A Method and System for Optimizing the Operation of Power Grids in Coal Mining Based on Intelligent Analysis

Abstract

This invention discloses a method and system for optimizing the operation of a coal mine power grid based on intelligent analysis, belonging to the field of power system technology. The method includes: mapping the mine's power distribution network structure into a topology graph, which includes multiple nodes and edges. Nodes represent load devices, and edges represent power lines between nodes; determining the importance level of nodes and the weight of edges; treating each node and the edges connected to it as a connection network, sorting the connection network, and obtaining the priority of the load devices after sorting; predicting the energy storage capacity for multiple preset time periods based on the current energy storage capacity and power source; constructing an optimal operation plan based on the predicted energy storage capacity, priority, and power consumption of the load devices; and restoring power to the corresponding load devices according to the optimal operation plan if a mains power outage occurs in the mine. This solution can restore power to critical equipment and safety systems more promptly and orderly, ensuring normal operation of the mine.

Description

Technical Field

[0001] This invention belongs to the field of power system technology, specifically relating to a method and system for optimizing the operation of power grids in coal mines based on intelligent analysis. Background Technology

[0002] As a crucial infrastructure for mine production and safe operation, the stability and reliability of the coal mine power grid directly impact normal operations and personnel safety. However, due to the remote locations of mines and harsh power supply environments, power outages are frequent. Therefore, to ensure safe mine operation after a power outage, a backup power system is typically used to guarantee continuous power supply to critical mine equipment and safety systems. To ensure the quality of backup power, various methods for stabilizing power supply quality have been proposed in existing technologies. For example, Chinese patent document CN120073670A discloses a coal mine power supply system and a smart power generation method and system for resuming operation. This method deploys voltage monitoring sensors in the coal mine power supply system to collect voltage data in real time. Upon detecting voltage anomalies, it can adjust the energy storage power supply strategy in real time, ensuring continuous power supply to the coal mine under various power anomaly conditions while avoiding excessive or insufficient energy consumption.

[0003] However, the aforementioned existing technologies focus on maintaining mine power supply by adjusting energy storage power supply strategies in real time after voltage anomalies or power outages occur. These are reactive measures and lack advance prediction and analysis of grid operation status and energy storage capacity. Therefore, if reasonable power supply measures can be formulated before a power outage, the occurrence of abnormal situations can be further reduced. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a method and system for optimizing the operation of power grids in coal mines based on intelligent analysis, thereby resolving the issues present in the background art.

[0005] To achieve the aforementioned objectives, this invention proposes a method for optimizing the operation of power grids in coal mines based on intelligent analysis, comprising:

[0006] The power distribution network structure of the mine is mapped as a topology graph, which includes multiple nodes and edges. Nodes represent load devices, and edges represent power lines between nodes.

[0007] The importance level of nodes is determined based on expert knowledge, and the weight of edges is determined based on the length of the power line, the area it passes through, and the supporting structure.

[0008] Each node and the edges connected to it are considered as a connection network. The connection networks in the topology are sorted based on the importance level of the nodes and the weight of the edges. The sorted network includes the priority of the load devices.

[0009] At preset intervals, the current energy storage capacity of the independent power source in the mine and the power source of the energy storage capacity are obtained. Based on the current energy storage capacity and the power source, the energy storage capacity for multiple preset time periods in the future is predicted.

[0010] Based on the predicted energy storage capacity, priority, and power consumption of the load devices, the optimal operating scheme is constructed by extracting multiple load devices.

[0011] If the mine experiences a mains power outage, the optimal operating plan is obtained based on the preset time period in which the power outage occurs, and the power supply to the corresponding load equipment is restored using the stored energy based on the optimal operating plan.

[0012] Furthermore, obtaining the priority of the load devices after sorting includes the following steps:

[0013] Statistical features of the connection network are obtained, including temporal features of the number of edges and edge weights. A multidimensional feature vector of the connection network is constructed based on the importance level of the load device and the statistical features. The connection network is clustered based on the similarity between the multidimensional feature vectors to obtain multiple clusters.

[0014] Based on the multidimensional feature vectors of the connected networks, a comprehensive feature of the class groups is constructed. After normalization of the comprehensive feature, standard features are obtained. The standard features are then subjected to dimensionality reduction processing based on the PCA dimensionality reduction algorithm, and principal components with a cumulative contribution rate greater than a first threshold are selected. A projection matrix is ​​constructed based on the principal components. The standard features of each class group are projected onto the principal component space based on the projection matrix to obtain the score of each standard feature of the class group on each principal component. The scores are weighted and summed based on the contribution rate of the principal components to obtain the total score of the class group. The class groups are ranked based on the total score, and each connected network within the class group is assigned a priority based on the ranking.

[0015] Furthermore, based on the predicted energy storage capacity, priority, and power consumption of the load devices, the optimal operating scheme for constructing multiple load devices includes the following steps:

[0016] A preset power outage duration is set, and a production plan for the preset time period is obtained. Based on the production plan, production equipment and necessary equipment are determined from the load equipment. Production equipment is the load equipment required by the production plan, and necessary equipment is the equipment that must be powered in the production plan. The necessary power of the necessary equipment is determined according to the power outage duration, and the part of the energy storage power that is greater than the necessary power is defined as redundant power.

