A multi-type power supply collaborative correction control method using prediction information

Abstract

The application discloses a multi-type power supply collaborative correction control method using prediction information and relates to the technical field of power supply control.The method comprises the following steps: collecting real-time power supply data, pre-processing the data and collecting long and short term data sets; inputting the long and short term data sets into a prediction model to output short term power supply power prediction values; obtaining power supply preliminary correction power adjustment values through a nonlinear mapping function; screening adjustment values meeting multi-power supply comprehensive optimization goals to generate power supply final correction strategies; executing the power supply final correction strategies, collecting and processing feedback information in the execution process; dynamically updating short term prediction results and power adjustment values to form a short term real-time adjustment mechanism; and formulating a long term global scheduling scheme to form a multi-type power supply collaborative correction closed loop control.The application solves the real-time feedback and nonlinear adjustment problems in multi-power supply collaborative scheduling, can dynamically adjust based on real-time data, optimizes power supply power output, and balances system economy and stability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power supply control, in particular to a multi-type power supply collaborative correction control method using prediction information. BACKGROUND

[0002] With the increasing demand for renewable energy worldwide, the power system is gradually transitioning from traditional fossil fuel power generation to clean energy such as wind and solar power. At the same time, the volatility of power grid load and the instability of renewable energy generation have brought great challenges to the scheduling and management of the power system. In order to achieve efficient operation and stability of the power grid, the use of multi-type power supply collaborative scheduling method has become an important research direction. Traditional power scheduling methods rely on real-time monitoring and fixed scheduling strategies, usually only considering the power output adjustment of a single power source, ignoring the coordination and comprehensive optimization between power sources. However, with the continuous expansion of the power grid and the access of multi-type power sources, how to achieve efficient collaborative scheduling of different power sources while ensuring system stability has become the focus of current research.

[0003] The existing multi-power supply collaborative control technology still has some obvious shortcomings in practical application. First, most traditional scheduling methods are based on historical data for static optimization, lacking dynamic processing capability for real-time feedback information, resulting in scheduling results lagging behind actual demand changes. Second, although the access of multi-type power sources provides more adjustment means for the system, existing collaborative control methods often fail to fully consider the characteristics of each type of power source and their nonlinear relationship with each other. Specifically, the power output of renewable energy sources such as wind and solar power is greatly affected by weather conditions, while energy storage devices can be quickly adjusted through charging and discharging. Existing technical methods are difficult to accurately handle the coordination of various types of power sources, resulting in slow system response speed and inability to achieve optimal scheduling without increasing costs. In addition, in traditional methods, most optimization models rely only on direct adjustment of power, ignoring system economics and long-term operating cost balance. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a multi-type power supply collaborative correction control method using prediction information to solve the problem that most optimization models rely only on direct adjustment of power, ignoring system economics and long-term operating cost balance.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a multi-type power supply collaborative correction control method using prediction information, which includes collecting real-time power supply data, preprocessing the data and summarizing it into long and short-term data sets;

[0008] inputting the long-term and short-term data set into the prediction model to output short-term power supply power prediction values; mapping the similarity between the short-term power supply power prediction values through a nonlinear mapping function to obtain power preliminary correction power adjustment values; screening adjustment values meeting a multi-power supply comprehensive optimization goal based on the power preliminary correction power adjustment values in combination with costs to generate a power final correction strategy; executing the power final correction strategy, collecting and processing feedback information in the execution process; inputting the feedback information into the prediction model to dynamically update the short-term prediction results and the power adjustment values to form a short-term real-time adjustment mechanism; formulating a long-term global scheduling scheme, executing the global scheduling scheme through the short-term real-time adjustment mechanism to form a multi-type power supply collaborative correction closed-loop control.

[0009] As a preferred scheme of the multi-type power supply collaborative correction control method using prediction information, the specific steps of collecting real-time power supply data, pre-processing the data and collecting the data into long-term and short-term data sets are as follows:

[0010] The real-time power supply data is collected through a sensor network, including the output power of the multi-type power supply, load demand data and environmental variables.

[0011] The output power of the multi-type power supply includes the charge and discharge power of the energy storage system, photovoltaic power and wind power.

[0012] The load demand data refers to the power demand of users in the power grid.

[0013] The environmental variables include wind speed, light intensity and temperature.

[0014] The collected real-time power supply data is subjected to denoising processing, outlier rejection and missing value filling.

[0015] The real-time power supply data is divided into long-term and short-term data sets according to the time period.

[0016] As a preferred scheme of the multi-type power supply collaborative correction control method using prediction information, the specific steps of inputting the long-term and short-term data set into the prediction model to output short-term power supply power prediction values are as follows:

[0017] Data features related to the current power supply state are extracted from the long-term and short-term data sets.

[0018] The short-term power supply power prediction values are calculated according to the extracted data features.

[0019] The expression of the short-term power supply power prediction model is as follows:

[0020] P(t)=W·h t +d;

[0021] where t is the time index, P((t) is the short-term power prediction value of the power source at time point t, h t is the data feature related to the power source state at time point t, W is the weight vector of the data feature, and d is the bias term constant.

