Outdoor fully-enclosed small high-voltage isolated load switch intelligent control method and system

Abstract

The application discloses an outdoor fully-enclosed small high-voltage isolated load switch intelligent control method and system, and relates to the technical field of intelligent control systems. The method comprises the following steps: collecting a historical switch operation data sequence set of a target switch in a historical time zone through a multi-source sensing monitoring array of an edge perception layer, wherein the target switch is an outdoor fully-enclosed small high-voltage isolated load switch; acquiring a predicted switch operation data sequence set in a preset time zone according to the historical switch operation data sequence set by using a long short-term memory network; based on the predicted switch operation data sequence set, taking the minimization of line loss, fault outage minutes, voltage out-of-limit time and voltage sag times as multi-optimization targets, iteratively optimizing and searching the control parameters of the target switch, determining an adaptive control strategy, and intelligently controlling the target switch in the preset time zone according to the adaptive control strategy. The application effectively improves the operation efficiency and power supply reliability of a remote area power distribution network.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control systems, in particular to an intelligent control method and system for an outdoor fully-enclosed small high-voltage isolation load switch. BACKGROUND

[0002] With the intelligent construction of distribution networks extending to remote areas such as suburbs and rural areas, the operation stability, intelligent level and outdoor environmental adaptability of high-voltage isolation load switches for 10kV distribution networks are required to be higher. The outdoor fully-enclosed small high-voltage isolation load switch is still widely used due to its compact structure and moderate cost.

[0003] However, the traditional outdoor high-voltage isolation load switch still has technical problems: on the one hand, the device generally has the problems of transmission jamming and poor outdoor weather resistance, which is difficult to adapt to the complex natural environment in remote areas; on the other hand, it mostly relies on regular manual inspection and manual operation, which not only has high operation and maintenance cost, but also cannot avoid the risk of device failure in time, thereby resulting in insufficient operation efficiency and power supply reliability of the distribution network in remote areas. SUMMARY

[0004] The present application provides an intelligent control method and system for an outdoor fully-enclosed small high-voltage isolation load switch, aiming to solve the technical problem of insufficient operation efficiency and power supply reliability of the distribution network in remote areas in the prior art.

[0005] In view of the above problems, the present application provides an intelligent control method and system for an outdoor fully-enclosed small high-voltage isolation load switch.

[0006] In a first aspect, the present application provides an intelligent control method for an outdoor fully-enclosed small high-voltage isolation load switch, comprising:

[0007] A multi-source sensing monitoring array of an edge perception layer collects a historical switch operation data sequence set of a target switch in a historical time zone, wherein the target switch is an outdoor fully-enclosed small high-voltage isolation load switch;

[0008] A long short-term memory network is used to predict and obtain a predicted switch operation data sequence set in a preset time zone according to the historical switch operation data sequence set;

[0009] Based on the predicted switch operation data sequence set, a control parameter of the target switch is iteratively optimized and searched with the minimum line loss, fault outage minutes, voltage out-of-limit time and voltage sag frequency as multi-optimization targets, an adaptive control strategy is determined, and the target switch is intelligently controlled in the preset time zone according to the adaptive control strategy.

[0010] In a second aspect, the present application provides an intelligent control system for an outdoor fully-enclosed small high-voltage isolation load switch, comprising:

[0011] An edge-aware acquisition module is configured to acquire a historical switch operation data sequence set of a target switch in a historical time zone through a multi-source sensing monitoring array of an edge-aware layer, wherein the target switch is an outdoor fully-enclosed small high-voltage isolation load switch.

[0012] A data prediction module is configured to acquire a predicted switch operation data sequence set in a preset time zone by using a long short-term memory network and based on the historical switch operation data sequence set.

[0013] A multi-objective optimization control module is configured to perform iterative optimization search on a control parameter of the target switch based on the predicted switch operation data sequence set, to minimize line loss, fault outage minutes, voltage out-of-limit time and voltage sag frequency as multi-optimization objectives, to determine an adaptive control strategy, and to intelligently control the target switch in the preset time zone according to the adaptive control strategy.

[0014] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0015] The present application provides an intelligent control method and system for an outdoor fully-enclosed small high-voltage isolation load switch, which precisely acquires a historical operation data sequence set of an outdoor fully-enclosed small high-voltage isolation load switch through a multi-source sensing monitoring array of an edge-aware layer, to provide comprehensive and reliable data support for switch state analysis; then precisely predicts switch operation data in a preset time zone by using a long short-term memory network, to realize early prediction of future operation state of the switch; and then iteratively optimizes a control parameter based on the predicted data, to minimize line loss, fault outage minutes, voltage out-of-limit time and voltage sag frequency as objectives, to determine an adaptive control strategy, to realize dynamic intelligent control of the switch, and to form a complete data acquisition, state prediction and precise control closed loop, thereby effectively improving operation efficiency and power supply reliability of a distribution network in a remote area. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0017] Figure 1 A flowchart of an intelligent control method for an outdoor fully-enclosed small high-voltage isolation load switch provided by the embodiments of the present application is shown.

[0018] Figure 2 A structure diagram of an intelligent control system for an outdoor fully-enclosed small high-voltage isolation load switch provided by the embodiments of the present application is shown.

[0019] In the drawings, the components represented by the reference numerals are explained as follows:

[0020] Edge-aware acquisition module 11, data prediction module 12, multi-objective optimization control module 13. DETAILED DESCRIPTION

[0021] The application provides an outdoor fully-enclosed small high-voltage isolation load switch intelligent control method and system, and aims to solve the technical problems of insufficient operation efficiency and power supply reliability of the existing technology in remote areas.

[0022] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the application.

[0023] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server comprising a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or modules that are not clearly listed or inherent to the process, method, product or device.

[0024] Embodiment one, as shown in the application provides an outdoor fully-enclosed small high-voltage isolation load switch intelligent control method, which comprises: Figure 1

[0025] S100: Collect a historical switch operation data sequence set of a target switch in a historical time zone through a multi-source sensing monitoring array of an edge-aware layer, wherein the target switch is an outdoor fully-enclosed small high-voltage isolation load switch.

[0026] ​In the embodiment of the application, a multi-source sensing monitoring array of an edge perception layer is used to collect a historical switch operation data sequence set of a target switch in a historical time zone, wherein the target switch is an outdoor fully-enclosed small high-voltage isolation load switch. As a key device of a 10kV distribution network in suburbs and rural areas, the operation state of the outdoor fully-enclosed small high-voltage isolation load switch directly determines the power supply stability of the distribution network. However, in remote areas, the outdoor environment is complex, and the traditional switch data collection has three major pain points: first, the collection hardware is scattered and not integrated with the switch body, which is poor in weather resistance and easy to be disturbed; second, the monitoring index is single, and only the basic electrical parameters such as line current and voltage are concerned, and the health state of the switch itself and environmental factors are ignored; third, the data lacks time sequence continuity, and a complete historical operation data sequence cannot be formed, which leads to a lack of comprehensive and reliable data support for subsequent control strategy development. Therefore, by integrating the switch hardware structure design and multi-dimensional index monitoring, a data collection system with hardware integration, comprehensive index and time sequence collection is constructed, which provides a high-quality data source for subsequent operation state prediction based on long short-term memory network and multi-objective control parameter optimization.

[0027] The step S100 in the method provided by the embodiment of the application comprises:

[0028] The outdoor fully-enclosed small high-voltage isolation load switch comprises a fully-enclosed shell, a three-phase double-break structure, a three-phase linkage spring operating mechanism, an edge perception layer and an intelligent control unit, wherein the three-phase double-break structure integrates a vacuum arc-extinguishing chamber and a visible isolation break in series in each phase circuit, the three-phase linkage spring operating mechanism adopts a two-station spring mechanism and drives three-phase contacts to be synchronously opened and closed through a mechanical connecting rod, and the edge perception layer is internally provided with a multi-source sensing monitoring array.

[0029] The operation monitoring index set of the target switch is acquired, wherein the operation monitoring index set comprises a switch operation state, electrical parameters, health indexes and environmental parameters, the switch operation state comprises an opening and closing position and an opening and closing operation time, the electrical parameters comprise a line current, a power factor and a three-phase voltage, the health indexes comprise a contact temperature and a contact contact resistance, and the environmental parameters comprise an environmental temperature and an environmental humidity.

[0030] According to a preset data monitoring time interval and the operation monitoring index set, the target switch is monitored in real time in the historical time zone through the multi-source sensing monitoring array of the edge perception layer, and a historical switch operation data sequence set is acquired.

[0031] Firstly, the hardware structure of the target switch is determined. The target switch is an outdoor fully-enclosed small high-voltage isolation load switch. The outdoor fully-enclosed small high-voltage isolation load switch comprises a fully-enclosed shell, a three-phase double-break structure, a three-phase linkage spring operating mechanism, an edge perception layer and an intelligent control unit. The three-phase double-break structure integrates a vacuum arc-extinguishing chamber and a visible isolation break in series in each phase circuit. The three-phase linkage spring operating mechanism adopts a two-position spring mechanism and drives three-phase contacts to be synchronously opened and closed through a mechanical connecting rod. The edge perception layer is internally provided with a multi-source sensing monitoring array. The outdoor fully-enclosed small high-voltage isolation load switch refers to a miniaturized power device suitable for 10kV distribution networks, has double functions of isolation circuit and load current interruption, and adopts a fully-enclosed structure to adapt to complex outdoor environments.

