Power utilization evaluation-based grid supply load regulation and control method and system

Abstract

The invention provides a power utilization evaluation-based network supply load regulation and control method and system, and the method is applied to a regulation and control system comprising a load analysis module, a power utilization evaluation module, a model construction module and a load regulation and control module, and specifically comprises the steps: calling historical load data of a target regulation and control region, constructing a typical daily power consumption curve and an atypical daily power consumption curve in combination with the corresponding acquisition time, and extracting daily load characteristic parameters; establishing a power utilization evaluation index system according to the daily load characteristic parameters, and combining the daily load characteristic parameters to construct a load management and control model comprising a load prediction sub-model and a regulation and control index decomposition sub-model; inputting the real-time network supply load data into the load control model to obtain a load control strategy; and setting a regulation and control time period and a corresponding regulation and control quantity of each network supply load in the target regulation and control area according to the load regulation and control strategy. According to the method, the power utilization evaluation index system is integrated on the basis of load prediction, subjective verification judgment is replaced by quantitative parameters, and the regulation and control accuracy of the network supply load is improved.

Description

Technical Field

[0001] This invention relates to the field of load regulation technology, and in particular to a grid-based load regulation method and system based on electricity consumption assessment. Background Technology

[0002] With the continuous expansion of installed capacity of new energy sources such as wind power and photovoltaics, and the rapid growth of new loads such as flexible industrial and commercial loads and residential distributed photovoltaics, the fluctuation characteristics of the power grid's source and load sides are becoming increasingly significant. Traditional grid-supply load regulation mainly relies on passive response and experience-driven methods, setting fixed load thresholds based on the grid's carrying capacity. When the real-time load exceeds the threshold, a blanket power curtailment or peak-shaving order is triggered. Although this approach has low deployment costs, it lacks flexibility and adaptability. Alternatively, it relies on the historical experience of dispatchers, combined with macroeconomic factors such as seasons and weather, to formulate control plans. However, this lack of quantitative data support makes it prone to errors and unable to cope with complex and ever-changing electricity demand. Under the influence of these two-sided fluctuation characteristics, the traditional grid-supply load regulation model is no longer suitable for the current power system's requirements for safe and stable operation.

[0003] To improve adaptability to the two-end fluctuation characteristics of the power grid and optimize the accuracy and efficiency of load regulation, most current methods adopt time-series load forecasting models combined with basic load classification regulation to achieve grid-supply load regulation. The time-series load forecasting model, which takes into account multiple influencing factors such as meteorological data, load type, and seasonal characteristics, is the core, replacing simple human experience judgment to achieve quantitative load forecasting. At the same time, basic classification is completed based on the importance of the load. When the predicted load exceeds the threshold, non-critical flexible loads are regulated first to ensure the normal operation of critical loads. This breaks away from the dependence of traditional grid-supply load regulation on fixed thresholds and pure experience, reduces the probability of human error, and improves the flexibility and adaptability of regulation.

[0004] However, in this grid-supply load control method, load classification and control potential judgment are subjective, relying only on vague dimensions such as importance for basic classification, lacking quantitative parameters such as peak-valley difference rate and load elasticity coefficient. This can easily lead to the wrong inclusion of key rigid loads in the control or the omission of highly elastic loads. When facing grid-supply load control in complex scenarios with obvious fluctuation characteristics, the accuracy of load control remains low. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies that use time-series load forecasting models in conjunction with basic load classification and control for grid-controlled load regulation. These shortcomings include subjective judgment of load classification and control potential, lack of corresponding quantitative parameters, inability to adapt to complex scenarios with significant fluctuations in both source and load characteristics, and low accuracy of load regulation. This invention provides a grid-based load regulation method and system based on electricity consumption assessment. By extracting daily load characteristic parameters and constructing an electricity consumption assessment index system, quantitative parameters are used to replace subjective experience judgments. This determines the quantitative definition standards for load classification and the quantitative assessment basis for control potential. In this way, combined with load forecasting, a rapid response to fluctuations in both source and load is achieved, improving the adaptability of load regulation in complex scenarios and ensuring the accuracy of grid-based load regulation.

[0006] The objective of this invention is achieved through the following technical solution: Grid-based load control methods based on electricity consumption assessment include: Historical load data of the target control area is retrieved, and typical and atypical daily electricity consumption curves are constructed by combining the corresponding collection time. Daily load characteristic parameters are then extracted. A power consumption assessment index system is established based on daily load characteristic parameters, and a load control model including a load forecasting sub-model and a control index decomposition sub-model is constructed in combination with daily load characteristic parameters. Input the real-time grid supply load data of the target control area into the load control model to obtain the load control strategy; Based on the load control strategy, set the control period and corresponding control amount for each grid supply load within the target control area.

