-voltage DC power supply has an abnormal fault; otherwise, the high-voltage DC power supply has not an abnormal fault.
[0013] In a second aspect, an embodiment of the present application also provides a detection device for a medium- and high-voltage DC power supply for mining, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of any one of the above methods when executing the computer program.
[0014] The embodiments of the present application have at least the following beneficial effects: The present application realizes the detection of abnormal faults occurring during the operation of the mine high-voltage DC power supply using the LOF anomaly detection algorithm. Compared with the traditional LOF anomaly detection algorithm, which needs to traverse the entire data set to calculate the LOF value of each sample point, the LOF anomaly detection algorithm runs slowly, resulting in the problem of low efficiency in abnormal fault detection of the mine high-voltage DC power supply. The present application combines the severity of the load generated by the mine high-voltage DC power supply when multiple high-power production equipment runs in parallel during mining to construct a comprehensive impact index for each acquisition moment, calculates the impact of harmonic components, constructs a power supply reliability coefficient and extracts the time when power supply is unreliable, and then uses the LOF anomaly detection algorithm to detect abnormal faults of the load current at the time of unreliable power supply. In the present application, only the LOF value of the load current at the time of unreliable power supply is calculated, avoiding the LOF anomaly detection algorithm from calculating the LOF value of the load current at all times, solving the problem of slow running speed of the LOF anomaly detection algorithm and improving the efficiency of abnormal fault detection of the mine high-voltage DC power supply. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.
[0016] Figure 1 A flowchart of a method for detecting a medium- and high-voltage DC power supply for mining provided in one embodiment of the present application; Figure 2 Schematic diagram of k-means clustering results; Figure 3 Schematic diagram of the process of obtaining the comprehensive impact index. DETAILED DESCRIPTION
[0017] To further illustrate the technical means and effectiveness of this application to achieve the intended invention objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a method and device for detecting a medium- and high-voltage DC power supply for mining, including its specific implementation, structure, features, and effectiveness. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
[0019] The following describes in detail a method and device for detecting a medium- and high-voltage DC power supply for mining provided by the present application with reference to the accompanying drawings.
[0020] See also Figure 1 , which shows a flowchart of a method for detecting a medium- and high-voltage DC power supply for mining provided by an embodiment of the present application, the method comprising the following steps: Step S1: Using sensors to measure the load current and load voltage during the operation of the high-voltage direct current power supply, and performing preprocessing to obtain a load current monitoring sequence and a load voltage monitoring sequence.
[0021] The purpose of this solution is to improve the efficiency of abnormal fault detection of mine-used high-voltage DC power supplies by increasing the speed of the LOF anomaly detection algorithm, which is conducive to timely repair of abnormal faults caused by the high-voltage DC power supply, thereby avoiding safety hazards in the mining process.
[0022] (1) During the mining process, the load current and load voltage of the high-voltage DC power supply are collected in real time using a current sensor and a voltage sensor. In this embodiment, the collection time interval is 1 second and the collection time length is 30 minutes. It should be noted that the collection time interval and the collection time length can be set by the implementer, and this embodiment does not impose any specific restrictions.
[0023] (2) Preprocess the collected load current and load voltage: A sequence of all collected load currents in ascending time order is used as a load current time series, and a sequence of all collected load voltages in ascending time order is used as a load voltage time series.
[0024] (3) To avoid the influence of external interference signals on data acquisition, the load current time series and the load voltage time series are used as the input of the adaptive filtering algorithm, and the output of the adaptive filtering algorithm is used as the load current monitoring sequence and the load voltage monitoring sequence, respectively. The adaptive filtering algorithm is a well-known technology, and the specific process is not described in detail. It should be noted that this embodiment only provides one filtering algorithm for denoising. The implementer may also use other filtering algorithms for denoising, and this embodiment does not impose specific restrictions.
[0025] Step S2: constructing a comprehensive impact index at each acquisition moment based on data changes in the load current monitoring sequence and the load voltage monitoring sequence.