[0032] The hardware of the target switch is a basic carrier for data acquisition, and its integrated structure provides hardware support for multi-source data acquisition. The specific structure and functions are as follows: the fully-enclosed shell is made of corrosion-resistant aluminum alloy material to realize dustproof, waterproof and condensation-proof, and adapt to the humid and dusty environment in rural areas; the three-phase double-break structure integrates a vacuum arc-extinguishing chamber and a visible isolation break in series, and each phase circuit independently completes the arc-extinguishing and isolation functions; the three-phase linkage spring operating mechanism drives three-phase contacts to be synchronously opened and closed through a mechanical connecting rod to ensure the consistency of opening and closing actions; the edge perception layer is an internal module, and the multi-source sensing monitoring array is the core functional component of this layer; and the intelligent control unit is responsible for receiving and temporarily storing the collected sensing data. For example, the GW4-10 series outdoor high-voltage isolation switch is taken as a unified example object, G represents an isolation switch, W represents an outdoor type, 4 represents a design serial number, and 10 represents a rated voltage of 10kV, which is a general standard model for 10kV distribution networks. The fully-enclosed shell of the GW4-10 series outdoor high-voltage isolation switch has a sealing level of IP67, the opening and closing synchronization error of the three-phase linkage spring operating mechanism is ≤2ms, and the multi-source sensing monitoring array internally provided in the edge perception layer includes eight different types of sensing elements, which can synchronously collect electrical, health and environmental data, and adapt to the large temperature difference and high humidity operating environment in rural areas.

[0033] Secondly, the operation monitoring index set of the target switch is acquired, wherein the operation monitoring index set comprises a switch operation state, electrical parameters, health indexes and environmental parameters, the switch operation state comprises a switching position and a switching operation time, the electrical parameters comprise a line current, a power factor and a three-phase voltage, the health indexes comprise a contact temperature and a contact contact resistance, and the environmental parameters comprise an environmental temperature and an environmental humidity. The operation monitoring index set is a parameter set for comprehensively characterizing a switch operation state, covering four dimensions of a switch state, electrical performance, device health and an external environment. Based on a switch operation requirement and a distribution network operation focus, a multi-dimensional and fully-covered operation monitoring index set is constructed, data collection is avoided, the effectiveness of long-term data is ensured, and the index classification and specific contents are as follows: the switch operation state reflects a switch operation condition, the electrical parameters reflect a line power supply performance, the health indexes reflect a switch main component aging degree, and the environmental parameters reflect a long-term influence of an outdoor environment on the switch.

[0034] For example, for a GW4-10 series switch, the operation monitoring index set is determined as follows: the switch operation state: a switching position (1 for closing and 0 for opening) and a switching operation time (accurate to milliseconds); the electrical parameters: a line current (range 0-630 A), a power factor (range 0.8-1.0) and a three-phase voltage (range 10 kV±10%); the health indexes: a contact temperature (range -40℃-150℃) and a contact contact resistance (range 0-100 mΩ); and the environmental parameters: an environmental temperature (range -40℃-85℃) and an environmental humidity (range 0-100%RH), covering key monitoring dimensions of a switch long-term operation full life cycle, and can capture long-term change trends such as contact aging and insulation performance decline.

[0035] Further, according to the preset data monitoring time interval and the set of running monitoring indicators, the target switch is monitored in real time by the multi-source sensing monitoring array of the edge perception layer in the historical time zone, and a set of historical switch running data sequences is obtained. The multi-source sensing monitoring array refers to a combined monitoring module integrating current sensors, temperature sensors, humidity sensors, displacement sensors and other sensing elements, which is built into the switch edge perception layer and can synchronously collect multi-dimensional running data. The historical time zone refers to the time range for collecting historical data, which is uniformly set as the last three years in this embodiment, meeting the historical data storage period requirement of the power industry. The set of historical switch running data sequences refers to a continuous set of switch running data arranged in chronological order in the historical time zone, with each data point corresponding to full-dimensional indicator data of a monitoring time point, and having the characteristics of not losing data when power is off and supporting remote access. According to the preset time interval and the determined monitoring indicators, the multi-source sensing monitoring array of the edge perception layer continuously collects data and arranges them in chronological order to form the set of historical switch running data sequences. The preset data monitoring time interval needs to balance data density and storage pressure, and an interval of 5 minutes / time can ensure data accuracy without causing storage overload; the historical time zone is selected as the last three years, which can completely cover a complete seasonal running cycle and a short-term aging process of the device; during the collection process, the sensing array converts analog signals into digital signals, which are stored by the intelligent control unit in chronological order and support subsequent remote access.

[0036] For example, for the GW4-10 series switch, real-time monitoring is performed in the historical time zone of the last three years at a preset interval of 5 minutes / time. Taking the starting time of data collection as January 1, 2020, 00:00 as an example, the data collected at this time point are: closing position one (closed state), closing operation time 20 ms, line current 320 A, power factor 0.92, A-phase voltage 10.1 kV / B-phase voltage 10.0 kV / C-phase voltage 10.2 kV, contact temperature 45℃, contact resistance 15 mΩ, ambient temperature 25℃, and ambient humidity 60%RH; the latest set of data collected at 00:05 on the same date three years later are: closing position one (closed state), closing operation time 21 ms, line current 318 A, power factor 0.91, A-phase voltage 10.0 kV / B-phase voltage 9.9 kV / C-phase voltage 10.1 kV, contact temperature 47℃, contact resistance 17 mΩ, ambient temperature 26℃, and ambient humidity 58%RH; and so on. A set of historical switch running data sequences with 315360 data points are formed in three years, each data point containing eleven specific indicator data, and completely recording the running state changes and environmental influence of the switch in the past three years.

[0037] In the embodiment of the present application, by embedding the multi-source sensing monitoring array in the switch edge perception layer and combining the fully enclosed shell, the problems of poor weather resistance and easy interference of traditional distributed sensors are solved, and the outdoor complex environment in remote areas and long-term operation requirements are adapted; The specific indicators of four dimensions are covered, the collected data is comprehensive, accurate and compliant; The historical switch operation data sequence set formed has completeness, time sequence and long periodicity, which can not only reflect the short-term operation state of the equipment, but also capture seasonal load fluctuations, component aging and other long-term regularities, providing high-quality and high-density training data sources for the operation state prediction of the long short-term memory network in the subsequent S200 step, and avoiding the deviation of the prediction result caused by data loss, distortion or short cycle.

[0038] S200: using a long short-term memory network, predicting and obtaining a predicted switch operation data sequence set in a preset time zone according to the historical switch operation data sequence set.

[0039] In the embodiment of the present application, a long short-term memory network is used to predict and obtain a predicted switch operation data sequence set in a preset time zone according to the historical switch operation data sequence set. In order to master the future operation state of the switch in advance for distribution network operation and maintenance to develop predictive strategies, this step is based on the historical data obtained in S100, combines data volatility analysis and the time sequence prediction advantage of the long short-term memory network (LSTM), and constructs a technical system of volatility perception, dynamic adaptation and accurate prediction, to ensure that the predicted data can truly reflect the future operation trend of the switch.

[0040] The step S200 in the method provided by the embodiment of the present application comprises:

[0041] Respectively, data volatility analysis is performed on a plurality of historical switch operation data sequences in the historical switch operation data sequence set, and a plurality of operation data variation coefficients are obtained, wherein the operation data variation coefficient is the ratio of the data standard deviation to the data mean value in the historical switch operation data sequence;

[0042] According to the plurality of operation data variation coefficients, the switch operation fluctuation degree in the historical time zone is determined by weighted evaluation;

[0043] Based on the switch operation fluctuation degree, a switch operation state prediction engine is called, and a predicted switch operation data sequence set in a preset time zone is predicted and obtained according to the historical switch operation data sequence set, wherein the switch operation state prediction engine is constructed based on a long short-term memory network.

[0044] First, data volatility analysis is performed on several historical switch operation data sequences in the historical switch operation data sequence set to obtain several coefficients of variation (CVs). The CV is the ratio of the standard deviation to the mean of the data in the historical switch operation data sequence. The CV measures the degree of volatility of a single operation data sequence, and its calculation formula is: CV = Standard Deviation of Data / Data Mean The value ranges from ≥0, with larger values ​​indicating more volatile sequences. Volatility analysis is performed on the core indicator sequences that have the greatest impact on the switch's operating state, avoiding interference from irrelevant indicators. The volatility characteristics of a single indicator are quantified using the coefficient of variation.

[0045] For example, from the eleven indicators collected by S100, four series are selected: line current, contact temperature, three-phase voltage, and ambient temperature, covering electrical parameters, health indicators, and environmental parameters. The mean and standard deviation of three years' data for each series are calculated, and then substituted into the formula for the coefficient of variation. For the three-year historical data of the GW4-10 series switch, the calculation is as follows: Line current series: mean Standard deviation coefficient of variation Contact temperature sequence: mean Standard deviation coefficient of variation Phase A voltage sequence: mean Standard deviation coefficient of variation Ambient temperature series: mean Standard deviation coefficient of variation The results showed that the A-phase voltage was most stable during the most drastic fluctuations in ambient temperature.

[0046] Secondly, the operational volatility of the switch within the historical time zone is determined by a weighted evaluation based on the coefficients of variation of the aforementioned operational data. Switch operational volatility refers to the weighted evaluation value of the coefficients of variation of all monitored indicators, reflecting the overall operational stability of the switch within the historical time zone. The value ranges from 0 to 1, where 0 represents no volatility and 1 represents extremely drastic volatility. Based on the weighted allocation coefficients of different indicators' influence on switch operation, the overall volatility is obtained through weighted summation, avoiding misjudgment of the overall state due to fluctuations in a single indicator.