[0022] As a preferred scheme of the multi-type power source collaborative correction control method using prediction information, the similarity between short-term power prediction values is mapped by a nonlinear mapping function to obtain a preliminary correction power adjustment value of the power source, and the specific steps are as follows:

[0023] The Gaussian kernel function is selected as the nonlinear mapping method, and the expression is as follows:

[0024]

[0025] where K(P i ,P j ) is the similarity between sample points P i and P j , P i is the short-term power prediction value of sample point i, P j is the short-term power prediction value of sample point j, ||P i -P j || 2 is the Euclidean distance between sample points P i and P j , σ is the kernel width parameter, and i and j are the indices of two different sample points.

[0026] The similarity mapping value is calculated, and the expression is as follows:

[0027]

[0028] where P n (t) is the predicted power of the power source n at time point t, P m (t) is the predicted power of the power source m at time point t, N is the total number of power sources, m is the power source index coefficient, K n (t) is the similarity mapping value of the power source P n (t) and other power sources, K(P n (t), P m (t)) is the similarity between the power source P n (t) and the power source P m (t) at time point t.

[0029] The similarity mapping value K n (t) is used to perform weighted average on the power of different power sources to obtain the preliminary correction power value of the power source, and the expression is as follows:

[0030]

[0031] wherein, is the preliminary correction power value of power supply n at time t, a m is the similarity mapping value weight coefficient, β n is the bias term coefficient of power supply n.

[0032] As a preferred scheme of the multi-type power supply collaborative correction control method using prediction information, wherein: based on the power supply preliminary correction power adjustment value, combined with the cost, the adjustment value meeting the multi-power supply comprehensive optimization goal is screened out to generate the power supply final correction strategy, and the specific steps are,

[0033] Based on the power supply preliminary correction power adjustment value, combined with the power generation cost required by the power supply in the adjustment process, the total correction power adjustment cost is calculated, and the expression is:

[0034]

[0035] wherein, C co is the total correction power adjustment cost, γ m is the cost coefficient of the mth power supply;

[0036] Based on the resistance power loss generated by the current in the flow process, the total network loss cost is calculated, and the expression is:

[0037]

[0038] wherein, L is the total network loss cost, δ is the unit network loss cost coefficient, k is the index of the transmission line, R k is the resistance of the kth transmission line, I k is the current of the kth transmission line;

[0039] Based on the economic loss caused by wear and tear and aging of the equipment in the running process, the total life loss cost is calculated, and the expression is:

[0040]

[0041] wherein, m is the power supply index coefficient, N is the total number of power supplies, T is the total life loss cost, κ m is the economic loss coefficient of the mth power supply, is the preliminary correction power adjustment value of power supply n at time t;

[0042] The total correction power adjustment cost, the total network loss cost and the total life loss cost are weighted and summed to obtain the comprehensive cost, and the multi-power supply comprehensive optimization goal meeting the minimum comprehensive cost is formulated;

[0043] The power supply preliminary correction power adjustment value is input into the comprehensive cost, the costs are compared, and the costs meeting the comprehensive optimization target are screened out;

[0044] Based on the cost meeting the comprehensive optimization target, the power supply final correction strategy is generated in combination with the power supply preliminary correction power adjustment value.

[0045] As a preferred scheme of the multi-type power supply collaborative correction control method using prediction information, the power supply final correction strategy is executed, feedback information in the execution process is collected and processed, and the specific steps are as follows,

[0046] Based on the power supply preliminary correction power adjustment value in the power supply final correction strategy, the correction instruction of the power supply equipment is obtained;

[0047] The correction instruction is issued to the device control unit through the power dispatching communication network;

[0048] After the device control unit receives the instruction, the power output is adjusted in real time;

[0049] The power sensor at the output end of the power supply equipment collects the actual output power and actual load demand in the execution process;

[0050] The collected feedback data is uploaded to the central control unit through the optical fiber network.

[0051] As a preferred scheme of the multi-type power supply collaborative correction control method using prediction information, the feedback information is input into the short-term prediction model, the short-term prediction result and the power adjustment value are dynamically updated, the short-term real-time adjustment mechanism is formed, and the specific steps are as follows,

[0052] The central control unit inputs the received feedback information into the short-term power supply power prediction model, and dynamically updates the short-term prediction result;

[0053] Based on the dynamically updated short-term prediction result, the power supply correction power is dynamically adjusted, and a new correction strategy is continuously screened out;

[0054] The new correction strategy is issued and executed through the communication network, the feedback information is updated, the next round of optimization is entered, and the short-term real-time control mechanism is formed.

[0055] As a preferred scheme of the multi-type power supply collaborative correction control method using prediction information, the long-term global scheduling scheme is formulated, the global scheduling scheme is executed through the short-term real-time adjustment mechanism, the multi-type power supply collaborative correction closed-loop control is formed, and the specific steps are as follows,

[0056] The long-term load trend prediction value is obtained from the long-term and short-term data set, and the long-term trend curve of the load is drawn;

[0057] Find the maximum load trend forecast value from the long-term load trend curve as the peak load and the minimum load trend forecast value as the valley load;

[0058] Load balance is obtained based on the difference between power generation and peak load trend forecast and valley load trend forecast within the same time period;

[0059] Based on the load balance between peak and valley loads, an optimization objective is obtained, and a long-term global scheduling scheme is formulated.