[0017] The optimal operating scheme is obtained based on the genetic algorithm. The chromosome in the genetic algorithm is a binary string, which represents whether the load device is powered or not. When constructing the chromosome, the chromosome is filled based on the necessary devices. The unfilled part of the chromosome is randomly filled by the production equipment. When filling, the non-necessary power of the production equipment is determined based on the power outage duration. Filling stops when the non-necessary power is greater than or equal to the redundant power, or when the total power consumption of the chromosome exceeds the total output power of the independent power supply.

[0018] The fitness function of the genetic algorithm is a weighted sum of a first score and a second score, wherein the first score is calculated based on the number of production equipment included in the chromosome, and the second score is calculated based on the priority of the production equipment included in the chromosome.

[0019] After the genetic algorithm meets the termination condition, the chromosome with the highest fitness score is obtained, and the optimal running scheme is determined based on the distribution of its binary values.

[0020] Furthermore, after determining the optimal operating scheme, historical operating data of the production equipment is obtained. Based on the historical operating data, the impact of an unexpected power outage on the independent power supply voltage of the production equipment during operation is predicted. If the impact value is greater than a second threshold, the corresponding production equipment is defined as a risky equipment. When the independent power supply supplies power to the risky equipment, its operating power is suppressed, or the risky equipment is not supplied with power.

[0021] Furthermore, based on the historical operating data, the prediction of the impact of an unexpected power outage on the independent power supply voltage during the operation of the production equipment includes the following steps:

[0022] The historical operating data includes the power grid environment data of the production equipment and the impact value of power failure on the independent power supply voltage. The power grid environment data includes the voltage distribution and power distribution in the power grid. The production equipment to be predicted is defined as the target equipment, and the corresponding power grid environment data is the first data.

[0023] Multiple historical operating data that are similar to the first data are obtained from the power grid environment data and defined as the second data. The impact values ​​corresponding to the second data are aggregated into a first sequence. The first element in the first sequence is defined as the first element, and the first element is used as the reference data of the first data.

[0024] Obtain third data, which is historical operating data that previously used the first element as reference data. Obtain the influence value of the third data and aggregate it into a second sequence. Calculate the first cumulative probability of the first element based on the first sequence. Calculate the second cumulative probability of each element in the second sequence. Locate the element in the second sequence whose second cumulative probability is closest to the first cumulative probability and replace the first element with this element as the prediction result. Repeat this step until all elements in the first sequence are corrected. Average the prediction results in the corrected first sequence to obtain the influence value of the target device.

[0025] Furthermore, the first score and the second score correspond to the first weight and the second weight, respectively. If the energy storage capacity is less than the third threshold, the first weight is set to be less than the second weight; otherwise, the first weight is set to be greater than the second weight.

[0026] Furthermore, in the initial population and the iterative population obtained after crossover and mutation, if the total power consumption of the chromosomes exceeds the total output power of the independent power supply, the binary values ​​of the chromosomes belonging to the production equipment are corrected to not supply power until the power constraint is met.

[0027] This invention also provides a coal mine power grid operation optimization system based on intelligent analysis. This system is used to implement the methods described above, and includes:

[0028] The mapping module maps the power distribution network structure of the mine into a topology graph, which includes multiple nodes and edges. Nodes represent load devices, and edges represent power lines between nodes.

[0029] The hierarchical module determines the importance level of nodes based on expert knowledge, determines the weight of edges based on the length of power lines, the areas they pass through, and the supporting structures, and treats each node and the edges connected to it as a connection network. Based on the importance level of nodes and the weight of edges, the connection networks in the topology graph are sorted, and the sorted results include the priority of load devices.

[0030] The prediction module acquires the current energy storage capacity of the independent power source in the mine and the power source of the energy storage capacity at preset time intervals, and predicts the energy storage capacity for multiple preset time periods in the future based on the current energy storage capacity and the power source.

[0031] The execution module extracts multiple load devices to construct the optimal operating plan based on the predicted energy storage capacity, priority, and power consumption of the load devices. If the mine experiences a mains power outage, the optimal operating plan is obtained according to the preset time period in which the power outage occurs, and the energy storage capacity is used to restore the power supply to the corresponding load devices based on the optimal operating plan.

[0032] Beneficial effects: This invention maps the mine power distribution network structure into a topology graph. By determining the importance level of nodes and considering the length of power lines, the areas they pass through, and the supporting structures to determine the weight of edges, the priority of load equipment can be obtained. This allows for the rational allocation of energy storage capacity based on priority when formulating power supply strategies, ensuring priority power supply to important equipment. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating the steps of a method for optimizing the operation of a coal mine power grid based on intelligent analysis, as described in this invention.