[0026] Typically, mining operations require a large number of high-power production equipment. Simultaneous operation of these devices requires significant amounts of power, which can cause the HVDC power supply to remain in a high-load state for extended periods. This means that the load current remains high for extended periods. To improve the efficiency of fault detection for HVDC power supplies used in mines, it is necessary to analyze the characteristics of these long-term high-load conditions.
[0027] (1) Based on the load current time series and the load voltage time series, the long-term characteristic sequence of each load current and the long-term characteristic sequence of each load voltage are constructed respectively: In order to analyze the long-term load characteristics of the high-voltage DC power supply at different acquisition times, each load current in the load current monitoring sequence is set as the center. A window of different sizes is used as a sliding window for each load current in the load current monitoring sequence; a sequence composed of all load currents in the sliding window of each load current in ascending time order is used as a long-term characteristic sequence of the current for each load current; It should be noted that when the load currents in the sliding window are less than 31, the load currents in the sliding window are completed by using the mean filling method; Similarly, for each load voltage in the load voltage monitoring sequence, the same acquisition method as the current long-term characteristic sequence is adopted to obtain the voltage long-term characteristic sequence of each load voltage in the load voltage monitoring sequence.
[0028] (2) Calculate the long-term high load characteristic value of the load current at each acquisition moment based on the data discreteness and overall numerical value in the long-term characteristic sequence of the current: Calculate the mean and dispersion of all elements in the long-term characteristic sequence of the load current at each acquisition moment, and record them as the first current mean and the first dispersion respectively; A long-term high-load characteristic value of the load current at each acquisition moment is determined based on the first current mean value and the first dispersion, wherein the long-term high-load characteristic value of the load current at each acquisition moment is positively correlated with the first current mean value and negatively correlated with the first dispersion.
[0029] Preferably, as an embodiment of the present application, the first dispersion may be the variance of all load currents in the long-term characteristic sequence of the load current at each acquisition moment; a preset error parameter is provided. It should be noted that the error parameter can be set by the implementer. In this embodiment, the error parameter is set to 1, and the sum of the first dispersion and the error parameter is calculated. The long-term high load characteristic value of the load current at each acquisition moment may be the ratio of the first current mean to the sum; In other embodiments of the present application, the first dispersion may be the standard deviation of all load currents in the long-term characteristic sequence of the load current at each acquisition moment.
[0030] The larger the first current mean value is, the more it can reflect that the load current is at a higher level at this time, and the stronger the high load feature of the high-voltage DC power supply is at this time. At the same time, the smaller the first discreteness is, the stronger the concentration of the load current in the local range is, that is, all load currents in the local range tend to the same level at this time. Therefore, when the load current is at a higher level at this time, and all load currents in the local range tend to the same level at this time, it can be said that the high-voltage DC power supply for mining is in a long-term high load state at this time, and the long-term high load characteristic value is larger.
[0031] (3) Calculate the load severity characteristic value of the load voltage at each acquisition moment based on the data changes in the long-term voltage characteristic sequence: Generally speaking, the deeper the high-voltage DC power supply is in a high-load operating state, the higher the load current is, and the greater the voltage drop characteristic generated by the higher level of load current through the load line is, that is, the load voltage shows a decreasing characteristic, and the stronger the load voltage decrease characteristic is, the heavier the load state of the high-voltage DC power supply is.
[0032] Calculate the extreme difference value in the long-term characteristic sequence of the load voltage at each acquisition moment; Calculate the mean of all elements in the long-term characteristic sequence of the load voltage at each acquisition moment, and record it as the first voltage mean; A load severity characteristic value of the load voltage at each acquisition moment is determined based on the extreme difference value and the first voltage mean value, wherein the load severity characteristic value of the load voltage at each acquisition moment is positively correlated with the extreme difference value and negatively correlated with the first voltage mean value.
[0033] Preferably, as an embodiment of the present application, the load severity characteristic value of the load voltage at each acquisition moment may be a ratio of the extreme difference value to the first voltage mean value; In other embodiments of the present application, an exponential function with a natural constant as the base and a negative number of the first voltage mean as the exponent is calculated and recorded as a first exponential function; the load severity characteristic value of the load voltage at each acquisition moment can be the product of the extreme difference value and the first exponential function.
