Power supply equipment early warning system and method based on electric power data analysis

Abstract

The invention discloses a power supply equipment early warning system and method based on electric power data analysis, and relates to the technical field of power supply equipment early warning. The distance between each power supply device and the center early warning point is extracted, each edge early warning node is calculated by using a delay threshold, and a distributed early warning network is constructed; constructing an early warning model at each early warning node; calculating by using a linear regression algorithm to obtain the influence of the environmental data on the power data, and constructing a self-adaptive optimization model of the power supply equipment; in the distributed early warning network, each early warning node is used to collect the power data of the power supply equipment in the range, and distributed early warning is carried out on the power supply equipment in the power supply area. And when real-time early warning is carried out on the power supply equipment, real-time monitoring is carried out on the environment data of each power supply equipment, and the real-time environment data is input into the adaptive optimization model to carry out adaptive adjustment and optimization on early warning parameters in the early warning model.

Description

Technical Field

[0001] The present invention relates to the technical field of power supply equipment early warning, and in particular to a power supply equipment early warning system and method based on power data analysis. Background Art

[0002] The power system is the cornerstone of the operation of modern society. It includes multiple links such as power generation, transmission, distribution and consumption, covering all processes from power plants to user ends. With the increasing demand for industrialization and electrification of society, the scale of the power system has become increasingly large, with a wide variety of equipment and a complex operating environment. Once a power supply equipment (such as transformers, switches, circuit breakers, generators, etc.) fails, it may cause a large-scale power outage or even a safety accident. Therefore, ensuring the safe and stable operation of power supply equipment has become an important task for the power industry. As the scale of the power system expands, simple manual monitoring and alarm devices have gradually been introduced, such as real-time monitoring meters for current and voltage and basic overload protection. However, these devices can usually only trigger alarms when equipment fails and lack predictive capabilities. With the development of the Internet of Things technology, a large number of sensors are installed on power supply equipment, which can collect massive data such as temperature, voltage, current, and power of the equipment in real time. The introduction of big data technology makes it possible to fully monitor and analyze the status of power supply equipment. However, since the power supply area is relatively wide, when warning the power supply equipment in the marginal area, the data needs to be transmitted to the central warning point for warning, which causes a large warning delay and cannot be warned in time for the marginal power supply equipment. In addition, due to the different environmental data in different regions, the same warning model cannot accurately warn every power supply equipment, which is easy to cause misjudgment. Summary of the invention

[0003] The object of the present invention is to provide a power supply equipment early warning system and method based on power data analysis to solve the problems raised in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A power supply equipment early warning method based on power data analysis, the method comprising the following steps:

[0006] S100, collecting historical records of warning delays when power anomaly warnings are issued to power supply equipment in different regions, extracting the data transmission distance between the central warning point and the power supply equipment in the records, and calculating the delay threshold using the data transmission distance;

[0007] Furthermore, the specific steps of calculating the delay threshold using the data transmission distance are as follows:

[0008] S101. Collect the historical records of warning delays when power anomaly warnings are issued to power supply equipment in different regions, and extract the data transmission distance between the central warning point and the power supply equipment in the records.

[0009] {L1, L2, L3, ..., L n}, L1, L2, L3, ..., L n Indicates the data transmission distance between the central warning point and the 1st, 2nd, 3rd, ..., nth power supply equipment in the extracted record, where n is a positive integer;

[0010] S102, using the data transmission distance between the central warning point and the power supply equipment in the extracted records to calculate the delay threshold, the formula is:

[0011]

[0012] In the formula, Ly represents the calculated delay threshold, L min Indicates the minimum value of data transmission distance in the extracted records, L max Indicates the maximum value of the data transmission distance in the extracted records.

[0013] S200, collect the locations of all power supply equipment in the power supply area, extract the distance between each power supply equipment and the central warning point, calculate each edge warning node using the delay threshold, set a warning node in each edge warning node, and build a distributed warning network;

[0014] Furthermore, the specific steps for building a distributed early warning network are:

[0015] S201, collect the locations of all power supply equipment in the power supply area, and extract the distance between each power supply equipment and the central warning point as {Ld1, Ld2, Ld3, ..., Ld p}, Ld1, Ld2, Ld3,..., Ld p Indicates the distance between the 1st, 2nd, 3rd, ..., pth power supply equipment and the central warning point in the extracted power supply area, where p is a positive integer; the distance between each power supply equipment and the central warning point is judged by using the calculated delay threshold, and the power supply equipment responsible for the central warning point is screened. The formula is:

[0016]

[0017] In the formula, A represents the set of power supply equipment that the central warning point is responsible for after screening using the delay threshold, and the central warning circle Q0 is constructed using the set of power supply equipment that the central warning point is responsible for after screening;

[0018] S202: The power supply devices excluded after screening are regarded as edge power supply devices, and an edge power supply device set B is constructed. The edge warning circles {Q1, Q2, Q3, ..., Q m}, Q1, Q2, Q3, ..., Q m Indicates the 1st, 2nd, 3rd, ..., mth edge warning circles set, and the number of edge warning circles is calculated using the formula: In the formula, Ld max Indicates the distance of the power supply equipment farthest from the central warning point in the power supply area; suppose the distances between m edge warning circles and the central warning point are {2×Ly, 4×Ly, 6×Ly, ..., (2×m)×Ly}, 2×Ly, 4×Ly, 6×Ly, ..., (2×m)×Ly represent the distances between the 1st, 2nd, 3rd, ..., mth edge warning circles and the central warning point; the distance between each edge warning circle and the central warning point is the product of twice the number of edge warning circles and the delay threshold;