[0034] Figure 2 This is a schematic diagram of the structure of a coal mine power grid operation optimization system based on intelligent analysis according to the present invention. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0036] It is understood that the terms "first," "second," etc., used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are used only to distinguish one element from another. For example, without departing from the scope of this application, a first script may be referred to as a second script, and similarly, a second script may be referred to as a first script.

[0037] like Figure 1 As shown, a method for optimizing the operation of a coal mine power grid based on intelligent analysis includes:

[0038] S1: Map the power distribution network structure of the mine into a topology graph. The topology graph includes multiple nodes and edges. Nodes represent load devices, and edges represent power lines between nodes.

[0039] First, the load devices included in the mine are collected as nodes, and the power lines connecting these devices are identified as edges. For example, a mine has 3 ventilation fans A, B, and C, 2 main conveyor belts D and E, and 3 underground drainage pumps F, G, and H. These load devices are connected via independent or shared high- and low-voltage cables. After collecting the basic information of the equipment, the following nodes are obtained: A, B, C, D, E, F, G, and H, and the following edges are obtained: AB, BE, CDF, CGH, etc. The edges include their lengths, such as AD being 210 meters and BE being 185 meters. For the areas traversed, such as edge AB crossing the transport roadway, and for the supporting structures, such as edge AB being fixed by cable racks. By mapping the mine's power distribution network structure to a topology graph, the connection status of the load devices in the power distribution network can be visually displayed, facilitating subsequent priority ranking of the load devices.

[0040] S2: Determine the importance level of nodes based on expert knowledge, and determine the weight of edges based on the length of power lines, the areas they pass through, and the supporting structures.

[0041] S3: Treat each node and the edges connected to it as a connection network. Sort the connection networks in the topology graph based on the importance level of the nodes and the weight of the edges. After sorting, obtain the priority of the load devices.

[0042] The importance level of load equipment is determined through expert evaluation. For example, the importance level of load equipment is divided into 1-10 levels, with level 1 equipment being the most important, such as roadway ventilation fans and lighting equipment. The lower the importance level, the higher the priority of power supply. In the event of a mains power outage, priority should be given to supplying power to these equipment.

[0043] The weights of edges can be obtained as follows: First, a scoring rule is set, which includes setting an initial value of 1. For every 100 meters increase in the line length between two nodes, the initial value decreases by 0.1. Multiple hazardous areas are selected within the mine, and scores are assigned to these areas. The reduction in the initial value is determined based on the hazardous areas the power line passes through and their corresponding scores. For example, if the line between two nodes passes through a return air shaft, which is considered a hazardous area with a score of 0.1, the initial value is reduced by 0.1. The robustness score is determined based on the support structure used for the power line. If a cable tray is used, the initial value increases by 0.2. Under this rule, a smaller weight indicates a greater risk of problems with the power line. Therefore, when considering supplying power to the corresponding load equipment through this power line, the likelihood of unexpected power outages is higher. Power outages not only affect production but also the stability of the bus voltage and output power. Thus, the higher the risk of a power line supplying power to a load equipment, the lower the power supply priority for that load equipment.

[0044] To comprehensively assess the importance of nodes, i.e., the importance of load devices, by combining importance level and weight, this embodiment treats nodes and their surrounding edges as a whole connection network. The priority is determined by rating the connection network. The higher the priority, the higher the priority of power supply to the load devices included. When determining the priority, the importance level of the node, the weight of the power lines connected to the node, and the number of edges are used for calculation. The more edges there are, the higher the power supply redundancy of the load devices. The specific determination method will be introduced later.

[0045] S4: At preset intervals, obtain the current energy storage capacity of the independent power source in the mine, as well as the power source of the energy storage capacity, and predict the energy storage capacity for multiple preset time periods in the future based on the current energy storage capacity and power source.

[0046] The preset timeframe is one day, meaning the energy storage capacity of the independent power source is acquired and calculated every other day at 00:00 each day. Then, the future energy storage capacity for each hour within the next 24 hours is predicted, resulting in a prediction period of one hour. This prediction can be combined with production plans; for example, based on the production plan, it might be determined that an additional 300m³ of energy will be added in the 6th hour. 3 The gas, then the above 300m 3 The gas is converted into electricity, and this electricity is added to the energy storage capacity measured at 00:00 to obtain the energy storage capacity for the next 6 hours. In particular, if the independent power source has already reached full storage at 00:00, meaning that it cannot store more electricity after that, the energy storage capacity at 00:00 is directly used as the predicted energy capacity.

[0047] S5: Based on the predicted energy storage capacity, priority, and power consumption of the load devices, extract multiple load devices to construct the optimal operating scheme.

[0048] S6: If the mine experiences a mains power outage, the optimal operating plan will be obtained based on the preset time period in which the power outage occurs, and the power supply to the corresponding load equipment will be restored using the stored energy based on the optimal operating plan.

[0049] The optimal operating plan includes multiple load devices that require power, such as load devices 1 to 3. If a mains power outage occurs, the optimal operating plan will be followed to restore power to load devices 1 to 3 in the mine using an independent power source.