[0060] The global scheduling scheme is executed through the power communication network, and long-term forecast data is dynamically corrected to form a closed-loop control for collaborative correction of multiple types of power sources.

[0061] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of a multi-type power supply cooperative correction control method utilizing predictive information as described in the first aspect of the present invention.

[0062] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the multi-type power supply cooperative correction control method utilizing predictive information as described in the first aspect of the present invention.

[0063] The beneficial effects of this invention are as follows: By introducing predictive information and combining it with a collaborative correction control strategy for multiple types of power sources, this invention solves the problems of low dynamic adjustment accuracy and insufficient utilization of predictive information in existing technologies, thereby improving the dynamic response capability and operating efficiency of the power grid. This invention fully utilizes short-term predictive information from renewable energy generation to estimate power fluctuation trends, reducing the risk of grid imbalance caused by renewable energy fluctuations. By designing a collaborative correction mechanism for multiple types of power sources, this invention can comprehensively consider the dynamic characteristics of traditional power sources, energy storage systems, and renewable energy generation, adjusting power output according to real-time load demand to achieve efficient collaborative regulation. Finally, this invention introduces a dynamic feedback correction strategy, which can promptly correct the control scheme based on predictive deviations, reducing resource waste that may occur in fixed-rule control methods and improving power utilization efficiency and adjustment accuracy. Attached Figure Description

[0064] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0065] Figure 1 A flow chart of a multi-type power source collaborative correction control method using prediction information in Example 1.

[0066] Figure 2 A flow chart of a multi-type power source collaborative correction control method using prediction information in Example 1. DETAILED DESCRIPTION

[0067] In order to make the above objectives, features and advantages of the present application more apparent, specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0068] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. However, it will be apparent to one skilled in the art that the present application can be practiced without the specific details given below. In other instances, well-known methods have not been described in detail in order to avoid unnecessarily obscuring the present application.

[0069] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. The "in one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments.

[0070] Example 1, with reference to Figure 1 and Figure 2 , the first embodiment of the present application provides a multi-type power source collaborative correction control method using prediction information, comprising the following steps:

[0071] S1: Collect real-time power data, pre-process the data and aggregate into a long-short term data set.

[0072] Specifically, it includes the following steps:

[0073] S1.1: Collect real-time power data through a sensor network, including output power of multi-type power sources, load demand data and environmental variables.

[0074] S1.1.1: The output power of the multi-type power source includes the charge and discharge power of the energy storage system, photovoltaic power and wind power.

[0075] Specifically, the battery charge and discharge power is collected, the sampling interval is 5 seconds, and the data format is timestamp + power value.

[0076] The current DC input power and AC output power are obtained from the photovoltaic inverter or the data acquisition terminal (DAS), the sampling interval is 1 minute, and the average value is recorded.

[0077] The wind turbine output power is recorded, and the data of synchronous speed and blade angle is recorded with a sampling interval of 10 seconds.

[0078] S1.1.2: The load demand data refers to the power demand of users in the power grid.

[0079] Specifically, the intelligent electric meter is used to collect the user's power load demand data, and the total regional power load is obtained every 15 minutes. The dynamic power demand change data of industrial users is collected, and the real-time power demand of production equipment is recorded by minute. The hourly power consumption record is obtained through the user management system of the power company as indirect load data.

[0080] S1.1.3: The environmental variables include wind speed, light intensity and temperature.

[0081] Specifically, the ultrasonic anemometer is used to collect wind speed and direction data, with a recording frequency of every second, and the 10-minute average value is calculated. The light intensity is recorded every minute through the light sensor. The distributed temperature sensor network is used to record the environmental temperature, with a sampling interval of 5 minutes.

[0082] S1.2: The collected real-time power data is denoised, outliers are removed and missing values are filled.

[0083] Among them, the denoising process refers to using wavelet filtering method to remove high-frequency noise for all data, and using local regression smoothing algorithm to remove abnormal mutation points for photovoltaic power and wind speed data.

[0084] Outlier removal refers to using the 3σ principle to identify outliers, calculating the mean and standard deviation for each data, and removing data points outside the range of [mean ± 3σ]. For load demand data, use rule setting (such as negative load, abnormal high power) to remove data that does not conform to physical meaning.

[0085] Missing value filling refers to using time interpolation method to fill short time data missing (such as 5 minutes). The weighted average method based on historical similar time period is used to predict and fill long time data missing (such as more than 1 hour).

[0086] S1.3: According to the time period, the real-time power data is divided into long and short term data sets.

[0087] It should be understood that the time period is divided into short term and long term, and is divided into short term data set and long term data set.

[0088] Short term data set: time span is from hours to 1 day, suitable for real-time scheduling and short-term prediction, containing hourly average value and fluctuation range of energy storage power, photovoltaic power, wind power and load demand.

[0089] Long-term data set: Time span from 1 day to 1 month, suitable for trend analysis and long-term planning, containing daily average, maximum, minimum and cumulative statistics of each variable.

