A method for predicting the application volume of inter-provincial high-voltage direct current transmission channels based on multi-source fusion and probability.

Abstract

This invention discloses a method for predicting the application volume of inter-provincial high-voltage direct current (HVDC) channels based on multi-source fusion and probability, belonging to the field of data processing technology. The method includes the following steps: S1, determining the prediction period and basic parameters; S2, determining the final available capacity based on the basic parameters; S3, calculating energy priority based on the final available capacity; S4, constructing a pass rate prediction model based on the final available capacity; and S5, determining the output results for the prediction period based on the pass rate prediction model and energy priority. This invention is the first to model channel capacity calculation as a deterministic optimization problem with physical constraints, combining a probabilistic correction factor for meteorological risk to achieve a hybrid modeling of determinism and stochasticity. It supports risk attribution, outperforms black-box prediction, and improves the application pass rate.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability. Background Technology

[0002] In inter-provincial electricity spot trading, market participants need to predict the available capacity of transmission channels during the application stage. Existing methods mostly use the static capacity deduction method, which ignores the impact of meteorological disturbances on transmission capacity and the nonlinear constraints of grid safety verification, resulting in significant deviations between the application volume and the final verification results. Especially with the high proportion of renewable energy, competition for transmission channel resources has intensified, urgently requiring a prediction method with mathematical rigor, configurability, and risk quantification capabilities. Summary of the Invention

[0003] To address the above problems, this invention proposes a method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability.

[0004] The technical solution of this invention is: a method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability, comprising the following steps: S1. Determine the forecast period and basic parameters; S2. Determine the final available capacity based on the basic parameters; S3. Calculate energy priority based on the final available capacity; S4. Based on the final available capacity, construct a throughput prediction model; S5. Determine the output results for the forecast period based on the throughput prediction model and energy priority.

[0005] Furthermore, in S1, the basic parameters include the total capacity of the input channels.

[0006] Furthermore, S2 includes the following sub-steps: S21. Construct the locked capacity; S22. Calculate the capacity of contracts in transit; S23. Calculate the configurable constraint parameters; S24. Calculate the theoretical available capacity based on the basic parameters, locked capacity, in-transit contract capacity, and configurable constraint parameters. S25. The theoretical available capacity is corrected using safety constraints to obtain the final available capacity.

[0007] Furthermore, in S22, the capacity of contracts in transit The expression is: ; in, For the first The same period of the month Daily application volume This is the average review period. As weight, This represents the number of historical months. The current moment; In S23, configurable constraint parameters include spot reserve margin, minimum safety check margin, and meteorological risk factor; Meteorological risk factors The expression is: ; in, As the first adjustable weight, As the second adjustable weight, This represents the probability of extreme weather events. The index is eroded by electricity consumption for people's livelihood or to ensure power supply. In S24, the theoretical available capacity The expression is: ; in, This represents the total capacity of the input channels. Capacity is already locked. Reserve a margin for spot goods; In S25, the final available capacity The expression is: ; in, Minimum margin for safety verification.

[0008] Furthermore, in S3, energy priority The expression is: ; in, For the final available capacity, For the specific energy type or energy unit whose priority is being calculated, The parameters are used to minimize the function. As the optimal energy type or energy unit, For energy Belonging to the same category or having the same attribute The collection of all energy sources.

[0009] Furthermore, S4 includes the following sub-steps: S41. Construct feature vectors based on the final available capacity; S42. Construct a pass rate prediction model based on the feature vector.

[0010] Furthermore, in S41, the eigenvector The expression is: ; ; ; ; ; in, The deviation from the historical mean. This represents the probability of an extreme event occurring. The intensity of competition during the journey, For maintenance indicator variables, For the remaining relevant features, For the final available capacity, This represents the historical average available capacity. This is an indicator of the probability of extreme events. For the capacity of contracts in transit, For the maintenance plan of power transmission channels, This represents the total capacity of the input channels.

[0011] Furthermore, in S42, the pass rate prediction model The expression is: ; in, For the Sigmoid function, For feature vectors, For the weight vector, For bias terms, This is the transpose operator.

[0012] Furthermore, in S5, if the pass probability determined by the pass rate prediction model is greater than or equal to the first threshold, then energy priority is output; if the pass probability determined by the pass rate prediction model is greater than or equal to the second threshold and less than the first threshold, then risk warning is output; if the pass probability determined by the pass rate prediction model is greater than or equal to the third threshold and less than the second threshold, then recommended declaration quantity is output.

[0013] Furthermore, the recommended number of applications The expression is: ; in, The pass rate prediction model determines the pass probability. This refers to the amount of energy declared.

