A conventional power output data fitting method and system based on cloud transformation fitting

Abstract

The application discloses a conventional power output data fitting method and system based on cloud transformation fitting, and belongs to the technical field of power system prediction. The method comprises the following steps: acquiring conventional power output data of a conventional power source of a power system, performing statistics on the conventional power output data, and obtaining an output distribution curve of the conventional power output data; based on cloud transformation, fitting the conventional power output data according to the output distribution curve, and obtaining a distribution function of the output distribution curve; and taking the distribution function as a curve fitting equation for fitting the conventional power output data. The application has obvious advantages in improving prediction accuracy, saving resources, simplifying operation, promoting environmental protection and the like by fitting data based on cloud transformation.

Description

Technical Field

[0001] The present invention relates to the technical field of power system prediction, and more particularly to a conventional power output data fitting method and system based on cloud transformation fitting. Background Art

[0002] With the global emphasis on energy transition, the operational dispatch and economic efficiency of traditional power generation methods, such as thermal and hydropower, are receiving increasing attention. Medium- and long-term demand fluctuations, climate change, and water resource availability all impact power generation capacity. A deeper understanding of minimum output patterns can help power dispatchers more effectively plan power generation strategies and ensure safe and stable grid operation.

[0003] Furthermore, national policies regulating thermal and hydropower, as well as environmental protection pressures, have made optimizing their output characteristics particularly important. Modern data analysis techniques and models can more accurately capture the minimum output characteristics of generators under different operating conditions. This not only helps improve resource utilization efficiency but also provides a scientific basis for market pricing and electricity trading. Summary of the Invention

[0004] To address the above problems, the present invention proposes a conventional power output data fitting method based on cloud transformation fitting, comprising:

[0005] Obtaining conventional power output data of a conventional power source of the power system, performing statistics on the conventional power output data, and obtaining an output distribution curve of the conventional power output data;

[0006] Based on the cloud transformation, the conventional power supply output data is fitted according to the output distribution curve to obtain a distribution function of the output distribution curve;

[0007] The distribution function is used as a curve fitting equation for fitting conventional power output data.

[0008] Optionally, based on cloud transformation, fitting the conventional power supply output data according to the output distribution curve to obtain a distribution function of the distribution curve includes:

[0009] Based on the cloud transformation, a cloud model that fits the output distribution curve is generated according to the output distribution curve, and based on the cloud model, the residual data distribution is determined. Based on the residual data distribution, the cloud model distribution function is repeatedly constructed until all the peaks of the output distribution curve are fitted, and the distribution function of the distribution curve is generated.

[0010] Optionally, based on the cloud transformation, generating a cloud model distribution function that fits the output distribution curve according to the output distribution curve includes:

[0011] Based on cloud transformation, the output distribution curve is identified, and the peak position of each conventional power output data is determined. The peak position is used as the center of gravity position of the cloud. The entropy of the cloud model is calculated based on the center of gravity position. Based on the center of gravity position and entropy, the normal cloud model formula is applied to generate the cloud model.

[0012] Optionally, determining the residual data distribution based on the cloud model distribution function includes:

[0013] The data values ​​of the part of the output distribution curve that is fitted with the cloud model are removed, and the data values ​​of the remaining part of the output distribution curve are used as the residual distribution.

[0014] Optionally, based on the residual data distribution, repeatedly constructing a number of cloud models until all peaks of the output distribution curve are fitted, including:

[0015] Repeating the peak identification on the output distribution curve for the residual data distribution to construct a cloud model until all peaks are fitted;

[0016] After all peaks are fitted, multiple corresponding cloud model distribution functions are obtained.

[0017] Optionally, generate a distribution function for the distribution curve, including:

[0018] The distribution functions of the distribution curve are generated by superimposing and fitting multiple corresponding cloud model distribution functions.

[0019] Optionally, the method further includes: when generating a cloud model for fitting the output distribution curve, using an expected curve of the cloud model for fitting.

[0020] Optionally, when fitting the expected curve of the cloud model, determine the peak position, take the peak position as the center, screen out the first trough and the first peak on the left and right sides of the center respectively, calculate the difference between the first trough and the first peak, and compare the difference with a preset threshold. If the difference is greater than the preset threshold, record the value of the first peak. If the difference is less than the preset threshold, look for the next peak until the difference is greater than the preset threshold. Construct a value range of data points based on the recorded value of the first peak, and use the data points within the value range as cloud droplets. Based on the cloud droplets, use the inverse cloud algorithm to calculate the initial value of entropy.