[0007] Furthermore, the process of retrieving historical load data for the target control area, constructing typical and atypical daily electricity consumption curves based on the corresponding collection time, and extracting daily load characteristic parameters includes: Retrieve historical load data for the target control area and synchronously associate it with the corresponding collection time; Historical load data are classified according to collection time and scenario attributes, and typical daily electricity consumption curves and atypical daily electricity consumption curves are constructed based on the classified historical load data. Feature extraction is performed on typical and atypical daily electricity consumption curves to obtain characteristic parameters of typical and atypical daily loads; Correlation analysis is performed on typical daily load characteristic parameters and atypical daily load characteristic parameters to identify corresponding parameter correlation terms, and targeted corrections are made based on the parameter correlation terms; A set of daily load characteristic parameters is established based on the revised typical daily load characteristic parameters and atypical daily load characteristic parameters.

[0008] Furthermore, the establishment of an electricity consumption assessment index system based on daily load characteristic parameters includes: The evaluation index dimensions are set according to the parameter type of the daily load characteristic parameters, and the mapping relationship between the daily load characteristic parameters and the evaluation index is established. Based on the established mapping relationship, the quantitative formulas and corresponding scenario thresholds for each evaluation indicator are set, and the weights of each evaluation indicator are allocated using the analytic hierarchy process.

[0009] Furthermore, the establishment of an electricity consumption assessment index system based on daily load characteristic parameters, and the construction of a load control model including a load forecasting sub-model and a control index decomposition sub-model based on the daily load characteristic parameters, includes: The daily load characteristic parameters are correlated with the electricity consumption assessment index system to form a training sample set; Prediction constraints are set based on the scenario thresholds of each evaluation indicator. The load prediction sub-model is constructed using the daily load characteristic parameters as input and a long short-term memory network. Based on the scenario thresholds of each evaluation indicator, control constraints are set. The output of the load forecasting sub-model, daily load characteristic parameters, and each evaluation indicator are used as inputs. The control indicator decomposition sub-model is constructed by combining the indicator weights of each evaluation indicator through a multi-objective optimization algorithm. The load control model is obtained by training the load forecasting sub-model and the regulation index decomposition sub-model using a training sample set.

[0010] Furthermore, the step of inputting real-time grid supply load data of the target control area into the load management model to obtain load control strategies includes: Collect real-time grid load data for the target control area, extract corresponding real-time daily load characteristic parameters, and combine the associated auxiliary data from the collection time to determine the current electricity consumption scenario; Input the real-time daily load characteristic parameters, associated auxiliary data and the current electricity consumption scenario into the load prediction sub-model in the load management model to obtain the load prediction results; The load forecast results and real-time daily load characteristic parameters are input into the control index decomposition sub-model of the load management model. The control index decomposition sub-model associates the real-time daily load characteristic parameters with the corresponding scenario constraints, solves and outputs the optimal load control strategy that satisfies all scenario constraints of the current electricity consumption scenario.

[0011] Furthermore, the load control strategy includes time-based total control amount, user-type-based control instructions, and control execution priority.

[0012] Furthermore, the step of setting the control period and corresponding control amount for each grid supply load within the target control area according to the load control strategy includes: The control period range is divided according to the total control volume of each time period, and the control period to which each grid supply load belongs is determined by combining real-time daily load characteristic parameters and control execution priority. The control ratio of each user type is obtained based on the control instructions for different user types, and the basic control amount of each network supply load is obtained by combining the real-time daily load characteristic parameters. Based on the adjustable capacity of each grid supply load, and combined with the total control quantity of the corresponding control period, the basic control quantity of each grid supply load is adjusted to obtain the control quantity of each grid supply load.

[0013] Furthermore, after setting the control period and corresponding control amount for each grid supply load within the target control area according to the load control strategy, the following is also executed: The power supply loads of each grid are scheduled according to the corresponding control period and control amount, and real-time control data of each grid power supply loads are collected in real time. The control execution deviation rate is calculated based on the corresponding control quantity and real-time control data. The scenario thresholds of the corresponding assessment indicators in the electricity consumption assessment indicator system and the corresponding model parameters of the load control model are adjusted according to the control execution deviation rate.