[0090] Preferably, accurate and reliable data support is provided for the coordinated scheduling of multiple types of power sources through precise data collection and processing. Real-time power data is collected through a sensor network, including the power output of energy storage, photovoltaic, wind power and other power sources, as well as load demand and environmental variables, ensuring comprehensive monitoring of the power system. Through preprocessing methods such as denoising, outlier removal and missing value filling, the accuracy and integrity of the data are guaranteed, providing a high-quality data foundation for subsequent prediction and optimization. The data is divided into short-term and long-term data sets according to the time period, with the short-term data set suitable for real-time scheduling and short-term prediction, and the long-term data set supporting trend analysis and long-term planning, thereby improving the intelligent management and operation efficiency of the power system and ensuring the flexibility and reliability of power scheduling.

[0091] S2: Input the long-term and short-term data set into the prediction model, output the short-term power prediction value.

[0092] Specifically, the following steps are included:

[0093] S2.1: Extract data features related to the current power state from the long-term and short-term data set.

[0094] S2.2: Calculate the short-term power prediction value based on the extracted data features.

[0095] The expression of the short-term power prediction model is:

[0096] P(t) = W h + d t ;

[0097] Where t is the time index, P((t) is the short-term power prediction value at time point t, h t is the data feature related to the power state at time point t, W is the weight vector of the data feature, and d is the bias term constant.

[0098] Preferably, through comprehensive collection of real-time data of multiple types of power sources (energy storage, photovoltaic, wind power), load demand and environmental variables, a multi-dimensional data framework is constructed, and the data quality and reliability are improved through denoising, outlier removal and missing value filling. The division of short-term and long-term data sets meets the different application requirements of real-time scheduling, short-term prediction and long-term planning, providing a solid foundation for accurate analysis, prediction modeling and operation optimization of the power system. Overall, this step covers the dynamic characteristics and long-term trend characteristics of the power system data, effectively supporting the improvement of power grid operation efficiency and stability.

[0099] S3: mapping the similarity between short-term power predictions by a nonlinear mapping function to obtain preliminary power correction power adjustment values.

[0100] Specifically, the method comprises the following steps:

[0101] S3.1: selecting a Gaussian kernel function as the nonlinear mapping method, and the expression is:

[0102]

[0103] wherein, K(P i ,P j ) is the similarity between sample points P i and P j , P i is the short-term power prediction value of sample point i, P j is the short-term power prediction value of sample point j, ||P i -P j || 2 is the Euclidean distance between sample points P i and P j , and σ is the kernel width parameter, and i and j are the indices of two different sample points.

[0104] It should be understood that when P i and P j are closer, K(P i ,P j ) tends to 1, and when P i and P j are farther apart, K(P i ,P j ) tends to 0. The kernel width parameter σ determines the sensitivity of the similarity value, and a smaller σ will make the Gaussian kernel function pay more attention to the similarity of the local neighborhood, and a larger σ will pay more attention to the overall trend.

[0105] S3.2: calculating the similarity mapping value, and the expression is:

[0106]

[0107] wherein, P n (t) is the predicted power of power source n at time point t, P m (t) is the predicted power of power source m at time point t, N is the total number of power sources, m is the power source index coefficient, K n (t) is the similarity mapping value of power source P n (t) and other power sources, K(P n (t), P m (t)) is the similarity mapping value of power source P n (t) and power source P m(t) similarity between time points t.

[0108] S3.3: similarity mapping value K n (t) is weighted average of power of different power sources, to get preliminary correction power value of power source, expression is:

[0109]

[0110] wherein, is preliminary correction power value of power source n at time t, α m is similarity mapping value weight coefficient, β n is bias term coefficient of power source n.

[0111] It should be understood that weight coefficient α m is usually optimized through data training or validation set, to ensure rationality of weighted average. Bias term coefficient β n is adjusted according to deviation of historical actual power and predicted power.

[0112] Preferably, similarity between short-term power source power prediction values is nonlinearly mapped through Gaussian kernel function, to capture correlation between power prediction values, and preliminary correction power value is generated by using similarity mapping value to correct power prediction. This method can effectively improve prediction accuracy, fully consider synergistic relationship between power sources, reduce influence of single prediction error, and at the same time, by adjusting kernel width σ and optimizing weight coefficient α m , adaptability and robustness of model are enhanced. Overall, above process short-term power prediction, to provide more reliable data support for power grid dispatching and operation.

[0113] S4: based on power preliminary correction power adjustment value, combined with cost, to screen adjustment value meeting multi-power source comprehensive optimization target, to generate power final correction strategy.

[0114] Specifically, including following steps:

[0115] S4.1: based on power preliminary correction power adjustment value, combined with power generation cost required by power source in adjustment process, to calculate total correction power adjustment cost, expression is:

[0116]

[0117] wherein, C co is total correction power adjustment cost, γ m is cost coefficient of mth power source.

[0118] It should be understood that cost coefficient γ mThe cost coefficients of different power sources should be set according to actual conditions, for example, the γ of renewable energy such as wind power and photovoltaic power is relatively low. The charge and discharge cost of energy storage system is relatively high. The γ of traditional energy such as thermal power needs to consider fuel cost and carbon emission penalty. m m The cost coefficients of different power sources should be set according to actual conditions, for example, the γ of renewable energy such as wind power and photovoltaic power is relatively low. The charge and discharge cost of energy storage system is relatively high. The γ of traditional energy such as thermal power needs to consider fuel cost and carbon emission penalty.