[0021] On the other hand, the present invention also proposes a conventional power output data fitting system based on cloud transformation fitting, comprising:

[0022] a statistical unit, configured to obtain conventional power output data of a conventional power source of the power system, perform statistics on the conventional power output data, and obtain an output distribution curve of the conventional power output data;

[0023] a fitting unit, configured to fit the conventional power supply output data according to the output distribution curve based on cloud transformation to obtain a distribution function of the output distribution curve;

[0024] The output unit is used to use the distribution function as a curve fitting equation for fitting conventional power supply output data.

[0025] Optionally, based on cloud transformation, fitting the conventional power supply output data according to the output distribution curve to obtain a distribution function of the distribution curve includes:

[0026] Based on the cloud transformation, a cloud model that fits the output distribution curve is generated according to the output distribution curve, and based on the cloud model, the residual data distribution is determined. Based on the residual data distribution, the cloud model distribution function is repeatedly constructed until all the peaks of the output distribution curve are fitted, and the distribution function of the distribution curve is generated.

[0027] Optionally, based on the cloud transformation, generating a cloud model distribution function that fits the output distribution curve according to the output distribution curve includes:

[0028] Based on cloud transformation, the output distribution curve is identified, and the peak position of each conventional power output data is determined. The peak position is used as the center of gravity position of the cloud. The entropy of the cloud model is calculated based on the center of gravity position. Based on the center of gravity position and entropy, the normal cloud model formula is applied to generate the cloud model.

[0029] Optionally, determining the residual data distribution based on the cloud model distribution function includes:

[0030] The data values ​​of the part of the output distribution curve that is fitted with the cloud model are removed, and the data values ​​of the remaining part of the output distribution curve are used as the residual distribution.

[0031] Optionally, based on the residual data distribution, repeatedly constructing a number of cloud models until all peaks of the output distribution curve are fitted, including:

[0032] Repeating the peak identification on the output distribution curve for the residual data distribution to construct a cloud model until all peaks are fitted;

[0033] After all peaks are fitted, multiple corresponding cloud model distribution functions are obtained.

[0034] Optionally, generate a distribution function for the distribution curve, including:

[0035] The distribution functions of the distribution curve are generated by superimposing and fitting multiple corresponding cloud model distribution functions.

[0036] Optionally, when generating a cloud model for fitting the output distribution curve, an expected curve of the cloud model is used for fitting.

[0037] Optionally, when fitting the expected curve of the cloud model, determine the peak position, take the peak position as the center, screen out the first trough and the first peak on the left and right sides of the center respectively, calculate the difference between the first trough and the first peak, and compare the difference with a preset threshold. If the difference is greater than the preset threshold, record the value of the first peak. If the difference is less than the preset threshold, look for the next peak until the difference is greater than the preset threshold. Construct a value range of data points based on the recorded value of the first peak, and use the data points within the value range as cloud droplets. Based on the cloud droplets, use the inverse cloud algorithm to calculate the initial value of entropy.

[0038] In yet another aspect, the present invention further provides a computing device comprising: one or more processors;

[0039] a processor for executing one or more programs;

[0040] When the one or more programs are executed by the one or more processors, the above-described method is implemented.

[0041] In another aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed, the method described above is implemented.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention provides a method for fitting conventional power output data based on cloud transformation fitting, comprising: obtaining conventional power output data of conventional power sources in a power system, statistically analyzing the conventional power output data to obtain an output distribution curve for the conventional power output data; fitting the conventional power output data based on the output distribution curve based on cloud transformation to obtain a distribution function for the output distribution curve; and using the distribution function as a curve fitting equation for fitting the conventional power output data. This method, which uses cloud transformation to fit data, demonstrates significant advantages in improving prediction accuracy, saving resources, simplifying operations, and promoting environmental protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a flow chart of the method of the present invention;

[0045] Figure 2 This is a framework diagram of cloud transformation fitting conventional power output in the present invention;