[0019] S203, an edge warning node is set at a distance of 2×Ly on the first layer edge warning circle, and all edge warning nodes set on the first layer edge warning circle are represents the 1st, 2nd, 3rd, ..., hth edge warning nodes set on the first layer edge warning circle, where h is a positive integer; then edge warning nodes are set every 2×Ly on each layer edge warning circle; the number of edge warning nodes set is calculated to be h×m;

[0020] S204, after setting the edge warning node on the edge warning circle, find the power supply equipment that each edge warning node needs to be responsible for warning. When the edge warning node needs to be responsible for the power supply equipment, collect the data of each power supply equipment and edge warning node in each set B. The distance between {Lq1, Lq2, Lq3, ..., Lq g}, Lq1, Lq2, Lq3,..., Lq g Represents the 1st, 2nd, 3rd, ..., gth power supply equipment and edge warning nodes in set B distance, g is a positive integer; when Lq≤Ly, the edge warning node is judged Responsible for the early warning of the corresponding power supply equipment; when Lq>Ly, determine the edge early warning node It is not responsible for the early warning of the corresponding power supply equipment; the power supply equipment that each edge early warning node is responsible for warning is searched in turn, and the area that each edge power supply node is responsible for warning is approximated as a regular hexagon; finally, a distributed early warning network consisting of h×m+1 regular hexagons is formed.

[0021] By distributing computing power to edge nodes of power supply equipment (such as substations or distribution stations), the system can process and analyze device data locally in real time, reducing the latency and bandwidth requirements for data transmission to the central server. This not only improves the response speed of the early warning system, but also reduces the computing burden on the central server.

[0022] S300, collecting power data and working status of each power supply device in history, using machine learning to train the collected data, extracting power data associated with power supply device failures, and using the associated power data as warning parameters for each warning node; and constructing a warning model at each warning node;

[0023] Furthermore, the specific steps of constructing the early warning model at each early warning node are:

[0024] S301, collecting the power data and working status of each power supply device in history

[0025] {(S, Z)1, (S, Z)2, (S, Z)3, ..., (S, Z) j}, (S, Z)1, (S, Z)2, (S, Z)3, ..., (S, Z) j represents the power data of each power supply device in the 1st, 2nd, 3rd, ..., jth working state in the collected history, S represents the power data of the power supply device, Z represents the working state of the power supply device, and j is a positive integer; suppose the power data contained in the power supply device is {S1, S2, S3, ..., S f}, S1, S2, S3, ..., S f represents the 1st, 2nd, 3rd, ..., fth type of power equipment included in the power supply equipment, where f is a positive integer;

[0026] S302, using machine learning to train the collected data, extracting power data associated with power supply equipment failures, and when training the collected power data and the corresponding working status, comparing when the working status is abnormal, looking for the types of power data that have changed; using the types of power data that have changed as the power data associated with the power supply equipment abnormality, assuming that the associated power data is {Si1, Si2, Si3, ..., Si e}, Si1, Si2, Si3,..., Si e represents the 1st, 2nd, 3rd, ..., eth type of associated power data in the power supply equipment, where e is a positive integer; the type of associated power data that changes in the power supply equipment when each different abnormal state in the history is R = (Si1, Si2, Si3, ..., Si q ), Si1, Si2, Si3,..., Si qIt represents the 1st, 2nd, 3rd, ..., qth types of associated power data that change when the power supply equipment is in different abnormal states, R represents the set of associated power data that change; q is a positive integer; the abnormal threshold corresponding to the abnormal state is calculated using the associated power data that change, and the formula is: Y = Si1 + Si2 + Si3 + ... + Si q ,In the formula, Y represents the calculated abnormal threshold corresponding to the abnormal state;

[0027] S303, searching for the types of related power data that have changed in each abnormal state in history, and calculating the abnormal threshold of each abnormal state accordingly, and obtaining the standard of each abnormal state as follows: Indicate the abnormal standards of the 1st, 2nd, 3rd, ..., cth abnormal states of the power supply equipment; and build an abnormal database, use the power data of each abnormal state change as an early warning parameter, and use the abnormal database to generate an early warning model at each edge early warning node.

[0028] By analyzing the power data of different abnormal states of power supply equipment in history and building an early warning model, we can make early warning judgments on different abnormal states of power supply equipment, and make abnormal warnings for different abnormal change data, thus ensuring the accuracy of abnormal warnings. By integrating multiple data sources, the system can capture more comprehensive equipment status information, thereby improving the accuracy and robustness of predictions.

[0029] S400, collecting environmental data and power data of each power supply device in history, using a linear regression algorithm to calculate the impact of environmental data on power data, and building an adaptive optimization model for power supply devices;

[0030] Furthermore, the specific steps of constructing the adaptive optimization model of the power supply equipment are:

[0031] S401, collecting the environmental data and power data of each power supply device in history

[0032] {(S, H)1, (S, H)2, (S, H)3, ..., (S, H) w}, (S, H)1, (S, H)2, (S, H)3, ..., (S, H) w represents the 1st, 2nd, 3rd, ..., wth power equipment value of the power supply equipment under different environmental data, w is a positive integer; H represents different environmental data, S represents the power data of the power supply equipment; suppose the number of environmental data types is D;

[0033] S402, using a linear regression algorithm to calculate the collected environmental data and power data, and obtain an optimization function for each power data and all environmental data of the power supply equipment, the function is:

[0034] SST=a1×H1+a2×H2+a3×H3+…+a w ×H w ;

[0035] In the formula, SST represents the optimized power data, {a1, a2, a3, ..., a w} represents the coefficients of the 1st, 2nd, 3rd, ..., wth types of environmental data. The optimization function of each power data of the power supply equipment is calculated, and the adaptive optimization model of the power supply equipment is constructed using all the calculated optimization functions.