[0047] For example, the analytic hierarchy process is used to determine the index weight: the health index (contact temperature) is directly related to the risk of equipment failure, with the highest weight, for example, 0.4; the electrical parameters (line current, A-phase voltage) are related to the power supply quality, with the second highest weight, for example, 0.25 and 0.2 respectively; the environmental parameter (ambient temperature) is an external influencing factor, with the lowest weight, for example, 0.15. The weighted sum of the four types of index variation coefficients of the GW4-10 series switch is: fluctuation degree = (0.0469 x 0.25) + (0.1667 x 0.4) + (0.0199 x 0.2) + (0.4545 x 0.15) ≈ 0.1506. For example, the grading standard is: 0-0.1 low fluctuation, 0.1-0.3 medium fluctuation, and above 0.3 high fluctuation. The running fluctuation degree of the switch is medium, and a medium number of prediction plug-ins need to be selected to balance accuracy and efficiency.

[0048] Further, based on the switch running fluctuation degree, a switch running state prediction engine is called to predict a set of predicted switch running data sequences in a preset time zone according to the set of historical switch running data sequences, wherein the switch running state prediction engine is constructed based on a long short-term memory network.

[0049] Wherein, based on the switch running fluctuation degree, a switch running state prediction engine is called, comprising:

[0050] Based on the historical running monitoring records of the same type of switches of the target switch, with the time span of the historical time zone as a constraint, a plurality of sample switch running data sequence sets are collected, and the historical switch running data sequence set in the preset time zone of different sample switch running data sequence sets is taken as a sample predicted switch running data sequence set, to obtain a plurality of sample predicted switch running data sequence sets, wherein the time span of the historical time zone and the preset time zone is the same;

[0051] The plurality of sample switch running data sequence sets and the plurality of sample predicted switch running data sequence sets are used as training data sets, and a with-replacement random selection strategy is used to construct K sample training sets using the training data sets, wherein K is an integer greater than or equal to 20;

[0052] The K sample training sets are used to train a long short-term memory network to convergence respectively with the sample switch running data sequence set as input and the sample predicted switch running data sequence set as supervision, to generate K switch running state prediction plug-ins;

[0053] Based on the principle of ensemble learning, the K switch running state prediction plug-ins are integrated and fused according to the mean fusion strategy to generate a switch running state prediction engine;

[0054] The ratio of the switch operation fluctuation to the historical maximum switch operation fluctuation in the historical time range is multiplied by K to obtain the number of adaptive plug-in selection J, wherein J is greater than or equal to 3 and less than or equal to K.

[0055] Randomly select J prediction plug-ins from the K switch operation state prediction plug-ins of the switch operation state prediction engine, and perform switch operation data prediction in the preset time zone according to the historical switch operation data sequence set.

[0056] First, based on the historical operation monitoring records of the same type of switch of the target switch, a plurality of sample switch operation data sequence sets are collected with the time span of the historical time zone as a constraint, and different sample switch operation data sequence sets in the historical switch operation data sequence set in the preset time zone are taken as sample prediction switch operation data sequence sets, thereby obtaining a plurality of sample prediction switch operation data sequence sets. The time span of the historical time zone and the preset time zone is the same. The same type of switch refers to a high-voltage disconnector with the same model number, consistent deployment scene, and similar operating environment as the target switch. The sample switch operation data sequence set refers to the historical time zone operation data of the same type of switch, which has the same time span as the target switch historical time zone and is used as input data for model training. The sample prediction switch operation data sequence set refers to the real data of the same type of switch in the preset time zone, which is used as the supervision label for model training.

[0057] For example, ensure that the time dimension of the sample data is completely aligned with the target switch, select three hundred groups of GW4-10 series same type switches deployed in rural 10kV distribution network, collect three years of historical operation data sequence set for each group of samples as sample input, collect historical data of each seven-day preset time zone in the corresponding three years as sample supervision label, and form a one-to-one correspondence between input and supervision. Select three hundred switches of the same type in a certain county rural distribution network as samples. For the first sample switch, collect historical data from January 1, 2021 to December 31, 2023 (three years) as sample input set, and collect historical data of the switch from January 8 to 14, 2021, from January 15 to 21, 2021, etc. The historical data of all seven-day time periods are taken as sample prediction set, and finally three groups of sample data pairs of three-year input + multi-segment seven-day supervision are formed.

[0058] Secondly, the plurality of sample switch operation data sequence sets and the plurality of sample predicted switch operation data sequence sets are used as a training data set, and a random selection strategy with replacement is used to construct K sample training sets by using the training data set, wherein K is an integer greater than or equal to 20. The random selection strategy with replacement refers to that when a new training set is constructed by randomly extracting samples from the training data set, the samples are put back into the original data set after each extraction, and repeated extraction is allowed, so that multiple independent training sets can be efficiently constructed. By constructing multiple independent training sets through random sampling with replacement, diversity basis is provided for subsequent ensemble learning, and the risk of overfitting is reduced.

[0059] For example, three hundred samples are randomly extracted from three hundred groups of samples, such as sample one, sample two, sample three,..., sample three hundred are extracted for the first time to construct training set one; three hundred samples are randomly extracted again for the second time, which may include sample three, sample three, sample five,..., to construct training set two; and the like, K = 30 training sets are generated, and each training set includes three hundred input-supervision sample pairs.

[0060] Further, the plurality of sample switch operation data sequence sets are used as input, the plurality of sample predicted switch operation data sequence sets are used as supervision, and the K sample training sets are used to train a long short-term memory network to convergence, to generate K switch operation state prediction plugins. The long short-term memory network (LSTM) is an improved recurrent neural network that solves the gradient disappearance problem in long sequence data training through a gating mechanism, and is suitable for predicting long-period time series data of power equipment operation. The prediction plugin refers to an independent LSTM model trained to convergence based on a single sample training set, has independent prediction capability, and is a core component unit of the prediction engine. Multiple LSTM sub-models are trained based on independent training sets, and the stability of ensemble prediction is improved through model diversity.

[0061] For example, 30 training sets are respectively configured with LSTM networks of the same structure, the input layer has eleven neurons corresponding to eleven operation monitoring indicators, two hidden layers each have 64 neurons, and the output layer has eleven neurons. The sample switch operation data sequence set is used as input, the sample predicted switch operation data sequence set is used as supervision, the Adam optimizer is used for training until the loss function (MAE) is less than 0.05 to converge, and 30 prediction plugins are generated. Plugin one is trained using training set one, the input is the index data of a certain sample switch for three years, the supervision is the true data of the corresponding seven days of the switch, and the loss function is reduced to 0.048 after one hundred rounds of training to converge; similarly, plugin one, plugin two,..., and plugin thirty are respectively trained using thirty training sets, and each plugin has the ability to independently predict seven-day switch operation data.

[0062] Subsequently, based on the principle of ensemble learning, the K switch operation state prediction plug-ins are integrated and fused according to the mean fusion strategy to generate a switch operation state prediction engine. Through the integration of multiple plug-ins by ensemble learning, the random error problem of single model prediction is solved. Based on the idea of majority over minority of ensemble learning, the K prediction plug-ins are integrated into a prediction engine, and the mean fusion strategy is adopted, that is, for the same prediction time point, the arithmetic mean of the output results of all plug-ins is taken as the temporary prediction value to ensure the stability of the prediction result. For example, taking the test data of a certain type of switch as input into the prediction engine, plug-in one predicts that the line current at a certain time point is 350A, plug-in two is 354A, …, and plug-in thirty is 348A. After mean fusion, the prediction value at this time point is (350+354+…+348) / 30=351A, with an error of only 0.29% from the true value 350A, which is 1.5% lower than the average error of a single plug-in.

[0063] Next, the ratio of the switch operation fluctuation degree to the historical maximum switch operation fluctuation degree in the historical time range is multiplied by K to obtain the number of adaptive plug-in selected J, wherein J is greater than or equal to 3 and less than or equal to K. The historical maximum switch operation fluctuation degree refers to the extreme fluctuation upper limit value obtained by statistical analysis of the operation data of the target switch and similar switches in the past three years, and the value in this embodiment is 0.8, which corresponds to the operation state under extremely severe weather. Dynamically matching the number of plug-ins avoids the waste of computing power caused by running all plug-ins. The calculation formula is: adaptive plug-in selection number J=round[(target switch fluctuation degree / historical maximum fluctuation degree)×K], while being forced to constrain 3≤J≤K to ensure prediction stability and adaptability. For example, the target switch fluctuation degree is 0.1506, the historical maximum fluctuation degree is 0.8, and K=30. Substituting into the formula gives J=round[(0.1506 / 0.8)×30]=round[5.6475]=6, which satisfies the constraint 3≤6≤30, that is, six plug-ins are selected for prediction, and only 20% of the computing power of the thirty plug-ins is started.

[0064] Finally, J prediction plugins are randomly selected from the K switch operation state prediction plugins of the switch operation state prediction engine, and switch operation data in the preset time zone is predicted according to the historical switch operation data sequence set. After J plugins are randomly selected and predicted in parallel, the prediction accuracy and efficiency are taken into account, the prediction efficiency is improved under the premise of ensuring the prediction accuracy, and the waste of computing resources is reduced. For example, five plugins 5, 12, 18, 22, 25 and 29 are randomly selected, and the historical data of the GW4-10 target switch in the past three years are input, and the six plugins output twelve index data such as line current and contact temperature in the future seven days; for the contact temperature at 14:00 on the fifth day, the predicted values of the six plugins are 53℃, 51℃, 54℃, 52℃, 53℃ and 52℃, and the final predicted value after mean fusion is 52.5℃, and the error with the subsequent actual operation value 53℃ is only 0.94%, and the computing resource consumption is saved by 80% compared with the full amount of thirty plugins.