[0119] S4.2: Calculate the total network loss cost based on the resistance power loss generated by the current during flow, the expression is:

[0120]

[0121] Where, L is the total network loss cost, δ is the unit network loss cost coefficient, k is the index of transmission line, R k is the resistance of the kth transmission line, I k is the current of the kth transmission line.

[0122] It should be understood that the unit network loss cost coefficient δ is set according to the network loss cost (such as electricity price or transmission efficiency penalty).

[0123] S4.3: Calculate the total life loss cost based on the economic loss caused by wear and tear and aging of equipment during operation, the expression is:

[0124]

[0125] Where, m is the power index coefficient, N is the total number of power sources, T is the total life loss cost, κ m is the economic loss coefficient of the mth power source, is the preliminary correction power adjustment value of power source n at time t.

[0126] It should be understood that the loss cost of different power equipment due to wear and tear and aging is different, for example, the number of charge and discharge cycles of energy storage equipment determines its life loss, the start-stop frequency of thermal power equipment and the operation time of photovoltaic inverter affect its economic loss. κ m It should be evaluated through historical operation data and equipment maintenance cost.

[0127] S4.4: Weighted sum of total correction power adjustment cost, total network loss cost and total life loss cost to get comprehensive cost, and formulate multi-power comprehensive optimization goal to minimize comprehensive cost.

[0128] Specifically, the comprehensive cost expression is:

[0129] C total = w1·C co +w2·L+w3·T;

[0130] Where, C total ​For comprehensive cost, w1 is the weight coefficient of correcting power adjustment cost, w2 is the weight coefficient of network loss cost, and w3 is the weight coefficient of life consumption cost, and w1+w2+w3=1 should be met.

[0131] According to the specific optimization target, the weights of w1, w2 and w3 are reasonably distributed: if the power generation economy is more important, w1 is increased; if the power transmission efficiency is more important, w2 is increased; and if the equipment life is more important, w3 is increased.

[0132] S4.5: The power preliminary correction power adjustment value is input into the comprehensive cost, the costs are compared, and the cost meeting the comprehensive optimization target is screened out.

[0133] Specifically, the preliminary correction power adjustment value is brought into the comprehensive cost formula, and the C total of all schemes is calculated.

[0134] S4.6: Based on the cost meeting the comprehensive optimization target, the power final correction strategy is generated in combination with the power preliminary correction power adjustment value.

[0135] Specifically, according to the optimal comprehensive cost scheme screened out, the correction power value of each power source is determined. The final correction power value of each power source and the corresponding power generation plan are output. The actual output power of the power source is adjusted to meet the correction strategy. If there is a deviation in real-time operation, the correction strategy is dynamically adjusted in combination with real-time data.

[0136] Preferably, through comprehensive quantitative analysis and optimization of power generation cost, power transmission loss and equipment life loss, the weighted comprehensive cost function realizes the dynamic balance of economy, efficiency and equipment reliability. By flexibly adjusting the weight coefficient to adapt to different operation requirements, the optimal correction strategy is screened out and dynamically adjusted in combination with real-time data, which not only reduces the total power generation and dispatching cost, but also prolongs the service life of the key equipment, improves the operation efficiency and stability of the power grid. Preferably, clean energy is used preferentially, carbon emissions are reduced, and the green and low-carbon development goal is met, which shows an intelligent, efficient and environmentally friendly multi-power comprehensive optimization solution.

[0137] S5: The power final correction strategy is executed, and feedback information in the execution process is collected and processed.

[0138] Specifically, the following steps are included:

[0139] S5.1: Based on the power preliminary correction power adjustment value in the power final correction strategy, the correction instruction of the power equipment is obtained.

[0140] Specifically, read the adjustment value in the power supply final correction strategy, including the power adjustment requirements of specific power supply equipment. Convert the adjustment requirements into device recognizable control instructions. Generate instruction content, including power supply number, adjustment target value, execution time window, and other key information. Combine with the current grid load demand, and perform multiple verifications on the instructions to ensure that the adjustment instructions do not affect system safety.

[0141] S5.2: Distribute the correction instructions to the device control unit through the power dispatching communication network.

[0142] Specifically, select the IEC 61850 communication network protocol, encapsulate the instructions as standard data packets, verify the availability of the communication network to ensure smooth communication link, transmit the instructions to each device control unit through the communication network, and monitor the sending status of the instructions in real time to ensure successful sending and record the instruction distribution log.

[0143] S5.3: After receiving the instructions, the device control unit adjusts the power output in real time.

[0144] Specifically, the control unit parses the received correction instructions and verifies their integrity and validity. According to the correction instructions, adjust the power output in real time, adjust the charge and discharge power of the energy storage device, and modify the power factor of the inverter.

[0145] S5.4: The power sensor at the output end of the power supply equipment collects the actual output power and actual load demand during the execution process.

[0146] S5.5: Upload the collected feedback data to the central control unit through the optical fiber network.

[0147] Specifically, establish a data receiving interface in the central control unit to receive and parse the uploaded feedback data. Update and store the received feedback data in real time to form a historical operation database for subsequent analysis and optimization.