[0036] The early warning model is dynamically adjusted according to the environmental conditions (such as temperature, humidity, etc.) and historical operating modes of the power supply equipment. Over time, the system can automatically update the model through a feedback mechanism to improve the accuracy and adaptability of fault prediction; when the working conditions of a certain device change significantly (such as load fluctuations or drastic changes in the external environment), the system will automatically adjust the parameters of the prediction model to ensure the real-time prediction results. Through federated learning, the system can combine the operating data of equipment in different regions to improve the intelligence and generalization capabilities of the overall prediction model.

[0037] S500, using each warning node in the distributed warning network to collect power data of power supply equipment within the range, input the power equipment into the warning model, and perform distributed warning on the power supply equipment in the power supply area;

[0038] Furthermore, the specific steps for performing distributed early warning on power supply equipment in the power supply area are as follows:

[0039] S501, using the distributed early warning network to collect the power data of each power supply device in real time, and send the collected real-time power data to the early warning node responsible for the corresponding power supply device. The early warning node inputs the received real-time power data into the early warning model. The early warning model traverses and calculates in turn according to each abnormal power data type in the abnormal database constructed in S303. The formula is:

[0040]

[0041] In the formula, Indicates the abnormal index of the power supply equipment, {Sx1, Sx2, Sx3, ..., Sx q} represents the 1st, 2nd, 3rd, ..., qth types of power data that change in real time;

[0042] S502, the early warning model calculates each abnormal state in the abnormal database in turn to obtain an abnormal index. The early warning model output is 1, when When , the warning model output is 0; then the abnormal index is judged as:

[0043]

[0044] Through calculation, it is determined that there is an abnormal risk in the power supply equipment, and the type of abnormality is determined by using the type of changed power data.

[0045] S600: When performing real-time early warning on power supply equipment, the environmental data of each power supply equipment is monitored in real time, and the real-time environmental data is input into the adaptive optimization model to adaptively adjust and optimize the early warning parameters in the early warning model.

[0046] Furthermore, the specific steps of inputting the real-time environmental data into the adaptive optimization model to adaptively adjust and optimize the warning parameters in the warning model are as follows:

[0047] S601, when performing real-time early warning on power supply equipment, real-time monitoring is performed on the environmental data of each power supply equipment, the real-time environmental data of the power supply equipment is collected, the real-time environmental data is input into the adaptive optimization model, and the early warning parameters in the early warning model are optimized. The formula is:

[0048] SST=a1×Hs1+a2×Hs2+a3×Hs3+…+a w ×Hs w ;

[0049] In the formula, {Hs1, Hs2, Hs3, ..., Hs w} represents the 1st, 2nd, 3rd, ..., wth types of power equipment collected in real time; SST represents the optimized warning parameters; and the adaptive optimization model is used to optimize each warning parameter in the warning model in real time.

[0050] A power supply equipment early warning system based on power data analysis, the power supply equipment early warning system includes a data collection module, an edge early warning module, a model building module, an early warning module and an optimization module;

[0051] The data collection module is used to collect power data, environmental data, working status and distance data of the power supply equipment in history;

[0052] The edge warning module is used to build a distributed warning network in the power supply area, set warning nodes at the edge, and perform edge warning;

[0053] The model building module is used to build an early warning model based on the power data and working status of the power supply equipment in history, and to build an adaptive optimization model using the power data and environmental data in history;

[0054] The early warning module is used to collect power data of the power supply equipment and use the distributed early warning network to issue real-time early warning to the power supply equipment;

[0055] The optimization module is used to collect environmental data of the power supply equipment in real time and optimize the warning parameters in the warning model using the adaptive optimization model.

[0056] The edge warning module includes a warning node setting unit and a distributed warning network unit;

[0057] The early warning node setting unit is used to calculate the delay threshold according to the data transmission distance of historical delay, and use the delay threshold to set the early warning node in the power supply area;

[0058] The distributed early warning network unit is used to construct a distributed early warning network by utilizing early warning nodes in the power supply area to be responsible for power supply equipment in the corresponding area respectively.

[0059] The model building module includes an early warning model unit and an adaptive optimization model unit;

[0060] The early warning model unit is used to analyze the power data and working status of the power supply equipment in history and build an early warning model;

[0061] The adaptive optimization model unit constructs an adaptive optimization model using historical power data and environmental data.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] 1. By distributing computing power to the edge nodes of power supply equipment, the system can process and analyze equipment data locally in real time, reducing the delay and bandwidth requirements for data transmission to the central server. This not only improves the response speed of the early warning system, but also reduces the computing burden of the central server.

[0064] 2. The early warning model constructed by the present invention can make early warning judgments on different abnormal states of power supply equipment, and make abnormal warnings for different abnormal change data, thereby ensuring the accuracy of abnormal warnings. Through the integration of multiple data sources, the system can capture more comprehensive equipment status information, thereby improving the accuracy and robustness of the prediction.