[0034] The larger the range value in the long-term characteristic sequence of voltage, the stronger the characteristic of the decrease in load voltage in the local range, that is, the greater the voltage drop effect. At the same time, the smaller the first voltage mean value, reflecting that the average level of load voltage is lower after the voltage drop effect occurs in the load circuit of the high-voltage DC power supply. To a certain extent, it shows that the higher the severity of the load state of the high-voltage DC power supply, the greater the load severity characteristic value.
[0035] (4) Calculate the comprehensive impact index of each collection moment based on the long-term high load characteristic value and the load severity characteristic value: During the mining process, when the high-voltage DC power supply is in a high-load state for a long time, and the severity of the load state is higher, since multiple high-power production equipment require a large amount of power resources when running simultaneously during the mining process, the high-voltage DC power supply will be in a serious load state for a long time. Therefore, the greater the characteristic of the high-voltage DC power supply being in a serious load state for a long time, the more it can be explained that high-power production equipment in the mine is running simultaneously, and the high-voltage DC power supply is more likely to have an abnormal failure at this time.
[0036] The fusion value of the long-term high load characteristic value of the load current and the load severity characteristic value of the load voltage at each acquisition moment is used as the comprehensive impact index of each acquisition moment, wherein the comprehensive impact index of each acquisition moment is positively correlated with the long-term high load characteristic value of the load current and the load severity characteristic value of the load voltage at each acquisition moment, respectively.
[0037] Preferably, as an embodiment of the present application, the comprehensive impact index at each acquisition moment may be the product of the long-term high load characteristic value of the load current at each acquisition moment and the load severity characteristic value of the load voltage at each acquisition moment; In other embodiments of the present application, the comprehensive impact index at each acquisition moment may be the sum of the normalized value of the long-term high load characteristic value of the load current at each acquisition moment and the normalized value of the load severity characteristic value of the load voltage at each acquisition moment.
[0038] The larger the long-term high load characteristic value and the larger the load severity characteristic value, to a certain extent, it indicates that the high-voltage DC power supply is in a high-load state for a long time, and the higher the severity of the load state, the greater the impact of the simultaneous operation of high-power production equipment on the high-voltage DC power supply during mining operations, and the greater the comprehensive impact index, that is, the more likely the high-voltage DC power supply is to have an abnormal failure.
[0039] Step S3: calculating the harmonic component action intensity at each acquisition moment based on the difference between the comprehensive impact indices at different acquisition moments, and determining the power supply reliability coefficient at each acquisition moment based on the harmonic component action intensity at each acquisition moment.
[0040] In mines, the simultaneous and prolonged operation of multiple high-power production equipment can impact the reliability of the mine's high-voltage DC power supply. For example, when multiple high-power production equipment operate simultaneously for extended periods, they generate strong harmonic components. These strong harmonic components enter the high-voltage DC power supply system through the lines, causing fluctuations in the quality of the high-voltage DC power supply. This reduces power supply reliability and can easily create safety hazards. Therefore, to analyze the reliability of mine-used high-voltage DC power supply, it is necessary to continue the analysis based on the comprehensive impact index.
[0041] (1) A set consisting of a preset first number of collection moments closest to any collection moment is used as a neighboring time set of the said collection moment. It should be noted that the value of the first number can be set by the implementer. In this embodiment, the first number is set to 30; a sequence consisting of the comprehensive influence indexes of all collection moments in the neighboring time set of each collection moment in ascending time order is used as a parallel influence feature sequence of each collection moment.
[0042] (2) Calculate the harmonic component intensity at each acquisition moment based on the difference between adjacent elements in the parallel impact feature sequence: Calculate the difference between any two adjacent elements in the parallel influence feature sequence at each acquisition moment, and record it as the first difference; Calculate the mean of all elements in the parallel influence feature sequence at each acquisition moment, and record it as the second mean; The harmonic component intensity at each acquisition moment is determined based on all the first differences and the second mean values in the parallel impact feature sequence at each acquisition moment, wherein the harmonic component intensity at each acquisition moment is positively correlated with the second mean value and negatively correlated with the first difference.