[0148] Preferably, by accurately converting the power correction strategy into control instructions and issuing them through the IEC 61850 protocol, the strategy can efficiently adjust the output of the power supply equipment. Real-time monitoring and feedback mechanism ensures that the response of the power supply equipment meets the expectations, and the feedback data is uploaded to the central control unit through the optical fiber network, ensuring the adaptive ability of the strategy. Multiple verifications and real-time monitoring ensure the safety of the power grid, avoiding the impact on the stability of the power grid. Feedback data is stored as a historical database, supporting subsequent optimization and adjustment, improving the intelligent level of the strategy, and promoting the continuous improvement of power dispatching and management.

[0149] S6: Input the feedback information into the prediction model to dynamically update the short-term prediction results and power adjustment values, forming a short-term real-time adjustment mechanism.

[0150] Specifically, the following steps are included:

[0151] S6.1: The central control unit inputs the received feedback information into the short-term power supply power prediction model to dynamically update the short-term prediction results.

[0152] Specifically, the feedback data is checked for format, outliers and redundant data are removed, missing data is processed using time series filling method to ensure data integrity. Key features such as power supply actual power output, load demand, environmental variables are extracted, and input data is standardized to meet the input requirements of the prediction model. The processed feedback information is input into the short-term power supply power prediction model to dynamically calculate the short-term power prediction value, and the power prediction results of each power supply at the next time are generated. The dynamic prediction results are verified and compared with historical data and current feedback data; if the prediction deviation is large, the model parameters are adjusted to optimize the prediction performance.

[0153] S6.2: Based on the dynamically updated short-term prediction results, the power supply correction power is dynamically adjusted to continuously select new correction strategies.

[0154] Specifically, according to the updated prediction results, the power correction value of each power supply is recalculated, the newly calculated correction power adjustment value is input into the comprehensive cost model, the correction strategy is re-evaluated, the optimized power adjustment value is converted into power correction instructions, and the adjustment target value of each power supply device is determined.

[0155] S6.3: The new correction strategy is issued and executed through the communication network, the feedback information is updated, and the next round of optimization is entered to form a short-term real-time control mechanism.

[0156] Specifically, the correction strategy is issued to each device control unit through the communication network, the device control unit receives and executes the correction instructions in real time to adjust the power output. The execution effect of the device is monitored in real time, and the updated power output and load data are collected through the sensor network. The new feedback information is input into the prediction model again to enter the next round of short-term prediction and correction. The prediction results and correction strategies are continuously optimized to realize real-time dynamic adjustment.

[0157] Preferably, by dynamically receiving feedback information and optimizing the short-term prediction model, the prediction accuracy is improved, the system's flexible response capability to load changes and external disturbances is enhanced, a closed-loop optimization control mechanism is formed, and efficient cooperation of real-time data collection, prediction adjustment, strategy optimization and feedback update is realized. The application of the comprehensive cost model ensures the economy of the correction strategy, effectively reduces the operating cost, and at the same time, the execution effect of the device and the load condition are monitored in real time to ensure the safe and stable operation of the power grid. The overall process promotes the intelligent development of power supply power dispatching, improves the system operation efficiency and resource utilization rate, and provides certain technical support for modern power system management.

[0158] S7: Formulate a long-term global scheduling scheme, execute the global scheduling scheme through a short-term real-time adjustment mechanism, and form a multi-type power source collaborative correction closed-loop control.

[0159] Specifically, the following steps are included:

[0160] S7.1: Obtain a long-term load trend prediction value from the long-term and short-term data set, and draw a long-term trend curve of the load.

[0161] Specifically, the load trend (including the load demand of the power grid and the power output, etc.) is extracted from the long-term historical data. The time series analysis method is used to predict the future load demand. Based on the long-term load prediction data, a load trend curve is drawn to reflect the change law of the power grid load demand in each time period. This curve can be used to determine the peak period, valley period and stable period of the load to ensure accurate planning in subsequent steps.

[0162] S7.2: Find the maximum load trend prediction value from the long-term trend curve of the load as the peak load, and the minimum load trend prediction value as the valley load.

[0163] It should be noted that on the long-term load trend curve, the highest point of the load is identified as the peak load, and the lowest point is identified as the valley load. The peak load usually occurs during the peak power demand period, and the valley load occurs during the trough period of power demand.

[0164] S7.3: Based on the difference between the peak load trend prediction and the valley load trend prediction in the same time period, the load balance is obtained.

[0165] It should be understood that by analyzing the difference, it is evaluated whether the load can be balanced. If the difference is large, power scheduling or energy storage means may need to be taken to adjust the power output to maintain load balance. For periods with small differences, unnecessary power generation can be reduced to save costs.

[0166] S7.4: Based on the load balance of the peak load and the valley load, an optimization target is obtained, and a long-term global scheduling scheme is formulated.

[0167] Specifically, according to the result of the load balance, a global optimization target is set. The target includes maximizing energy utilization, minimizing cost, reducing emissions, and ensuring system stability, etc. Based on the load balance and the optimization target, a long-term global scheduling scheme is formulated. The scheme should include the output plan of each power source and how to adjust the power combination during the peak and trough periods of the load to ensure real-time response when demand fluctuates.