[0065] 3. The present invention dynamically adjusts the early warning model according to the environmental conditions and historical operation modes of the power supply equipment. Over time, the system can automatically update the model through a feedback mechanism to improve the accuracy and adaptability of fault prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a module distribution diagram of a power supply equipment early warning system based on power data analysis of the present invention;

[0067] Figure 2A schematic diagram of the steps of a power supply equipment early warning method based on power data analysis according to the present invention;

[0068] Figure 3 A schematic diagram of a distributed early warning network of a power supply equipment early warning method based on power data analysis according to the present invention. DETAILED DESCRIPTION

[0069] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0070] Example: Figure 1-Figure 3 As shown, the present invention provides a technical solution.

[0071] A power supply equipment early warning method based on power data analysis, the method comprising the following steps:

[0072] S100, collecting historical records of warning delays when power anomaly warnings are issued to power supply equipment in different regions, extracting the data transmission distance between the central warning point and the power supply equipment in the records, and calculating the delay threshold using the data transmission distance;

[0073] The specific steps for calculating the delay threshold using the data transmission distance are as follows:

[0074] S101. Collect the historical records of warning delays when power anomaly warnings are issued to power supply equipment in different regions, and extract the data transmission distance between the central warning point and the power supply equipment in the records.

[0075] {L1, L2, L3, ..., L n}, L1, L2, L3, ..., L n Indicates the data transmission distance between the central warning point and the 1st, 2nd, 3rd, ..., nth power supply equipment in the extracted record, where n is a positive integer;

[0076] S102, using the data transmission distance between the central warning point and the power supply equipment in the extracted records to calculate the delay threshold, the formula is:

[0077]

[0078] In the formula, Ly represents the calculated delay threshold, L min Indicates the minimum value of data transmission distance in the extracted records, L max Indicates the maximum value of the data transmission distance in the extracted records.

[0079] S200, collect the locations of all power supply equipment in the power supply area, extract the distance between each power supply equipment and the central warning point, calculate each edge warning node using the delay threshold, set a warning node in each edge warning node, and build a distributed warning network;

[0080] The specific steps to build a distributed early warning network are:

[0081] S201, collect the locations of all power supply equipment in the power supply area, and extract the distance between each power supply equipment and the central warning point as {Ld1, Ld2, Ld3, ..., Ld p}, Ld1, Ld2, Ld3,..., Ld p Indicates the distance between the 1st, 2nd, 3rd, ..., pth power supply equipment and the central warning point in the extracted power supply area, where p is a positive integer; the distance between each power supply equipment and the central warning point is judged by using the calculated delay threshold, and the power supply equipment responsible for the central warning point is screened. The formula is:

[0082]

[0083] In the formula, A represents the set of power supply equipment that the central warning point is responsible for after screening using the delay threshold, and the central warning circle Q0 is constructed using the set of power supply equipment that the central warning point is responsible for after screening;

[0084] S202: The power supply devices excluded after screening are regarded as edge power supply devices, and an edge power supply device set B is constructed. The edge warning circles {Q1, Q2, Q3, ..., Q m}, Q1, Q2, Q3, ..., Q m Indicates the 1st, 2nd, 3rd, ..., mth edge warning circles set, and the number of edge warning circles is calculated using the formula: In the formula, Ld max Indicates the distance of the power supply equipment farthest from the central warning point in the power supply area; suppose the distances between m edge warning circles and the central warning point are {2×Ly, 4×Ly, 6×Ly, ..., (2×m)×Ly}, 2×Ly, 4×Ly, 6×Ly, ..., (2×m)×Ly represent the distances between the 1st, 2nd, 3rd, ..., mth edge warning circles and the central warning point; the distance between each edge warning circle and the central warning point is the product of twice the number of edge warning circles and the delay threshold;

[0085] S203, an edge warning node is set at a distance of 2×Ly on the first layer edge warning circle, and all edge warning nodes set on the first layer edge warning circle are represents the 1st, 2nd, 3rd, ..., hth edge warning nodes set on the first layer edge warning circle, where h is a positive integer; then edge warning nodes are set every 2×Ly on each layer edge warning circle; the number of edge warning nodes set is calculated to be h×m;

[0086] S204, after setting the edge warning node on the edge warning circle, find the power supply equipment that each edge warning node needs to be responsible for warning. When the edge warning node needs to be responsible for the power supply equipment, collect the data of each power supply equipment and edge warning node in each set B. The distance between {Lq1, Lq2, Lq3, ..., Lq g}, Lq1, Lq2, Lq3,..., Lq g Represents the 1st, 2nd, 3rd, ..., gth power supply equipment and edge warning nodes in set B distance, g is a positive integer; when Lq≤Ly, the edge warning node is judged Responsible for the early warning of the corresponding power supply equipment; when Lq>Ly, determine the edge early warning node It is not responsible for the early warning of the corresponding power supply equipment; search for the power supply equipment that each edge early warning node is responsible for in turn, and approximate the area that each edge power supply node is responsible for early warning to a regular hexagon; finally, a distributed early warning network consisting of h×m+1 regular hexagons is formed. For example, the delay threshold calculated in the power supply area is 1km, and an edge early warning node is set for every 1km to issue early warnings to the power supply equipment in the area. Figure 3 As shown;

[0087] By distributing computing power to edge nodes of power supply equipment (such as substations or distribution stations), the system can process and analyze device data locally in real time, reducing the latency and bandwidth requirements for data transmission to the central server. This not only improves the response speed of the early warning system, but also reduces the computing burden on the central server.