[0014] A grid-connected load control system based on electricity consumption assessment, used to perform any of the above-mentioned tasks, includes: The load analysis module is used to retrieve historical load data of the target control area, construct typical daily electricity consumption curves and atypical daily electricity consumption curves by combining the corresponding collection time, and extract daily load characteristic parameters; The electricity consumption assessment module is connected to the load analysis module and is used to establish an electricity consumption assessment index system based on daily load characteristic parameters. The model building module, connected to the electricity consumption assessment module, is used to build a load control model, including a load forecasting sub-model and a control index decomposition sub-model, based on the electricity consumption assessment index system and daily load characteristic parameters. The load control module, connected to the model building module, is used to input real-time grid supply load data of the target control area into the load management model, obtain the load control strategy, and set the control period and corresponding control amount for each grid supply load in the target control area according to the load control strategy.

[0015] Furthermore, it also includes: The feedback adjustment module is connected to the electricity consumption assessment module, the model building module, and the load control module, respectively. It is used to collect real-time control data of the load supplied by each grid in real time, calculate the control execution deviation rate in combination with the load control strategy, and adjust the scenario threshold of the corresponding assessment index in the electricity consumption assessment index system and the corresponding model parameters of the load control model according to the control execution deviation rate.

[0016] The beneficial effects of this invention are: (1) Based on the extraction of daily load characteristic parameters and the construction of electricity consumption assessment index system, load classification and control potential assessment are standardized and objective by replacing subjective experience judgment with quantitative parameters. Furthermore, combined with load forecasting, sudden events caused by source and load fluctuations can be quickly identified, and corresponding index thresholds can be matched to formulate corresponding load control strategies, so as to achieve rapid response to fluctuations and improve adaptability to complex scenarios.

[0017] (2) Based on the synergistic linkage between the load forecasting sub-model and the control index decomposition sub-model, the load change trend is first accurately predicted by real-time grid supply load, and the corresponding threshold is quickly matched with the corresponding scenario to decompose the control target, and then the load control plan is executed to maximize the release of grid supply capacity, quickly adapt to complex scenarios, and ensure the accuracy of load control.

[0018] (3) By setting up a dynamic feedback loop, the evaluation index threshold and model parameters are adjusted in reverse to regulate the execution deviation rate, so as to ensure that the load regulation accuracy can be continuously improved with the iteration of operation data and to ensure the accuracy of grid supply load regulation. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a process of the present invention; Figure 2 This is a schematic diagram of a structure according to an embodiment of the present invention.

[0020] The modules include: 1. Load analysis module, 2. Electricity consumption assessment module, 3. Model building module, 4. Load control module, and 5. Feedback adjustment module. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] Example: Grid-based load control methods based on electricity consumption assessment, such as Figure 1 As shown, it includes: Historical load data of the target control area is retrieved, and typical and atypical daily electricity consumption curves are constructed by combining the corresponding collection time. Daily load characteristic parameters are then extracted. A power consumption assessment index system is established based on daily load characteristic parameters, and a load control model including a load forecasting sub-model and a control index decomposition sub-model is constructed in combination with daily load characteristic parameters. Input the real-time grid supply load data of the target control area into the load control model to obtain the load control strategy; Based on the load control strategy, set the control period and corresponding control amount for each grid supply load within the target control area.

[0023] Historical load data can reflect load operation patterns, and by associating the data with the collection time, the data can acquire spatiotemporal attributes, such as load differences in different seasons and peak / valley periods. This can effectively avoid the limitation that isolated data cannot reflect the changing patterns, thus providing relevant data support for the construction of subsequent electricity consumption assessment index systems and the training of load management models.

[0024] However, with the increasing penetration of new energy sources and the frequent occurrence of extreme weather, load fluctuations in atypical scenarios, such as extreme weather, major events, and power grid failures, are becoming more and more significant. If only a single electricity consumption curve is constructed from historical load data to extract the corresponding load characteristic parameters, it is difficult to reflect the load patterns under such sudden scenarios. This will affect the accuracy of the subsequent electricity consumption assessment index system and load control model construction, and will not be able to adapt to the grid supply load control needs of complex scenarios with fluctuations at both the source and load ends.

[0025] Therefore, before carrying out corresponding grid load regulation, historical load data of the target regulation area should be retrieved first. Typical and atypical daily electricity consumption curves should be constructed by combining the corresponding data collection time, and daily load characteristic parameters should be extracted, including: Retrieve historical load data for the target control area and synchronously associate it with the corresponding collection time; Historical load data are classified according to collection time and scenario attributes, and typical daily electricity consumption curves and atypical daily electricity consumption curves are constructed based on the classified historical load data. Feature extraction is performed on typical and atypical daily electricity consumption curves to obtain characteristic parameters of typical and atypical daily loads; Correlation analysis is performed on typical daily load characteristic parameters and atypical daily load characteristic parameters to identify corresponding parameter correlation terms, and targeted corrections are made based on the parameter correlation terms; A set of daily load characteristic parameters is established based on the revised typical daily load characteristic parameters and atypical daily load characteristic parameters.

