Checking method and system for random production simulation and multi-scene load prediction

Abstract

The invention discloses a checking method for random production simulation and multi-scene load prediction, and the method comprises the steps: obtaining load prediction influence factors of a selected to-be-predicted region, carrying out the sorting of the influence degrees of the load prediction influence factors, obtaining the historical data of the load prediction influence factors according to the sorted load prediction influence factors, and carrying out the verification of the historical data of the load prediction influence factors. Selecting a load prediction model of the to-be-predicted area by using the historical data of the load prediction influence factors, and predicting the power load of the to-be-predicted area to obtain a first predicted load value; judging the load level index of the first predicted load value and the total number of the load levels, and obtaining a reliability index of the power generation system when judging that the load level index of the first predicted load value is smaller than the total number of the load levels; according to the reliability index of the power generation system, the monthly electric quantity deviation correction model is used for adjusting loads in different scenes, the operation reliability of the system is improved, and the power generation cost and monthly electric quantity deviation can be considered.

Description

Technical Field

[0001] This invention relates to the field of capacity market technology, specifically to a verification method and system for stochastic production simulation and multi-scenario load forecasting. Background Technology

[0002] With the rapid development of the economy and technology and the improvement of people's living standards, electricity has become indispensable in modern society. It not only provides continuous power for production activities but also brings numerous conveniences to people's daily lives. Against this backdrop, people's demand for electricity is constantly growing, and the need for efficient utilization and optimized allocation of electricity is becoming increasingly important. Current technological advancements enable the rapid selection of optimal cooling load configuration schemes, especially in multi-cooling load scenarios. These schemes not only consider the operational requirements of multiple cooling load scenarios but also the optimized configuration of various refrigeration unit combinations and the comprehensive consideration of multiple optimization objectives. This allows for better fulfillment of the requirements under different cooling load scenarios, maximizing energy utilization efficiency and system performance. Through the application of these technologies, we can better cope with diverse and complex cooling load demands and achieve a higher level of optimized electricity allocation. This will be of great significance for improving energy utilization efficiency, reducing energy consumption, and minimizing adverse environmental impacts. Therefore, continuing to promote the research and application of optimized electricity allocation technologies will not only improve electricity utilization efficiency but also be of great significance for sustainable development and environmental protection.

[0003] However, current optimization scheduling methods that only study a single time scale are insufficient to address the power deviations caused by the uncertainties of large-scale renewable energy grid integration. They lack effective coordination between medium- and long-term power generation and short-term power balance, consider only a single power source structure, and the analysis method that transforms the uncertainties of new energy into qualitative variables yields inaccurate results. Therefore, those skilled in the art provide a system reliability verification method based on stochastic production simulation and multi-scenario load forecasting to address the problems mentioned in the background. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is that existing technologies are unable to cope with the power deviation caused by the uncertainty of large-scale renewable energy grid connection, lack effective connection between medium- and long-term power volume and short-term power balance, and the results obtained by transforming the uncertainty factors of new energy into qualitative variables are inaccurate.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a verification method for stochastic production simulation and multi-scenario load forecasting, comprising:

[0008] Obtain the load forecasting influencing factors of the selected area to be predicted, sort them by the degree of influence of the load forecasting influencing factors, obtain the historical data of the load forecasting influencing factors based on the sorted load forecasting influencing factors, select the load forecasting model of the area to be predicted using the historical data of the load forecasting influencing factors, predict the power load of the area to be predicted, and obtain the first predicted load value.

[0009] Determine the load level index of the first predicted load value and the total number of load levels. The determination includes finding that the load level index of the first predicted load value is less than the total number of load levels to obtain a power generation system reliability index.

[0010] Based on the reliability indicators of the power generation system, the monthly power deviation correction model is used to adjust the load for different scenarios.

[0011] As a preferred embodiment of the verification method for stochastic production simulation and multi-scenario load forecasting described in this invention, wherein:

[0012] Obtain the load forecasting influencing factors for the selected area to be forecasted, and rank them according to their degree of influence, including:

[0013] Load forecasting under the influence of various factors is expressed as follows:

[0014] X=x0+(1+a1)x1+(1+a2)x2+(1+a3)x3+(1+a4)x4

[0015] Where X represents the load forecast under the influence of various factors, x0 is the base load, x1, x2, x3 and x4 are the load components of weather, environment, holidays and remaining influencing factors, and a1, a2, a3 and a4 are the influence magnitudes of weather, environment, holidays and remaining influencing factors.

[0016] Assign initial values ​​to the impact magnitude of each influencing factor. Based on the base load and the load components of each factor, use the least squares method to obtain the impact magnitude of each factor's value at each time compared with the previous value, and the relative difference is within the threshold.

[0017] Based on the magnitude of each factor's influence, the influencing factors are ranked.

[0018] As a preferred embodiment of the verification method for stochastic production simulation and multi-scenario load forecasting described in this invention, wherein:

[0019] Selecting a load forecasting model for the region to be forecasted using historical data of the factors influencing load forecasting includes:

[0020] If the historical data is greater than the first threshold, there is a non-linear relationship between the changes in the historical data and the input variables, and the tolerance for the deviation between the prediction result and the actual observed value is less than the second threshold, then a neural network is selected as the load prediction model.

