Intelligent management and control system and method for environmental factors of transformer substation based on side cloud cooperation

Abstract

According to the invention, the system carries out the multi-dimensional data collection of the factory boundary and in-station environment of a transformer substation through an end-side sensing layer, and generates an environment factor change trend in combination with an environment prediction model preset by an edge control layer. And gathering multi-station data through a cloud collaboration layer to generate an optimization strategy. An environment prediction model of the edge management and control layer realizes advanced pre-judgment of environment risks, and the problem that passive response measures are taken only after environment indexes exceed the standard is solved; the edge layer generates a regulation and control instruction, and the cloud layer is globally optimized, so that the problem of disjunction of monitoring and management and control is solved, and an efficient closed loop is formed; the end-edge-cloud layered architecture not only ensures the real-time performance of local management and control, but also realizes collaborative optimization of a plurality of substations.

Description

Technical Field

[0001] This invention relates to the field of environmental management technology, and in particular to an intelligent control system and method for environmental factors in substations that integrates edge and cloud computing. Background Technology

[0002] As the hub of power grid transmission and distribution, substations are increasingly widely constructed, with some even extending into environmentally sensitive areas such as city centers and residential areas. During operation, substations continuously generate environmental impacts such as noise, power frequency electric fields, and power frequency magnetic fields. With increasingly stringent national environmental regulations, the effectiveness of controlling these environmental factors directly affects the compliant operation of power grid companies, and is also closely related to the quality of life of surrounding residents and the occupational health and safety of on-site workers. Therefore, accurate monitoring and intelligent management of substation environmental factors have become a crucial requirement in the field of power grid environmental management.

[0003] Currently, the industry's control over environmental factors in substations mainly relies on traditional monitoring and management models: the monitoring stage mostly adopts fixed-point sensor deployment or manual periodic inspection, which can only obtain environmental data from discrete points; in terms of data processing, most existing systems only realize basic data acquisition, storage, and alarm functions for exceeding thresholds, lacking the ability to deeply integrate and analyze massive amounts of real-time and historical data; in the control stage, there is a common problem of disconnect between monitoring and control, failing to form a complete control loop, and only taking passive response measures after environmental indicators exceed the standards. Summary of the Invention

[0004] This invention provides an intelligent management and control system and method for substation environmental factors through edge-cloud collaboration, which solves the problem that existing substation environmental monitoring cannot predict environmental risks in advance.

[0005] On the one hand, the present invention provides an edge-cloud collaborative intelligent management and control system for substation environmental factors, comprising: The edge sensing layer is used to collect multi-dimensional data on the substation boundary and the substation interior environment, obtain environmental data including noise, power frequency electric field and power frequency magnetic field, and send them to the edge control layer. The edge control layer is pre-set with an environmental prediction model; the edge control layer is used to: input the environmental data into the environmental prediction model, generate environmental factor change trend data for a preset time period, and generate early warning signals and / or generate equipment control commands based on the comparison results of the environmental factor change trend data and preset thresholds. The cloud collaboration layer is communicatively connected to at least one of the edge control layers, and is used to collect and analyze early warning signals from multiple edge control layers and / or generate device control commands, and generate optimization strategies based on the analysis results and send them to the edge control layers.

[0006] Optionally, the end-side sensing layer includes: A fixed environmental monitoring module is deployed at preset fixed points at the substation boundary and equipment area within the substation for continuous fixed-point monitoring and acquisition of the environmental data. A mobile environmental monitoring robot is deployed along a closed track laid on the perimeter wall of a substation; the mobile environmental monitoring robot is used to: move along the track according to a preset strategy or receive instructions from the edge control layer, and collect environmental data.

[0007] Optionally, the edge sensing layer further includes a portable intelligent warning module; The portable intelligent alert module is used for: Used to acquire and display the location coordinates of a person and the corresponding location environment parameters; Calculate and display the remaining safe time based on the location and environmental data and the preset time-dose safety limit; When the remaining safety time is lower than different level thresholds, tiered sound, light, and / or vibration alarms are triggered.

[0008] Optionally, the edge control layer includes a data processing module; The data processing module is used to perform timestamp alignment, outlier removal based on the isolated forest algorithm, and Kalman filtering on environmental data from the edge perception layer to generate a fused time-series dataset.