[0026] First, determine the target control area for grid-connected load regulation. Then, retrieve the corresponding historical load data through the power grid's existing data acquisition systems, such as SCADA systems. Simultaneously, associate the corresponding data timestamps with the acquisition time to determine the spatiotemporal attributes of the historical load data, such as date type, season affiliation, and time period type. By correlating the acquisition time with the historical load data, isolated load values ​​are transformed into structured data labeled with corresponding spatiotemporal tags.

[0027] Based on this, and combining the data collection time attribute with the preset scenario attributes, historical load data for typical days and atypical days are selected by using either K-means clustering algorithm based on the similarity of load change patterns or decision tree classification model based on scenario feature rules. Typical days represent normal, stable scenarios during power grid operation, while atypical days represent sudden fluctuations such as extreme weather, major events, or power grid failures.

[0028] After selecting historical load data for typical and atypical days, further preprocessing of each historical load data is required, including filling in missing data and handling outliers. Then, based on the corresponding historical load data, electricity consumption curves for the two scenarios are constructed through time series fitting.

[0029] Feature extraction is performed based on the constructed typical and atypical daily electricity consumption curves. The extracted daily load features include basic statistical features such as daily load peak value, daily load valley value, and peak-valley difference; fluctuation features such as load fluctuation coefficient, peak duration, and fluctuation frequency; and controllability potential features such as load elasticity coefficient, controllable load range, and continuously adjustable duration.

[0030] Typical and atypical daily load characteristic parameters are not isolated; they are logically related. For example, the peak value of an atypical day is an extension of the peak value of a typical day, and the fluctuation coefficient has a certain proportional relationship. During historical data collection and curve construction, anomalies in parameters from a single scenario or logical contradictions across scenarios may occur, which are difficult to detect using only single-scenario data cleaning. Therefore, by utilizing the logical relationship between typical and atypical daily load characteristic parameters, these problems can be identified and located. Then, distorted data can be corrected through targeted adjustments, ensuring the accuracy of the final obtained daily load characteristic parameters.

[0031] Specifically, correlation analysis algorithms and causal inference models can be used to identify correlations between typical and atypical daily load characteristic parameters. The logical consistency between these correlations can then be verified, and logically contradictory abnormal parameters can be identified. Finally, by considering actual constraints such as grid equipment rating standards and minimum power supply requirements for users, the corresponding parameters can be adjusted to meet the relevant correlation rules and the actual operating patterns of the power grid.

[0032] Finally, the corrected typical and atypical daily load characteristic parameters are integrated according to the structure of scenario type, parameter type, and parameter value to establish a daily load characteristic parameter set.

[0033] After obtaining the daily load characteristic parameter set for the entire scenario of the power grid supply load operation, a corresponding electricity consumption assessment index system can be established based on the included daily load characteristic parameters.

[0034] Specifically, the establishment of an electricity consumption assessment index system based on daily load characteristic parameters includes: The evaluation index dimensions are set according to the parameter type of the daily load characteristic parameters, and the mapping relationship between the daily load characteristic parameters and the evaluation index is established. Based on the established mapping relationship, the quantitative formulas and corresponding scenario thresholds for each evaluation indicator are set, and the weights of each evaluation indicator are allocated using the analytic hierarchy process.

[0035] First, based on the parameter types of daily load characteristics, corresponding evaluation index dimensions are set: safety, efficiency, fairness, and potential. On this basis, a mapping relationship between parameters and indicators is established, clarifying the data support sources for each evaluation index. For example, daily load characteristics such as peak load and peak-valley difference should be assigned a grid carrying capacity safety index within the safety dimension. Subsequently, the index values ​​of the grid carrying capacity safety index can be calculated based on the daily load peak and peak-valley difference.

[0036] Furthermore, for each evaluation indicator, a corresponding quantitative formula is set for all daily load characteristic parameters. The specific quantitative formula can be set according to actual needs or empirical formulas. For example, for the grid carrying capacity safety factor of the safety dimension, it can be calculated by the ratio of daily load peak to grid rated carrying capacity.

[0037] However, the operating patterns of loads, the safety boundaries of the power grid, and the control objectives vary significantly across different scenarios. Using a uniform threshold can easily lead to control deviations, resulting in problems such as missed risk assessments or over-control. Therefore, based on a quantitative formula, scenario thresholds are set differently, ensuring that each evaluation indicator has a corresponding standard in different scenarios. This guarantees that the evaluation results accurately reflect the control status of the current scenario, and the scenario thresholds can be set according to actual needs.