[0021] Let the time series be {Xi, i = n, n+1, ..., n+m};

[0022] Among them, historical data Xn, Xn+1, ..., Xn+m are used to predict the values ​​at future times n+m+k (k>=1);

[0023] A neural network is used to fit a function f through a set of data points Xn, Xn+1, ..., Xn+m, which is expressed as: Xn+m+k=f(Xn,Xn+1,...,Xn+m), to obtain the predicted value of the data at time n+m+k (k>=1).

[0024] As a preferred embodiment of the verification method for stochastic production simulation and multi-scenario load forecasting described in this invention, it further includes:

[0025] If the historical data is less than the first threshold, and the tolerance for the deviation between the predicted result and the actual observed value is greater than the second threshold, then the time series is selected as the load forecasting model.

[0026] Historical data is arranged chronologically into a time series to obtain the trend of historical data over time. A quantitative forecast is then performed using the quadratic moving average method. The linear model of the quadratic moving average method is expressed as follows:

[0027]

[0028]

[0029]

[0030]

[0031]

[0032] Among them, X t The actual value for period t. This is the predicted value for period t+T, where t is the current period and T is the period from t to the prediction period. This is the moving average for the first stage. The moving average for the second stage, a t For the intercept parameter, b t is the slope parameter, and N is the size of the moving average window.

[0033] As a preferred embodiment of the verification method for stochastic production simulation and multi-scenario load forecasting described in this invention, wherein: the load level index of the first predicted load value is less than the total number of load levels, the resulting power generation system reliability indicators include:

[0034] Set the system generator state sequence index wc = 0; load level index wk = 0;

[0035] The system generator state sequence index wc and load level index wk are gradually increased, i.e., wk = wk + 1, wc = wc + 1;

[0036] Select the system generator state sequence from the system generator state matrix and the corresponding load level to obtain the corresponding load probability;

[0037] The index function for updating the system reliability index at the wk-th load level is expressed as:

[0038] FLOLP(WXwc)=WW(WXwc)×WPwkFEENS(WXwc)=8760×SD

[0039] ×WW(WXwc)×WPwk

[0040] Where SD is the load shedding amount, WXwc is the wc-th sequence in the generator state matrix WX, WPwk is the load probability, and FLOLP(WXwc) and FEENS(WXwc) are index functions.

[0041] Determine if the load level index wk is less than the total number of load levels Nw. If the condition is met, update the system reliability index and variance coefficient under the multi-level load level. If the condition is not met, continue to gradually increase wc and wk.

[0042] If the obtained variance coefficients meet the convergence condition, the reliability index of the power generation system is output; otherwise, the judgment operation continues.

[0043] As a preferred embodiment of the verification method for stochastic production simulation and multi-scenario load forecasting described in this invention, the adjustment of load for different scenarios based on the predicted power load using a monthly power deviation correction model includes:

[0044] Based on the revised monthly power generation plan, the contracted power volume of the power plants is broken down into daily amounts. Each power plant then uses the contracted power volume for the next day obtained from the breakdown, combined with load demand forecasts, to obtain the daily power generation plan for its units.

[0045] Using the minimum adjustment amount of each unit's output as the objective function, a day-ahead optimal scheduling model is established to adjust the unit output, expressed as:

[0046]

[0047] Where T is the optimization time range, N is the number of units, and λ i Let P be the weighting coefficient for the i-th unit. i,t Let i be the output of thermal power unit i during time period t. For the pre-decomposed power of unit i during the t period on the next day, λ w Let w be the weighting coefficient for the w-th unit. Let w be the amount of wind curtailed by the wind farm during time period t. Let ΔP be the amount of solar power curtailed by photovoltaic power station p during time period t. h,t N represents the amount of water discharged by the hydropower unit h during time period t. w λ represents the number of load levels. pv λ is the weighting coefficient for the photovoltaic power generation system. h N represents the weighting coefficient for the wind power generation system. pv N represents the number of photovoltaic power generation systems. h This refers to the number of wind power generation systems.

[0048] As a preferred embodiment of the verification method for stochastic production simulation and multi-scenario load forecasting described in this invention, the adjustment of load for different scenarios using a monthly electricity deviation correction model includes:

[0049] With the objective function of minimizing the monthly electricity consumption deviation adjustment cost within the system, a monthly electricity consumption deviation correction model is established, expressed as follows:

[0050]

[0051] Where C is the scene number, w c For scene C weights, These are the bid prices for increasing and decreasing power generation for unit i, respectively. Let's define the increased and decreased power generation allocated to unit i under scenario c.

[0052] Secondly, the present invention provides a system for verifying stochastic production simulation and multi-scenario load forecasting, comprising:

[0053] The model selection module is used to obtain the load forecasting influencing factors of the selected area to be predicted, sort the load forecasting influencing factors according to their degree of influence, obtain historical data of the load forecasting influencing factors based on the sorted load forecasting influencing factors, select the load forecasting model of the area to be predicted using the historical data of the load forecasting influencing factors, predict the power load of the area to be predicted, and obtain the first predicted load value.