[0046] Figure 3 Schematic diagram of correlation analysis of factors influencing conventional power output prediction in the present invention;

[0047] Figure 4 This is a polynomial fitting diagram of the conventional power output in the present invention;

[0048] Figure 5 This is a random forest fitting diagram of the conventional power output in the present invention;

[0049] Figure 6 This is a cloud transformation fitting diagram of conventional power output in the present invention;

[0050] Figure 7 It is a structural diagram of the system of the present invention. DETAILED DESCRIPTION

[0051] Exemplary embodiments of the present invention will now be described with reference to the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to provide a thorough and complete disclosure of the present invention and to fully convey the scope of the present invention to those skilled in the art. The terminology used in the exemplary embodiments shown in the accompanying drawings is not intended to limit the present invention. In the accompanying drawings, identical elements are denoted by the same reference numerals.

[0052] Unless otherwise specified, the terms used herein (including technical terms) have the meanings commonly understood by those skilled in the art. In addition, it is understood that terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0053] Example 1:

[0054] Accurate prediction of conventional power output in power systems has a direct impact on the dispatching plan of power systems. To improve prediction accuracy, scholars have extensively studied and proposed a variety of related models. However, these methods still have significant shortcomings when dealing with data with complex time-frequency characteristics, noise interference, and local mutations. Traditional fitting methods, such as polynomial regression and linear regression, usually assume that the data is stationary and are suitable for describing overall trends. However, power output data is often affected by various factors such as weather changes, load fluctuations, and equipment operating status, and exhibits significant non-stationarity and local fluctuations. Traditional methods based on the stationary assumption have difficulty accurately capturing these non-stationary characteristics, resulting in poor prediction of short-term fluctuations and emergencies, thereby increasing prediction errors.

[0055] On the other hand, traditional fitting methods, which mostly rely on global models, often struggle to accurately handle local fluctuations and transient changes in the data. Power output data often contains short-term, sudden fluctuations, such as load fluctuations or sudden changes in equipment output. Global models are generally unable to accurately capture these local features, resulting in low fitting accuracy. In contrast, cloud transform, with its unique time-frequency localization characteristics and multi-resolution analysis capabilities, overcomes the shortcomings of these traditional methods and demonstrates significant advantages.

[0056] The cloud transform decomposes signals at different scales, extracting both high-frequency details and capturing overall low-frequency trends, adapting to the volatile nature of power output data. Through multi-scale analysis, the cloud transform can simultaneously identify both short-term fluctuations and long-term trends in the signal, effectively overcoming the limitation of traditional methods in processing non-stationary signals. When fitting power output data, the cloud transform adaptively adjusts the decomposition scale based on local variations in the data, accurately capturing sudden changes or local fluctuations in the data. This allows for more detailed fitting results, especially when faced with complex dynamic changes.

[0057] Furthermore, Cloud Transform's ability to capture local features enables it to accurately fit transient changes and sudden fluctuations in power output data, a feat difficult to achieve with traditional methods. By meticulously analyzing signals at multiple scales, Cloud Transform efficiently extracts key features from power output data, reducing the risk of overfitting and improving the model's generalization. This adaptive analysis approach enables Cloud Transform to provide higher fitting accuracy and stronger predictive capabilities when processing power system data with complex dynamic characteristics.

[0058] Therefore, considering the advantages of cloud transform in multi-resolution analysis, local feature capture, noise suppression and adaptive fitting, compared with traditional methods, cloud transform can better handle non-stationary data, local mutations and noise interference, and significantly improve the accuracy and stability of fitting results.

[0059] Analysis of conventional power output data reveals that climate, primary energy prices, dates, and conventional power output data are related by a series of complex nonlinear differential and integral equations. Optimizing production processes involving numerous differential and integral equations faces the "curse of dimensionality" challenge, making it unsuitable for conventional power output forecasting. Therefore, the relationship model between these factors and conventional power output data requires a fundamental transformation from differential and integral equations to algebraic equations to reduce computational complexity.

[0060] Cloud transforms are a process that converts quantitative data into qualitative concepts, identifying regularities and statistical characteristics within massive amounts of data. Cloud transforms offer the advantage of comprehensively analyzing ambiguity, randomness, and the correlations between them. They are suitable for curve fitting problems involving ambiguity and randomness during sample collection. While one-dimensional cloud transforms can only perform single-factor analysis, conventional power output is influenced by multiple factors, necessitating the use of multidimensional cloud transform theory.