[0043] Preferably, as an embodiment of the present application, the first difference may be the absolute value of the difference between any two adjacent elements in the parallel influence feature sequence at each acquisition moment; the average value of all the first differences in the parallel influence feature sequence at each acquisition moment is calculated, the sum of the average value of the first differences and a preset error parameter is used as the denominator, and the second mean value is used as the numerator; the harmonic component action intensity at each acquisition moment may be the ratio of the numerator to the denominator; the error parameter is used to prevent the denominator from being 0; In other embodiments of the present application, the first difference may be the square of the difference between any two adjacent elements in the parallel impact feature sequence at each acquisition moment.
[0044] The larger the second mean value is and the smaller the difference between adjacent elements in the parallel influence characteristic sequence is, the more likely it is that multiple large-scale mining production equipment will operate in parallel and continuously at the overall level within the local time range. During this process, each large-scale mining production equipment generates harmonic components, and multiple large-scale mining production equipment generate harmonic components at the same time, resulting in a superposition effect of harmonic components. The greater the intensity of the harmonic component, the more likely it is to damage the high-voltage DC power supply.
[0045] (3) The higher the intensity of the harmonic component, the more likely it is to damage the high-voltage DC power supply, thereby affecting the reliability of the high-voltage DC power supply. Therefore, the power supply reliability coefficient at each acquisition moment is determined based on the intensity of the harmonic component at each acquisition moment. The power supply reliability coefficient at each acquisition moment is negatively correlated with the intensity of the harmonic component at each acquisition moment.
[0046] Preferably, as an embodiment of the present application, the power supply reliability coefficient at each acquisition moment may be a calculation result of an exponential function with a natural constant as the base and a negative number of the harmonic component action intensity at each acquisition moment as the exponent; In other embodiments of the present application, the power supply reliability coefficient at each acquisition moment may be the inverse of the harmonic component action intensity at each acquisition moment.
[0047] The greater the intensity of the harmonic component, the greater the intensity of the harmonic component generated by the parallel operation of multiple high-power production equipment in mineral resource development. At this time, it is more likely to affect the quality of the high-voltage DC power supply, resulting in reduced reliability of the high-voltage DC power supply, and the smaller the power supply reliability coefficient.
[0048] Step S4: obtaining the power supply unreliable moment based on the power supply reliability coefficient in combination with the clustering algorithm, and performing high-voltage DC power supply abnormality detection based on the abnormality degree of the load current at the power supply unreliable moment.
[0049] By increasing the speed of the LOF anomaly detection algorithm, the efficiency of abnormal fault detection for mine high-voltage DC power supplies is improved.
[0050] (1) Based on the power supply reliability coefficient and clustering algorithm, the unreliable power supply moment is obtained: The power supply reliability coefficients at all acquisition moments are used as input to the k-means clustering algorithm. In this embodiment, the number of clusters is 2. It should be noted that the number of clusters can be set by the implementer and is not specifically limited in this embodiment. The metric distance used for clustering is Euclidean distance, and the output of the k-means clustering algorithm is 2 clusters. The k-means clustering algorithm is a well-known technology, and the specific process will not be repeated here. For the clustering of power supply reliability coefficients, this embodiment of the application only provides one clustering method. There are many existing clustering methods, and implementers can also use other clustering algorithms for clustering, which is not specifically limited in this application.
[0051] The k-means clustering result diagram is as follows Figure 2 shown.
[0052] The mean of all elements in each cluster is calculated and recorded as the third mean. Since the elements in the cluster with a smaller third mean represent the lower reliability of the mine high-voltage DC power supply, the collection time corresponding to all elements in the cluster with the smallest third mean is taken as the time of unreliable power supply.
[0053] (2) Detection of high-voltage DC power supply anomalies based on the abnormality of the load current at the time of unreliable power supply: In order to realize abnormal fault detection of the high-voltage DC power supply for mining, the load current at all times of unreliable power supply is used as the input of the LOF anomaly detection algorithm, wherein the distance neighborhood parameter is taken as 10, and the output of the LOF anomaly detection algorithm is the LOF value of each abnormal load current. The LOF anomaly detection algorithm is a well-known technology, and the specific process is not repeated here. It should be noted that the implementer of the distance neighborhood parameter can set it by himself, and this embodiment does not impose any specific restrictions.