[0054] The judgment module is used to judge the load level index of the first predicted load value and the total number of load levels. The judgment includes determining that the load level index of the first predicted load value is less than the total number of load levels, thereby obtaining the power generation system reliability index.

[0055] The adjustment module is used to adjust the load for different scenarios based on the reliability index of the power generation system and using a monthly power deviation correction model.

[0056] Thirdly, the present invention provides a computing device, comprising:

[0057] Memory and processor;

[0058] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the verification method for random production simulation and multi-scenario load prediction.

[0059] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the verification method for random production simulation and multi-scenario load prediction.

[0060] The beneficial effects of this invention are as follows: The stochastic production simulation and multi-scenario load forecasting verification method proposed in this invention can more accurately and reliably predict future loads. It is simple, reliable, scientific, and practical, and can adjust the power plan for different scheduling scenarios, thereby effectively adapting to the stochasticity of new energy output such as wind, solar, and hydropower. It has strong operational reliability. The constructed monthly power and day-ahead power connection optimization model can ensure the effective decomposition of the corrected monthly power, promote the balance of power, improve the reliability of system operation, and has good economic efficiency. Based on the corrected monthly power plan, the optimized output of each unit can better balance the power generation cost and the monthly power deviation. Attached Figure Description

[0061] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0062] Figure 1 The overall flowchart of the verification method for random production simulation and multi-scenario load prediction provided by the present invention is shown. Figure 2 The voltage trend graph over time is provided for the verification method of random production simulation and multi-scenario load prediction provided by the present invention. Detailed Implementation

[0063] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0064] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0065] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0066] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0067] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0068] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0069] Example 1

[0070] Reference Figure 1 As an embodiment of the present invention, a verification method for stochastic production simulation and multi-scenario load forecasting is provided, comprising:

[0071] S100: Obtain the load forecasting influencing factors of the selected area to be predicted, sort the load forecasting influencing factors according to their degree of influence, obtain the historical data of the load forecasting influencing factors according to the sorted load forecasting influencing factors, select the load forecasting model of the area to be predicted using the historical data of the load forecasting influencing factors, predict the power load of the area to be predicted, and obtain the first predicted load value.

[0072] Furthermore, the load forecast under the influence of various factors is expressed as follows:

[0073] X=x0+(1+a1)x1+(1+a2)x2+(1+a3)x3+(1+a4)x4

[0074] Where X represents the load forecast under the influence of various factors, x0 is the base load, x1, x2, x3 and x4 are the load components of weather, environment, holidays and remaining influencing factors, and a1, a2, a3 and a4 are the influence magnitudes of weather, environment, holidays and remaining influencing factors.

[0075] Assign initial values ​​to the impact magnitude of each influencing factor. Based on the base load and the load components of each factor, use the least squares method to obtain the impact magnitude of each factor's value at each time compared with the previous value, and the relative difference is within the threshold.

[0076] Based on the magnitude of each factor's influence, the influencing factors are ranked.

[0077] It should be noted that the selected area to be predicted is divided into regions according to the provincial, municipal, county, and district levels to clarify the boundaries and scope of the calculation. The load prediction influencing factors within the area to be predicted are obtained and sorted. The sorting steps are repeated until a stable influence amplitude of each factor is obtained. The relative difference between the current value and the previous value is within a threshold of 5%, which is considered to be a stable calculated value. A threshold of 5% can provide sufficient accuracy and reliability to judge the stability of the prediction results. The influence amplitude of each factor is obtained, and each influencing factor is sorted. Influencing factors with an influence amplitude less than the set threshold are ignored. The set threshold is set by the user according to each influencing factor.

[0078] Furthermore, if the historical data exceeds the first threshold, there is a non-linear relationship between the changes in the historical data and the input variables, and the tolerance for the deviation between the prediction result and the actual observed value is less than the second threshold, then a neural network is selected as the load prediction model.

[0079] Let the time series be {Xi, i = n, n+1, ..., n+m};

[0080] Among them, historical data Xn, Xn+1, ..., Xn+m are used to predict the values ​​at future times n+m+k (k>=1);

[0081] A neural network is used to fit a function f through a set of data points Xn, Xn+1, ..., Xn+m, which is expressed as: Xn+m+k=f(Xn,Xn+1,...,Xn+m), to obtain the predicted value of the data at time n+m+k (k>=1).

[0082] If the historical data is less than the first threshold, and the tolerance for the deviation between the predicted result and the actual observed value is greater than the second threshold, then the time series is selected as the load forecasting model.

[0083] Historical data is arranged chronologically into a time series to obtain the trend of historical data over time. A quantitative forecast is then performed using the quadratic moving average method. The linear model of the quadratic moving average method is expressed as follows:

[0084]

[0085]

[0086]

[0087]

[0088]

[0089] Among them, X t The actual value for period t. This is the predicted value for period t+T, where t is the current period and T is the period from t to the prediction period. This is the moving average for the first stage. The moving average for the second stage, a t For the intercept parameter, b t is the slope parameter, and N is the size of the moving average window.