[0061] Based on this, the present invention proposes a conventional power output data fitting method based on cloud transformation fitting, such as Figure 1 Shown, including:

[0062] Step 1: obtaining conventional power output data of a conventional power source of a power system, performing statistics on the conventional power output data, and obtaining an output distribution curve of the conventional power output data;

[0063] Step 2: Based on the cloud transformation, the conventional power supply output data is fitted according to the output distribution curve to obtain a distribution function of the output distribution curve;

[0064] Step 3: Using the distribution function as a curve fitting equation for fitting conventional power output data.

[0065] Wherein, based on the cloud transformation, the conventional power supply output data is fitted according to the output distribution curve to obtain the distribution function of the distribution curve, including:

[0066] Based on the cloud transformation, a cloud model that fits the output distribution curve is generated according to the output distribution curve, and based on the cloud model, the residual data distribution is determined. Based on the residual data distribution, the cloud model distribution function is repeatedly constructed until all the peaks of the output distribution curve are fitted, and the distribution function of the distribution curve is generated.

[0067] Wherein, based on the cloud transformation, generating a cloud model distribution function fitting the output distribution curve according to the output distribution curve includes:

[0068] Based on cloud transformation, the output distribution curve is identified, and the peak position of each conventional power output data is determined. The peak position is used as the center of gravity position of the cloud. The entropy of the cloud model is calculated based on the center of gravity position. Based on the center of gravity position and entropy, the normal cloud model formula is applied to generate the cloud model.

[0069] Wherein, determining the residual data distribution based on the cloud model distribution function includes:

[0070] The data values ​​of the part of the output distribution curve that is fitted with the cloud model are removed, and the data values ​​of the remaining part of the output distribution curve are used as the residual distribution.

[0071] Wherein, based on the residual data distribution, repeatedly constructing the cloud model number until all the peaks of the output distribution curve are fitted includes:

[0072] Repeating the peak identification on the output distribution curve for the residual data distribution to construct a cloud model until all peaks are fitted;

[0073] After all peaks are fitted, multiple corresponding cloud model distribution functions are obtained.

[0074] The distribution function for generating the distribution curve includes:

[0075] The distribution functions of the distribution curve are generated by superimposing and fitting multiple corresponding cloud model distribution functions.

[0076] The method further includes: when generating a cloud model for fitting the output distribution curve, using the expected curve of the cloud model for fitting.

[0077] Among them, when the expected curve of the cloud model is used for fitting, the peak position is determined, and with the peak position as the center, the first trough and the first peak are respectively screened out on the left and right sides around the center, and the difference between the first trough and the first peak is calculated, and the difference is compared with the preset threshold. If the difference is greater than the preset threshold, the value of the first peak is recorded. If the difference is less than the preset threshold, the next peak is searched until the difference is greater than the preset threshold. The value range of the data point is constructed based on the recorded value of the first peak, and the data points within the value range are used as cloud droplets. Based on the cloud droplets, the initial value of entropy is calculated using the inverse cloud algorithm.

[0078] The present invention will be further described below with reference to specific cases:

[0079] The specific case step framework is as follows Figure 2 Shown, including:

[0080] (1) Count the output of conventional power sources and obtain the distribution curve f[i].

[0081] (2) Fitting the obtained force data based on cloud transformation:

[0082] ① In the statistically obtained conventional power output distribution curve, identify the position of each data peak one by one. The peak value is used to determine the center of gravity position (expectation) Ex[i] (i=0,…,m-1) of the cloud model; the value of the peak position is recorded as the center of gravity (expectation) of the cloud. The determination of the center of gravity position plays a decisive role in the subsequent entropy calculation and model fitting, so the accuracy of the peak selection and position should be ensured. After the peak position has been determined, the entropy En of the cloud model is calculated based on this position. iThe entropy calculation process includes analyzing the volatility and dispersion of the cloud model near the center of gravity to build a model suitable for describing the distribution characteristics of the peak data. Based on the calculated expectation and entropy values, the normal cloud model formula is applied to generate a distribution function f suitable for fitting the peak. i (x), used to characterize the local features of the data.