[0038] In the process of grid load regulation, there are multiple regulation objectives with different priorities. Therefore, to ensure the accuracy of subsequent load regulation, the analytic hierarchy process (AHP) is used to allocate the weights of each evaluation indicator, so that high-priority indicators have a greater weight in the evaluation and their constraints can be met first. Similarly, when allocating indicator weights using the AHP, corresponding settings are still made for different scenarios on typical and atypical days to ensure that the weight allocation conforms to the actual operation of the power grid.

[0039] After establishing an electricity consumption assessment index system based on daily load characteristic parameters, a load control model can be constructed by combining the daily load characteristic parameters, including a load forecasting sub-model and a control index decomposition sub-model, specifically: The daily load characteristic parameters are correlated with the electricity consumption assessment index system to form a training sample set; Prediction constraints are set based on the scenario thresholds of each evaluation indicator. The load prediction sub-model is constructed using the daily load characteristic parameters as input and a long short-term memory network. Based on the scenario thresholds of each evaluation indicator, control constraints are set. The output of the load forecasting sub-model, daily load characteristic parameters, and each evaluation indicator are used as inputs. The control indicator decomposition sub-model is constructed by combining the indicator weights of each evaluation indicator through a multi-objective optimization algorithm. The load control model is obtained by training the load forecasting sub-model and the regulation index decomposition sub-model using a training sample set.

[0040] The extracted daily load characteristic parameters are used as input features of the samples, and the quantitative calculation results of each indicator in the electricity evaluation index system are used as labels for the samples. At the same time, scenario attribute labels are added to each sample, and finally integrated to form a structured training sample set.

[0041] A load forecasting sub-model is constructed with the goal of achieving accurate future load prediction based on time-series patterns, while ensuring the prediction results conform to scenario constraints. Specifically, a Long Short-Term Memory (LSTM) network is selected as the corresponding model architecture to effectively capture the temporal continuity of load changes. The input to the load forecasting sub-model consists of daily load characteristic parameters from the training sample set. Simultaneously, scenario thresholds for various evaluation indicators are transformed into prediction constraints, such as ensuring that the predicted future daily load peak does not exceed the scenario upper limit threshold for safety, and that fluctuation amplitude remains within the acceptable range of stability indicators, thus preventing the prediction results from deviating from actual control boundaries. The output of the load forecasting sub-model is the load forecast value for the future period, providing accurate load baseline data for subsequent decomposition of control indicators.

[0042] To achieve a reasonable allocation of control quantities under multi-objective constraints, a control index decomposition sub-model is constructed with weight priority as the objective. The inputs to this sub-model include the future load value output from the load forecasting sub-model, daily load characteristic parameters, and the quantified results of each evaluation index. Simultaneously, the scenario thresholds of the evaluation indices are transformed into control constraints, such as ensuring that user control quantities do not exceed their controllable load range thresholds. Furthermore, the rule that the constraint satisfaction of high-weighted indices occupies higher priority in optimization is applied, integrating the weights of each index into the optimization objective function. Finally, by solving the optimization objective function, the overall control objective is decomposed into specific control instructions for individual users or time periods, ensuring that the results satisfy all scenario constraints while achieving optimal trade-offs among multiple objectives.

[0043] Finally, the load forecasting sub-model and the regulation index decomposition sub-model are trained using the constructed training sample set. Specifically, the weight parameters of the long short-term memory network are first iteratively optimized to minimize the error between the predicted value and the actual load-related indicators in the sample labels, while ensuring that the prediction results always meet the scenario constraints. Then, the output of the trained load forecasting sub-model, combined with the sample set data, is used as input to train the regulation index decomposition sub-model. A multi-objective optimization algorithm is used iteratively to optimize the matching degree between the decomposition results and the regulation effect evaluation indicators in the sample labels, ensuring that the decomposition logic conforms to the weight priority and constraint requirements. Finally, the final load management model is obtained through training.

[0044] Then, the real-time grid supply load data of the target control area is input into the load management model to obtain load control strategies, including: Collect real-time grid load data for the target control area, extract corresponding real-time daily load characteristic parameters, and combine the associated auxiliary data from the collection time to determine the current electricity consumption scenario; Input the real-time daily load characteristic parameters, associated auxiliary data and the current electricity consumption scenario into the load prediction sub-model in the load management model to obtain the load prediction results; The load forecast results and real-time daily load characteristic parameters are input into the control index decomposition sub-model of the load management model. The control index decomposition sub-model associates the real-time daily load characteristic parameters with the corresponding scenario constraints, solves and outputs the optimal load control strategy that satisfies all scenario constraints of the current electricity consumption scenario.