[0014] The beneficial effects of this invention are as follows: This invention integrates the modeling method of channel physical constraints, historical behavior data, external environmental risks, and scheduling rules to predict the amount of electricity that can be declared for different power sources (new energy, thermal power, and hydropower) and their safe passage probability for inter-provincial high-voltage DC channels in a specific future period; This invention is the first to model channel capacity calculation as a deterministic optimization problem with physical constraints, and combines the probability correction factor of meteorological risks to achieve a hybrid modeling of determinism and randomness, supporting risk attribution, which is superior to black-box prediction and improves the declaration pass rate. Attached Figure Description

[0015] Figure 1 The flowchart shows a method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability. Figure 2 This is a time series trend chart. Detailed Implementation

[0016] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0017] like Figure 1 As shown, this invention provides a method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability, including the following steps: S1. Determine the forecast period and basic parameters; S2. Determine the final available capacity based on the basic parameters; S3. Calculate energy priority based on the final available capacity; S4. Based on the final available capacity, construct a throughput prediction model; S5. Determine the output results for the forecast period based on the throughput prediction model and energy priority.

[0018] In this embodiment of the invention, S1 includes the total capacity of the input channels.

[0019] Let the prediction target be the future number... The day of Scheduling time intervals (e.g., 15-minute or 1-hour granularity).

[0020] In this embodiment of the invention, S2 includes the following sub-steps: S21. Construct the locked capacity; S22. Calculate the capacity of contracts in transit; S23. Calculate the configurable constraint parameters; S24. Calculate the theoretical available capacity based on the basic parameters, locked capacity, in-transit contract capacity, and configurable constraint parameters. S25. The theoretical available capacity is corrected using safety constraints to obtain the final available capacity.

[0021] In this embodiment of the invention, in S21, the capacity has been locked. The expression is: ; in, To ensure a reliable power supply, For special-purpose electricity use, For cross-regional priority transmission protocol, Medium- to long-term contracts have been signed with large industrial and commercial users; In S22, the capacity of contracts in transit The expression is: ; in, For the first The same period of the month Daily application volume This is the average review period. As weight, This represents the number of historical months. The current moment; (or use exponentially decaying weights) This reflects a higher confidence level in recent data.

[0022] In S23, configurable constraint parameters include spot reserve margin, minimum safety check margin, and meteorological risk factor; Spot reserves (Configurable; minimum margin for safety check) (Configurable).

[0023] Meteorological risk factors The expression is: ; in, As the first adjustable weight, As the second adjustable weight, This represents the probability of extreme weather events. The index is eroded by electricity consumption for people's livelihood or to ensure power supply. In S24, the theoretical available capacity The expression is: ; in, This represents the total capacity of the input channels. Capacity is already locked. Reserve a margin for spot goods; If a comprehensive overhaul plan exists, then If partial maintenance is required, multiply by the maintenance derating factor. .

[0024] In S25, the final available capacity The expression is: ; in, Minimum margin for safety verification.

[0025] In this embodiment of the invention, in S3, let the energy type set be... Its priority sequence is The smaller the value, the higher the priority. For renewable energy, For thermal power, It is for hydroelectric power.

[0026] Energy Priority The expression is: ; in, For the final available capacity, For the specific energy type or energy unit whose priority is being calculated, The parameters are used to minimize the function. As the optimal energy type or energy unit, For energy Belonging to the same category or having the same attribute The collection of all energy sources.

[0027] Within each province, select the energy source with the lowest available cost (e.g., the lowest available cost) and assign it high priority (equal to its available capacity), while other energy sources within the province have a priority of 0.

[0028] In this embodiment of the invention, S4 includes the following sub-steps: S41. Construct feature vectors based on the final available capacity; S42. Construct a pass rate prediction model based on the feature vector.

[0029] In this embodiment of the invention, in S41, the feature vector The expression is: ; ; ; ; ; in, The deviation from the historical mean. This represents the probability of an extreme event occurring. The intensity of competition during the journey, For maintenance indicator variables, For the remaining relevant features, For the final available capacity, This represents the historical average available capacity. This is an indicator of the probability of extreme events. For the capacity of contracts in transit, For the maintenance plan of power transmission channels, This represents the total capacity of the input channels.

[0030] This feature vector The inputs used to build predictive models (such as machine learning models) are each feature that characterizes the state of the high-voltage direct current channel at time (t, h) from different perspectives: This reflects the degree to which the voltage load deviates from the historical average at a given time point (t, h).

[0031] This reflects the probability of extreme weather (such as storms, extreme temperatures), extreme loads, or major power grid events occurring at a given time point (t, h).