[0090] Specifically, if the historical data exceeds the first threshold by 10%, there is a non-linear relationship between the changes in the historical data and the input variables, and the tolerance for deviation between the predicted results and the actual observed values ​​is less than the second threshold of 15%, then a neural network is selected as the load forecasting model. If the historical data is less than the first threshold by 10%, and the tolerance for deviation between the predicted results and the actual observed values ​​is greater than the second threshold of 15%, then a time series model is selected as the load forecasting model. The first threshold of 10% is based on the amount of historical data, while the second threshold of 15% is based on the requirement for the tolerance of deviation between the predicted results and the actual observed values. When the amount of historical data exceeds the threshold of 10%, and there is a non-linear relationship between the historical data and the input variables, neural networks are usually better able to capture the non-linear relationship. When the amount of historical data is less than the threshold of 10%, the reason for selecting time series as the load forecasting model is that the amount of data is too small to support a complex neural network model. The tolerance for deviation between the predicted results and the actual observed values ​​is greater than 15%, indicating that the model needs to have high stability and reliability.

[0091] It should be noted that choosing a neural network as the load forecasting model includes both single-step and multi-step forecasting. Single-step forecasting occurs when k=1 and all input data to the network are actual time-series observations. The actual time-series observations Xn, Xn+1, ..., Xn+m are input into the network, and the predicted value for the next time step is output as X. n+m+1 If you want to continue with X n+m+2 To predict the value, the actual observed values ​​Xn, Xn+1, ..., Xn+m+1 are used as input data to obtain the predicted value Xn+m+2. When k>1, multi-step prediction is performed, with the network inputting m+1 historical data points and outputting predicted values ​​Xn+m+1, Xn+m+2, ..., Xn+m+k. Multi-step prediction is used when adjusting weights and thresholds during network operation. Each iteration accumulates the error of the previous k predicted values. However, factors such as weather changes have a certain degree of randomness, which can cause the network to have difficulty converging or even oscillate.

[0092] Time series forecasting is a quantitative forecasting method that arranges the historical data of the target in chronological order into a time series, analyzes its trend over time, and establishes a mathematical model for extrapolation.

[0093] It should also be noted that, based on the statistical patterns of weather, environmental, and other data, the range and proportion of different influencing factors are obtained. Optimal load forecast values ​​under different factor combinations are calculated, forming multiple load forecast sensitive schemes. The standard deviation of the optimal load forecast value obtained from each calculation and the standard deviation of the range of variation of different influencing factors are multiplied together to obtain the load forecast error for that set of influencing factors. The optimal load forecast value under different combinations of influencing factors is multiplied by the proportion of the corresponding factor, and then summed to obtain the recommended load forecast scheme.

[0094] Specifically, based on the statistical regularities of influencing factors, the predicted values, ranges of change, and proportions of influence of each influencing factor for the area to be predicted over the prediction period are obtained; based on the selected load prediction model, the optimal values ​​and error range of the load prediction model are calculated under different combinations of influencing factors.

[0095] Historical load data is used to predict load data, and the accuracy of the prediction model is obtained, including the mean error and standard deviation. The prediction model and corresponding prediction steps are automatically selected according to the principle of minimizing the standard deviation. For factors with large influence, the load of each component needs to be predicted separately and then summed to obtain the total load prediction value. When predicting the load value of unknown years in different scenarios, the magnitude and order of the obtained influencing factors are considered, and the load prediction value is calculated by the aforementioned automatically selected prediction model, including the total load prediction and the load prediction of each component with large influence. The optimal load prediction value and its error range are obtained by multiplying the prediction standard deviation of historical data and the load prediction value of unknown years.

[0096] It should be noted that the selected load forecasting model is trained using the obtained historical data to obtain a determined load forecasting model and its corresponding error range; the load forecasting results for each load influencing factor are calculated based on the load forecasting influencing factors and their ranking results; the final load forecasting value and its range of variation are calculated based on the error range obtained from the obtained load forecasting results; and the final load forecasting value and its range of variation are calculated based on the influence ratio of each influencing factor and the optimal value and its error range of the load forecasting model under different combinations of influencing factors.

[0097] S200: Determine the load level index of the first predicted load value and the total number of load levels. The determination includes that the load level index of the first predicted load value is less than the total number of load levels, thereby obtaining the power generation system reliability index.

[0098] Furthermore, set the system generator state sequence index wc = 0; load level index wk = 0;

[0099] The system generator state sequence index wc and load level index wk are gradually increased, i.e., wk = wk + 1, wc = wc + 1;

[0100] Select the system generator state sequence from the system generator state matrix and the corresponding load level to obtain the corresponding load probability;

[0101] The index function for updating the system reliability index at the wk-th load level is expressed as:

[0102] FLOLP(WXw)=WW(WXwc)×WPwkFEENS(WXwc)=8760×SD

[0103] ×WW(WXwc)×WPwk

[0104] Where SD is the load shedding amount, WXwc is the wc-th sequence in the generator state matrix WX, WPwk is the load probability, and FLOLP(WXwc) and FEENS(WXwc) are index functions.