[0083] ② On the distribution curve, use the cloud model distribution function calculated in the previous step to convert the data value f[i] in the current data distribution f[i] that matches the cloud model fitting part into i (x) is removed, and the remaining part f′[i] is retained as the new residual data distribution. For the new residual data, the peak identification and cloud model construction process in step 2 are repeated again, and new cloud models are continuously generated until all peaks are effectively fitted, and multiple corresponding cloud model distribution functions f are obtained. i (x), through multiple iterations, a comprehensive data distribution model is formed by superimposing multiple cloud model distribution functions. Each cloud model f i (x) represents a specific local data feature, which meticulously depicts the overall trend of the data curve layer by layer. By superimposing and adjusting the cloud model distribution function, a complete curve fitting equation is ultimately formed, integrating all local features into the overall distribution, and fitting the overall distribution of conventional power output data with optimal accuracy.

[0084] (3) Obtain the distribution function of the curve And use it as the curve fitting equation.

[0085] In step (2), when calculating the cloud model of the fitted distribution curve, the expected curve of the cloud model is used for fitting. The expected curve formula is: When using the cloud model for curve fitting, the determination of En is crucial. Ex can be determined in the jth step of the iteration. j After that, according to the actual situation, j Select several representative points around and use these representative points to obtain En through the reverse cloud algorithm j The specific implementation steps are as follows:

[0086] (1) Find the peak value Ex of the curve.

[0087] (2) Take n points on both sides of Ex (the size of n can be selected experimentally). With Ex as the center, find the value of the first trough and the value of the first peak in the range (Ex-n, Ex), and calculate whether the difference between the two is greater than a certain threshold. If it is greater than the threshold, the value of this point is recorded as x left If it is less than the specific value, ignore the change and continue to look for the next peak until the condition is met; similarly, find x on the right right Compare Ex-x leftwith x righ -Ex value, take the smaller one, assuming it is x right -Ex.

[0088] (3) [Ex-fabs(x righ -Ex),Ex+fabs(x right -Ex)] as cloud droplets, and the initial value of En is calculated using the reverse cloud algorithm that does not require certainty information.

[0089] This paper selects the public operation data of the Belgian power grid in 2018 as the data set and compares the effects of polynomial fitting, random forest fitting and cloud transformation fitting. The fitting results of each method are shown in Figure 2. Figure 3 As shown. Obviously, the described embodiments are some embodiments of the present invention, not all embodiments. Generally, the components of the embodiments of the present invention described and shown in the figures herein can be arranged and designed in a variety of different configurations, including the following:

[0090] (1) Identify key factors from a wide variety of influencing factors.

[0091] In actual production, many factors influence power grid output, including temperature, dew point, wind speed, cloud cover, crude oil prices, industrial production, and the date. To develop an analytical algebraic model for conventional power output data and related factors, it is necessary to first identify key factors from the myriad influencing factors. This method uses weighted analysis using Pearson and Spearman correlation coefficients to identify key influencing factors.

[0092] The formula for calculating the Pearson correlation coefficient is as follows:

[0093]

[0094] in, represents the sum of squared deviations from the mean of X;

[0095] represents the sum of squared deviations from the mean of Y;

[0096] It represents the sum of squared deviations from the mean of X and Y. The formula for calculating the Spearman correlation coefficient is as follows:

[0097]

[0098] Among them, R(x) and R(y) are the ranks of x and y respectively; and Represent the average ranking respectively.

[0099] The simplified Spearman formula is as follows:

[0100]

[0101] Among them, d i It represents the difference in the rank values ​​of the i-th data pair, and n represents the total number of observed samples.

[0102] The weighted formulas for Pearson and Spearman correlation coefficients are as follows:

[0103]

[0104] (2) The temperature, dew point temperature, and date are identified as key influencing factors, and their corresponding relationships with conventional power output data are stored in a set Ξ:

[0105]

[0106] Among them, x1, x2, …, x L is the key influencing factor, y is the conventional power output data, (x1, x2, ... x L ,y) k is the value correspondence of the kth group, N s is the total number of elements in the set. Draw a distribution curve for the set Ξ, and let m = 1.