[0032] It reflects the proportion of capacity that has already been pre-occupied, thus indicating the degree of resource scarcity.

[0033] This reflects the impact of planned maintenance on the available capacity of the channel.

[0034] In practical applications, this indicates the inclusion of more relevant features (such as seasonal indicators, market quotations, inter-provincial transmission policies, economic factors, and other power grid state variables), and the selection of variable factors to predict the results is considered based on specific circumstances. These features together provide multi-dimensional information support for predicting the application volume of inter-provincial high-voltage direct current channels (i.e., the application value of future actual usage capacity).

[0035] It is usually a binary variable (0 indicates no maintenance, 1 indicates maintenance) or a continuous variable (representing the capacity reduction ratio caused by maintenance).

[0036] In this embodiment of the invention, in S42, the pass rate prediction model The expression is: ; in, For the Sigmoid function, For feature vectors, For the weight vector, For bias terms, This is the transpose operator.

[0037] In this embodiment of the invention, in step S5, if the pass probability determined by the pass rate prediction model is greater than or equal to the first threshold, then energy priority is output; if the pass probability determined by the pass rate prediction model is greater than or equal to the second threshold and less than the first threshold, then risk warning is output; if the pass probability determined by the pass rate prediction model is greater than or equal to the third threshold and less than the second threshold, then recommended declaration quantity is output.

[0038] In this embodiment of the invention, the recommended reporting quantity is... The expression is: ; in, The pass rate prediction model determines the pass probability. This refers to the amount of energy declared.

[0039] In this embodiment of the invention, the single-time multi-channel scheme is shown in Table 1.

[0040] Table 1 like Figure 2 The diagram illustrates a single-channel, multi-period scenario, demonstrating the actual forecast and application recommendations generated after applying the above method to a specific HVDC channel and executing it continuously for multiple periods (T0-T7). The solid line represents the predicted available capacity, and the dashed line represents the recommended application amount. Applications in the red area carry a higher risk, those in the yellow area carry a lower risk, and those in the green area are likely to be approved safely. When the predicted capacity is low (e.g., T2, T3), the recommended application amount is usually more conservative (with a larger difference); when the predicted capacity is sufficient (e.g., T1, T6, T7), the recommended application amount is closer to the predicted value, indicating a more proactive strategy.

[0041] by Figure 2 Taking the T3 time period as an example, the predicted available capacity is 270MW, and the recommended application amount is 260MW. The decision-making process is as follows: The first step of this invention involves multi-source fusion and capacity prediction, integrating information from multiple sources such as meteorology, maintenance plans, in-transit contracts, and historical data to predict the final available capacity of the channel during time period T3. Figure 2 Of the 270MW.

[0042] The second step of this invention involves feature extraction and risk level assessment. The system calculates the feature vector: (Relative historical deviation) = 270 / historical mean (In-transit competition intensity) = In-transit contracts / Total capacity, and (Extreme probability) (Maintenance indicator variables), etc. The system will refer to the risk level framework in Table 1 to determine which level the T3 period falls under. Assuming that due to intense competition (high... If the safety margin is low, the system determines that it is in the "higher risk (red)" level.

[0043] The third step involves applying decision-making rules and providing recommendations. The logic for the declared capacity is: "Recommended declaration: 300MW (Original available capacity: 500MW)". This means that under red-level risk, the recommended declaration should be lower than the predicted available capacity to reserve a safety buffer. The predicted available capacity is 270MW; following the conservative strategy for red-level risk, the system recommends a lower declaration value of 260MW. This 10MW difference is the buffer space reserved for risks such as "intense competition in transit" and "safety margin approaching the threshold". Simultaneously, the system's internal probability model calculates the expected approval rate corresponding to this declaration (260MW). The reference approval rate for red-level risk is approximately 63%, meaning that this declaration carries a certain degree of uncertainty.

[0044] Fourth, the invention will generate risk warnings, and generate qualitative warnings based on the rules in Table 1, such as "Intense competition in transit, safety margin close to the threshold", for decision-makers to refer to.

[0045] In this invention, the system corresponding to the inter-provincial high-voltage direct current channel application quantity prediction method based on multi-source fusion and probability adopts a five-layer hierarchical design, including a five-layer architecture of "data access → data processing → core calculation → model service → application output". The modules are loosely coupled and highly cohesive. The specific architecture design is as follows: In the external data source layer, all input data comes from authoritative systems or third-party platforms and is accessed through standardized interfaces, as shown in Table 2.