[0009] Optionally, the edge control layer includes a data analysis module; The data analysis module is used for: Perform 1 / 3 octave band spectrum analysis on the noise data in the fused time series dataset, extract the energy distribution characteristics of the preset frequency band, and generate noise source type identification data; The historical sequences, meteorological data, equipment load data, and noise source type identification data in the fused time series dataset are input into the environmental prediction model, and the environmental factor change trend data for a preset time period are output. The fused time-series dataset is compared with the environmental factor change trend data, the residuals are calculated, and abnormal operating condition identification data is generated based on the residuals.

[0010] Optionally, the edge management layer further includes a collaborative control module; The collaborative control module is used for: The environmental factor change trend data are compared with the dynamic early warning threshold, and an early warning signal corresponding to the early warning level is generated and output based on the comparison result. When the abnormal operating condition identification data is received or the warning level reaches the preset level, the system combines the noise source type identification data to generate and output control commands for the devices associated with the noise source type identification data.

[0011] Optionally, the cloud collaboration layer includes a data storage and computing module, a model training module, and a visualization module; The data storage and computing module is used for: Receive and store the fused time-series dataset, environmental factor change trend data, early warning signals and device control logs uploaded from each of the edge control layers, and generate global data; The model training module is used for: The environmental prediction model is trained and optimized based on the global data; The visualization module is used for: The global data is then visualized.

[0012] Optionally, the model training module is further configured to: According to a preset cycle, the fused time-series datasets of each substation and the associated early warning event tags are retrieved from the data storage and computing module. Using the fused time-series dataset and early warning event labels, the environmental prediction model is retrained and optimized through gradient descent or backpropagation algorithms to explore the coupling relationship and change pattern among environmental factors, equipment load, meteorological parameters and noise source types, and output a set of model parameters that can characterize the prediction capability of the optimized environmental factors. The model parameter set is distributed to the corresponding edge intelligent management and control layers through a secure channel to replace or update the corresponding parameters in the environmental prediction model.

[0013] Optionally, after generating the warning signal, the collaborative control module is further used for: A detection command is issued to the mobile environmental monitoring robot, which is then directed to the area where the early warning signal occurs to perform encrypted detection. The data generated by the encrypted detection is then returned to the data processing module to form the fused time-series dataset.

[0014] On the other hand, the present invention also provides an intelligent management and control method for environmental factors in substations using edge-cloud collaboration, comprising: Multi-dimensional data collection was conducted on the substation boundary and internal environment to obtain environmental data including noise, power frequency electric field and power frequency magnetic field. Input the environmental data into the environmental prediction model to generate environmental factor change trend data for a preset time period, and generate early warning signals and / or generate equipment control commands based on the comparison results of the environmental factor change trend data and preset thresholds; Collect and analyze the warning signals and / or generate equipment control commands, and generate optimization strategies based on the analysis results.

[0015] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the edge-cloud collaborative intelligent management and control method for substation environmental factors as described above.

[0016] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the intelligent control method for substation environmental factors in an edge-cloud collaborative manner as described above.

[0017] On the other hand, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the edge-cloud collaborative intelligent management and control method for substation environmental factors as described above.

[0018] This invention discloses an intelligent environmental factor control system and method for substations with edge-cloud collaboration. The system collects multi-dimensional data on the substation boundary and internal environment through an edge-side sensing layer. It then combines this data with a pre-set environmental prediction model in the edge control layer to generate trends in environmental factor changes. Finally, a cloud-based collaboration layer aggregates data from multiple substations to generate optimization strategies. The prediction model in the edge control layer enables proactive prediction of environmental risks, overcoming the passive alarm mode that only triggers alarms after exceeding limits, and allowing sufficient time for control. The linked design of edge layer generating control commands and cloud layer globally optimizing solves the problem of disconnect between monitoring and control, forming an efficient closed loop. The edge-cloud layered architecture ensures both the real-time nature of local control and the collaborative optimization of multiple substations. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the structure of the intelligent control system for substation environmental factors with edge-cloud collaboration provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the end-side sensing layer structure provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the edge control layer structure provided in an embodiment of the present invention; Figure 4This is a schematic diagram of the cloud collaboration layer structure provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] Figure 1 This is a schematic diagram of the structure of the intelligent management and control system for substation environmental factors with edge-cloud collaboration provided in an embodiment of the present invention.

[0023] like Figure 1 As shown in the embodiment of the present invention, the intelligent management and control system for substation environmental factors based on edge-cloud collaboration includes: The edge sensing layer 110 is used to collect multi-dimensional data on the substation boundary and the substation internal environment, obtain environmental data including noise, power frequency electric field and power frequency magnetic field, and send them to the edge control layer 120.