[0054] The larger the LOF value of the abnormal load current, the higher the degree of abnormality of the abnormal load current. At this time, the more likely it is that the high-voltage DC power supply for mining has an abnormal fault. Therefore, the average of the LOF values of all abnormal load currents is used as the abnormal fault metric value. When the abnormal fault metric value is greater than the abnormal fault determination threshold, it is determined that the high-voltage DC power supply has an abnormal fault and the abnormal fault caused by the high-voltage DC power supply should be repaired. It should be noted that the abnormal fault determination threshold can be set by the implementer. In this embodiment, the abnormal fault determination threshold is set to 2. If the abnormal fault metric value is less than the abnormal fault determination threshold, it is determined that the high-voltage DC power supply has not an abnormal fault and does not need to be repaired.
[0055] The schematic diagram of the process of obtaining the comprehensive impact index is as follows: Figure 3 shown.
[0056] Based on the same inventive concept as the above method, an embodiment of the present application also provides a detection device for a medium- and high-voltage DC power supply for mining, comprising a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of any one of the above-mentioned methods for detecting a medium- and high-voltage DC power supply for mining are implemented.
[0057] In summary, the embodiment of the present application provides a detection method for a medium and high voltage DC power supply for mining. The present application realizes the use of a LOF anomaly detection algorithm to detect abnormal faults that occur during the operation of a high voltage DC power supply for mining. Compared with the traditional LOF anomaly detection algorithm, it is necessary to traverse the entire data set to calculate the LOF value of each sample point, which makes the LOF anomaly detection algorithm run slowly, resulting in low efficiency in detecting abnormal faults for the high voltage DC power supply for mining. The present application combines the severity of the load generated by the high voltage DC power supply for mining when multiple high-power production equipment runs in parallel during mining to construct a comprehensive impact index for each acquisition moment, calculate the impact of harmonic components, construct a power supply reliability coefficient and extract the time when power supply is unreliable, and then use the LOF anomaly detection algorithm to detect abnormal faults of the load current at the time when power supply is unreliable. In the present application, only the LOF value of the load current at the time of unreliable power supply is calculated, avoiding the LOF anomaly detection algorithm from calculating the LOF value of the load current at all times, solving the problem of slow running of the LOF anomaly detection algorithm, and improving the efficiency of abnormal fault detection for the high voltage DC power supply for mining.
[0058] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the above descriptions are of specific embodiments of the present application. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0059] The various embodiments in this application are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.
[0060] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present application shall be included in the scope of protection of the present application.
Claims
1. A method for detecting a medium- and high-voltage DC power supply for mining, characterized in that: The method comprises the following steps: Collect the load current and load voltage at each moment during the operation of the high-voltage DC power supply, and construct a load current monitoring sequence and a load voltage monitoring sequence; Based on the data changes in the load current monitoring sequence and the load voltage monitoring sequence, a comprehensive impact index of each acquisition moment is constructed; Based on the difference between the comprehensive impact indexes at different collection times, the harmonic component action intensity at each collection time is calculated, and the power supply reliability coefficient at each collection time is determined; The power supply unreliable moment is obtained based on the power supply reliability coefficient combined with the clustering algorithm, and the high-voltage DC power supply anomaly detection is performed based on the abnormal degree of load current at the moment of unreliable power supply.
2. A method for detecting a medium- and high-voltage DC power supply for mining as claimed in claim 1, characterized in that: The process of obtaining the comprehensive impact index is as follows: Based on the load current time series, a current long-term characteristic sequence of each load current is constructed; based on the load voltage monitoring sequence, the same acquisition method as the current long-term characteristic sequence is adopted to obtain the voltage long-term characteristic sequence of each load voltage; The long-term high load characteristic value of the load current at each acquisition moment is calculated based on the data discreteness and overall value size in the long-term characteristic sequence of current; the load severity characteristic value of the load voltage at each acquisition moment is calculated based on the data change in the long-term characteristic sequence of voltage; The fusion value of the long-term high load characteristic value of the load current at each acquisition moment and the load severity characteristic value of the load voltage is used as the comprehensive impact index at each acquisition moment.