[0107] (3) Determine the location of the mth peak point of the distribution curve, and find the nearest trough on the left and right sides with the peak point as the center. Calculate the independent variable x according to the following formula n Expectations n And the entropy Yx corresponding to this expectation n .

[0108]

[0109] in, and are the independent variables x that can be obtained between the two troughs n The data distribution function f is calculated using the normal cloud model m (x1,x2,…,x L ).

[0110]

[0111] The normal cloud model is based on the normal distribution and is universal. Let m = m + 1.

[0112] (4) Repeat step (3) until m=N p , where Np is the total number of peak points in the distribution curve. The curve equation after fitting is shown as follows

[0113]

[0114] Where: a m is the amplitude coefficient, and its value can be determined by least squares parameter estimation; is the value of each controllable factor at time t during the operation process; is the corresponding electrical power.

[0115] During the specific implementation process, all historical data for the whole year of 2022 were selected as the data set based on open source meteorological factors, industrial production and measured data of conventional power output, with a step size of 1 hour.

[0116] In order to fully evaluate the effect of the present invention, the following three fitting methods are set for comparison:

[0117] Method 1: Polynomial fitting

[0118] Method 2: Random Forest Fitting

[0119] Method 3: Cloud Transform Fitting

[0120] Method 1 uses a polynomial fitting method, with the independent variables set as temperature and time data, and the dependent variable set as conventional power output data. First, the highest order independent variable is selected, with the temperature independent variable order being 3 and the time independent variable order being 2, to construct a polynomial model. The data points are then substituted into the polynomial model to obtain a system of linear equations, which are then solved using the least squares method. The resulting model is shown in Equation (1.10), with the parameters shown in Table 1.

[0121]

[0122] Table 1

[0123]

[0124] Method 2 uses a random forest fitting method, defining the feature variables as temperature, dew point temperature, date, and day of the week, and the target variable as conventional power load. The dataset is divided into a training set of 292 samples and a test set of the remaining 73 samples. The feature and target variables are normalized to the interval [0, 1]. A random forest model containing 40 decision trees is created. The trained model is used to predict the test set, and the prediction results are denormalized to the original target variable range.

[0125] Compare the fitting curves drawn by the three fitting methods with the true value data, such as Figure 4 、 5, 6. Comparing the results and characteristics reveals that polynomial fitting is computationally simple, efficient, intuitive, and easy to interpret and understand. However, it is not suitable for nonlinear relationships and produces large errors near the endpoints of the data. Random forest fitting provides very high prediction accuracy, can handle data with missing values, and is not prone to overfitting. However, its model is complex, difficult to interpret, and lacks a clear mathematical expression. Cloud transform fitting, on the other hand, can simulate complex data by combining multiple normal distributions. Furthermore, each normal distribution corresponds to a feature in the data, making the model output easier to interpret and having a clear mathematical expression.

[0126] Table 2 compares the fitting errors generated by the three fitting methods, in which two error indicators are considered, namely R 2 (Coefficient of Determination) and RMSE (Root Mean Squared Error), the calculation formulas of each indicator are shown in formula (11) and formula (12). i is the true value, is the predicted value, is the mean.

[0127]

[0128] Table 2

[0129]

[0130] As shown in Table 2, the R 2 The coefficient is 0.17803 higher than the traditional polynomial fitting method, and the RMSE coefficient is 31.88% lower than the polynomial fitting method. Compared with advanced machine learning methods such as random forest, R 2 The coefficient is not much different from the RMSE coefficient. Figure 4 、 5 As shown in Figure 6, cloud transform fitting demonstrates its unique advantages. It not only provides higher fitting accuracy, but also has a clear mathematical expression compared to machine learning methods such as random forests, which makes the model results easier to interpret and understand. This clear expression facilitates the analysis and optimization of the model, and also makes the model's prediction process more transparent, helping users to verify and trust the model's prediction results. In addition, this feature of the cloud transform fitting method is particularly important in practical applications. It allows engineers and decision makers to better understand the working principles of the model, thereby making more scientific and reasonable decisions in actual power system planning and scheduling. In summary, the cloud transform fitting method demonstrates its superiority in both accuracy and interpretability, which makes it have important application value in the field of power system prediction.