[0065] In the embodiment of the application, based on the full-dimensional index long-term historical data collected by S100, the operation characteristics are quantified through fluctuation analysis, the long short-term memory network and the ensemble learning are combined, the adaptive number of prediction plugins is dynamically selected according to the operation fluctuation degree for prediction, only part of the plugins are started in the medium and low fluctuation scene, the prediction efficiency is effectively improved and the computing power redundancy is effectively reduced, and the plugins are increased in the high fluctuation to guarantee the accuracy. The full-dimensional prediction data in the preset time zone is finally output, and a reliable basis is provided for the optimization of the control parameters of S300.

[0066] S300: Based on the predicted switch operation data sequence set, the control parameters of the target switch are iteratively optimized and searched to minimize the line loss, the fault outage minutes, the voltage limit time and the voltage sag times, the adaptive control strategy is determined, and the target switch is intelligently controlled in the preset time zone according to the adaptive control strategy.

[0067] The method provided in the embodiment of the application comprises the following steps S300:

[0068] The control parameter adjustment space of the target switch is obtained, wherein the control parameters comprise protection parameters, operation parameters, operation parameters and system parameters;

[0069] A plurality of initial control parameters are randomly generated in the control parameter adjustment space;

[0070] The predicted switch operation data sequence set is fused with the plurality of initial control parameters to generate a plurality of switch operation control schemes;

[0071] According to the several control schemes of the switch operation, control simulation is performed in a digital operation simulation space of the target switch, and several simulated line losses, several simulated fault outage minutes, several simulated voltage over-limit times and several simulated voltage sag times are outputted;

[0072] The several simulated line losses, the several simulated fault outage minutes, the several simulated voltage over-limit times and the several simulated voltage sag times are weighted and evaluated to determine several parameter fitness degrees, wherein the parameter fitness degrees are negatively correlated with the simulated line losses, the simulated fault outage minutes, the simulated voltage over-limit times and the simulated voltage sag times.

[0073] The control parameter adjustment space is taken as an optimization space, and the control parameters of the target switch are iteratively optimized and searched based on the several initial control parameters and the several parameter fitness degrees to determine an adaptive control strategy.

[0074] The protection class parameters include instantaneous overcurrent action current threshold, timing limit overcurrent action current threshold, zero sequence current action threshold, overcurrent protection action delay and reclosing action delay, the operation class parameters include target closing phase angle, target opening phase angle and closing / opening operation speed, the operation class parameters include dynamic load current upper limit, alarm temperature threshold and forced air cooling start temperature threshold, and the system class parameters include reactive power compensation switching threshold and network reconstruction logic state.

[0075] Firstly, a control parameter adjustment space of the target switch is acquired, wherein the control parameters include protection class parameters, operation class parameters, operation class parameters and system class parameters. The protection class parameters include instantaneous overcurrent action current threshold, timing limit overcurrent action current threshold, zero sequence current action threshold, overcurrent protection action delay and reclosing action delay, the operation class parameters include target closing phase angle, target opening phase angle and closing / opening operation speed, the operation class parameters include dynamic load current upper limit, alarm temperature threshold and forced air cooling start temperature threshold, and the system class parameters include reactive power compensation switching threshold and network reconstruction logic state. The control parameter adjustment space refers to a legal value range of the control parameters of the target switch, which is determined according to the hardware performance limit of the switch, the operation and maintenance standard of the power industry and the actual operation demand of the distribution network, and is the boundary of the subsequent parameter optimization search, so as to avoid that the parameter setting exceeds the tolerance of the equipment or does not meet the industry standard.

[0076] Among them, the protection class parameter is used to realize the switch failure protection function, avoid the core parameter of equipment damage and line failure expansion, and is the protective barrier of safe operation of the switch. Specifically, it includes: instantaneous overcurrent action current threshold: the instantaneous overcurrent protection starting current standard set, when the line current instantaneously exceeds the threshold, the switch will quickly trigger the trip action to prevent short circuit and other sudden failure damage to the equipment; timing limit overcurrent action current threshold: the current threshold set for continuous overcurrent fault, which is used in combination with the action delay, when the line current exceeds the threshold and lasts for a set time, the switch will execute trip to avoid misoperation; zero sequence current action threshold: the current threshold for detecting line ground fault, when the line has single-phase ground fault, the zero sequence current will increase significantly, and the protection action will be triggered when the threshold is exceeded to ensure the safety of the distribution network grounding; overcurrent protection action delay: the time for the switch to delay the protection action after the overcurrent fault occurs, which can be adjusted according to the load characteristics of the distribution network to balance the demand for rapid fault removal and avoidance of instantaneous fault mis-trip; reclosing action delay: the time for the switch to delay the reclosing operation after tripping, which reserves the time for the fault to be eliminated by itself, reduces unnecessary power outage, and avoids reclosing on permanent faults.

[0077] Among them, the operation class parameter is the parameter for controlling the core action of switch closing and opening, which directly affects the stability and accuracy of switch operation. Specifically, it includes: target closing phase angle: the voltage or current phase angle set when the switch performs closing operation, selecting appropriate phase angle closing can reduce the closing inrush current, prolong the service life of the switch contact, and ensure the stability of the line; target opening phase angle: the voltage or current phase angle set when the switch performs opening operation, opening near the current zero crossing point can reduce the difficulty of arc extinction and reduce the loss of switch components caused by arc; closing and opening operation speed: the rate at which the switch contacts complete closing or opening action, too fast speed is easy to produce mechanical impact, and too slow speed will prolong the arc extinction time, which needs to be adapted to the mechanical performance of the spring operating mechanism.

[0078] Among them, the operation class parameter is the parameter for controlling the core action of switch closing and opening, which directly affects the stability and accuracy of switch operation. Specifically, it includes: target closing phase angle: the voltage or current phase angle set when the switch performs closing operation, selecting appropriate phase angle closing can reduce the closing inrush current, prolong the service life of the switch contact, and ensure the stability of the line; target opening phase angle: the voltage or current phase angle set when the switch performs opening operation, opening near the current zero crossing point can reduce the difficulty of arc extinction and reduce the loss of switch components caused by arc; closing and opening operation speed: the rate at which the switch contacts complete closing or opening action, too fast speed is easy to produce mechanical impact, and too slow speed will prolong the arc extinction time, which needs to be adapted to the mechanical performance of the spring operating mechanism.

[0079] Among them, the system class parameter refers to the parameter that adapts to the overall operation demand of the distribution network and realizes the optimization of the switch and the distribution network, specifically including: reactive power compensation switching threshold: based on the line power factor, the reactive power compensation device start-stop standard is set, when the power factor is lower or higher than the threshold, the reactive power compensation device is automatically switched, the line loss is reduced, and the power supply quality is improved; Network reconstruction logic state: set the trigger condition and execution logic of the switch participating in the distribution network topology reconstruction, when the distribution network fails or the load is uneven, the switch acts according to the preset logic to assist in reconstructing the distribution network structure and ensuring power supply continuity.

[0080] For example, the control parameter adjustment space details of the GW4-10 series switch: protection class: instantaneous overcurrent action current threshold (1.2-1.5 times rated current, i.e. 756-945A), time limit overcurrent action current threshold (1.1-1.3 times rated current, i.e. 693-819A), zero sequence current action threshold (5-10A, adaptive to distribution network grounding fault detection), overcurrent protection action delay (0.5-3s), reclosing action delay (0.5-3s); operation class: target closing phase angle (0-360°, voltage zero crossing 0° optimal), target opening phase angle (current zero crossing 180° optimal), closing and opening operation speed (0.8-1.2m / s, adaptive to spring mechanism mechanical limit); running class: dynamic load current upper limit (630-800A, adjusted according to load fluctuation), alarm temperature threshold (60-75℃, lower than contact resistance temperature 80℃), forced air cooling start temperature (70-75℃); system class: reactive power compensation switching threshold (power factor 0.85-0.95), network reconstruction logic state (load deviation > 20% automatically switched).

[0081] Secondly, a plurality of initial control parameters are randomly generated in the control parameter adjustment space. Five groups of parameters are generated in the adjustment space by using uniform sampling method, covering high, medium and low value intervals, to avoid the optimization being trapped in local optimum due to initial sample concentration. For example, group one: instantaneous overcurrent = 1.3x630 = 819A, closing phase angle = 90°, alarm temperature = 65℃, reactive power compensation switching threshold = 0.9; group two: instantaneous overcurrent = 1.5x630 = 945A, closing phase angle = 0°, alarm temperature = 70℃, switching threshold = 0.85; group three: instantaneous overcurrent = 1.2x630 = 756A, closing phase angle = 180°, alarm temperature = 60℃, switching threshold = 0.95; group four and group five: cross combination of the above parameters, covering the value scenes not involved.