[0105] Determine if the load level index wk is less than the total number of load levels Nw. If the condition is met, update the system reliability index and variance coefficient under the multi-level load level. If the condition is not met, continue to gradually increase wc and wk.

[0106] If the obtained variance coefficients meet the convergence condition, the reliability index of the power generation system is output; otherwise, the judgment operation continues.

[0107] It should be noted that the updated system reliability index and variance coefficient under multi-level load conditions are expressed as follows:

[0108]

[0109]

[0110]

[0111] Where LOLP is the load mismatch probability, EENS is the expected unsupplied energy, and β EENS Here, Wc is the variance coefficient, and F is the number of load levels. LOLP (wx cc ) for given conditions (wx cc The distribution function of the probability of load mismatch, F. EENS (wx cc ) for given conditions (wx cc The expected distribution function of unsupplied energy. for variance for The expected value.

[0112] S300: Based on the reliability indicators of the power generation system, adjust the load for different scenarios using the monthly power deviation correction model;

[0113] Furthermore, based on the revised monthly power generation plan, the contracted power volume of the power plants is broken down into daily amounts. Each power plant then uses the contracted power volume for the next day obtained from the breakdown, combined with load demand forecasts, to obtain the daily power generation plan for its units.

[0114] Using the minimum adjustment amount of each unit's output as the objective function, a day-ahead optimal scheduling model is established to adjust the unit output, expressed as:

[0115]

[0116] Where T is the optimization time range, N is the number of units, and λ i Let P be the weighting coefficient for the i-th unit. i,t Let i be the output of thermal power unit i during time period t. For the pre-decomposed power of unit i during the t period on the next day, λ w Let w be the weighting coefficient for the w-th unit. Let w be the amount of wind curtailed by the wind farm during time period t. Let ΔP be the amount of solar power curtailed by photovoltaic power station p during time period t. h,t N represents the amount of water discharged by the hydropower unit h during time period t. w λ represents the number of load levels. pv λ is the weighting coefficient for the photovoltaic power generation system. h N represents the weighting coefficient for the wind power generation system. pv N represents the number of photovoltaic power generation systems. h This refers to the number of wind power generation systems.

[0117] Adjusting load for different scenarios using a monthly electricity consumption deviation correction model includes:

[0118] With the objective function of minimizing the monthly electricity consumption deviation adjustment cost within the system, a monthly electricity consumption deviation correction model is established, expressed as follows:

[0119]

[0120] Where C is the scene number, w c For scene C weights, These are the bid prices for increasing and decreasing power generation for unit i, respectively. Let's define the increased and decreased power generation allocated to unit i under scenario c.

[0121] Specifically, during the month's operation, the dispatching agency compares the monthly electricity load demand forecast, renewable energy power generation forecast, and monthly contracted power volume to determine the increased and decreased power generation provided by the generating units. It then establishes a monthly power deviation correction model that takes into account the uncertainty of renewable energy generation in the system. The model aims to minimize the monthly power deviation adjustment cost within the system, and the constraints include power supply and demand balance constraints, upper and lower limits constraints on increased and decreased power generation, and upper and lower limits constraints on power generation based on the remaining days of the generating units.

[0122] It should be noted that the electricity supply and demand balance constraint is expressed as:

[0123]

[0124] Among them, W m W represents the predicted monthly load demand of the system for that day. h,c W represents the predicted monthly power generation of the hydropower unit under scenario c. pv,c W represents the predicted monthly power generation from photovoltaic power generation. w,c W represents the predicted monthly wind power generation. i The monthly contracted electricity volume for unit i;

[0125] Upper and lower limits constraints on increased and decreased power generation:

[0126]

[0127]

[0128] Among them, W i max W i min These are the upper limit of unit i's monthly power generation capacity and the lower limit of its required power generation, respectively.

[0129] Remaining days of generating capacity upper and lower limits constraints:

[0130]

[0131] Among them, P i min Let r be the upper limit of the maximum output and the lower limit of the minimum output of unit i, and r be the number of days remaining in the month.

[0132] The monthly power generation plan is revised using a monthly power deviation correction model and constraints.

[0133] It should also be noted that the monthly contract electricity volume decomposition and day-ahead tie-in model is represented as follows:

[0134]

[0135]

[0136]

[0137]

[0138]

[0139] Among them, W i 0 For the contracted electricity volume already completed, W i,d k represents the daily contracted electricity generated by unit i on day d. i This represents the percentage of the monthly pre-generated electricity plan for unit i. Pre-decomposed power for unit i during period t on the next day;

[0140] The established day-ahead optimization scheduling model satisfies the following constraints:

[0141] Thermal power unit output constraints:

[0142] P i,min ·y i,t ≤P i,t ≤P i,max ·y i,t

[0143] Among them, P i,max and P i,min These are the upper and lower limits of the output of thermal power unit i, respectively;

[0144] Thermal power unit ramping constraints:

[0145] -P i,down ≤P i,t -P i,t -1≤P i,up

[0146] Among them, P i,up P i,down These represent the upper and lower limits of the ramp power of thermal power unit i during time period t;

[0147] Start / stop logic constraints:

[0148] U i,t -D i,t =d i,t -d i,t-1

[0149] Where, d i,t U is a 0 / 1 variable representing the operating state of thermal power unit i during time period t. i,t D i,t For the start-up and shutdown status of thermal power unit i during time period t, there are 0 / 1 variables.