[0131] Compared with the existing technology, the present invention has the following advantages: it realizes the transformation of conventional power supply output data fitting from differential and integral models to algebraic models, solves the difficulties of analytical modeling and model dimensionality reduction, and provides a methodological basis for the conventional power supply output law in the medium and long term, including the confidence interval of the minimum output of conventional power supply.

[0132] Example 2:

[0133] The present invention also proposes a conventional power output data fitting system 200 based on cloud transformation fitting, such as Figure 7 Shown, including:

[0134] A statistical unit 201 is configured to obtain conventional power output data of conventional power sources of the power system, perform statistics on the conventional power output data, and obtain an output distribution curve of the conventional power output data;

[0135] A fitting unit 202 is configured to fit the conventional power supply output data according to the output distribution curve based on cloud transformation to obtain a distribution function of the output distribution curve;

[0136] The output unit 203 is configured to use the distribution function as a curve fitting equation for fitting conventional power supply output data.

[0137] Wherein, based on the cloud transformation, the conventional power supply output data is fitted according to the output distribution curve to obtain the distribution function of the distribution curve, including:

[0138] Based on the cloud transformation, a cloud model that fits the output distribution curve is generated according to the output distribution curve, and based on the cloud model, the residual data distribution is determined. Based on the residual data distribution, the cloud model distribution function is repeatedly constructed until all the peaks of the output distribution curve are fitted, and the distribution function of the distribution curve is generated.

[0139] Wherein, based on the cloud transformation, generating a cloud model distribution function fitting the output distribution curve according to the output distribution curve includes:

[0140] Based on cloud transformation, the output distribution curve is identified, and the peak position of each conventional power output data is determined. The peak position is used as the center of gravity position of the cloud. The entropy of the cloud model is calculated based on the center of gravity position. Based on the center of gravity position and entropy, the normal cloud model formula is applied to generate the cloud model.

[0141] Wherein, determining the residual data distribution based on the cloud model distribution function includes:

[0142] The data values ​​of the part of the output distribution curve that is fitted with the cloud model are removed, and the data values ​​of the remaining part of the output distribution curve are used as the residual distribution.

[0143] Wherein, based on the residual data distribution, repeatedly constructing the cloud model number until all the peaks of the output distribution curve are fitted includes:

[0144] Repeating the peak identification on the output distribution curve for the residual data distribution to construct a cloud model until all peaks are fitted;

[0145] After all peaks are fitted, multiple corresponding cloud model distribution functions are obtained.

[0146] The distribution function for generating the distribution curve includes:

[0147] The distribution functions of the distribution curve are generated by superimposing and fitting multiple corresponding cloud model distribution functions.

[0148] When generating a cloud model for fitting the output distribution curve, the expected curve of the cloud model is used for fitting.

[0149] Among them, when the expected curve of the cloud model is used for fitting, the peak position is determined, and with the peak position as the center, the first trough and the first peak are respectively screened out on the left and right sides around the center, and the difference between the first trough and the first peak is calculated, and the difference is compared with the preset threshold. If the difference is greater than the preset threshold, the value of the first peak is recorded. If the difference is less than the preset threshold, the next peak is searched until the difference is greater than the preset threshold. The value range of the data point is constructed based on the recorded value of the first peak, and the data points within the value range are used as cloud droplets. Based on the cloud droplets, the initial value of entropy is calculated using the inverse cloud algorithm.

[0150] In Examples 1 and 2, a comparison of the output forecasting capabilities of a cloud transformation-based method and a traditional linear regression method within a specific region revealed that the cloud transformation model reduced prediction errors by 15% to 25%. Compared to the traditional linear regression model, it offers higher fitting accuracy, better capturing the nonlinear characteristics of output data and understanding the changing patterns of conventional power output levels. By improving prediction accuracy, the cloud transformation-based fitting method enables the dispatch center to more effectively schedule generator operations and avoid over- or under-generation.

[0151] Furthermore, the cloud transformation model is relatively simple to build and implement, especially during the data preprocessing phase. Through automated algorithms, operators can quickly obtain the required forecast results, reducing reliance on senior technical personnel and operational complexity. Accurate output forecasts can reduce reliance on fossil fuels and promote the efficient use of renewable energy.

[0152] Power plants that use cloud transformation methods to fit conventional power output data have reduced a large amount of carbon emissions, and their environmental benefits are significantly improved compared to traditional methods.