3. A method for detecting a medium- and high-voltage DC power supply for mining as claimed in claim 2, characterized in that: The acquisition process of the current long-term characteristic sequence is as follows: A sliding window of each load current is set with each load current in the load current monitoring sequence as the center; and a sequence composed of all load currents in the sliding window is used as a current long-term characteristic sequence of each load current.
4. A method for detecting a medium- and high-voltage DC power supply for mining as claimed in claim 2, characterized in that: The process of obtaining the long-term high load characteristic value is as follows: Calculate the mean and dispersion of all elements in the long-term characteristic sequence of the load current at each acquisition moment, and record them as the first current mean and the first dispersion respectively; The long-term high-load characteristic value of the load current at each acquisition moment is positively correlated with the first current mean value, and negatively correlated with the first dispersion.
5. A method for detecting a medium- and high-voltage DC power supply for mining as claimed in claim 2, characterized in that: The process of obtaining the load severity characteristic value is as follows: The extreme difference value in the voltage long-term characteristic sequence of the load voltage at each acquisition moment is calculated; the mean value of all elements in the voltage long-term characteristic sequence of the load voltage at each acquisition moment is calculated and recorded as the first voltage mean value; the load severity characteristic value of the load voltage at each acquisition moment is positively correlated with the extreme difference value and negatively correlated with the first voltage mean value.
6. A method for detecting a medium- and high-voltage direct current power supply for mining as claimed in claim 1, characterized in that: The process of obtaining the intensity of the harmonic component is as follows: A sequence composed of comprehensive influence indexes of a preset first number of collection moments closest to any collection moment is used as a parallel influence feature sequence of the any collection moment; The difference between any two adjacent elements in the parallel influence feature sequence at each acquisition moment is calculated, which is recorded as the first difference; the mean of all elements in the parallel influence feature sequence at each acquisition moment is calculated, which is recorded as the second mean; the intensity of the harmonic component at each acquisition moment is positively correlated with the second mean, and negatively correlated with the first difference.
7. A method for detecting a medium- and high-voltage direct current power supply for mining as claimed in claim 1, characterized in that: The power supply reliability coefficient is a calculation result of an exponential function with a natural constant as the base and the negative number of the intensity of the harmonic component at each acquisition moment as the exponent.
8. A method for detecting a medium- and high-voltage DC power supply for mining as claimed in claim 1, characterized in that: The process of obtaining the time when power supply is unreliable is as follows: The power supply reliability coefficients of all collection moments are used as the input of the clustering algorithm, and each cluster is output; the mean of all elements in each cluster is calculated and recorded as the third mean, and the collection moment corresponding to all elements in the cluster with the smallest third mean is taken as the power supply unreliable moment.
9. A method for detecting a medium- and high-voltage direct current power supply for mining as claimed in claim 1, characterized in that: The abnormality detection of the high-voltage DC power supply based on the abnormality degree of the load current at the time when the power supply is unreliable is specifically as follows: The load current at all times when power supply is unreliable is used as the input of the abnormality detection algorithm, and the output is the abnormal value of each abnormal load current; the average of the abnormal values of all abnormal load currents is used as the abnormal fault metric value. If the abnormal fault metric value is greater than the abnormal fault judgment threshold, an abnormal fault occurs in the high-voltage DC power supply; otherwise, no abnormal fault occurs in the high-voltage DC power supply.
10. A detection device for a medium- and high-voltage DC power supply for mining, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.
Citation Information
Patent Citations
Green building energy consumption monitoring method and system based on Internet of Things
CN117294019A
Chip detection method and system based on FPGA technology
CN118228006A
Third-party load aggregation platform malicious behavior detection method and device, equipment and medium
CN118378191A
A method for identifying a fault event in an electric power distribution grid sector
EP3396802A1
Lithium battery data acquisition anomaly identification method and system, storage medium and device
WO2024212403A1
Cited By
Electrical engineering high-voltage line fault monitoring system
CN119335299A
Method for monitoring working state of power supply in high-temperature environment and high-temperature power supply
CN120595183A
Intelligent equipment management method and system
CN120634053A
Voltage fluctuation suppression method and system for mobile module energy storage
CN121238571A