[0082] Then, the predicted switch operation data sequence set is fused with the initial control parameters to generate a plurality of switch operation control schemes. Each set of initial parameters is bound to the working condition characteristics predicted by S200 to determine the parameter execution logic at different time periods, so that the scheme has timing adaptability. For example, after the first set of parameters is fused with the predicted data, the scheme is generated: morning peak (6:00-9:00, predicted current 800A): instantaneous overcurrent threshold maintained at 819A (slightly higher than the predicted peak value), forced air cooling start temperature reduced to 70°C (earlier heat dissipation), dynamic load current upper limit set to 800A; rainy day (3rd day, humidity 85%): closing phase angle fixed at 0° (voltage zero-crossing point, reducing the closing impact), reclosing delay extended to 2s (leaving time for fault self-elimination); night valley (0:00-5:00, predicted current 300A): reactive power compensation switching threshold raised to 0.95 (reducing invalid compensation), alarm temperature raised to 70°C (reducing false alarms).

[0083] In addition, in the digital operation simulation space of the target switch, control simulation is performed according to the plurality of switch operation control schemes, and a plurality of simulated line losses, a plurality of simulated fault outage minutes, a plurality of simulated voltage out-of-limit times, and a plurality of simulated voltage sag times are output. The digital operation simulation space refers to a virtual environment calibrated with historical data, taking the physical structure of the target switch and the distribution network topology as the prototype, which can reproduce the switch action and line response under load fluctuation and temperature and humidity changes, and output quantitative results. The control scheme is input into the digital operation simulation space, the switch operation is driven according to the predicted time sequence, the load fluctuation and environmental change scenarios are reproduced, and the quantitative results of the four types of targets are output. For example, five sets of schemes are simulated for seven days of operation: set one simulation result: simulated line loss is at a medium level; simulated fault outage minutes is 0 minutes; simulated voltage out-of-limit time is 10 minutes; simulated voltage sag times is 2 times. Set two simulation result: simulated line loss is at a lower level; simulated fault outage minutes is 30 minutes; simulated voltage out-of-limit time is 30 minutes; simulated voltage sag times is 3 times.

[0084] Subsequently, a plurality of parameter fitnesses are determined according to the plurality of simulated line losses, the plurality of simulated fault outage minutes, the plurality of simulated voltage out-of-limit times, and the plurality of simulated voltage sag times, with the minimization of line loss, fault outage minutes, voltage out-of-limit time, and voltage sag times as the multi-optimization target, wherein the parameter fitness is negatively correlated with the simulated line loss, the simulated fault outage minutes, the simulated voltage out-of-limit time, and the simulated voltage sag times. The parameter fitness refers to a comprehensive score allocated with weights according to the operation and maintenance priority, with a full score of 10, and is negatively correlated with the optimization target, such as no fault outage for a full score, and lower loss for a higher score.

[0085] For example, the weight is set according to the operation and maintenance priority, wherein the weight of the fault power failure is 40%, the weight of the voltage sag is 25%, the weight of the voltage overrun is 20%, and the weight of the line loss is 15%, and the score is calculated according to the full score of 10. Group one: no fault power failure occurs, 4 points are obtained according to the weight of 40% of the fault power failure; 2 times of voltage sag occurs, 3 points are obtained according to the weight of 25% of the voltage sag; the voltage overrun time is 10 minutes, 1.8 points are obtained according to the weight of 20% of the voltage overrun; the line loss is at a medium level, 1.2 points are obtained according to the weight of 15% of the line loss; the total score is 10 by adding the scores of the four items. Group two: 1 time of fault power failure occurs and lasts for 30 minutes, 4 points are deducted according to the weight of 40% of the fault power failure, and 0 points are obtained for this item; 3 times of voltage sag occurs, 2 points are obtained according to the weight of 25% of the voltage sag; the voltage overrun time is 30 minutes, 1 point is obtained according to the weight of 20% of the voltage overrun; the line loss is at a low level, 1.5 points are obtained according to the weight of 15% of the line loss; the total score is 4.5 by adding the scores of the four items.

[0086] Further, the control parameter adjustment space is taken as an optimization space, and the control parameter of the target switch is iteratively optimized and searched based on the plurality of initial control parameters and the plurality of parameter fitnesses to determine an adaptive control strategy.

[0087] Further, the control parameter adjustment space is taken as an optimization space, and the control parameter of the target switch is iteratively optimized and searched based on the plurality of initial control parameters and the plurality of parameter fitnesses to determine an adaptive control strategy, including:

[0088] The initial control parameters are set as initial solutions, the plurality of initial solutions are arranged in descending order of parameter fitness to generate an initial solution sequence;

[0089] The first solution of the initial solution sequence is selected as a good solution, the last solution is selected as a poor solution, and the remaining initial solutions except the good solution and the poor solution are selected as inferior solutions to obtain the good solution, the plurality of inferior solutions, and the poor solution;

[0090] The plurality of inferior solutions are updated and adjusted according to a preset optimization direction and a preset optimization step, and are reordered to obtain an updated solution sequence;

[0091] A preset number of solutions in the updated solution sequence are eliminated, and a control parameter that does not appear in the optimization process is randomly selected in the control parameter adjustment space for equivalent supplement to generate an updated solution sequence, wherein the preset number decreases as the number of optimization times increases;

[0092] Based on the optimization mechanism of the optimization direction selection-inferior solution adjustment-solution elimination-solution supplement, the good solution of the updated solution sequence is output as the adaptive control strategy according to the updated solution sequence.

[0093] Firstly, the initial control parameters are set as initial solutions, and the initial solutions are arranged according to the parameter fitness from large to small to generate an initial solution sequence. The initial solution refers to the generated initial control parameter combination, and each initial solution corresponds to a group of control parameters and the corresponding parameter fitness. The initial solutions are sorted according to the fitness from large to small to determine the adaptation level of each solution. For example, five groups of initial solutions are sorted according to the fitness as follows: group one (10 points), group four (8.5 points), group three (7.8 points), group two (6.2 points), and group five (4.5 points), forming an initial solution sequence: [group one, group four, group three, group one, group five].

[0094] Secondly, the first solution of the initial solution sequence is selected as the optimal solution, the last solution is selected as the poor solution, and the remaining initial solutions except the optimal solution and the poor solution are selected as the inferior solutions, obtaining the optimal solution, the plurality of inferior solutions and the poor solution. The first and last solutions of the sequence are extracted as the optimal solution and the poor solution, and the middle solutions are extracted as the inferior solutions, thereby determining the adjustment object and the reference. For example, the first solution group one of the sequence is selected as the optimal solution, the last solution group five of the sequence is selected as the poor solution, and the remaining groups four, three and two are selected as the inferior solutions.

[0095] Further, the plurality of inferior solutions are updated and adjusted according to a preset optimization direction and a preset optimization step, and a new solution sequence is obtained by reordering.

[0096] The preset optimization direction setting method comprises:

[0097] The optimal solution parameter fitness of the optimal solution, the poor solution parameter fitness of the poor solution, and the plurality of inferior solution parameter fitnesses of the plurality of inferior solutions are obtained, and the mean value of the inferior solution parameter fitness is calculated;

[0098] The optimal solution parameter fitness and the mean value of the inferior solution parameter fitness are subjected to deviation calculation, and the absolute value of the deviation is taken as the optimal solution fitness deviation;

[0099] The mean value of the inferior solution parameter fitness and the poor solution parameter fitness are subjected to deviation calculation, and the absolute value of the deviation is taken as the poor solution fitness deviation;

[0100] If the optimal solution fitness deviation is greater than or equal to the poor solution fitness deviation, the optimization direction of this optimization is set as the optimization strategy, wherein the optimization strategy is the direction close to the optimal solution;

[0101] If the optimal solution fitness deviation is less than the poor solution fitness deviation, the optimization direction of this optimization is set as the poor solution avoidance strategy, wherein the poor solution avoidance strategy is the direction away from the poor solution.

[0102] First, the optimal solution parameter fitness, the poor solution parameter fitness and the plurality of poor solution parameter fitnesses of the plurality of poor solutions are obtained, and the poor solution parameter fitness average is calculated. The poor solution parameter fitness average is the sum of the parameter fitness values of all poor solutions divided by the number of poor solutions, reflecting the average performance of the medium adaptation parameters. Extract the fitness values of the optimal solution, the poor solution and all poor solutions, calculate the poor solution fitness average by arithmetic mean, and determine the medium adaptation level benchmark. For example, the fitness of each solution is obtained: optimal solution fitness = 10 points, poor solution fitness = 4.5 points, and the fitness of the three poor solutions is 8.5 points, 7.8 points and 6.2 points respectively. The poor solution fitness average = (8.5+7.8+6.2) / 3 = 7.5 points is calculated.

[0103] Secondly, the deviation of the optimal solution parameter fitness and the poor solution parameter fitness average is calculated, and the absolute value of the deviation is taken as the optimal solution fitness deviation. The optimal solution fitness deviation is the absolute difference between the optimal solution fitness and the poor solution fitness average, which quantifies the gap between the optimal solution and the medium adaptation level. The optimal solution fitness deviation is obtained by subtracting the poor solution fitness average from the optimal solution fitness and taking the absolute value, which measures the advantage of the optimal solution. For example, the optimal solution fitness deviation = | optimal solution fitness - poor solution fitness average | = | 10 - 7.5 | = 2.5 points, indicating that the optimal solution is 2.5 points higher than the medium adaptation level.

[0104] Further, the deviation of the poor solution parameter fitness average and the poor solution parameter fitness is calculated, and the absolute value of the deviation is taken as the poor solution fitness deviation. The poor solution fitness deviation is the absolute difference between the poor solution fitness average and the poor solution fitness, which quantifies the gap between the poor solution and the medium adaptation level. The poor solution fitness deviation is obtained by subtracting the poor solution fitness from the poor solution fitness average and taking the absolute value, which measures the disadvantage of the poor solution. For example, the poor solution fitness deviation = | poor solution fitness average - poor solution fitness | = | 7.5 - 4.5 | = 3.0 points, indicating that the poor solution is 3.0 points lower than the medium adaptation level.