[0024] Among them, the edge sensing layer 110, through the coordinated operation of the fixed environmental monitoring module 1101, the mobile environmental monitoring robot 1102 and the portable intelligent warning module 1103, conducts comprehensive and multi-dimensional data collection on the environmental conditions of the entire substation boundary and the area inside the station, accurately acquires environmental data including noise, power frequency electric field and power frequency magnetic field, and simultaneously collects relevant auxiliary parameters to ensure data integrity. Then, all the collected environmental data is transmitted to the edge control layer 120 in real time and stably.

[0025] Specifically, such as Figure 2 As shown, the end-side sensing layer 110 includes: The fixed environmental monitoring module 1101 is deployed at preset fixed points at the substation boundary and equipment area within the station for continuous fixed-point monitoring and acquisition of environmental data.

[0026] The fixed environmental monitoring module 1101 is deployed at preset fixed points at the substation boundary and within the substation to achieve uninterrupted continuous fixed-point monitoring. The deployment logic of the fixed environmental monitoring module 1101 focuses on sensitive areas and core pollution sources to ensure the benchmark and continuity of data. For example, in a 220kV substation, fixed environmental monitoring modules 1101 are deployed at the midpoint of the east side of the substation boundary, the sensitive side facing the newly built residential area, the northeast corner, and the southeast corner of the substation boundary where the maximum noise is predicted. At the same time, monitoring points are also set at the A / C phases of the #1 and #2 main transformers, 5 meters away from the equipment casing. The data sampling frequency is set to 1Hz to continuously capture real-time changes in noise, power frequency electric field, and power frequency magnetic field.

[0027] A mobile environmental monitoring robot 1102 is deployed along a closed track laid on the substation perimeter wall; the mobile environmental monitoring robot 1102 is used to: move along the track according to a preset strategy or receive instructions from the edge control layer 120, and collect environmental data.

[0028] Among them, the mobile environmental monitoring robot 1102 is deployed along a closed track laid on the substation wall, combining mobility and precision. Driven by a servo motor, the mobile environmental monitoring robot 1102 can move flexibly according to a preset strategy or by receiving adaptive commands from the edge control layer 120. Combined with encoder and inertial navigation fusion positioning technology, it ensures precise and controllable movement trajectory and data collection location. For example, a suburban substation has a 500-meter-long closed aluminum alloy sliding track laid along the top of the inner side of the wall, with charging and communication nodes set every 50 meters. During the day, the robot moves 20 meters along the track every 2 hours and stops for 3 minutes to collect data; at night, the scanning density doubles. When the edge control layer 120 predicts that the noise at the eastern boundary is approaching the limit, the robot will automatically go to the eastern boundary area and cycle through monitoring every 15 minutes, effectively filling the blind spots of fixed-point monitoring and achieving continuous full-area coverage of boundary environmental data.

[0029] The portable smart warning module 1103 is used for: Used to obtain and display the location coordinates of personnel and the corresponding location environment parameters.

[0030] Calculate and display the remaining safe time based on location and environmental data and preset time-dose safety limits.

[0031] When the remaining safe time is less than the threshold for different levels, trigger tiered sound, light, and / or vibration alarms.

[0032] The portable intelligent warning module 1103 is a protective tool for ensuring the occupational health and safety of on-site personnel. It is equipped to inspection or operation personnel in the form of a wearable device or a handheld device. The portable intelligent warning module 1103 has built-in positioning sensors, environmental sensors, and wireless communication capabilities. It links with environmental data in real time, first acquiring the personnel's current location coordinates and corresponding environmental parameters, then calculating the remaining safe time based on preset time-dose safety limits, and finally alerting personnel to risks through tiered alarms. For example, when a worker enters an area with a power frequency electric field strength of 10kV / m, the portable intelligent warning module 1103 presets a 4-hour allowable stay in this scenario and starts a countdown. When less than 30 minutes remain, it triggers a slow beep and a flashing yellow light; when less than 10 minutes remain, it switches to a rapid beep and a flashing red light, providing real-time warnings for timely evacuation to avoid occupational health risks.

[0033] The edge control layer 120 is pre-set with an environmental prediction model; the edge control layer 120 is used to: input environmental data into the environmental prediction model, generate environmental factor change trend data for a preset time period, and generate early warning signals and / or generate equipment control commands based on the comparison results of the environmental factor change trend data and preset thresholds.