[0045] First, real-time grid load data for the target control area is collected. Then, using the same feature extraction method as for typical and atypical daily load characteristics, the corresponding real-time daily load characteristic parameters are extracted. Simultaneously, relevant auxiliary data directly affecting electricity consumption patterns are collected, such as real-time weather, holiday status, and information on major events. The current electricity consumption scenario is then determined by combining the real-time load characteristic parameters and relevant auxiliary data. The current grid operating environment is reconstructed using the real-time daily load characteristic parameters, relevant auxiliary data, and the current electricity consumption scenario. Finally, the load management model performs corresponding load forecasting and load control planning based on the current grid operating environment.

[0046] The load forecasting sub-model in the load management model is input into the real-time daily load characteristic parameters, associated auxiliary data, and the current electricity consumption scenario. The load forecasting sub-model will output the load forecast results for a certain period of time in the future based on the daily load characteristic parameters and the scenario constraints of the application electricity scenario, combined with the learned load characteristics and the pattern of time series changes.

[0047] The load forecasting results output by the load forecasting sub-model are then combined with real-time daily load characteristic parameters and input into the control index decomposition sub-model. The control index decomposition sub-model will automatically associate the evaluation index threshold of the current electricity consumption scenario with the scenario constraint conditions, and then combine the corresponding preset evaluation index weights under the scenario. The model will solve the problem within the range of satisfying all scenario constraints through a multi-objective optimization algorithm, and finally output the optimal load control strategy.

[0048] The output load control strategy includes time-based total control amount, user-type control instructions, and control execution priority.

[0049] Considering that the output load control strategy is a macro-level framework, and that multiple grid-supplied loads with different operating states exist within the target control area, fine-grained control of these loads cannot be achieved solely through a macro-level strategy. Therefore, a further step is taken: by assigning a corresponding control amount to each grid-supplied load, the total control amount for each time period is precisely broken down into individual loads. This ensures that the sum of all load control amounts matches the overall target, preventing control failure due to total deviation. Simultaneously, dedicated time periods are set based on the electricity consumption characteristics of each load to reduce the impact of control on users' production and daily lives, and improve compliance.

[0050] Specifically, based on the load control strategy, the control period and corresponding control amount for each grid supply load within the target control area are set, including: The control period range is divided according to the total control volume of each time period, and the control period to which each grid supply load belongs is determined by combining real-time daily load characteristic parameters and control execution priority. The control ratio of each user type is obtained based on the control instructions for different user types, and the basic control amount of each network supply load is obtained by combining the real-time daily load characteristic parameters. Based on the adjustable capacity of each grid supply load, and combined with the total control quantity of the corresponding control period, the basic control quantity of each grid supply load is adjusted to obtain the control quantity of each grid supply load.

[0051] First, the control period range is divided based on the distribution of the total control volume in different time periods and the real-time daily load characteristic parameters. Then, the control period corresponding to each load is determined based on the control execution priority and the electricity consumption characteristics of each grid load. For example, industrial loads with high control execution priority are allocated to their adjustable lunch break time, while residential loads with low control execution priority are allocated to the non-peak electricity consumption period at night.

[0052] Then, based on the user type control instructions, determine the control ratio of different user types, and combine the current load scale of each network supply load and the load ratio of the same historical period in the real-time daily load characteristic parameters to initially allocate the basic control amount according to the user type control ratio.

[0053] Finally, based on the extracted real-time adjustable capacity parameters of each grid-supplied load, it is determined whether the basic control quantity is within the adjustable range. Simultaneously, the allocation of the basic control quantity is verified in conjunction with the total control quantity for the corresponding control period. If the sum of the basic control quantities for all loads in that period deviates from the total control quantity, a secondary adjustment is made according to the adjustable capacity ratio of each load. Ultimately, this ensures that the control quantity for each grid-supplied load neither exceeds the adjustable limit nor fails to match the total control quantity for the time period, thus achieving precise control of the grid-supplied load.

[0054] After setting the control period and corresponding control amount for each grid supply load within the target control area according to the load control strategy, the following also applies: The power supply loads of each grid are scheduled according to the corresponding control period and control amount, and real-time control data of each grid power supply loads are collected in real time. The control execution deviation rate is calculated based on the corresponding control quantity and real-time control data. The scenario thresholds of the corresponding assessment indicators in the electricity consumption assessment indicator system and the corresponding model parameters of the load control model are adjusted according to the control execution deviation rate.

[0055] Real-time control data is collected during the control process through power grid dispatching systems and load monitoring terminals. This includes the actual start and end times of control for each grid-supplied load, the real-time controlled load amount, and the load fluctuation curve during control. Based on the collected real-time control data and control amounts, control execution deviations are calculated to identify substandard loads, over-adjusted loads, and periods of concentrated deviations, thereby pinpointing the causes of the deviations.