[0150] Minimum start-stop time constraints:

[0151] (d i,t-1 -d i,t (T) i,t -T ion )≥0

[0152] (d i,t -d i,t-1 (-T) i,t-1 -T ioff )≥0

[0153] Among them, T i,t-1 T is the continuous operating time of unit i. ion T ioff These are the minimum start-up and shutdown times for unit i, respectively.

[0154] Hydropower generation flow constraints:

[0155] y h,t Qh,min≤q h,t ≤y h,t Qh,max

[0156] Among them, y h,t q is a 0-1 variable representing the operating state of the hydropower unit h during time period t. h,t Let Qh be the power generation flow of hydropower unit h during time period t, and Qh,max and Qh,min be the upper and lower limits of the power generation flow of hydropower unit h.

[0157] Reservoir discharge flow constraints:

[0158]

[0159] ΔP h,t ≥0

[0160] Where h∈S represents the hydropower unit, and h belongs to the reservoir S. These are the upper and lower limits of the discharge flow from reservoir S;

[0161] Storage capacity constraints:

[0162] V s,min ≤V s,t ≤V s,max

[0163] Among them, V s,min V s,max This represents the allowable storage capacity of reservoir S during the scheduling period;

[0164] Initial and final storage capacity constraints:

[0165] V s,0 =V s,0 V s,T ≥V s,T

[0166] Among them, V s,0 V s,T These represent the reservoir capacity of reservoir S at the beginning and end of the scheduling period, respectively;

[0167] Reservoir water balance constraints:

[0168]

[0169] Among them, V s,t Let S be the reservoir capacity of reservoir S during time period t, and R be the capacity of reservoir S during time period t. s,t Let S be the inflow rate of reservoir S during time period t.

[0170] Wind power and solar power output constraints:

[0171]

[0172]

[0173] in, These are the projected output values ​​for wind power and solar power, respectively.

[0174] Hydropower output constraints:

[0175] P h,t =eh,rqh,t+fh,r

[0176] P h,min ≤P h,t ≤P h,max

[0177] The established monthly power deviation correction model that considers the uncertainty of new energy sources can adjust the power plan for different dispatch scenarios, thereby effectively adapting to the randomness of the output of new energy sources such as wind, solar, and hydropower. It has strong operational reliability. The constructed monthly power and day-ahead power connection optimization model can ensure the effective decomposition of the corrected monthly power, promote the balance of power, improve the reliability of system operation, and has good economic efficiency. Based on the corrected monthly power plan, the optimized output of each unit can better balance the power generation cost and the monthly power deviation.

[0178] The above is an illustrative scheme of a verification method for stochastic production simulation and multi-scenario load forecasting according to this embodiment. It should be noted that the technical solution of the device for verifying stochastic production simulation and multi-scenario load forecasting belongs to the same concept as the technical solution of the above-described verification method for stochastic production simulation and multi-scenario load forecasting. Details not described in detail in the technical solution of the device for verifying stochastic production simulation and multi-scenario load forecasting in this embodiment can be found in the description of the technical solution of the above-described verification method for stochastic production simulation and multi-scenario load forecasting.

[0179] The apparatus for verifying random production simulation and multi-scenario load forecasting in this embodiment includes:

[0180] The model selection module is used to obtain the load forecasting influencing factors of the selected area to be predicted, sort the load forecasting influencing factors according to their degree of influence, obtain historical data of the load forecasting influencing factors based on the sorted load forecasting influencing factors, select the load forecasting model of the area to be predicted using the historical data of the load forecasting influencing factors, predict the power load of the area to be predicted, and obtain the first predicted load value.

[0181] The judgment module is used to judge the load level index of the first predicted load value and the total number of load levels. The judgment includes determining that the load level index of the first predicted load value is less than the total number of load levels, thereby obtaining the power generation system reliability index.

[0182] The adjustment module is used to adjust the load for different scenarios based on the reliability index of the power generation system and using a monthly power deviation correction model.

[0183] This embodiment also provides a computing device suitable for verifying stochastic production simulations and multi-scenario load forecasting, including:

[0184] The system includes a memory and a processor. The memory stores computer-executable instructions, and the processor executes these instructions to implement the verification method for random production simulation and multi-scenario load prediction proposed in the above embodiments.

[0185] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the verification method for random production simulation and multi-scenario load forecasting as proposed in the above embodiments.

[0186] The storage medium proposed in this embodiment belongs to the same inventive concept as the verification method for implementing random production simulation and multi-scenario load prediction proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0187] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, 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 a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0188] Example 2

[0189] Refer to Table 1, Figure 2 As an embodiment of the present invention, a verification method for stochastic production simulation and multi-scenario load forecasting is provided to verify and illustrate the technical effects adopted in this method.