[0034] The environmental prediction model is constructed using a long short-term memory network. When constructing the environmental prediction model, the collected environmental data is first preprocessed, including filling in missing values, removing outliers, and unifying the data time granularity. Then, key information in the environmental prediction model is extracted through feature engineering, such as the trend characteristics of historical sequences, the gradient of equipment load rate changes, the spatiotemporal correlation characteristics of meteorological parameters, and the conversion of noise source type identifiers into one-hot encoded feature vectors.

[0035] Subsequently, using the preprocessed fused data as training samples and the monitoring data of environmental factors within the corresponding time period as labels, the training set, validation set, and test set were divided. The number of hidden layers, the number of neurons, and the learning rate of the Long Short-Term Memory (LSTM) network were iteratively adjusted. The LSTM network was optimized by minimizing the mean square error between the predicted and actual values. At the same time, in combination with the requirements of lightweight deployment on the edge, the LSTM network structure was pruned and compressed. Finally, an environmental prediction model that balances prediction accuracy and operational efficiency and can accurately capture the correlation between changes in multi-source influencing factors and environmental factors was constructed.

[0036] After the environmental prediction model is built, environmental data is input into the environmental prediction model. The environmental prediction model utilizes its internal long short-term memory network structure to perform in-depth analysis and processing of the input data. It identifies historical trend characteristics in the environmental data, combining this with the gradient of equipment load rate changes and the spatiotemporal correlation characteristics of meteorological parameters for comprehensive evaluation. Simultaneously, the model employs one-hot encoded feature vectors to accurately identify and classify noise source types, outputting environmental factor change trend data for future time periods. Finally, it compares the environmental factor change trend data with preset thresholds to generate early warning signals and / or equipment control commands.

[0037] Specifically, such as Figure 3 As shown, the edge control layer 120 includes a data processing module 1201, a data analysis module 1202, and a collaborative control module 1203.

[0038] The data processing module 1201 is used to: perform timestamp alignment, outlier removal based on the isolated forest algorithm, and Kalman filtering on environmental data from the edge perception layer 110 to generate a fused time series dataset.

[0039] Specifically, the data processing module 1201 is responsible for converting the multi-source heterogeneous environmental data uploaded by the edge sensing layer 110 into a standardized fused time-series dataset. During the conversion process, timestamp alignment is first performed. Since there are differences in the sampling frequency and transmission delay of different edge devices, the data processing module 1201 uses the clock of the edge sensing layer 110 as a reference to uniformly calibrate all data to the same time granularity, ensuring the spatiotemporal consistency of multi-source data.

[0040] Outliers were then removed using the Isolation Forest algorithm, which effectively identifies isolated outlier data caused by momentary sensor failures, robot movement jitter, or electromagnetic interference, thus preventing outlier data from affecting the analysis results.

[0041] Finally, Kalman filtering is used to smooth and reduce noise in the timestamp-aligned data, thus reducing random errors. For example, when a mobile environmental monitoring robot 1102 of a 220kV substation was collecting data along the track from 14:00 to 14:03, three sets of noise anomalies were generated due to track joint vibration, with values ​​of 85dB, 28dB, and 92dB respectively, far exceeding the normal range of 40-60dB. The data processing module 1201 first aligned the anomaly data with the sampled data to one record per second, then identified and removed the three sets of anomalies using the isolated forest algorithm, and finally smoothed the data using Kalman filtering to generate a continuous and stable fused time-series dataset.

[0042] The data analysis module 1202 is used to: perform 1 / 3 octave band spectrum analysis on the noise data in the fused time series dataset, extract the energy distribution characteristics of the preset frequency band, and generate noise source type identification data.

[0043] In the process of performing 1 / 3 octave band spectrum analysis on the noise data in the fused time series dataset, it is necessary to extract the energy distribution characteristics of different frequency bands. If the energy proportion of low frequency bands such as 100Hz and 200Hz is significant, a transformer excitation noise label is generated; if the energy of mid-to-high frequency bands such as 500Hz-2000Hz is dominant, a cooling fan noise label is generated to accurately locate the source of noise.

[0044] For example, after noise data was collected near the #2 main transformer of a substation, spectrum analysis revealed prominent energy peaks in the 630Hz and 1250Hz frequency bands. The data analysis module 1202 then generated noise identification data for the cooling fan.