[0056] Based on the cause of the deviation, determine the assessment indicators or model parameters of the load control model that need adjustment, and then determine the specific adjustment amount in conjunction with the control execution deviation rate. For example, if the deviation is caused by a certain type of load failing to reach the predetermined control amount, it proves that the load adjustment potential threshold in the original scenario threshold is too high, and the corresponding threshold needs to be lowered. If the deviation is caused by an inaccurate load forecasting sub-model, then the time-series weight parameters of the load forecasting sub-model need to be fine-tuned based on real-time control data and the control execution deviation rate to improve forecast accuracy.

[0057] By adjusting the feedback of the deviation execution rate, the load control accuracy can be continuously improved as the operating data iterates, thus ensuring the accuracy of the control of the grid-supplied load.

[0058] Another aspect of this embodiment also provides a grid-based load control system based on electricity consumption assessment, such as... Figure 2 As shown, it includes: Load analysis module 1 is used to retrieve historical load data of the target control area, construct typical daily electricity consumption curves and atypical daily electricity consumption curves by combining the corresponding collection time, and extract daily load characteristic parameters; Electricity assessment module 2 is connected to the load analysis module and is used to establish an electricity assessment index system based on daily load characteristic parameters. Model building module 3, connected to the electricity consumption assessment module, is used to build a load control model including a load prediction sub-model and a control index decomposition sub-model based on the electricity consumption assessment index system and daily load characteristic parameters; The load control module 4 is connected to the model building module. It is used to input the real-time grid supply load data of the target control area into the load management model, obtain the load control strategy, and set the control period and corresponding control amount of each grid supply load in the target control area according to the load control strategy.

[0059] The grid-connected load control system also includes: Feedback adjustment module 5 is connected to the electricity consumption assessment module, the model building module, and the load control module, respectively. It is used to collect real-time control data of the load supplied by each grid in real time, calculate the control execution deviation rate in combination with the load control strategy, and adjust the scenario threshold of the corresponding assessment index in the electricity consumption assessment index system and the corresponding model parameters of the load control model according to the control execution deviation rate.

[0060] The load analysis module, electricity consumption assessment module, model building module, load control module, and feedback regulation module are all data processing devices such as microprocessors and computers with corresponding data processing capabilities and built-in relevant algorithms.

[0061] The load control module can be equipped with an analysis unit that extracts features from real-time network load data, or it can directly call the load analysis module to extract features.

[0062] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Other variations and modifications are possible without departing from the technical solutions described in the claims.

Claims

1. A grid-based load control method based on electricity consumption assessment, characterized in that, include: Historical load data of the target control area is retrieved, and typical and atypical daily electricity consumption curves are constructed by combining the corresponding collection time. Daily load characteristic parameters are then extracted. A power consumption assessment index system is established based on daily load characteristic parameters, and a load control model including a load forecasting sub-model and a control index decomposition sub-model is constructed in combination with daily load characteristic parameters. Input the real-time grid supply load data of the target control area into the load control model to obtain the load control strategy; Based on the load control strategy, set the control period and corresponding control amount for each grid supply load within the target control area.

2. The grid-supply load control method based on electricity consumption assessment according to claim 1, characterized in that, The process involves retrieving historical load data for the target control area, constructing typical and atypical daily electricity consumption curves based on the corresponding collection time, and extracting daily load characteristic parameters, including: Retrieve historical load data for the target control area and synchronously associate it with the corresponding collection time; Historical load data are classified according to collection time and scenario attributes, and typical daily electricity consumption curves and atypical daily electricity consumption curves are constructed based on the classified historical load data. Feature extraction is performed on typical and atypical daily electricity consumption curves to obtain characteristic parameters of typical and atypical daily loads; Correlation analysis is performed on typical daily load characteristic parameters and atypical daily load characteristic parameters to identify corresponding parameter correlation terms, and targeted corrections are made based on the parameter correlation terms; A set of daily load characteristic parameters is established based on the revised typical daily load characteristic parameters and atypical daily load characteristic parameters.

3. The grid-supply load control method based on electricity consumption assessment according to claim 1, characterized in that, The establishment of an electricity consumption assessment index system based on daily load characteristic parameters includes: The evaluation index dimensions are set according to the parameter type of the daily load characteristic parameters, and the mapping relationship between the daily load characteristic parameters and the evaluation index is established. Based on the established mapping relationship, the quantitative formulas and corresponding scenario thresholds for each evaluation indicator are set, and the weights of each evaluation indicator are allocated using the analytic hierarchy process.