[0046] Table 2 The data integration and preprocessing layer cleanses, aligns, and extracts features from multi-source heterogeneous data. The data integration gateway uniformly receives various inputs and performs authentication, format conversion (JSON / XML), and timestamp normalization.

[0047] The channel status parser parses the maintenance plan and outputs a Boolean flag. If it is a partial overhaul, output the derating factor. .

[0048] The locked capacity calculator calculates the locked capacity. The in-transit contract prediction engine uses a sliding time window weighted model to calculate the in-transit contract capacity and supports online updates of weight parameters. The weather risk quantification module outputs weather risk factors.

[0049] In the feature engineering service, the feature vectors required to construct the logistic regression model are generated.

[0050] The core computing engine layer performs deterministic capacity calculations and priority allocation.

[0051] The capacity constraint solver executes the master formula; if maintenance is required, then... Alternatively, set it to 0 and perform a security check and judgment.

[0052] The energy type allocator outputs priority order allocation.

[0053] The machine learning model service layer provides pass rate prediction capabilities and supports hot model updates.

[0054] The pass rate prediction service includes model types such as Logistic Regression, XGBoost, and LightGBM (LR is used by default to ensure interpretability). Prediction results are automatically recorded for subsequent feedback learning.

[0055] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A method for predicting the application volume of inter-provincial high-voltage direct current (HVDC) channels based on multi-source fusion and probability, characterized in that, Includes the following steps: S1. Determine the forecast period and basic parameters; S2. Determine the final available capacity based on the basic parameters; S3. Calculate energy priority based on the final available capacity; S4. Based on the final available capacity, construct a throughput prediction model; S5. Determine the output results for the forecast period based on the throughput prediction model and energy priority; S4 includes the following sub-steps: S41. Construct feature vectors based on the final available capacity; S42. Construct a pass rate prediction model based on the feature vectors; In S41, the feature vector The expression is: ； ； ； ； ； in, The deviation from the historical mean. This represents the probability of an extreme event occurring. The intensity of competition during the journey, For maintenance indicator variables, For the remaining relevant features, For the final available capacity, This represents the historical average available capacity. This is an indicator of the probability of extreme events. For the capacity of contracts in transit, For the maintenance plan of power transmission channels, This represents the total capacity of the input channels. In step S5, if the pass probability determined by the pass rate prediction model is greater than or equal to the first threshold, then energy priority is output; if the pass probability determined by the pass rate prediction model is greater than or equal to the second threshold and less than the first threshold, then risk warning is output. If the pass rate prediction model determines a pass probability that is greater than or equal to the third threshold and less than the second threshold, then the recommended number of applications will be output.

2. The method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability as described in claim 1, characterized in that, In S1, the basic parameters include the total capacity of the input channels.

3. The method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability as described in claim 1, characterized in that, S2 includes the following sub-steps: S21. Construct the locked capacity; S22. Calculate the capacity of contracts in transit; S23. Calculate the configurable constraint parameters; S24. Calculate the theoretical available capacity based on the basic parameters, locked capacity, in-transit contract capacity, and configurable constraint parameters. S25. The theoretical available capacity is corrected using safety constraints to obtain the final available capacity.

4. The method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability according to claim 3, characterized in that, In S22, the in-transit contract capacity The expression is: ； in, For the first The same period of the month Daily application volume This is the average review period. As weight, This represents the number of historical months. The current moment; In S23, the configurable constraint parameters include spot reserve margin, minimum safety check margin, and meteorological risk factor; The meteorological risk factors The expression is: ； in, As the first adjustable weight, As the second adjustable weight, This represents the probability of extreme weather events. The index is eroded by electricity consumption for people's livelihood or to ensure power supply. In S24, the theoretical available capacity The expression is: ； in, This represents the total capacity of the input channels. Capacity is already locked. Reserve a margin for spot goods; In S25, the final available capacity The expression is: ； in, Minimum margin for safety verification.

5. The method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability as described in claim 1, characterized in that, In S3, energy priority The expression is: ； in, For the final available capacity, For the specific energy type or energy unit whose priority is being calculated, The parameters are used to minimize the function. As the optimal energy type or energy unit, For energy Belonging to the same category or having the same attribute The collection of all energy sources.

6. The method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability according to claim 1, characterized in that, In S42, the pass rate prediction model The expression is: ； in, For the Sigmoid function, For feature vectors, For the weight vector, For bias terms, This is the transpose operator.

7. The method for predicting the application volume of inter-provincial high-voltage direct current channels based on multi-source fusion and probability according to claim 1, characterized in that, The recommended reporting volume The expression is: ； in, The pass rate prediction model determines the pass probability. This refers to the amount of energy declared.

Images