[0045] The historical sequences, meteorological data, equipment load data, and noise source type identification data from the integrated time series dataset are input into the environmental prediction model, which outputs environmental factor change trend data for a preset time period.

[0046] This involves integrating historical sequences, synchronous meteorological data, equipment load data, and noise source type identification data from a time-series dataset, all of which are then input into a pre-defined environmental prediction model. Based on learned multi-source data correlation patterns, the model outputs noise change trend data for a predetermined future time. For example, it predicts that the noise level will reach 53.5 dB at 15:30.

[0047] The fused time-series dataset is compared with the environmental factor change trend data to calculate the residuals, and abnormal operating condition identification data is generated based on the residuals.

[0048] The process involves comparing environmental data from the fused time-series dataset with trend data output by the prediction model point by point, calculating the residual between the two. The residual is the difference between the actual and predicted values. When the residual exceeds a preset threshold for three consecutive time points, an abnormal operating condition indicator is generated, suggesting potential issues such as abnormal equipment startup or sudden weather changes. For example, if the main transformer suddenly increases its load, causing the actual noise level to reach 58 dB while the predicted value is 52 dB, the residual difference of 6 dB exceeds the threshold, and the data analysis module 1202 immediately generates an abnormal operating condition indicator.

[0049] The collaborative control module 1203 is used to compare environmental factor change trend data with dynamic early warning thresholds, so as to generate and output early warning signals of corresponding early warning levels based on the comparison results.

[0050] The collaborative control module 1203 has a built-in dynamic early warning threshold. For example, the daytime limit for noise at the plant boundary is 55 dB, and the early warning threshold is set to 52 dB; the nighttime limit is 45 dB, and the early warning threshold is set to 42 dB.

[0051] The environmental factor change trend data output by the data analysis module 1202 is compared with the dynamic early warning threshold to generate a four-level early warning signal. For example, a green normal warning is output when the trend is in the normal range, a yellow attention warning is output when the trend is close to the warning threshold, an orange warning is output when the trend exceeds the warning threshold but does not reach the limit, and a red exceedance warning is output when the trend exceeds the national standard limit. For example, if it is predicted that the noise at the eastern plant boundary will reach 53.5dB at 15:30, exceeding the daytime warning threshold of 52dB but not reaching the limit of 55dB, the collaborative control module 1203 will immediately generate an orange warning signal.

[0052] When abnormal operating condition identification data is received or the warning level reaches the preset level, the system combines the noise source type identification data to generate and output control commands for the devices associated with the noise source type identification data.

[0053] When abnormal operating condition identification data is received, or the warning level reaches orange or above, the collaborative control module 1203 combines the noise source type identification data output by the data analysis module 1202 to generate targeted control instructions. For example, if the noise source identification corresponding to the orange warning is cooling fan noise, the collaborative control module 1203 will then issue an instruction to the fan control system of the #2 main transformer to switch the two high-power fans to three low-power fans while keeping the total airflow unchanged, thereby reducing the contribution of mid-to-high frequency noise. If the noise source type identification data is transformer excitation noise, an instruction suggesting adjustment of the main transformer's operating mode will be generated and uploaded to the cloud collaborative layer 130 to provide a basis for manual decision-making, achieving precise control of different noise sources.

[0054] The cloud collaboration layer 130 is communicatively connected to at least one edge control layer 120, and is used to collect and analyze early warning signals from multiple edge control layers 120 and / or generate device control commands, and generate optimization strategies based on the analysis results and send them to the edge control layer 120.

[0055] The cloud-based collaboration layer 130 can be deployed in a remote data center or cloud server, and can receive early warning signals, equipment status data, and noise source type identification data uploaded from various edge control layers 120 in real time. Through data analysis algorithms and models, the data is deeply mined and analyzed to generate scientific and reasonable optimization strategies, such as adjusting the overall operation mode of the substation, optimizing equipment layout, and formulating more precise equipment control instructions. The optimization strategies are then promptly distributed to the corresponding edge control layers 120 to ensure the efficient operation of the entire substation environmental factor intelligent control system.

[0056] Specifically, such as Figure 4 As shown, the cloud collaboration layer 130 includes a data storage and computing module 1301, a model training module 1302, and a visualization module 1303; Data storage and computing module 1301 is used for: It receives and stores fused time-series datasets, environmental factor change trend data, early warning signals and equipment control logs uploaded from each edge control layer 120, and generates global data.