4. The grid-supply load control method based on electricity consumption assessment according to claim 3, characterized in that, The process involves establishing an electricity consumption assessment index system based on daily load characteristic parameters, and constructing a load control model that includes a load forecasting sub-model and a control index decomposition sub-model, based on the daily load characteristic parameters. The daily load characteristic parameters are correlated with the electricity consumption assessment index system to form a training sample set; Prediction constraints are set based on the scenario thresholds of each evaluation indicator. The load prediction sub-model is constructed using the daily load characteristic parameters as input and a long short-term memory network. Based on the scenario thresholds of each evaluation indicator, control constraints are set. The output of the load forecasting sub-model, daily load characteristic parameters, and each evaluation indicator are used as inputs. The control indicator decomposition sub-model is constructed by combining the indicator weights of each evaluation indicator through a multi-objective optimization algorithm. The load control model is obtained by training the load forecasting sub-model and the regulation index decomposition sub-model using a training sample set.

5. The grid-supply load control method based on electricity consumption assessment according to claim 1, characterized in that, The step of inputting real-time grid supply load data of the target control area into the load management model to obtain load control strategies includes: Collect real-time grid load data for the target control area, extract corresponding real-time daily load characteristic parameters, and combine the associated auxiliary data from the collection time to determine the current electricity consumption scenario; Input the real-time daily load characteristic parameters, associated auxiliary data and the current electricity consumption scenario into the load prediction sub-model in the load management model to obtain the load prediction results; The load forecast results and real-time daily load characteristic parameters are input into the control index decomposition sub-model of the load management model. The control index decomposition sub-model associates the real-time daily load characteristic parameters with the corresponding scenario constraints, solves and outputs the optimal load control strategy that satisfies all scenario constraints of the current electricity consumption scenario.

6. The grid-supply load regulation method based on electricity consumption assessment according to claim 5, characterized in that, The load control strategy includes time-based total control amount, user-type-based control instructions, and control execution priority.

7. The grid-supply load regulation method based on electricity consumption assessment according to claim 6, characterized in that, The step of setting the control period and corresponding control amount for each grid supply load within the target control area according to the load control strategy includes: The control period range is divided according to the total control volume of each time period, and the control period to which each grid supply load belongs is determined by combining real-time daily load characteristic parameters and control execution priority. The control ratio of each user type is obtained based on the control instructions for different user types, and the basic control amount of each network supply load is obtained by combining the real-time daily load characteristic parameters. Based on the adjustable capacity of each grid supply load, and combined with the total control quantity of the corresponding control period, the basic control quantity of each grid supply load is adjusted to obtain the control quantity of each grid supply load.

8. The grid-supply load control method based on electricity consumption assessment according to claim 1, characterized in that, After setting the control period and corresponding control amount for each grid supply load within the target control area according to the load control strategy, the following also applies: The power supply loads of each grid are scheduled according to the corresponding control period and control amount, and real-time control data of each grid power supply loads are collected in real time. The control execution deviation rate is calculated based on the corresponding control quantity and real-time control data. The scenario thresholds of the corresponding assessment indicators in the electricity consumption assessment indicator system and the corresponding model parameters of the load control model are adjusted according to the control execution deviation rate.

9. A grid-based load control system based on electricity consumption assessment, used to execute the grid-based load control system according to any one of claims 1 to 8, characterized in that, include: The load analysis module is used to retrieve historical load data of the target control area, construct typical daily electricity consumption curves and atypical daily electricity consumption curves by combining the corresponding collection time, and extract daily load characteristic parameters; The electricity consumption assessment module is connected to the load analysis module and is used to establish an electricity consumption assessment index system based on daily load characteristic parameters. The model building module, connected to the electricity consumption assessment module, is used to build a load control model, including a load forecasting sub-model and a control index decomposition sub-model, based on the electricity consumption assessment index system and daily load characteristic parameters. The load control module, connected to the model building module, is used to input real-time grid supply load data of the target control area into the load management model, obtain the load control strategy, and set the control period and corresponding control amount for each grid supply load in the target control area according to the load control strategy.

10. The grid-connected load control system based on electricity consumption assessment according to claim 9, characterized in that, Also includes: The feedback adjustment module is connected to the electricity consumption assessment module, the model building module, and the load control module, respectively. It is used to collect real-time control data of the load supplied by each grid in real time, calculate the control execution deviation rate in combination with the load control strategy, and adjust the scenario threshold of the corresponding assessment index in the electricity consumption assessment index system and the corresponding model parameters of the load control model according to the control execution deviation rate.

Images