[0050] Based on the above description, the process of this embodiment is as follows: First, the actual energy storage capacity is obtained at 00:00 every day. Then, the energy storage capacity for each hour is predicted according to the production plan. The optimal operating plan for each hour is generated by combining the predicted energy storage capacity for each hour. For example, the optimal operating plan 1 for 15:00 to 16:00 is generated based on the predicted energy storage capacity at 15:00. The optimal operating plan 1 includes load devices 1 to 3. The optimal operating plan 2 for 16:00 to 17:00 is generated based on the predicted energy storage capacity at 16:00. The optimal operating plan 2 includes load devices 4 to 6. If a mains power outage occurs between 15:00 and 16:00, the optimal operating plan 1 is used to restore the power supply to load devices 1 to 3. If a mains power outage occurs between 16:00 and 17:00, the optimal operating plan 2 is used to restore the power supply to load devices 4 to 6.

[0051] This invention maps the mine's power distribution network structure into a topology graph. By determining the importance level of nodes and considering the length of power lines, the areas they pass through, and the supporting structures to determine the weight of edges, the priority of load equipment is obtained. This allows for the rational allocation of energy storage capacity based on priority when formulating power supply strategies, ensuring priority power supply to critical equipment.

[0052] By acquiring the current energy storage capacity of independent power sources within the mine at preset intervals, and combining this with the power source to predict the energy storage capacity for multiple preset time periods in the future, the optimal operating plan can be made more reasonable, reducing the risk of power outages.

[0053] In the event of a mains power outage, the system can quickly restore power to the corresponding load equipment using stored energy according to a pre-defined plan. Compared to the reactive response of existing technologies, this solution can restore power to critical equipment and safety systems more promptly and systematically, reducing the time and scope of production interruptions and ensuring normal operation of the mine.

[0054] In this embodiment, obtaining the priority of the load devices after sorting includes the following steps:

[0055] Statistical features of the connected network are obtained, including the temporal features of the number of edges and edge weights. Multidimensional feature vectors of the connected network are constructed based on the importance level of the load devices and the statistical features. The connected network is clustered based on the similarity between the multidimensional feature vectors to obtain multiple groups.

[0056] The comprehensive features of the class groups are constructed based on the multidimensional feature vectors of the connection networks. After normalization, standard features are obtained. The standard features are then reduced in dimensionality using the PCA dimensionality reduction algorithm. Principal components with a cumulative contribution rate greater than a first threshold are selected. A projection matrix is ​​constructed based on the principal components. The standard features of each class group are projected onto the principal component space based on the projection matrix to obtain the score of each standard feature of the class group on each principal component. The scores are weighted and summed based on the contribution rate of the principal components to obtain the total score of the class group. The class groups are ranked based on the total score, and each connection network within the class group is assigned a priority based on the ranking.

[0057] To reduce computational load and avoid excessive prioritization of load devices after sorting, this embodiment uses the following method to determine priorities. First, statistical features of edges in each connected network are obtained. These features include the number of edges in the network and the temporal features of edge weights. Temporal features include the maximum, minimum, mean, standard deviation, peak-to-peak value, and root mean square value of the weights. All or some of these temporal features can be selected as statistical features. For example, if five temporal features are selected, along with the number of edges and the importance level of nodes, a total of seven features are generated, resulting in a 1*7 matrix of multidimensional feature vectors.

[0058] Next, the similarity between matrices is used as an indicator to cluster the connection network. In this embodiment, the Euclidean distance between matrices is used as the similarity metric; the smaller the Euclidean distance, the greater the similarity between the two matrices. This embodiment uses the K-means clustering algorithm to cluster the connection network, and the number of clusters is determined using the elbow rule. Through clustering, matrices with the same similarity are grouped into one cluster, that is, multiple similar connection networks are grouped into one cluster.

[0059] After that, by sorting the groups and assigning the priority of the sorted groups to each of the connected networks they include, the priority of each load device can be obtained. For example, if a group has a priority of 1 and includes 10 connected networks, then the priority of all 10 connected networks is 1.

[0060] The specific method for sorting the groups is as follows: First, a comprehensive feature of the groups is constructed. This comprehensive feature includes the average importance level of nodes within the group, the average number of edges in each connection network, and various temporal features of the edge weights of the connection networks. In other words, the various features of the groups are statistically analyzed on a connection network-by-connection basis, using the same statistical method as for the connection networks. Next, the standard features are normalized to eliminate dimensional differences between features. Then, the PCA algorithm is used to reduce the dimensionality of the comprehensive features, and features with a cumulative contribution rate of 95% are selected as principal components (i.e., the first threshold is 95%). The PCA method for dimensionality reduction is an existing technique and will not be described further here. A projection matrix is ​​constructed based on the selected principal components. The standard features are projected onto the principal component space using the projection matrix. This method allows us to obtain the score of each standard feature on the principal components. For example, if three principal components are selected, standard feature 1 scores 0.25 on principal component 1, 0.3 on principal component 2, and 0.4 on principal component 3. The same applies to other standard features.