[0190] Table 1 shows the increase or decrease in power generation of generating units in different scenarios.

[0191] Increased power generation on day 3 (MWh) Day 4 power generation reduction (MWh) Scene 5 22 Scene 9 23 Scene 10 26 Scene 12 24

[0192] Increased power generation on day 3: On day 3, the generating units in scenario 5 increased their power generation by 22 MWh, and the generating units in scenario 9 increased their power generation by 23 MWh. This indicates that the predicted electricity demand was high, requiring additional power generation to meet it. By increasing power generation, the system can ensure a balance between power supply and demand, avoid power shortages, and thus guarantee the stable operation of the power system.

[0193] Day 4 Power Generation Reduction: On Day 4, the generating units in Scenario 10 reduced their power generation by 26 MWh, and the generating units in Scenario 12 reduced their power generation by 24 MWh. This indicates that the electricity demand on Day 4 was lower than predicted, and the system needed to reduce power generation to avoid overcapacity. By reducing power generation, the system can adjust its supply level to match demand, improve the operating efficiency of the power system, and reduce energy waste.

[0194] By increasing and decreasing generator output on days 3 and 4 respectively, the system can better respond to fluctuations in electricity demand. This flexibility avoids data truncation, which prevents accurate demand forecasts from being met in cases of insufficient or excessive power supply. Therefore, by adjusting generator output, power supply and demand can be matched more accurately, improving the accuracy of reliability assessment calculations.

[0195] Because the system can adjust power generation according to actual needs, the data distribution more closely reflects reality. This further improves the sampling efficiency of system states, meaning that when assessing power system reliability, it can more accurately simulate and sample various power generation scenarios, helping to more precisely assess the system's reliability level. By avoiding data truncation and improving sampling efficiency, the system's calculation speed for power system reliability assessment is effectively improved. Since the data more accurately reflects reality, the simulation and sampling processes in the calculation are more effective and efficient. This means that reliability assessments can be conducted faster, enabling decision-makers to obtain assessment results promptly and take appropriate measures.

[0196] like Figure 2 As shown, P r For the power generation node at time r, v ct v r v co The figure shows the voltage at ct, r, and co. As can be seen from the graph, the voltage level increases accordingly with time, and the rate of increase is constant. This indicates that the proposed monthly power deviation correction model can adjust the power plan for different dispatch scenarios. This model can flexibly adjust the power plan according to actual conditions and changes in demand to ensure the stability and reliability of the power system. Especially when facing the random changes in the output of new energy sources such as wind, solar, and hydropower, this model can effectively adapt and adjust the power plan to meet actual demand, helping to improve the reliability of the power system, optimize energy utilization, and ensure the sustainability of power supply.

[0197] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method of checking for random production simulation and multi-scenario load forecasting, characterized by, The method comprises the following steps: obtaining load prediction influencing factors of a selected to-be-predicted area, sorting the load prediction influencing factors according to the influence degree of the load prediction influencing factors, obtaining historical data of the load prediction influencing factors according to the sorted load prediction influencing factors, selecting a load prediction model of the to-be-predicted area by using the historical data of the load prediction influencing factors, and predicting the power load of the to-be-predicted area to obtain a first predicted load value; judging the load level index of the first predicted load value and the total number of load levels, wherein the judgment comprises that the load level index of the first predicted load value is less than the total number of load levels, and a power generation system reliability index is obtained; according to the power generation system reliability index, adjusting different scene loads by using a monthly power quantity deviation correction model.

2. The method of claim 1, wherein the method further comprises: The method comprises the following steps: obtaining load prediction influencing factors of a selected to-be-predicted area, sorting the load prediction influencing factors according to the influence degree of the load prediction influencing factors, obtaining historical data of the load prediction influencing factors according to the sorted load prediction influencing factors, selecting a load prediction model of the to-be-predicted area by using the historical data of the load prediction influencing factors, and predicting the power load of the to-be-predicted area to obtain a first predicted load value; The load prediction under the influence of each factor is represented as: X = x0 + (1 + a1) x1 + (1 + a2) x2 + (1 + a3) x3 + (1 + a4) x4 wherein X is the load prediction under the influence of each factor, x0 is a basic load, x1, x2, x3 and x4 are load components of weather, environment, holidays and remaining influencing factors, and a1, a2, a3 and a4 are influence amplitudes of the weather, environment, holidays and remaining influencing factors; initial values are assigned to the influence amplitudes of each influencing factor, the influence amplitudes of each factor are obtained by using the least square method according to the basic load and the load components of each factor, and the relative difference between each time value of each factor and the value of the last time is within a threshold value; 3. The method of claim 2, wherein the method further comprises: according to the obtained influence amplitudes of each factor, each influencing factor is sorted. The method comprises the following steps: selecting a load prediction model of the to-be-predicted area by using the historical data of the load prediction influencing factors, wherein the method comprises the following steps: if the historical data is greater than a first threshold value, there is a nonlinear relationship between the change of the historical data and the input variable, and the deviation tolerance between the prediction result and the actual observation value is less than a second threshold value, a neural network is selected as the load prediction model; a time sequence {Xi, i = n, n+1, …, n+m} is set; 4. The method of claim 2, wherein the method further comprises: wherein the historical data Xn, Xn+1, …, Xn+m is used to predict the value at the future n+m+k (k >= 1) time; a function f is fitted by using the neural network through a group of data points Xn, Xn+1, …, Xn+m, which is represented as: Xn+m+k = f(Xn, Xn+1, …, Xn+m), and the predicted value of the data at the future n+m+k (k >= 1) time is obtained. The method further comprises the following steps: where X t is the actual value for period t, is the predicted value for period t+T, t is the current period number, and T is the number of periods from t to the predicted period, is the moving average for the first stage, is the moving average for the second stage, a t is the intercept parameter, b t is the slope parameter, and N is the size of the moving average window.