[0088] S300, collecting power data and working status of each power supply device in history, using machine learning to train the collected data, extracting power data associated with power supply device failures, and using the associated power data as warning parameters for each warning node; and constructing a warning model at each warning node;

[0089] The specific steps for building an early warning model at each early warning node are:

[0090] S301, collecting the power data and working status of each power supply device in history

[0091] {(S, Z)1, (S, Z)2, (S, Z)3, ..., (S, Z) j}, (S, Z)1, (S, Z)2, (S, Z)3, ..., (S, Z) j represents the power data of each power supply device in the 1st, 2nd, 3rd, ..., jth working state in the collected history, S represents the power data of the power supply device, Z represents the working state of the power supply device, and j is a positive integer; suppose the power data contained in the power supply device is {S1, S2, S3, ..., S f}, S1, S2, S3, ..., S f represents the 1st, 2nd, 3rd, ..., fth type of power equipment included in the power supply equipment, where f is a positive integer;

[0092] S302, using machine learning to train the collected data, extracting power data associated with power supply equipment failures, and when training the collected power data and the corresponding working status, comparing when the working status is abnormal, looking for the types of power data that have changed; using the types of power data that have changed as the power data associated with the power supply equipment abnormality, assuming that the associated power data is {Si1, Si2, Si3, ..., Si e}, Si1, Si2, Si3,..., Si e represents the 1st, 2nd, 3rd, ..., eth type of associated power data in the power supply equipment, where e is a positive integer; the type of associated power data that changes in the power supply equipment when each different abnormal state in the history is R = (Si1, Si2, Si3, ..., Si q ), Si1, Si2, Si3,..., Si q It represents the 1st, 2nd, 3rd, ..., qth types of associated power data that change when the power supply equipment is in different abnormal states, R represents the set of associated power data that change; q is a positive integer; the abnormal threshold corresponding to the abnormal state is calculated using the associated power data that change, and the formula is: Y = Si1 + Si2 + Si3 + ... + Si q ,In the formula, Y represents the calculated abnormal threshold corresponding to the abnormal state;

[0093] S303, searching for the types of related power data that have changed in each abnormal state in history, and calculating the abnormal threshold of each abnormal state accordingly, and obtaining the standard of each abnormal state as follows: Indicate the abnormal standards of the 1st, 2nd, 3rd, ..., cth abnormal states of the power supply equipment; and build an abnormal database, use the power data of each abnormal state change as an early warning parameter, and use the abnormal database to generate an early warning model at each edge early warning node.

[0094] By analyzing the power data of different abnormal states of power supply equipment in history and building an early warning model, we can make early warning judgments on different abnormal states of power supply equipment, and make abnormal warnings for different abnormal change data, thus ensuring the accuracy of abnormal warnings. By integrating multiple data sources, the system can capture more comprehensive equipment status information, thereby improving the accuracy and robustness of predictions.

[0095] S400, collecting environmental data and power data of each power supply device in history, using a linear regression algorithm to calculate the impact of environmental data on power data, and building an adaptive optimization model for power supply devices;

[0096] The specific steps of building an adaptive optimization model for power supply equipment are:

[0097] S401, collecting the environmental data and power data of each power supply device in history

[0098] {(S, H)1, (S, H)2, (S, H)3, ..., (S, H) w}, (S, H)1, (S, H)2, (S, H)3, ..., (S, H) w represents the 1st, 2nd, 3rd, ..., wth power equipment value of the power supply equipment under different environmental data, w is a positive integer; H represents different environmental data, S represents the power data of the power supply equipment; suppose the number of environmental data types is D;

[0099] S402, using a linear regression algorithm to calculate the collected environmental data and power data, and obtain an optimization function for each power data and all environmental data of the power supply equipment, the function is:

[0100] SST=a1×H1+a2×H2+a3×H3+…+a w ×H w ;

[0101] In the formula, SST represents the optimized power data, {a1, a2, a3, ..., a w} represents the coefficients of the 1st, 2nd, 3rd, ..., wth types of environmental data. The optimization function of each power data of the power supply equipment is calculated, and the adaptive optimization model of the power supply equipment is constructed using all the calculated optimization functions.

[0102] The early warning model is dynamically adjusted according to the environmental conditions (such as temperature, humidity, etc.) and historical operating modes of the power supply equipment. Over time, the system can automatically update the model through a feedback mechanism to improve the accuracy and adaptability of fault prediction; when the working conditions of a certain device change significantly (such as load fluctuations or drastic changes in the external environment), the system will automatically adjust the parameters of the prediction model to ensure the real-time prediction results. Through federated learning, the system can combine the operating data of equipment in different regions to improve the intelligence and generalization capabilities of the overall prediction model.