[0061] In weighted summation, the corresponding weights are determined based on the contribution rates of the principal components. For example, if the contribution rates of principal components 1 to 3 are 0.5, 0.4, and 0.05 respectively, then the corresponding weights are 0.53, 0.42, and 0.05 respectively. Therefore, the total score of standard feature 1 for a certain group is approximately 0.53*0.25 + 0.42*0.3 + 0.05*0.4 ≈ 0.28. The calculation method for standard features is similar. Finally, the total scores of all standard features are summed to obtain the total score of the group. The groups are then ranked according to the total score, and the priority is determined based on the ranking result.

[0062] This method assigns a subset of load devices the same priority. For example, if there are 100 load devices, they will be classified into only 10 priority levels instead of 100. The purpose is that when generating the optimal operating plan, if multiple factors such as the power consumption of the load devices and their impact on the stability of the independent power supply need to be considered, fewer priority levels help improve the decision-making efficiency for the optimal operating plan.

[0063] In this embodiment, the process of extracting the optimal operating scheme from multiple load devices based on the predicted energy storage capacity, priority, and power consumption of the load devices includes the following steps:

[0064] The system presets the power outage duration, obtains the production plan for the preset time period, and determines the production equipment and necessary equipment among the load equipment based on the production plan. The production equipment is the load equipment required by the production plan, and the necessary equipment is the equipment that must be powered in the production plan. The system determines the necessary power of the necessary equipment according to the power outage duration, and defines the portion of the energy storage power that is greater than the necessary power as the redundant power.

[0065] The preset power outage duration is determined based on historical power outage records. For example, the average of historical power outage durations can be used as the preset outage duration. The mine's production plan includes the mining or production operations to be carried out. Based on the production plan, the required load equipment can be determined. Necessary equipment is equipment designated by personnel and involved in personal safety. Production equipment is load equipment other than necessary equipment. For example, if production operations require ore hoists, and it is also necessary to ensure power supply to the corresponding mine's ventilation and lighting equipment, then the ore hoist is considered production equipment, while the ventilation and lighting equipment, which involve personal safety, are considered necessary equipment.

[0066] The required power consumption of the necessary equipment is calculated based on the preset power outage duration and the rated power of the necessary equipment. The portion of the energy stored by the independent power source in the corresponding predicted time period, excluding the required power consumption, is taken as the redundant power consumption.

[0067] The optimal operating scheme is obtained based on the genetic algorithm. The chromosome in the genetic algorithm is a binary string, which represents whether to supply power to the load device or not. When constructing the chromosome, the chromosome is filled based on the necessary devices. The unfilled part of the chromosome is randomly filled by the production equipment. During filling, the non-necessary power of the production equipment is determined based on the power outage duration. Filling stops when the non-necessary power is greater than or equal to the redundant power, or when the total power consumption of the chromosome exceeds the total output power of the independent power supply.

[0068] The fitness function of the genetic algorithm is a weighted sum of the first score and the second score, where the first score is calculated based on the number of production equipment included in the chromosome, and the second score is calculated based on the priority of the production equipment included in the chromosome.

[0069] After the genetic algorithm meets the termination condition, the chromosome with the highest fitness score is obtained, and the optimal running scheme is determined based on the distribution of its binary values.

[0070] This embodiment uses a genetic algorithm to generate the optimal running scheme. Population iteration, crossover, and mutation in the genetic algorithm are existing technologies; only the improvements to the adaptability of this invention are described here. In the genetic algorithm, chromosomes exist in the form of binary strings, and their length is equal to the number of all loads in the mining farm. For example, if the mining farm includes 100 load devices, the length of the chromosome is 100, meaning it includes 100 values. If power is to be supplied to a load device, the value at the corresponding position is 1; otherwise, it is 0. For example, load device 1 corresponds to the value at the first position in the chromosome. Finally, in the chromosome with the highest fitness score, if the value at the first position is 1, then load device 1 is supplied with power; if it is 0, then load device 1 is not supplied with power.

[0071] When constructing chromosomes, the chromosomes are first populated with necessary equipment to ensure that these equipment are the devices requiring power in the generated optimal operating scheme. The total required power for all necessary equipment during a power outage is obtained by summing the required power of all necessary equipment. Subtracting the total required power from the stored power yields the redundancy power. Next, chromosomes are randomly populated using production equipment, ensuring that the power consumption during population does not exceed the redundancy power. Furthermore, if the total power consumption of the load devices exceeds the total output power of the independent power supply, it will cause problems such as independent power supply overload and grid instability, affecting or even damaging the load devices. Therefore, power overload must be avoided during population. Random population generates a large number of chromosomes, thus creating an initial population. The fitness score of each chromosome in the initial population is then calculated using a fitness function, and subsequent iterative evolution is performed.

[0072] In this embodiment, the fitness function is a weighted sum of a first score and a second score. The first score is calculated based on the number of production devices in the chromosome, and the second score is determined based on the priority of the load devices in the chromosome. For example, the number of production devices in the chromosome can be directly used as the first score, and the priority of the load devices can be represented by values ​​such as 1, 2, 3, etc. Then the second score can be the average of the priority values. Regarding the weights of the first and second scores, when the weight of the first score is greater than the weight of the second score, it indicates that the evolutionary direction is more inclined to supply power to more production devices; conversely, the evolutionary direction is more inclined to supply power to higher-priority production devices.