[0168] S7.5: Execute the global scheduling scheme through the power communication network, dynamically correct the long-term prediction data, and form a multi-type power source collaborative correction closed-loop control.

[0169] Specifically, by means of the power dispatching communication network, dispatching instructions are sent to each power supply device in a timely manner to ensure that each power supply device adjusts the power generation according to the predetermined plan. Real-time collection of device operating status, load demand and environmental data, dynamic correction of long-term prediction data, updating of load prediction model and optimization of power generation scheme. Through the sensor network, feedback data and control unit, real-time adjustment is realized, and the power supply coordination and correction strategy is continuously optimized to ensure stable operation of the power grid and load balance, reduce energy waste and system failure risk.

[0170] Preferably, through fine long-term load prediction, load balance analysis and dynamic scheduling execution, efficient management and optimal control of the power grid are realized. By extracting load trends from long-term data and drawing load trend curves, the peak and valley periods of the power grid can be accurately identified, providing a scientific basis for power dispatching. Analysis of load balance helps to determine whether the power grid is in a stable state, and timely adjustment measures are taken to maintain load balance and avoid system overload or deficiency. By dynamically updating load prediction data and developing a global dispatching scheme, power output can be adjusted during peak and off-peak load periods to maximize energy utilization, reduce costs, reduce emissions, and improve overall system stability. Through the power communication network, dispatching instructions are executed, real-time data collection and feedback are performed, and power supply devices are coordinated to effectively optimize power grid operation, reduce energy waste and reduce system failure risk, forming a closed-loop control to continuously improve the efficiency and reliability of the power grid.

[0171] The embodiment also provides a computer device suitable for the case of the multi-type power supply collaborative correction control method using prediction information, comprising: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize the multi-type power supply collaborative correction control method using prediction information as proposed in the above embodiment.

[0172] The computer device can be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected by a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved by WIFI, a carrier network, NFC (Near Field Communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, a trackball or a touchpad arranged on the shell of the computer device, or an external keyboard, a touchpad or a mouse, etc.

[0173] The embodiment also provides a storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the method for cooperative correction control of multiple types of power supplies by using prediction information as described in the above embodiment. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disk.

[0174] In summary, the present application solves the problems of low dynamic adjustment precision and insufficient use of prediction information in the prior art by introducing prediction information and combining the collaborative correction control strategy of multiple types of power sources, thereby improving the dynamic response capability and operation efficiency of the power grid. The present application makes full use of the short-term prediction information of new energy power generation to estimate the power fluctuation trend, thereby reducing the imbalance risk of the power grid caused by new energy fluctuation. By designing a collaborative correction mechanism of multiple types of power sources, the present application can comprehensively consider the dynamic characteristics of traditional power sources, energy storage systems and new energy power generation, adjust the power output according to real-time load demand, and realize efficient collaborative regulation. Finally, the present application introduces a dynamic feedback correction strategy, which can timely correct the control scheme according to the prediction deviation, reduce the resource waste problem that may be caused by the fixed rule control method, and improve the power utilization efficiency and adjustment precision.

[0175] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A multi-type power supply cooperative correction control method utilizing predictive information, characterized in that: include, Collect real-time power data, preprocess the data, and summarize it into a short-term dataset; Input the long-term and short-term datasets into the prediction model and output the short-term power prediction value; By using a nonlinear mapping function to map the similarity between short-term power prediction values, preliminary power adjustment values ​​for power supply correction are obtained. The specific steps are as follows: Choosing the Gaussian kernel function as the nonlinear mapping method, the expression is: Wherein K(P) i ,P j ) represents sample point P i and P j The similarity between them, P i Let P be the short-term power prediction value for sample point i. j Let |P| be the short-term power prediction value for sample point j. i -P j || 2 For sample point P i and P j The Euclidean distance between them, where σ is the kernel width parameter, and i and j are the indices of two different sample points; The similarity mapping value is calculated using the following expression: Among them, P n (t) represents the predicted power of power source n at time t, P m (t) represents the predicted power of power source m at time t, N is the total number of power sources, m is the power source index coefficient, and K... n (t) represents the power supply P n (t) Similarity mapping value with other power sources, K(P) n (t),P m (t) is the power supply P n (t) and power supply P m (t) Similarity between time points t; Using similarity mapping value K n (t), the weighted average of the power from different power sources is used to obtain the preliminary corrected power value of the power source, expressed as: in, Let α be the initial corrected power value of power supply n at time t. m β is the weighting coefficient for the similarity mapping value. n For the bias term coefficients of power supply n; Based on the initial power adjustment value of the power supply, and combined with cost, the adjustment value that meets the comprehensive optimization goal of multiple power supplies is selected, and the final power supply adjustment strategy is generated. Execute the final power supply calibration strategy, and collect and process feedback information during the execution process; Feedback information is input into the prediction model to dynamically update short-term prediction results and power adjustment values, forming a short-term real-time adjustment mechanism. A long-term global scheduling plan is formulated, and the global scheduling plan is executed through a short-term real-time adjustment mechanism to form a closed-loop control for collaborative correction of multiple types of power sources.