[0103] S500, using each warning node in the distributed warning network to collect power data of power supply equipment within the range, input the power equipment into the warning model, and perform distributed warning on the power supply equipment in the power supply area;

[0104] The specific steps for distributed early warning of power supply equipment in the power supply area are:

[0105] S501, using the distributed early warning network to collect the power data of each power supply device in real time, and send the collected real-time power data to the early warning node responsible for the corresponding power supply device. The early warning node inputs the received real-time power data into the early warning model. The early warning model traverses and calculates in turn according to each abnormal power data type in the abnormal database constructed in S303. The formula is:

[0106]

[0107] In the formula, Indicates the abnormal index of the power supply equipment, {Sx1, Sx2, Sx3, ..., Sx q} represents the 1st, 2nd, 3rd, ..., qth types of power data that change in real time;

[0108] S502, the early warning model calculates each abnormal state in the abnormal database in turn to obtain an abnormal index. The early warning model output is 1, when When , the warning model output is 0; then the abnormal index is judged as:

[0109]

[0110] Through calculation, it is determined that there is an abnormal risk in the power supply equipment, and the type of abnormality is determined by using the type of changed power data.

[0111] S600: When performing real-time early warning on power supply equipment, the environmental data of each power supply equipment is monitored in real time, and the real-time environmental data is input into the adaptive optimization model to adaptively adjust and optimize the early warning parameters in the early warning model.

[0112] The specific steps of inputting real-time environmental data into the adaptive optimization model to adaptively adjust and optimize the warning parameters in the warning model are as follows:

[0113] S601, when performing real-time early warning on power supply equipment, real-time monitoring is performed on the environmental data of each power supply equipment, the real-time environmental data of the power supply equipment is collected, the real-time environmental data is input into the adaptive optimization model, and the early warning parameters in the early warning model are optimized. The formula is:

[0114] SST=a1×Hs1+a2×Hs2+a3×Hs3+…+a w ×Hs w ;

[0115] In the formula, {Hs1, Hs2, Hs3, ..., Hs w} represents the 1st, 2nd, 3rd, ..., wth types of power equipment collected in real time; SST represents the optimized warning parameters; and the adaptive optimization model is used to optimize each warning parameter in the warning model in real time.

[0116] A power supply equipment early warning system based on power data analysis, the power supply equipment early warning system includes a data collection module, an edge early warning module, a model building module, an early warning module and an optimization module;

[0117] The data collection module is used to collect power data, environmental data, working status and distance data of the power supply equipment in history;

[0118] The edge warning module is used to build a distributed warning network in the power supply area, set warning nodes at the edge, and perform edge warning;

[0119] The model building module is used to build an early warning model based on the power data and working status of the power supply equipment in history, and to build an adaptive optimization model using the power data and environmental data in history;

[0120] The early warning module is used to collect power data of the power supply equipment and use the distributed early warning network to issue real-time early warning to the power supply equipment;

[0121] The optimization module is used to collect environmental data of the power supply equipment in real time and optimize the warning parameters in the warning model using the adaptive optimization model.

[0122] The edge warning module includes a warning node setting unit and a distributed warning network unit;

[0123] The early warning node setting unit is used to calculate the delay threshold according to the data transmission distance of historical delay, and use the delay threshold to set the early warning node in the power supply area;

[0124] The distributed early warning network unit is used to construct a distributed early warning network by utilizing early warning nodes in the power supply area to be responsible for power supply equipment in the corresponding area respectively.

[0125] The model building module includes an early warning model unit and an adaptive optimization model unit;

[0126] The early warning model unit is used to analyze the power data and working status of the power supply equipment in history and build an early warning model;

[0127] The adaptive optimization model unit constructs an adaptive optimization model using historical power data and environmental data.

[0128] Embodiment: The associated power data of the power supply equipment obtained by calculating the historical data are voltage, current and power; the current abnormal state is two types D1 and D2, and the two abnormal change power data are current and power, voltage and current respectively; the early warning model calculated according to the historical data is D1: current + power = 150, D1: voltage + current = 200;

[0129] The environmental data that affect the power data in the power supply equipment are temperature and humidity; three optimization functions are calculated based on the historical environmental data and power data: current = 2×temperature + 3×humidity, voltage = 0.3×temperature + 1.2×humidity, power = 10×temperature + 18×humidity; an adaptive optimization model is constructed;

[0130] The three types of power data, voltage, current and power, are collected in real time, which are 50, 30 and 140 respectively. The power supply equipment is judged to be in abnormal state D1 through calculation;

[0131] The real-time collected environmental data are 20 and 8 respectively; the three types of power data are optimized, and the optimized data are 64, 14.6, and 344 respectively;

[0132] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A power supply equipment early warning method based on power data analysis, characterized in that: The method comprises the following steps: S100, collecting historical records of warning delays when power anomaly warnings are issued to power supply equipment in different regions, extracting the data transmission distance between the central warning point and the power supply equipment in the records, and calculating the delay threshold using the data transmission distance; S200, collecting the locations of all power supply equipment in the power supply area, extracting the distance between each power supply equipment and the central warning point, calculating each edge warning node using the delay threshold, and building a distributed warning network; S300, collecting power data and working status of each power supply device in history, using machine learning to train the collected data, extracting power data associated with power supply device failures, and using the associated power data as warning parameters for each warning node; and constructing a warning model at each warning node; S400, collecting environmental data and power data of each power supply device in history, using a linear regression algorithm to calculate the impact of environmental data on power data, and building an adaptive optimization model for power supply devices; S500, using each warning node in the distributed warning network to collect power data of power supply equipment within the range, input the power equipment into the warning model, and perform distributed warning on the power supply equipment in the power supply area; S600: When performing real-time early warning on power supply equipment, the environmental data of each power supply equipment is monitored in real time, and the real-time environmental data is input into the adaptive optimization model to adaptively adjust and optimize the early warning parameters in the early warning model.