[0073] Therefore, in this embodiment, the first score and the second score correspond to the first weight and the second weight, respectively. If the energy storage capacity is less than the third threshold, the first weight is set to be less than the second weight; otherwise, the first weight is set to be greater than the second weight.

[0074] According to the above technical solution, the third threshold can be determined based on the total energy storage of the independent power supply. For example, if 40% of the energy storage is used as the third threshold, when the energy storage is less than the third threshold, it indicates that the energy storage of the independent power supply is relatively small. Therefore, power should be supplied to high-priority load devices as much as possible. When the energy storage is greater than or equal to the third threshold, it indicates that the energy storage of the independent power supply is relatively large. Power can be supplied to more production equipment as much as possible to ensure production.

[0075] In this embodiment, if the total power consumption of chromosomes exceeds the total output power of independent power sources in the initial population and the iterative population obtained after crossover and mutation, the binary values ​​of chromosomes belonging to production equipment are corrected to not supply power until the power constraint is met.

[0076] This embodiment also performs a power constraint check on the chromosome. If it is found that the total power required by the equipment in the chromosome is greater than the total output power of the independent power supply, a binary value belonging to the production equipment in the chromosome is randomly selected and modified to 0 until the total power is less than the total output power of the independent power supply, that is, the power constraint is met.

[0077] After determining the optimal operating scheme, this implementation obtains the historical operating data of the production equipment. Based on the historical operating data, it predicts the impact of an unexpected power outage on the independent power supply voltage of the production equipment during operation. If the impact value is greater than a second threshold, the corresponding production equipment is defined as a risky equipment. When the independent power supply supplies power to the risky equipment, its operating power is suppressed, or the risky equipment is not supplied with power.

[0078] Specifically, in this embodiment, predicting the impact of an unexpected power outage on the independent power supply voltage of the production equipment during operation based on historical operating data includes the following steps:

[0079] Historical operating data includes power grid environment data for production equipment and the impact of power outages on the independent power supply voltage. Power grid environment data includes voltage and power distribution in the power grid. The production equipment to be predicted is defined as the target equipment, and the corresponding power grid environment data is the first data.

[0080] Multiple historical operational data points similar to the first data are obtained from the power grid environment data and defined as the second data. The impact values ​​corresponding to the second data are aggregated into a first sequence. The first element in the first sequence is defined as the first element, and the first element is used as the reference data for the first data.

[0081] Obtain the third data, which is historical operational data that used the first element as reference data. Obtain the influence value of the third data and aggregate it into a second sequence. Calculate the first cumulative probability of the first element based on the first sequence. Calculate the second cumulative probability of each element in the second sequence. Locate the element in the second sequence whose second cumulative probability is closest to the first cumulative probability, and use this element to replace the first element as the prediction result. Repeat this step until all elements in the first sequence are corrected. Average the prediction results in the corrected first sequence to obtain the influence value of the target device.

[0082] Historical operating data of the production equipment includes the voltage and power distribution at various nodes in the power grid. Specifically, power flow calculations can be performed using methods such as the Seidel method and the Newton-Raphson method to obtain the voltage and power distribution of the power grid. Then, based on the historical operating data, the impact of unexpected power outages of the production equipment on the output voltage of the independent power source under this power grid environment is obtained. By acquiring multiple historical power outage records of the production equipment and performing statistical analysis, the impact on the voltage of the independent power source should the production equipment, under optimal operating conditions, be predicted. If the impact value exceeds a second threshold (e.g., 5V), the production equipment is defined as a risky device because voltage fluctuations in the independent power source can affect other load devices in the power grid, thus requiring close monitoring. Then, based on actual needs, it is determined whether to suppress the operating power of the risky device or to directly withhold power to the production equipment. Suppressing the operating power of the risky device can reduce the probability of its unexpected power outage.

[0083] The following explanation uses the determination of the impact value of a certain production equipment R as an example. To determine the potential impact value of a power outage of production equipment R on the independent power supply, we first need to obtain the optimal operating scheme of production equipment R. Based on the optimal operating scheme, we can determine which load devices the independent power supply will supply. Based on the optimal operating scheme, we perform power flow calculation to obtain the voltage and power values ​​of each load device in the power grid, that is, to obtain the voltage and power distribution of the power grid. For ease of distinction, the power grid environment data corresponding to production equipment R is defined as the first data.

[0084] Then, multiple data points similar to the first data are obtained from historical operating data and defined as the second data. Euclidean distance can be used to determine the similarity between data. At the same time, two data points with a similarity greater than 90% can be defined as similar data. In other embodiments, they can also be defined as 95% or 80%, etc.