[0105] Further, if the optimal solution fitness deviation is greater than or equal to the poor solution fitness deviation, the optimization direction of this optimization is set to the optimization strategy, wherein the optimization strategy is the direction of approaching the optimal solution. The optimization strategy is the parameter value range of the optimal solution, which improves the adaptability by adjusting the poor solution to approach the optimal solution. The two deviations are compared, and if the optimal solution deviation is greater than or equal to the poor solution deviation, the optimization strategy is selected.

[0106] Thirdly, if the fitness deviation of the optimal solution is less than the fitness deviation of the poor solution, the optimization direction of this time is set as an avoiding poor solution strategy, wherein the avoiding poor solution strategy is a direction away from the poor solution. The avoiding poor solution strategy refers to that the optimization direction deviates from the parameter value range of the poor solution, and the effective search range is expanded by avoiding the inefficient parameter combination of the poor solution. By comparing the sizes of the two deviations, if the fitness deviation of the optimal solution is less than the fitness deviation of the poor solution, the avoiding poor solution strategy is selected. For example, the fitness deviation of the optimal solution is 2.5 points, which is less than the fitness deviation of the poor solution of 3.0 points, which indicates that the inefficient parameter combination of the poor solution interferes more with the optimization, and the avoiding poor solution strategy is selected. According to the preset optimization step, the parameters close to the poor solution in the poor solution are mainly adjusted, and the unreasonable values away from the poor solution are adjusted. The original instantaneous overcurrent 780A of group four is adjusted to 810A; the original closing phase angle of 120° of group three is adjusted to 30°; and the original alarm temperature of 60℃ of group two is adjusted to 65℃, so as to expand the effective search range by avoiding the shortcomings of the poor solution.

[0107] On this basis, the plurality of poor solutions are updated and adjusted according to the preset optimization direction and the preset optimization step, and are reordered to obtain an updated solution sequence. For example, the three groups of adjusted poor solutions are input into the simulation space for re-evaluation, and a new fitness is obtained: group four (9.2 points), group three (8.8 points), and group two (8.0 points). In combination with the original optimal solution and the poor solution, the updated solution sequence is generated by sorting according to the fitness from large to small: [group one (10 points), group four (9.2 points), group three (8.8 points), group two (8.0 points), and group five (4.5 points)].

[0108] Further, a preset number of solutions in the updated solution sequence are eliminated, and a control parameter that has not appeared in the optimization process is randomly selected in the control parameter adjustment space for equivalent supplement to generate an updated solution sequence, wherein the preset number decreases as the number of optimizations increases. The preset number refers to the number of solutions to be eliminated in each iteration, which decreases as the number of optimizations increases. In the early stage, the search range is expanded by eliminating inefficient solutions, and in the later stage, the optimal solution is stabilized by reducing the elimination. The equivalent supplement refers to that the number of new solutions supplemented is equal to the number of eliminated solutions, so as to keep the total number of solutions unchanged and ensure the stability of the optimization sample size. The control parameter that has not appeared refers to a control parameter combination that has never been included in any initial solution or updated solution and is strictly within the control parameter adjustment space, so as to avoid repeated search. The updated solution sequence refers to a new solution sequence formed after the poor solution is adjusted, reordered, eliminated, and supplemented, which is the basis sample for the next iteration.

[0109] For example, the more the number of optimization times, the less the number of elimination. The preset number is set according to the rule, and the preset elimination number is two for the first iteration. In the updated solution sequence, the two solutions with the worst adaptability at the end are eliminated, that is, group two (8.0 points) and group five (4.5 points). After eliminating the two solutions, two groups of new control parameter combinations that have never appeared are supplemented to maintain the size of five groups of solutions: supplemented solution one: instantaneous overcurrent 810A, closing phase angle 0°, alarm temperature 65°C; supplemented solution two: instantaneous overcurrent 820A, closing phase angle 60°, alarm temperature 70°C. The updated solution sequence is generated once: [group one (10 points), group four (9.2 points), group three (8.8 points), supplemented solution one, supplemented solution two].

[0110] Finally, based on the optimization mechanism of selecting the optimization direction, adjusting the inferior solution, eliminating the solution, and supplementing the solution, the iterative optimization is continued according to the updated solution sequence once until the preset convergence number is reached, and the optimal solution of the current updated solution sequence is output as the adaptive control strategy. The preset convergence number is the number of continuous iterations set, and when the fitness of the optimal solution fluctuates by no more than the preset threshold value in continuous iterations, the cumulative number reaches the value, and the optimization is stopped. The direction of each optimization may be different, and is set based on the actual state to maximize the convergence speed of the optimization. For example, the preset convergence number is three, and the fitness fluctuation threshold value is 0.5 points. According to the closed-loop mechanism, the second iteration is focused on avoiding the low-efficiency parameters of the poor solution: the optimization direction is still the difference avoidance strategy, the supplemented solution one and the supplemented solution two are adjusted, one solution at the end (supplemented solution two, fitness 8.2 points) is eliminated, one group of new solutions (supplemented solution three: instantaneous overcurrent 815A, closing phase angle 0°, alarm temperature 65°C) is supplemented, and the fitness of the optimal solution group one remains 10 points. In the third iteration, there is no elimination and supplementation, the fitness of the adjusted supplemented solution three is increased to 9.1 points, the fitness of group one is still stable at 10 points, the fitness of the optimal solution fluctuates by no more than 0.5 points in three consecutive iterations, and the preset convergence number is met. The adaptive control strategy is output: the current optimal solution is group one, and the core parameters of the final adaptive control strategy are: instantaneous overcurrent 819A, closing phase angle 0°, and alarm temperature 65°C.

[0111] In this embodiment of the invention, based on predicted switch operation data, the focus is on multi-objective optimization requirements. A dynamically adaptable optimization direction selection mechanism precisely balances the need to improve efficiency by seeking the best solution while avoiding errors and expanding the search range. Inferior solution updates and adjustments, combined with gradient step size, effectively correct the shortcomings of moderately adaptable parameters. Dynamically decreasing the number of eliminated solutions, coupled with the addition of equivalent new solutions, avoids inefficient solutions consuming computational power and introduces new parameter combinations to prevent the optimization from getting stuck in local optima. A closed-loop iterative mechanism, combined with a preset convergence standard, ensures the optimization process is efficient and stable, avoiding excessive iteration and wasted resources. The final output adaptive control strategy, verified through multiple rounds of simulation, can accurately adapt to the complex future operating conditions of the switch, reduce power outages and voltage anomalies, reduce line losses, achieve intelligent control without manual intervention, and is suitable for unattended distribution network scenarios in remote areas, effectively improving power supply stability and operating efficiency.

[0112] Through the specific implementation methods described above, the embodiments of the present invention achieve the following technical effects:

[0113] This invention provides an intelligent control method and system for outdoor fully enclosed small high-voltage disconnect load switches. It collects indicators covering all dimensions of switch operation, laying a solid data foundation for subsequent processes and ensuring data integrity and compliance. Combining long short-term memory networks and ensemble learning, it dynamically adapts the number of prediction plugins, improving prediction efficiency and reducing computational redundancy while ensuring the accuracy of core indicator predictions. This provides a reliable reference for future operating conditions in control strategy formulation. Guided by multi-objective optimization, it generates highly adaptable control strategies through dynamic selection of optimization directions, precise adjustment of inferior solutions, dynamic elimination of supplementary solutions, and closed-loop iterative convergence, effectively improving the operating efficiency and power supply reliability of distribution networks in remote areas.

[0114] Example 2, as Figure 2 As shown, this invention provides an intelligent control system for a fully enclosed outdoor miniature high-voltage disconnecting load switch, the system comprising:

[0115] The edge sensing acquisition module 11 is used to acquire a set of historical switch operation data sequences of the target switch in the historical time zone through the multi-source sensing monitoring array of the edge sensing layer, wherein the target switch is an outdoor fully enclosed small high-voltage isolation load switch;

[0116] Data prediction module 12 is used to use a long short-term memory network to predict and obtain a set of predicted switch operation data sequences within a preset time zone based on the historical switch operation data sequence set;

[0117] The multi-objective optimization control module 13 is configured to perform iterative optimization search on the control parameters of the target switch based on the set of predicted switch operation data sequences, with the multi-optimization objectives of minimizing line loss, fault outage minutes, voltage out-of-limit time, and voltage sag times, to determine an adaptive control strategy, and to perform intelligent control on the target switch in the preset time zone according to the adaptive control strategy.

[0118] In one embodiment, the edge-aware acquisition module 11 is further configured to:

[0119] The outdoor fully-enclosed small high-voltage isolation load switch includes a fully-enclosed shell, a three-phase double-break structure, a three-phase linkage spring operating mechanism, an edge-aware layer, and an intelligent control unit. The three-phase double-break structure integrates a vacuum arc-extinguishing chamber and a visible isolation break in series in each phase circuit. The three-phase linkage spring operating mechanism adopts a two-position spring mechanism and drives three-phase contacts to be synchronously opened and closed through a mechanical connecting rod. The edge-aware layer is internally provided with a multi-source sensing monitoring array.

[0120] The operation monitoring index set of the target switch is acquired, wherein the operation monitoring index set includes a switch operation state, electrical parameters, health indicators, and environmental parameters. The switch operation state includes opening and closing positions and opening and closing operation times. The electrical parameters include line current, power factor, and three-phase voltage. The health indicators include contact temperature and contact contact resistance. The environmental parameters include environmental temperature and environmental humidity.