5. The method of claim 3 or 4, wherein the method further comprises: if the historical data is less than the first threshold value, and the deviation tolerance between the prediction result and the actual observation value is greater than the second threshold value, a time sequence is selected as the load prediction model; the historical data is arranged into a time sequence according to the time sequence to obtain the change trend of the historical data with time, and a quantitative prediction is performed by using a twice moving average method, and a linear model of the twice moving average method is represented as: the load level index of the first predicted load value is less than the total number of load levels, and a power generation system reliability index is obtained, wherein the method comprises the following steps: let the system generator state sequence index wc = 0, and the load level index wk = 0; The system generator state sequence index wc and the load level index wk are increased step by step, that is, wk = wk + 1, wc = wc + 1; The system generator state sequence in the system generator state matrix is selected, and the corresponding load level is selected to obtain the corresponding load probability; The index function of the system reliability index under the first wk load level is updated, and is expressed as: FLOLP(WXw) = WW(WXwc) x WPwk FEENS(WXwc) = 8760 x SD x WW(WXwc) x WPwk Where, SD is the load shedding amount, WXwc is the first wc sequence in the generator state matrix WX, WPwk is the load probability, and FLOLP(WXwc) and FEENS(WXwc) are index functions. It is judged whether the load level index wk is less than the total number of load levels Nw, and if the judgment condition is met, the system reliability index and the variance coefficient under the multi-level load level are updated; if the judgment condition is not met, wc and wk are increased step by step. If the obtained variance coefficient meets the convergence condition, the power generation system reliability index is output, otherwise, the judgment operation is continued.

6. The method of claim 5, wherein the random production simulation and multi-scenario load forecast verification method is characterized by, The monthly power deviation correction model is used to adjust the different scene loads according to the predicted power load, including: According to the corrected monthly generation plan, the over-contract power of the power plant is decomposed to the day, and each power plant obtains the day-ahead generation plan of the unit according to the decomposed next-day contract power and the load demand prediction; A day-ahead optimal scheduling model is established to adjust the unit output with the minimum unit output adjustment amount as the objective function, and is expressed as: wherein, T is the optimized time range, N is the number of units, λ i is the weight coefficient of the i th unit, P i,t is the output of the thermal power unit i at the t period, is the pre-decomposition power of unit i at the t period of the next day, λ w is the weight coefficient of the w th unit, is the abandoned wind of the wind farm w at the t period, is the abandoned light of the photovoltaic power station p at the t period, ΔP h,t is the abandoned water of the hydropower unit h at the t period, N w is the number of load levels, λ pv is the weight coefficient of the photovoltaic power generation system, λ h is the weight coefficient of the wind power generation system, N pv is the number of photovoltaic power generation systems, N h is the number of wind power generation systems.

7. The method of claim 6, wherein the method further comprises: generating a plurality of random production scenarios; and generating a plurality of random load scenarios. The monthly power deviation correction model is used to adjust the different scene loads, including: A monthly power deviation correction model is established with the minimum adjustment cost of the monthly power deviation in the system as the objective function, and is expressed as: where C is the scenario number, w c is the c-scenario weight, are the unit i's up and down capacity offer, respectively, are the up and down capacity allocated to unit i in the c-scenario.

8. A system for checking methods of stochastic production simulation and multi-scenario load forecasting, characterized in that, Including: The model selection module is used to obtain the selected load prediction influencing factors of the to-be-predicted area, sort the load prediction influencing factors according to the influence degree, obtain the historical data of the load prediction influencing factors according to the sorted load prediction influencing factors, select the load prediction model of the to-be-predicted area by using the historical data of the load prediction influencing factors, and predict the power load of the to-be-predicted area to obtain a first predicted load value; The judgment module is used to judge the load level index of the first predicted load value and the total number of load levels, and the judgment includes that the load level index of the first predicted load value is less than the total number of load levels, and the power generation system reliability index is obtained. The adjustment module is used to adjust the different scene loads by using the monthly power deviation correction model according to the power generation system reliability index.

9. An electronic device, comprising: The device includes: A processor; A memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method of any one of claims 1-7.

10. A computer-readable storage medium having stored thereon computer program instructions, wherein, The computer program instructions are executed by the processor to implement the method of any one of claims 1-7. The computer program instructions are executed by the processor to implement the method of any one of claims 1-7.

Images