[0057] Among them, the data storage technology module is responsible for collecting various types of data uploaded by the edge control layer 120 of all substations, and provides efficient storage and multi-dimensional computing capabilities.

[0058] At the storage level, the data storage and computing module 1301 adopts a hybrid storage architecture of time-series database + relational database. The time-series database stores the fused time-series datasets and environmental factor change trend data uploaded from each edge layer, ensuring efficient writing and fast retrieval of time-series data; the relational database stores early warning signals and equipment control logs, realizing standardized management of non-time-series data.

[0059] At the computational level, noise data from the same type of equipment in different substations under the same meteorological conditions and different load rates are aggregated and statistically analyzed. Alternatively, the distribution patterns of daytime and nighttime early warning events and the average response efficiency of control measures across the entire substation network are calculated to provide data support for training environmental prediction models. For example, a power grid company's cloud platform accesses edge layer data from 50 220kV substations. The data storage and computation module 1301 aggregates fan noise data from all stations under summer high temperatures ≥35℃ and main transformer load rates ≥80%, generating a global dataset of fan noise under high-temperature and high-load conditions, providing targeted samples for optimizing environmental prediction models.

[0060] Model training module 1302 is used for: The environmental prediction model is trained and optimized based on global data.

[0061] The specific steps of training and optimizing the environmental prediction model in the model training module 1302 include: According to the preset cycle, the fused time series datasets of each substation and the associated early warning event tags are retrieved from the data storage and computing module 1301; By using a fusion of time-series datasets and early warning event labels, the environmental prediction model is retrained and optimized through gradient descent or backpropagation algorithms. The coupling relationship and variation law between environmental factors, equipment load, meteorological parameters and noise source types are explored, and a set of model parameters that can characterize the prediction ability of the optimized environmental factors is output. The model parameter set is distributed to the corresponding edge intelligent control layers through a secure channel to replace or update the corresponding parameters in the environmental prediction model.

[0062] The model training module 1302 conducts the training and optimization of the environmental prediction model in an orderly manner according to a preset cycle. For example, the preset cycle is once every seven days. During training and optimization, it first retrieves the fused time-series dataset of all substations and the warning event labels associated with the fused time-series dataset from the data storage and computing module 1301. The warning time label includes the warning trigger time, level, and corresponding operating conditions.

[0063] Subsequently, using a fused time-series dataset covering different regions, operating conditions, and noise source types as training samples, gradient descent was employed to retrain and optimize the environmental prediction model. During training and parameter optimization, the complex coupling relationships and dynamic changes between environmental factors and equipment load, meteorological parameters, and noise source types were explored. For example, the correlation was captured showing that when the ambient temperature is ≥35℃ and the main transformer load rate is ≥80%, the mid-to-high frequency noise of a certain type of cooling fan increases by 5-8dB compared to normal operating conditions. The final output was a model parameter set that improved the accuracy of environmental factor prediction.

[0064] Finally, the model parameter set is distributed to the corresponding edge control layers 120 via a secure channel. The edge control layers 120 automatically replace or update the corresponding parameters in the original environmental prediction model, realizing the iterative upgrade of the environmental prediction model. For example, the model training module 1302 of a power grid company adjusts the fused time-series dataset of 50 substations every week, optimizes the weight coefficients and bias terms of the environmental prediction model through the gradient descent algorithm, and discovers that transformer excitation noise is positively correlated with the main transformer load rate, and that for every 2 m / s increase in wind speed, the noise propagation attenuation increases by 3 dB. The optimized model parameter set is output and distributed through an encrypted channel, improving the prediction accuracy of the environmental prediction model of each edge control layer 120 for transformer excitation noise.

[0065] Visualization module 1303 is used for: Visualize and display global data.

[0066] Among them, the visualization module 1303 presents global data intuitively through multi-dimensional visualization, providing a visualization interface for operation and maintenance managers and environmental decision-makers.

[0067] Specifically, the visualization module 1303 includes a global overview, single-site details, and in-depth analysis.

[0068] At the overall overview level, the real-time status of all substations in the network is visualized through a GIS map, with different colored icons indicating the warning level of each station, making it easy for users to quickly locate risky stations.

[0069] At the single-station details level, when a user clicks on a substation icon, the page will display a real-time data list, dynamic cloud map, right trend curve, and management log in different areas.