2. The multi-type power supply cooperative correction control method utilizing predictive information as described in claim 1, characterized in that: The specific steps for collecting real-time power data, preprocessing the data, and summarizing it into short-term and long-term datasets are as follows: Real-time power data is collected through sensor networks, including the output power of various types of power sources, load demand data, and environmental variables. The output power of the various types of power sources includes the charging and discharging power of the energy storage system, photovoltaic power, and wind power. The load demand data refers to the electricity demand of users in the power grid. The environmental variables include wind speed, light intensity, and temperature; The collected real-time power data is subjected to noise reduction, outlier removal, and missing value imputation. Based on the time period, real-time power data is divided into short-term and long-term data sets.

3. The multi-type power supply cooperative correction control method utilizing predictive information as described in claim 1, characterized in that: The specific steps for inputting long-term and short-term datasets into the prediction model and outputting short-term power prediction values ​​are as follows: Extract data features related to the current power state from long-term and short-term data sets; Based on the extracted data characteristics, calculate the short-term power prediction value; The short-term power prediction model expression is: P(t)=W·h t +d; Where t is the time index, P(t) is the short-term power prediction value at time t, and h t Let W be the power state-related data feature at time t, where W is the weight vector of the data feature and d is the bias term constant.

4. The multi-type power supply cooperative correction control method utilizing predictive information as described in claim 1, characterized in that: The process involves using the initial power adjustment value for power supply calibration, combined with cost considerations to select adjustment values ​​that meet the comprehensive optimization objectives of multiple power supplies, and generating the final power supply calibration strategy. The specific steps are as follows: Based on the initial power adjustment value of the power source and considering the generation cost required during the adjustment process, the total power adjustment cost is calculated as follows: Among them, C co For the total correction power adjustment cost, γ m Let m be the cost coefficient of the m-th power source; The total network loss cost is calculated based on the resistive power loss generated during the flow of current, expressed as: Where L is the total network loss cost, δ is the unit network loss cost coefficient, k is the transmission line index, and R... k Let I be the resistance of the k-th transmission line. k Let be the current of the k-th transmission line; The total lifespan loss cost is calculated based on the economic losses caused by wear and aging during equipment operation. The expression is as follows: Where m is the power supply index coefficient, N is the total number of power supplies, T is the total lifetime loss cost, and κ is the total power supply index coefficient. m Let m be the economic loss coefficient of the m-th power source. This is the initial power adjustment value for power supply n at time t; The total correction power adjustment cost, total network loss cost, and total lifetime loss cost are weighted and summed to obtain the comprehensive cost, and a multi-power source comprehensive optimization objective that minimizes the comprehensive cost is formulated. Input the initial power adjustment value of the power supply into the comprehensive cost, compare the costs, and select the cost that meets the comprehensive optimization goal. Based on the cost required to meet the overall optimization objective, and combined with the initial power adjustment value of the power supply, a final power supply correction strategy is generated.

5. The multi-type power supply cooperative correction control method utilizing predictive information as described in claim 1, characterized in that: The final power supply calibration strategy involves collecting and processing feedback information during the execution process. The specific steps are as follows: Based on the power adjustment value of the initial power correction in the final power correction strategy, the correction command of the power supply device is obtained. The correction command is sent to the equipment control unit through the power dispatch communication network; After receiving the instruction, the equipment control unit adjusts the power output in real time; The power sensor at the output end of the power supply equipment collects the actual output power and actual load demand of the power supply during the execution process; The collected feedback data is uploaded to the central control unit via a fiber optic network.

6. The multi-type power supply cooperative correction control method utilizing predictive information as described in claim 1, characterized in that: The specific steps for inputting feedback information into the short-term prediction model, dynamically updating the short-term prediction results and power adjustment values, and forming a short-term real-time adjustment mechanism are as follows: The central control unit inputs the received feedback information into the short-term power prediction model and dynamically updates the short-term prediction results. Based on dynamically updated short-term forecast results, the power supply correction power is dynamically adjusted, and new correction strategies are continuously screened out. The new correction strategy is distributed and executed through the communication network, the feedback information is updated, and the next round of optimization is initiated, forming a short-term real-time control mechanism.

7. The multi-type power supply cooperative correction control method utilizing predictive information as described in claim 1, characterized in that: The process involves formulating a long-term global scheduling scheme, executing the global scheduling scheme through a short-term real-time adjustment mechanism, and forming a multi-type power supply collaborative correction closed-loop control. The specific steps are as follows: Obtain long-term load trend forecasts from short- and long-term datasets and plot the long-term load trend curve. Find the maximum load trend forecast value from the long-term load trend curve as the peak load and the minimum load trend forecast value as the valley load; Load balance is obtained based on the difference between power generation and peak load trend forecast and valley load trend forecast within the same time period; Based on the load balance between peak and valley loads, an optimization objective is obtained, and a long-term global scheduling scheme is formulated. The global scheduling scheme is executed through the power communication network, and long-term forecast data is dynamically corrected to form a closed-loop control for collaborative correction of multiple types of power sources.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the multi-type power supply cooperative correction control method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the multi-type power supply cooperative correction control method according to any one of claims 1 to 7.

Images