[0085] For example, if we obtain 10 historical operating data points similar to the first data, we define them as second data 1 to 10. In the second data 1 to 10, the impact value of the production equipment R on the independent power supply after the power failure is defined as impact value 1 to 10. The impact values ​​1 to 10 are combined into a first sequence, which is represented as [X1 X2 …… X10]. Then, we select the first element from the first sequence and define it as the first element, which is X1. Now we need to use the first element to predict the impact value of the production equipment. That is, the first element is used as the reference data for the first data, because the power grid environment data of the first data and the reference data are similar.

[0086] The system retrieves third-party data from historical operating data. This third-party data consists of historical operating data that previously used X1 as a reference data point. Initially, during the data accumulation phase, the system uses a different method to predict the impact value. Specifically, the system directly uses the impact value of the selected reference data as the impact value of the production equipment. For example, if the impact value of production equipment W needs to be predicted, the historical operating data most similar to its power grid environment data is Z. If the impact value of production equipment W in historical operating data Z is 3V, then 3V is directly used as the current impact value to be predicted for production equipment W. If production equipment W actually experiences a power outage, and the power outage value is 2.5V, then this data will be stored as historical operating data and defined as the third-party data.

[0087] Returning to X1 in the first sequence, assuming there are 20 corresponding third data points, a third sequence is constructed based on the impact values ​​from these 20 data points. Here, the impact value refers to the actual impact on the independent power supply after a power outage, such as the 2.5V mentioned above. Next, the first cumulative probability of X1 in the first sequence is calculated. If the first cumulative probability is 50%, the element in the second sequence with the second cumulative probability closest to 50% is found, for example, defined as the second element. This second element replaces the first element as the predicted value for the production equipment R. This process is repeated, replacing each element in the first sequence. Finally, the average of the elements in the first sequence after the first loop is calculated to obtain the impact value of the production equipment R. Simply using the values ​​of historical similar examples for prediction can easily lead to prediction distribution bias. This method introduces statistical principles, correcting the prediction results based on the cumulative probability during prediction, making the prediction results more consistent with the statistical distribution.

[0088] like Figure 2 As shown, the present invention also provides a coal mine power grid operation optimization system based on intelligent analysis. This system is used to implement the above-mentioned method and includes:

[0089] The mapping module maps the power distribution network structure of the mine into a topology graph, which includes multiple nodes and edges. Nodes represent load devices, and edges represent power lines between nodes.

[0090] The hierarchical module determines the importance level of nodes based on expert knowledge, and determines the weight of edges based on the length of power lines, the areas they pass through, and the supporting structures. Each node and the edges connected to it are regarded as a connection network. The connection networks in the topology graph are sorted based on the importance level of the nodes and the weight of the edges. After sorting, the priority of load devices is obtained.

[0091] The prediction module acquires the current energy storage capacity of independent power sources in the mine and the power source of the energy storage capacity at preset time intervals, and predicts the energy storage capacity for multiple preset time periods in the future based on the current energy storage capacity and power source.

[0092] The execution module extracts multiple load devices to construct the optimal operating plan based on the predicted energy storage capacity, priority, and power consumption of the load devices. If the mine experiences a mains power outage, the optimal operating plan is obtained according to the preset time period in which the power outage occurs. Based on the optimal operating plan, the energy storage capacity is used to restore the power supply to the corresponding load devices.

[0093] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0094] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this patent should be determined by the appended claims.

[0095] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing the operation of power grids in coal mines based on intelligent analysis, characterized in that, include: The power distribution network structure of the mine is mapped as a topology graph, which includes multiple nodes and edges. Nodes represent load devices, and edges represent power lines between nodes. The importance level of nodes is determined based on expert knowledge, and the weight of edges is determined based on the length of the power line, the area it passes through, and the supporting structure. Each node and the edges connected to it are considered as a connection network. The connection networks in the topology are sorted based on the importance level of the nodes and the weight of the edges. The sorted network includes the priority of the load devices. At preset intervals, the current energy storage capacity of the independent power source in the mine and the power source of the energy storage capacity are obtained. Based on the current energy storage capacity and the power source, the energy storage capacity for multiple preset time periods in the future is predicted. Based on the predicted energy storage capacity, priority, and power consumption of the load devices, the optimal operating scheme is constructed by extracting multiple load devices. If the mine experiences a mains power outage, the optimal operating plan is obtained based on the preset time period in which the power outage occurs, and the power supply to the corresponding load equipment is restored using the stored energy based on the optimal operating plan. The process of obtaining the priority of load devices after sorting includes the following steps: Statistical features of the connection network are obtained, including temporal features of the number of edges and edge weights. A multidimensional feature vector of the connection network is constructed based on the importance level of the load device and the statistical features. The connection network is clustered based on the similarity between the multidimensional feature vectors to obtain multiple clusters. Based on the multidimensional feature vectors of the connected networks, a comprehensive feature of the class groups is constructed. After normalization of the comprehensive feature, standard features are obtained. The standard features are then subjected to dimensionality reduction processing based on the PCA dimensionality reduction algorithm, and principal components with a cumulative contribution rate greater than a first threshold are selected. A projection matrix is ​​constructed based on the principal components. The standard features of each class group are projected onto the principal component space based on the projection matrix to obtain the score of each standard feature of the class group on each principal component. The scores are weighted and summed based on the contribution rate of the principal components to obtain the total score of the class group. The class groups are ranked based on the total score, and each connected network within the class group is assigned a priority based on the ranking.