[0070] At the in-depth analysis level, the system automatically generates multi-dimensional statistical reports and supports report export. For example, maintenance and management personnel can log in to the platform via the web and find three substations marked in orange on the GIS map. After clicking on a 220kV substation, they can intuitively see real-time noise data on the left, a red high-noise cloud map of the eastern boundary in the middle, and a forecast curve for the next hour and fan control logs on the right. At the same time, the platform automatically pushes a summary report of the station's early warning and handling for the past month, providing intuitive support for cross-station comparative analysis and equipment technical upgrade decisions.

[0071] In some embodiments, after the cooperative control module 1203 generates a warning signal, it is further used to: A detection command is issued to the mobile environmental monitoring robot 1102, which is then directed to the area where the warning signal occurs to perform encrypted detection. The data generated by the encrypted detection is then returned to the data processing module 1201 to form a fused time series dataset.

[0072] Specifically, when the collaborative control module 1203 determines that the level of the warning signal has reached the preset threshold, it immediately initiates the communication protocol with the mobile environmental monitoring robot 1102 and sends an instruction packet containing the target area coordinates, detection task type and encrypted detection parameters via wireless transmission.

[0073] After receiving instructions, the mobile environmental monitoring robot 1102 automatically plans the optimal path and starts a multi-sensor array to perform high-frequency, multi-dimensional sampling of environmental factors in the target area. The sampling frequency is 3-5 times higher than that of conventional inspection mode.

[0074] The encrypted detection data is transmitted back to the data processing module 1201 in real time via an encrypted link. The data processing module 1201 first performs check code verification and outlier removal on the original data. Then, it performs spatiotemporal alignment processing on the valid data and the regular monitoring data uploaded by the edge control layer 120. Finally, it merges the data to form a time series dataset containing high-precision spatiotemporal labels, providing richer feature dimensions for subsequent model training.

[0075] Based on the same general inventive concept, this invention also protects an intelligent management and control method for substation environmental factors in an edge-cloud collaborative manner. The intelligent management and control method for substation environmental factors in an edge-cloud collaborative manner provided by this invention will be described below. The intelligent management and control method for substation environmental factors in an edge-cloud collaborative manner described below can be referred to in correspondence with the intelligent management and control system for substation environmental factors in an edge-cloud collaborative manner described above.

[0076] In some embodiments, the present invention also provides an intelligent management and control method for environmental factors in substations using edge-cloud collaboration, comprising: Multi-dimensional data collection was conducted on the substation boundary and internal environment to obtain environmental data including noise, power frequency electric field and power frequency magnetic field. Input environmental data into the environmental prediction model to generate environmental factor change trend data for a preset time period, and generate early warning signals and / or generate equipment control commands based on the comparison results of environmental factor change trend data and preset thresholds. Collect and analyze early warning signals and / or generate equipment control commands, and generate optimization strategies based on the analysis results.

[0077] Figure 5 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention.

[0078] like Figure 5 As shown, the electronic device may include a processor 510, a communication interface 520, a memory 530, and a communication bus 540. The processor 510, communication interface 520, and memory 530 communicate with each other via the communication bus 540. The processor 510 can call logical instructions from the memory 530 to execute an edge-cloud collaborative intelligent control method for substation environmental factors.

[0079] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0080] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the edge-cloud collaborative intelligent management and control method for substation environmental factors provided by the above methods.

[0081] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the edge-cloud collaborative intelligent management and control method for substation environmental factors provided by the above methods.

[0082] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0083] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart control system for environmental factors in substations with edge-cloud collaboration, characterized in that, include: The edge sensing layer is used to collect multi-dimensional data on the substation boundary and the substation interior environment, obtaining environmental data including noise, power frequency electric field and power frequency magnetic field, and sending it to the edge control layer. The edge control layer is pre-set with an environmental prediction model; the edge control layer is used to: input the environmental data into the environmental prediction model, generate environmental factor change trend data for a preset time period, and generate early warning signals and / or generate equipment control commands based on the comparison results of the environmental factor change trend data and preset thresholds. The cloud collaboration layer is communicatively connected to at least one of the edge control layers, and is used to collect and analyze early warning signals from multiple edge control layers and / or generate device control commands, and generate optimization strategies based on the analysis results and send them to the edge control layers.