2. The power supply equipment early warning method based on power data analysis according to claim 1 is characterized in that: The specific steps of calculating the delay threshold using the data transmission distance in S100 are: S101, collecting historical records of warning delays when power anomaly warnings are issued to power supply equipment in different regions, extracting the data transmission distance between the central warning point and the power supply equipment as {L1, L2, L3, ..., L n }, L1, L2, L3, ..., L n Indicates the data transmission distance between the central warning point and the 1st, 2nd, 3rd, ..., nth power supply equipment in the extracted record, where n is a positive integer; S102, using the data transmission distance between the central warning point and the power supply equipment in the extracted records to calculate the delay threshold, the formula is: In the formula, Ly represents the calculated delay threshold, L min Indicates the minimum value of data transmission distance in the extracted records, L max Indicates the maximum value of the data transmission distance in the extracted records.

3. The power supply equipment early warning method based on power data analysis according to claim 2 is characterized in that: The specific steps of constructing the distributed early warning network in S200 are: S201, collect the locations of all power supply equipment in the power supply area, and extract the distance between each power supply equipment and the central warning point as {Ld1, Ld2, Ld3, ..., Ld p }, Ld1, Ld2, Ld3,..., Ld p Indicates the distance between the 1st, 2nd, 3rd, ..., pth power supply equipment and the central warning point in the extracted power supply area, where p is a positive integer; the distance between each power supply equipment and the central warning point is judged by using the calculated delay threshold, and the power supply equipment responsible for the central warning point is screened. The formula is: In the formula, A represents the set of power supply equipment that the central warning point is responsible for after screening using the delay threshold, and the central warning circle Q0 is constructed using the set of power supply equipment that the central warning point is responsible for after screening; S202: The power supply devices excluded after screening are regarded as edge power supply devices, and an edge power supply device set B is constructed. The edge warning circles {Q1, Q2, Q3, ..., Q m }, Q1, Q2, Q3, ..., Q m Indicates the 1st, 2nd, 3rd, ..., mth edge warning circles set, and the number of edge warning circles is calculated using the formula: In the formula, Ld max Indicates the distance of the power supply equipment farthest from the central warning point in the power supply area; suppose the distances between m edge warning circles and the central warning point are {2×Ly, 4×Ly, 6×Ly, ..., (2×m)×Ly}, 2×Ly, 4×Ly, 6×Ly, ..., (2×m)×Ly represent the distances between the 1st, 2nd, 3rd, ..., mth edge warning circles and the central warning point; the distance between each edge warning circle and the central warning point is the product of twice the number of edge warning circles and the delay threshold; S203, an edge warning node is set at a distance of 2×Ly on the first layer edge warning circle, and all edge warning nodes set on the first layer edge warning circle are represents the 1st, 2nd, 3rd, ..., hth edge warning nodes set on the first layer edge warning circle, where h is a positive integer; then edge warning nodes are set every 2×Ly on each layer edge warning circle; the number of edge warning nodes set is calculated to be h×m; S204, after setting the edge warning node on the edge warning circle, find the power supply equipment that each edge warning node needs to be responsible for warning. When the edge warning node needs to be responsible for the power supply equipment, collect the data of each power supply equipment and edge warning node in each set B. The distance between {Lq1, Lq2, Lq3, ..., Lq g }, Lq1, Lq2, Lq3,..., Lq g Represents the 1st, 2nd, 3rd, ..., gth power supply equipment and edge warning nodes in set B distance, g is a positive integer; when Lq≤Ly, the edge warning node is judged Responsible for the early warning of the corresponding power supply equipment; when Lq>Ly, determine the edge early warning node It is not responsible for the early warning of the corresponding power supply equipment; the power supply equipment that each edge early warning node is responsible for warning is searched in turn, and the area that each edge power supply node is responsible for warning is approximated as a regular hexagon; finally, a distributed early warning network consisting of h×m+1 regular hexagons is formed.

4. The power supply equipment early warning method based on power data analysis according to claim 1 is characterized in that: The specific steps of constructing the early warning model at each early warning node in S300 are: S301, collect the power data and working status of each power supply device in the history as {(S, Z)1, (S, Z)2, (S, Z)3, ..., (S, Z) j }, (S, Z)1, (S, Z)2, (S, Z)3,..., (S, Z) j represents the power data of each power supply device in the 1st, 2nd, 3rd, ..., jth working state in the collected history, S represents the power data of the power supply device, Z represents the working state of the power supply device, and j is a positive integer; suppose the power data contained in the power supply device is {S1, S2, S3, ..., S f }, S1, S2, S3, ..., S f represents the 1st, 2nd, 3rd, ..., fth type of power equipment included in the power supply equipment, where f is a positive integer; S302, using machine learning to train the collected data, extracting power data associated with power supply equipment failures, and when training the collected power data and the corresponding working status, comparing when the working status is abnormal, looking for the types of power data that have changed; using the types of power data that have changed as the power data associated with the power supply equipment abnormality, assuming that the associated power data is {Si1, Si2, Si3, ..., Si e }, Si1, Si2, Si3,..., Si e represents the 1st, 2nd, 3rd, ..., eth type of associated power data in the power supply equipment, where e is a positive integer; the type of associated power data that changes in the power supply equipment when each different abnormal state in the history is R = (Si1, Si2, Si3, ..., Si q ), Si1, Si2, Si3,..., Si q It represents the 1st, 2nd, 3rd, ..., qth types of associated power data that change when the power supply equipment is in different abnormal states, R represents the set of associated power data that change; q is a positive integer; the abnormal threshold corresponding to the abnormal state is calculated using the associated power data that change, and the formula is: Y = Si1 + Si2 + Si3 + ... + Si q ,In the formula, Y represents the calculated abnormal threshold corresponding to the abnormal state; S303, searching for the types of related power data that have changed in each abnormal state in history, and calculating the abnormal threshold of each abnormal state accordingly, and obtaining the standard of each abnormal state as follows: Indicates the abnormal standards of the 1st, 2nd, 3rd, ..., cth abnormal states of the power supply equipment; And build an abnormal database, take the power data of each abnormal state change as the warning parameter, and use the abnormal database to generate a warning model at each edge warning node.