[0121] In the historical time zone, the target switch is monitored in real time by the multi-source sensing monitoring array of the edge-aware layer according to a preset data monitoring time interval and the operation monitoring index set, to acquire a set of historical switch operation data sequences.

[0122] In one embodiment, the data prediction module 12 is further configured to:

[0123] Data fluctuation analysis is respectively performed on a plurality of historical switch operation data sequences in the set of historical switch operation data sequences, to acquire a plurality of operation data variation coefficients. The operation data variation coefficient is a ratio of a data standard deviation to a data mean value in the historical switch operation data sequence.

[0124] The switch operation fluctuation degree in the historical time zone is determined according to the weighted evaluation of the plurality of operation data variation coefficients.

[0125] The switch operation state prediction engine is called based on the switch operation fluctuation degree, to predict and acquire a set of predicted switch operation data sequences in the preset time zone according to the set of historical switch operation data sequences. The switch operation state prediction engine is constructed based on a long short-term memory network.

[0126] The switch operation fluctuation degree is used to call a switch operation state prediction engine, including:

[0127] Based on the historical operation monitoring records of the same type of switches of the target switch, a plurality of sample switch operation data sequence sets are collected within a time span of the historical time zone as a constraint, and different sample switch operation data sequence sets are used as sample prediction switch operation data sequence sets within a historical switch operation data sequence set in a preset time zone, to obtain a plurality of sample prediction switch operation data sequence sets, wherein the time span of the historical time zone and the preset time zone is the same.

[0128] The plurality of sample switch operation data sequence sets and the plurality of sample prediction switch operation data sequence sets are used as training data sets, and a random selection strategy is used to construct K sample training sets using the training data sets, wherein K is an integer greater than or equal to 20.

[0129] The sample switch operation data sequence set is used as input, and the sample prediction switch operation data sequence set is used as supervision, and the K sample training sets are used to train a long short-term memory network to convergence, to generate K switch operation state prediction plug-ins.

[0130] Based on the principle of ensemble learning, the K switch operation state prediction plug-ins are integrated and fused according to the mean fusion strategy to generate a switch operation state prediction engine.

[0131] The ratio of the switch operation fluctuation degree to the historical maximum switch operation fluctuation degree within the historical time range is multiplied by K to obtain an adaptive plug-in selection number J, wherein J is greater than or equal to 3 and less than or equal to K.

[0132] In the K switch operation state prediction plug-ins of the switch operation state prediction engine, J prediction plug-ins are randomly selected, and switch operation data within the preset time zone is predicted according to the historical switch operation data sequence set.

[0133] In one embodiment, the multi-objective optimization control module 13 is also used for:

[0134] Obtaining a control parameter adjustment space of the target switch, wherein the control parameters include protection type parameters, operation type parameters, operation type parameters, and system type parameters;

[0135] Randomly generating a plurality of initial control parameters within the control parameter adjustment space;

[0136] The prediction switch operation data sequence set is fused with the plurality of initial control parameters to generate a plurality of switch operation control schemes;

[0137] According to the several switch operation control schemes, control simulation is respectively performed in a digital operation simulation space of the target switch, and several simulated line losses, several simulated fault outage minutes, several simulated voltage over-limit times and several simulated voltage sag times are outputted;

[0138] The several simulated line losses, the several simulated fault outage minutes, the several simulated voltage over-limit times and the several simulated voltage sag times are weighted and evaluated to determine several parameter fitness degrees, wherein the parameter fitness degrees are negatively correlated with the simulated line losses, the simulated fault outage minutes, the simulated voltage over-limit times and the simulated voltage sag times.

[0139] The control parameter adjustment space is taken as an optimization space, and the control parameters of the target switch are iteratively optimized and searched based on the several initial control parameters and the several parameter fitness degrees to determine an adaptive control strategy.

[0140] The protection class parameters include an instantaneous overcurrent action current threshold, a time limit overcurrent action current threshold, a zero sequence current action threshold, an overcurrent protection action delay and a reclosing action delay, the operation class parameters include a target closing phase angle, a target opening phase angle and a closing / opening operation speed, the operation class parameters include a dynamic load current upper limit, an alarm temperature threshold and a forced air cooling start temperature threshold, and the system class parameters include a reactive power compensation switching threshold and a network reconstruction logic state.

[0141] The control parameter adjustment space is taken as an optimization space, and the control parameters of the target switch are iteratively optimized and searched based on the several initial control parameters and the several parameter fitness degrees to determine an adaptive control strategy, including:

[0142] The initial control parameters are set as initial solutions, the several initial solutions are arranged in descending order of parameter fitness degrees, and an initial solution sequence is generated;

[0143] The first solution of the initial solution sequence is set as a good solution, the last solution is set as a poor solution, and the remaining initial solutions except the good solution and the poor solution are set as inferior solutions, so that the good solution, the plurality of inferior solutions and the poor solution are obtained;

[0144] According to a preset optimization direction and a preset optimization step, the plurality of inferior solutions are updated and adjusted, and are reordered to obtain an updated solution sequence;

[0145] A preset number of solutions in the updated solution sequence are eliminated, control parameters that have not appeared in an optimization process are randomly selected and equivalently supplemented in the control parameter adjustment space to generate an updated solution sequence, and the preset number decreases with an increase in the number of optimization times;

[0146] Based on the optimization direction selection-weak solution adjustment-elimination of solution-supplement of solution optimization mechanism, according to the sequence of the updated solution, iterative optimization is continued until the preset convergence number is reached, and the optimal solution of the current updated solution sequence is output as the adaptive control strategy.

[0147] The preset optimization direction setting method comprises:

[0148] The optimal solution parameter fitness of the optimal solution, the difference solution parameter fitness of the difference solution, and the plurality of weak solution parameter fitnesses of the plurality of weak solutions are obtained, and the weak solution parameter fitness mean is calculated.

[0149] The optimal solution parameter fitness and the weak solution parameter fitness mean are subjected to deviation calculation, and the absolute value of the deviation is taken as the optimal solution fitness deviation.

[0150] The weak solution parameter fitness mean and the difference solution parameter fitness are subjected to deviation calculation, and the absolute value of the deviation is taken as the difference solution fitness deviation.

[0151] If the optimal solution fitness deviation is greater than or equal to the difference solution fitness deviation, the optimization direction of this optimization is set to the optimization strategy, wherein the optimization strategy is the direction close to the optimal solution.

[0152] If the optimal solution fitness deviation is less than the difference solution fitness deviation, the optimization direction of this optimization is set to the difference avoidance strategy, wherein the difference avoidance strategy is the direction away from the difference solution.

[0153] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the above describes a specific embodiment of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0154] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0155] The present application is only an exemplary description of the present application, and any and all modifications, changes, combinations or equivalents within the scope of the present application are considered to have been covered. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalent technology, the present application is intended to include these modifications and changes.

Claims

1. A method for intelligent control of outdoor fully enclosed miniature high-voltage disconnecting load switches, characterized in that: The methods include: The target switch is a small, fully enclosed outdoor high-voltage isolating load switch. The target switch is a set of historical switch operation data sequences in the historical time zone collected by the multi-source sensor monitoring array of the edge sensing layer. Using a long short-term memory network, a predicted set of switch operation data sequences within a preset time zone is obtained based on the historical switch operation data sequence set. Based on the predicted switch operation data sequence set, with minimizing line loss, fault outage minutes, voltage over-limit time and voltage sag count as multiple optimization objectives, the control parameters of the target switch are iteratively optimized and searched to determine the appropriate control strategy, and the target switch is intelligently controlled in the preset time zone according to the appropriate control strategy. Specifically, based on the predicted switch operation data sequence set, with minimizing line loss, fault outage minutes, voltage over-limit time, and voltage sag count as multiple optimization objectives, the control parameters of the target switch are iteratively optimized and searched to determine the suitable control strategy, including: Obtain the control parameter adjustment space of the target switch, wherein the control parameters include protection parameters, operation parameters, running parameters and system parameters; Several initial control parameters are randomly generated within the control parameter adjustment space; The predicted switch operation data sequence set is fused with the several initial control parameters to generate several switch operation control schemes; Within the digital operation simulation space of the target switch, control simulations are performed according to the several switch operation control schemes, and several simulated line losses, several simulated fault outage minutes, several simulated voltage over-limit times, and several simulated voltage sags are output. With minimizing line loss, fault outage minutes, voltage over-limit time, and voltage sag count as multiple optimization objectives, several parameter fitnesss are determined by weighted evaluation based on several simulated line losses, several simulated fault outage minutes, several simulated voltage over-limit time, and several simulated voltage sag counts. The parameter fitnesss are negatively correlated with simulated line losses, simulated fault outage minutes, simulated voltage over-limit time, and simulated voltage sag counts. Using the control parameter adjustment space as the search space, and based on the several initial control parameters and several parameter fitness values, the control parameters of the target switch are iteratively optimized and searched to determine the suitable control strategy, including: The initial control parameters are set as the initial solutions, and several initial solutions are arranged in descending order of parameter fitness to generate an initial solution sequence. The first solution of the initial solution sequence is selected as the optimal solution, the last solution is selected as the poor solution, and the remaining initial solutions other than the optimal and poor solutions are selected as inferior solutions, thus obtaining the optimal solution, multiple inferior solutions, and poor solutions; According to the preset optimization direction and the preset optimization step size, the multiple inferior solutions are updated and adjusted, and then reordered to obtain the updated solution sequence; After eliminating a preset number of solutions from the updated solution sequence, control parameters that did not appear during the optimization process are randomly selected within the control parameter adjustment space and supplemented with equivalent values ​​to generate an updated solution sequence. The preset number decreases as the number of optimization attempts increases. Based on the optimization mechanism of selecting optimization direction, adjusting inferior solutions, eliminating solutions, and supplementing solutions, iterative optimization is continued according to the updated solution sequence until the preset number of convergences is reached, and the optimal solution of the current updated solution sequence is output as the adaptation control strategy.