2. The intelligent control system for substation environmental factors based on edge-cloud collaboration as described in claim 1, characterized in that, The end-side sensing layer includes: A fixed environmental monitoring module is deployed at preset fixed points at the substation boundary and equipment area within the substation for continuous fixed-point monitoring and acquisition of the environmental data. A mobile environmental monitoring robot is deployed along a closed track laid on the perimeter wall of a substation; the mobile environmental monitoring robot is used to: move along the track according to a preset strategy or receive instructions from the edge control layer, and collect environmental data.

3. The intelligent control system for substation environmental factors based on edge-cloud collaboration as described in claim 1, characterized in that, The edge-side sensing layer also includes a portable intelligent warning module; The portable intelligent alert module is used for: Used to acquire and display the location coordinates of a person and the corresponding location environment parameters; Calculate and display the remaining safe time based on the location and environmental data and the preset time-dose safety limit; When the remaining safety time is lower than different level thresholds, tiered sound, light, and / or vibration alarms are triggered.

4. The intelligent control system for substation environmental factors based on edge-cloud collaboration according to claim 2, characterized in that, The edge control layer includes a data processing module; The data processing module is used to perform timestamp alignment, outlier removal based on the isolated forest algorithm, and Kalman filtering on the environmental data from the edge perception layer to generate a fused time-series dataset.

5. The intelligent control system for substation environmental factors based on edge-cloud collaboration according to claim 4, characterized in that, The edge control layer includes a data analysis module; The data analysis module is used for: Perform 1 / 3 octave band spectrum analysis on the noise data in the fused time series dataset, extract the energy distribution characteristics of the preset frequency band, and generate noise source type identification data; The historical sequences, meteorological data, equipment load data, and noise source type identification data in the fused time series dataset are input into the environmental prediction model, and the environmental factor change trend data for a preset time period are output. The fused time-series dataset is compared with the environmental factor change trend data, the residuals are calculated, and abnormal operating condition identification data is generated based on the residuals.

6. The intelligent control system for substation environmental factors based on edge-cloud collaboration according to claim 5, characterized in that, The edge control layer also includes a collaborative control module; The collaborative control module is used for: The environmental factor change trend data are compared with the dynamic early warning threshold, so as to generate and output the early warning signal corresponding to the early warning level based on the comparison result; When the abnormal operating condition identification data is received or the warning level reaches the preset level, the control command for the device associated with the noise source type identification data is generated and output in combination with the noise source type identification data.

7. The intelligent control system for substation environmental factors based on edge-cloud collaboration according to claim 6, characterized in that, The cloud collaboration layer includes a data storage and computing module, a model training module, and a visualization module; The data storage and computing module is used for: Receive and store the fused time-series dataset, environmental factor change trend data, early warning signals and device control logs uploaded from each of the edge control layers, and generate global data; The model training module is used for: The environmental prediction model is trained and optimized based on the global data; The visualization module is used for: The global data is then visualized.

8. The intelligent control system for substation environmental factors based on edge-cloud collaboration according to claim 7, characterized in that, The model training module is also used for: According to a preset cycle, the fused time-series datasets of each substation and the associated early warning event tags are retrieved from the data storage and computing module. Using the fused time-series dataset and early warning event labels, the environmental prediction model is retrained and optimized through gradient descent or backpropagation algorithms to explore the coupling relationship and change pattern among environmental factors, equipment load, meteorological parameters and noise source types, and output a set of model parameters that can characterize the prediction capability of the optimized environmental factors. The model parameter set is distributed to the corresponding edge intelligent management and control layers through a secure channel to replace or update the corresponding parameters in the environmental prediction model.

9. The intelligent control system for substation environmental factors based on edge-cloud collaboration according to claim 6, characterized in that, After generating the warning signal, the collaborative control module is also used for: A detection command is issued to the mobile environmental monitoring robot, which is then directed to the area where the early warning signal occurs to perform encrypted detection. The data generated by the encrypted detection is then returned to the data processing module to form the fused time-series dataset.

10. A method for intelligent management and control of environmental factors in substations through edge-cloud collaboration, characterized in that, include: Multi-dimensional data collection was conducted on the substation boundary and internal environment to obtain environmental data including noise, power frequency electric field and power frequency magnetic field. Input the environmental data into the environmental prediction model to generate environmental factor change trend data for a preset time period, and generate early warning signals and / or generate equipment control commands based on the comparison results of the environmental factor change trend data and preset thresholds; Collect and analyze the warning signals and / or generate equipment control commands, and generate optimization strategies based on the analysis results.

Images