2. The method according to claim 1, characterized in that, The optimal operating scheme for extracting multiple load devices based on predicted energy storage capacity, priority, and power consumption includes the following steps: A preset power outage duration is set, and a production plan for the preset time period is obtained. Based on the production plan, production equipment and necessary equipment are determined from the load equipment. Production equipment is the load equipment required by the production plan, and necessary equipment is the equipment that must be powered in the production plan. The necessary power of the necessary equipment is determined according to the power outage duration, and the part of the energy storage power that is greater than the necessary power is defined as redundant power. The optimal operating scheme is obtained based on the genetic algorithm. The chromosome in the genetic algorithm is a binary string, which represents whether the load device is powered or not. When constructing the chromosome, the chromosome is filled based on the necessary devices. The unfilled part of the chromosome is randomly filled by the production equipment. When filling, the non-necessary power of the production equipment is determined based on the power outage duration. Filling stops when the non-necessary power is greater than or equal to the redundant power, or when the total power consumption of the chromosome exceeds the total output power of the independent power supply. The fitness function of the genetic algorithm is a weighted sum of a first score and a second score, wherein the first score is calculated based on the number of production equipment included in the chromosome, and the second score is calculated based on the priority of the production equipment included in the chromosome. After the genetic algorithm meets the termination condition, the chromosome with the highest fitness score is obtained, and the optimal running scheme is determined based on the distribution of its binary values.

3. The method according to claim 2, characterized in that, After determining the optimal operating scheme, historical operating data of the production equipment is obtained. Based on the historical operating data, the impact of an unexpected power outage on the independent power supply voltage of the production equipment during operation is predicted. If the impact value is greater than a second threshold, the corresponding production equipment is defined as a risky equipment. When the independent power supply supplies power to the risky equipment, its operating power is suppressed, or the risky equipment is not supplied with power.

4. The method according to claim 3, characterized in that, Based on the historical operating data, the prediction of the impact of an unexpected power outage on the independent power supply voltage during the operation of the production equipment includes the following steps: The historical operating data includes the power grid environment data of the production equipment and the impact value of power failure on the independent power supply voltage. The power grid environment data includes the voltage distribution and power distribution in the power grid. The production equipment to be predicted is defined as the target equipment, and the corresponding power grid environment data is the first data. Multiple historical operating data that are similar to the first data are obtained from the power grid environment data and defined as the second data. The impact values ​​corresponding to the second data are aggregated into a first sequence. The first element in the first sequence is defined as the first element, and the first element is used as the reference data of the first data. Obtain third data, which is historical operating data that previously used the first element as reference data. Obtain the influence value of the third data and aggregate it into a second sequence. Calculate the first cumulative probability of the first element based on the first sequence. Calculate the second cumulative probability of each element in the second sequence. Locate the element in the second sequence whose second cumulative probability is closest to the first cumulative probability and replace the first element with this element as the prediction result. Repeat this step until all elements in the first sequence are corrected. Average the prediction results in the corrected first sequence to obtain the influence value of the target device.

5. The method according to claim 2, characterized in that, The first score and the second score correspond to the first weight and the second weight, respectively. If the energy storage capacity is less than the third threshold, the first weight is set to be less than the second weight; otherwise, the first weight is set to be greater than the second weight.

6. The method according to claim 2, characterized in that, In the initial population and the iterative population obtained after crossover and mutation, if the total power consumption of the chromosomes exceeds the total output power of the independent power sources, the binary values ​​of the chromosomes belonging to the production equipment will be corrected to not supply power until the power constraint is met.

7. A coal mine power grid operation optimization system based on intelligent analysis, used to implement the method as described in any one of claims 1-6, characterized in that, The system includes: The mapping module maps the power distribution network structure of the mine into a topology graph, which includes multiple nodes and edges. Nodes represent load devices, and edges represent power lines between nodes. The hierarchical module determines the importance level of nodes based on expert knowledge, determines the weight of edges based on the length of power lines, the areas they pass through, and the supporting structures, and treats each node and the edges connected to it as a connection network. Based on the importance level of nodes and the weight of edges, the connection networks in the topology graph are sorted, and the sorted results include the priority of load devices. The prediction module acquires the current energy storage capacity of the independent power source in the mine and the power source of the energy storage capacity at preset time intervals, and predicts the energy storage capacity for multiple preset time periods in the future based on the current energy storage capacity and the power source. The execution module extracts multiple load devices to construct the optimal operating plan based on the predicted energy storage capacity, priority, and power consumption of the load devices. If the mine experiences a mains power outage, the optimal operating plan is obtained according to the preset time period in which the power outage occurs, and the energy storage capacity is used to restore the power supply to the corresponding load devices based on the optimal operating plan.

Images