2. The intelligent control method for outdoor fully enclosed miniature high-voltage disconnector load switches according to claim 1, characterized in that, The outdoor fully enclosed miniature high-voltage disconnect load switch includes a fully enclosed housing, a three-phase double-break structure, a three-phase linkage spring operating mechanism, an edge sensing layer, and an intelligent control unit. The three-phase double-break structure integrates the vacuum interrupter and the visible isolation break in series in each phase circuit. The three-phase linkage spring operating mechanism adopts a two-position spring mechanism, which drives the three-phase contacts to open and close synchronously through mechanical linkages. The edge sensing layer has a built-in multi-source sensor monitoring array.

3. The intelligent control method for outdoor fully enclosed miniature high-voltage disconnecting load switch according to claim 1, characterized in that, The target switch's historical switch operation data sequence set within the historical time zone is collected through a multi-source sensing monitoring array at the edge sensing layer, including: Obtain the set of operation monitoring indicators for the target switch, wherein the set of operation monitoring indicators includes switch operation status, electrical parameters, health indicators and environmental parameters, the switch operation status includes opening and closing positions and opening and closing operation time, the electrical parameters include line current, power factor and three-phase voltage, the health indicators include contact temperature and contact resistance, and the environmental parameters include ambient temperature and ambient humidity; According to the preset data monitoring time interval and the set of operation monitoring indicators, the target switch is monitored in real time through the multi-source sensor monitoring array of the edge perception layer within the historical time zone to obtain the historical switch operation data sequence set.

4. The intelligent control method for outdoor fully enclosed miniature high-voltage disconnecting load switch according to claim 1, characterized in that, Using a Long Short-Term Memory (LSTM) network, a predicted set of switch operation data sequences within a preset time zone is obtained based on the historical switch operation data sequence set, including: Data volatility analysis is performed on several historical switch operation data sequences in the historical switch operation data sequence set to obtain several operation data variation coefficients, wherein the operation data variation coefficient is the ratio of the data standard deviation to the data mean in the historical switch operation data sequence; The switching operation fluctuation within the historical time zone is determined by weighted evaluation based on the coefficient of variation of the aforementioned operational data. Based on the fluctuation of the switch operation, the switch operation status prediction engine is invoked to predict the switch operation data sequence set within a preset time zone according to the historical switch operation data sequence set. The switch operation status prediction engine is constructed based on a long short-term memory network.

5. The intelligent control method for outdoor fully enclosed miniature high-voltage disconnecting load switch according to claim 4, characterized in that, The switch operation fluctuation is used to invoke the switch operation state prediction engine, including: Based on the historical operation monitoring records of similar switches of the target switch, and constrained by the time span of the historical time zone, several sample switch operation data sequence sets are collected. The historical switch operation data sequence sets within the preset time zone of different sample switch operation data sequence sets are used as sample prediction switch operation data sequence sets to obtain several sample prediction switch operation data sequence sets. The historical time zone and the preset time zone have the same time span. The training dataset is constructed using the aforementioned sample switch running data sequence set and the aforementioned sample prediction switch running data sequence set, and a random selection strategy with replacement is adopted. K sample training sets are constructed using the training dataset, where K is an integer greater than or equal to 20. Using the sample switch operation data sequence set as input and the sample predicted switch operation data sequence set as supervision, the long short-term memory network is trained to converge using the K sample training sets to generate K switch operation state prediction plug-ins. Based on the principle of ensemble learning, the K switch operation state prediction plug-ins are integrated and fused according to the mean fusion strategy to generate a switch operation state prediction engine. The number of compatible plug-ins selected, J, is obtained by multiplying the ratio of the switch operation fluctuation to the historical maximum switch operation fluctuation within the historical time range by K and rounding it down, where J is greater than or equal to 3 and less than or equal to K. J prediction plugins are randomly selected from the K switch operation status prediction plugins of the switch operation status prediction engine, and switch operation data within the preset time zone are predicted based on the historical switch operation data sequence set.

6. The intelligent control method for outdoor fully enclosed miniature high-voltage disconnecting load switch according to claim 1, characterized in that, The protection parameters include instantaneous overcurrent operating current threshold, definite-time overcurrent operating current threshold, zero-sequence current operating threshold, overcurrent protection operating delay, and reclosing operating delay. The operation parameters include target closing phase angle, target opening phase angle, and closing / opening operation speed. The running parameters include dynamic load current upper limit, alarm temperature threshold, and forced air-cooled start-up temperature threshold. The system parameters include reactive power compensation switching threshold and network reconfiguration logic status.

7. The intelligent control method for outdoor fully enclosed miniature high-voltage disconnecting load switch according to claim 1, characterized in that, The method for setting the preset optimization direction includes: Obtain the fitness of the optimal solution parameters, the fitness of the poor solution parameters, and the fitness of the multiple inferior solutions parameters, and calculate the average fitness of the inferior solutions parameters. The deviation of the fitness values ​​of the optimal solution parameters and the mean fitness values ​​of the inferior solution parameters is calculated, and the absolute value of the deviation is taken as the fitness deviation of the optimal solution. The deviation between the mean fitness of the inferior solution parameters and the fitness of the poor solution parameters is calculated, and the absolute value of the deviation is taken as the fitness deviation of the poor solution. If the fitness deviation of the optimal solution is greater than or equal to the fitness deviation of the poor solution, then the optimization direction for this optimization is set as the optimization strategy, wherein the optimization strategy is the direction that moves closer to the optimal solution; If the fitness deviation of the optimal solution is less than the fitness deviation of the poor solution, then the optimization direction for this optimization is set as the deviation avoidance strategy, wherein the deviation avoidance strategy is the direction away from the poor solution.

8. An intelligent control system for a fully enclosed outdoor miniature high-voltage disconnector load switch, characterized in that: The system is used to implement the intelligent control method for outdoor fully enclosed miniature high-voltage disconnect load switches according to any one of claims 1-7, the system comprising: The edge sensing acquisition module is used to acquire a set of historical switch operation data sequences of the target switch in the historical time zone through the multi-source sensor monitoring array of the edge sensing layer, wherein the target switch is an outdoor fully enclosed small high-voltage isolation load switch; The data prediction module is used to use a long short-term memory network to predict and obtain a set of predicted switch operation data sequences within a preset time zone based on the historical switch operation data sequence set. The multi-objective optimization control module is used to perform iterative optimization search on the control parameters of the target switch based on the predicted switch operation data sequence set, with minimizing line loss, fault outage minutes, voltage over-limit time and voltage sag count as multiple optimization objectives, to determine the appropriate control strategy, and to perform intelligent control on the target switch in the preset time zone according to the appropriate control strategy. Specifically, based on the predicted switch operation data sequence set, with minimizing line loss, fault outage minutes, voltage over-limit time, and voltage sag count as multiple optimization objectives, the control parameters of the target switch are iteratively optimized and searched to determine the suitable control strategy, including: Obtain the control parameter adjustment space of the target switch, wherein the control parameters include protection parameters, operation parameters, running parameters and system parameters; Several initial control parameters are randomly generated within the control parameter adjustment space; The predicted switch operation data sequence set is fused with the several initial control parameters to generate several switch operation control schemes; Within the digital operation simulation space of the target switch, control simulations are performed according to the several switch operation control schemes, and several simulated line losses, several simulated fault outage minutes, several simulated voltage over-limit times, and several simulated voltage sags are output. With minimizing line loss, fault outage minutes, voltage over-limit time, and voltage sag count as multiple optimization objectives, several parameter fitnesss are determined by weighted evaluation based on several simulated line losses, several simulated fault outage minutes, several simulated voltage over-limit time, and several simulated voltage sag counts. The parameter fitnesss are negatively correlated with simulated line losses, simulated fault outage minutes, simulated voltage over-limit time, and simulated voltage sag counts. Using the control parameter adjustment space as the search space, and based on the several initial control parameters and several parameter fitness values, the control parameters of the target switch are iteratively optimized and searched to determine the suitable control strategy, including: The initial control parameters are set as the initial solutions, and several initial solutions are arranged in descending order of parameter fitness to generate an initial solution sequence. The first solution of the initial solution sequence is selected as the optimal solution, the last solution is selected as the poor solution, and the remaining initial solutions other than the optimal and poor solutions are selected as inferior solutions, thus obtaining the optimal solution, multiple inferior solutions, and poor solutions; According to the preset optimization direction and the preset optimization step size, the multiple inferior solutions are updated and adjusted, and then reordered to obtain the updated solution sequence; After eliminating a preset number of solutions from the updated solution sequence, control parameters that did not appear during the optimization process are randomly selected within the control parameter adjustment space and supplemented with equivalent values ​​to generate an updated solution sequence. The preset number decreases as the number of optimization attempts increases. Based on the optimization mechanism of selecting optimization direction, adjusting inferior solutions, eliminating solutions, and supplementing solutions, iterative optimization is continued according to the updated solution sequence until the preset number of convergences is reached, and the optimal solution of the current updated solution sequence is output as the adaptation control strategy.

Images