5. The power supply equipment early warning method based on power data analysis according to claim 1 is characterized in that: The specific steps of constructing the adaptive optimization model of the power supply equipment in S400 are: S401, collect the environmental data and power data of each power supply device in the history as {(S, H)1, (S, H)2, (S, H)3, ..., (S, H) w }, (S, H)1, (S, H)2, (S, H)3, ..., (S, H) w represents the 1st, 2nd, 3rd, ..., wth power equipment value of the power supply equipment under different environmental data, w is a positive integer; H represents different environmental data, S represents the power data of the power supply equipment; suppose the number of environmental data types is D; S402, using a linear regression algorithm to calculate the collected environmental data and power data, and obtain an optimization function for each power data and all environmental data of the power supply equipment, the function is: SST=a1×H1+a2×H2+a3×H3+…+a w ×H w ; In the formula, SST represents the optimized power data, {a1, a2, a3, ..., a w } represents the coefficients of the 1st, 2nd, 3rd, ..., wth types of environmental data. The optimization function of each power data of the power supply equipment is calculated, and the adaptive optimization model of the power supply equipment is constructed using all the calculated optimization functions.

6. The power supply equipment early warning method based on power data analysis according to claim 4 is characterized in that: The specific steps of performing distributed early warning on the power supply equipment in the power supply area in S500 are: S501, using the distributed early warning network to collect the power data of each power supply device in real time, and send the collected real-time power data to the early warning node responsible for the corresponding power supply device. The early warning node inputs the received real-time power data into the early warning model. The early warning model traverses and calculates in turn according to each abnormal power data type in the abnormal database constructed in S303. The formula is: In the formula, Indicates the abnormal index of the power supply equipment, {Sx1, Sx2, Sx3, ..., Sx q } represents the 1st, 2nd, 3rd, ..., qth types of power data that change in real time; S502, the early warning model calculates each abnormal state in the abnormal database in turn to obtain an abnormal index. The early warning model output is 1, when When , the warning model output is 0; then the abnormal index is judged as: Through calculation, it is determined that there is an abnormal risk in the power supply equipment, and the type of abnormality is determined by using the type of changed power data.

7. The power supply equipment early warning method based on power data analysis according to claim 5 is characterized by: The specific steps of inputting the real-time environmental data into the adaptive optimization model to adaptively adjust and optimize the warning parameters in the warning model in S600 are: S601, when performing real-time early warning on power supply equipment, real-time monitoring is performed on the environmental data of each power supply equipment, the real-time environmental data of the power supply equipment is collected, the real-time environmental data is input into the adaptive optimization model, and the early warning parameters in the early warning model are optimized. The formula is: SST=a1×Hs1+a2×Hs2+a3×Hs3+…+a w ×Hs w ; In the formula, {Hs1, Hs2, Hs3, ..., Hs w } represents the 1st, 2nd, 3rd, ..., wth types of power equipment collected in real time; SST represents the optimized warning parameters; and the adaptive optimization model is used to optimize each warning parameter in the warning model in real time.

8. A power supply equipment early warning system based on power data analysis, characterized in that: The power supply equipment early warning system includes a data collection module, an edge early warning module, a model building module, an early warning module and an optimization module; The data collection module is used to collect power data, environmental data, working status and distance data of the power supply equipment in history; The edge warning module is used to build a distributed warning network in the power supply area, set warning nodes at the edge, and perform edge warning; The model building module is used to build an early warning model based on the power data and working status of the power supply equipment in history, and to build an adaptive optimization model using the power data and environmental data in history; The early warning module is used to collect power data of the power supply equipment and use the distributed early warning network to issue real-time early warning to the power supply equipment; The optimization module is used to collect environmental data of the power supply equipment in real time and optimize the warning parameters in the warning model using the adaptive optimization model.

9. The power supply equipment early warning system based on power data analysis according to claim 8, characterized in that: The edge warning module includes a warning node setting unit and a distributed warning network unit; The early warning node setting unit is used to calculate the delay threshold according to the data transmission distance of historical delay, and use the delay threshold to set the early warning node in the power supply area; The distributed early warning network unit is used to construct a distributed early warning network in the power supply area by using early warning nodes to be responsible for power supply equipment in the corresponding area respectively.

10. The power supply equipment early warning system based on power data analysis according to claim 8, characterized in that: The model building module includes an early warning model unit and an adaptive optimization model unit; The early warning model unit is used to analyze the power data and working status of the power supply equipment in history and build an early warning model; The adaptive optimization model unit constructs an adaptive optimization model using historical power data and environmental data.

Images