[0153] In addition, the present invention also enhances the ability to handle uncertainty.

[0154] Experimental results demonstrate that this method maintains excellent predictive performance in dynamic environments, particularly under extreme weather conditions, demonstrating enhanced adaptability and stability. Cloud transformation quantifies uncertainty, transforming complex nonlinear relationships into a manageable form, making the model more adaptable in complex environments. This method, based on fuzzy set theory and probability statistics, effectively integrates information from diverse sources, minimizing information loss and thus improving overall predictive capabilities.

[0155] In summary, the present invention has demonstrated significant advantages in improving prediction accuracy, saving resources, simplifying operations, and promoting environmental protection, fully demonstrating the practical value and potential impact of the invention.

[0156] Example 3:

[0157] Based on the same inventive concept, the present invention also provides a computer device, which includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of the method in the above embodiment.

[0158] Example 4:

[0159] Based on the same inventive concept, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It can be understood that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space that stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the steps of the method in the above embodiment.

[0160] It will be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention may be implemented in various computer languages, for example, the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0161] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0162] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0163] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0164] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0165] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A conventional power output data fitting method based on cloud transformation fitting, characterized in that: include: Obtaining conventional power output data of a conventional power source of the power system, performing statistics on the conventional power output data, and obtaining an output distribution curve of the conventional power output data; Based on the cloud transformation, the conventional power supply output data is fitted according to the output distribution curve to obtain a distribution function of the output distribution curve; Using the distribution function as a curve fitting equation for fitting conventional power output data; The cloud-based transformation, according to the output distribution curve, fits the conventional power output data to obtain a distribution function of the distribution curve, including: Based on the cloud transformation, a cloud model fitting the output distribution curve is generated according to the output distribution curve, and based on the cloud model, a residual data distribution is determined, and based on the residual data distribution, a cloud model distribution function is repeatedly constructed until all peaks of the output distribution curve are fitted, and a distribution function of the distribution curve is generated; The cloud-based transformation generates a cloud model distribution function that fits the output distribution curve according to the output distribution curve, including: Based on cloud transformation, the output distribution curve is identified to determine the peak position of each conventional power output data, the peak position is used as the center of gravity position of the cloud, the center of gravity position is used as a reference to calculate the entropy of the cloud model, and based on the center of gravity position and entropy, a normal cloud model formula is applied to generate a cloud model; The determining of the residual data distribution based on the cloud model distribution function includes: Eliminate the data values ​​of the output distribution curve that are consistent with the cloud model fitting part, and use the data values ​​of the remaining part of the output distribution curve as the residual distribution; The method of repeatedly constructing the cloud model number based on the residual data distribution until all peaks of the output distribution curve are fitted includes: Repeating the peak identification on the output distribution curve for the residual data distribution to construct a cloud model until all peaks are fitted; After all peaks are fitted, multiple corresponding cloud model distribution functions are obtained.

2. The method according to claim 1, characterized in that The distribution function for generating the distribution curve includes: The distribution functions of the distribution curve are generated by superimposing and fitting multiple corresponding cloud model distribution functions.

3. The method according to claim 1, characterized in that The method further includes: when generating a cloud model for fitting the output distribution curve, using an expected curve of the cloud model for fitting.

4. The method according to claim 3, characterized in that When fitting the expected curve of the cloud model, the peak position is determined, and with the peak position as the center, the first trough and the first peak are screened out on the left and right sides of the center respectively, and the difference between the first trough and the first peak is calculated, and the difference is compared with the preset threshold. If the difference is greater than the preset threshold, the value of the first peak is recorded. If the difference is less than the preset threshold, the next peak is searched until the difference is greater than the preset threshold. The value range of the data point is constructed based on the recorded value of the first peak, and the data points within the value range are used as cloud droplets. Based on the cloud droplets, the initial value of entropy is calculated using the inverse cloud algorithm.

5. A conventional power output data fitting system based on cloud transformation fitting, characterized in that: include: a statistical unit, configured to obtain conventional power output data of a conventional power source of the power system, perform statistics on the conventional power output data, and obtain an output distribution curve of the conventional power output data; a fitting unit, configured to fit the conventional power supply output data according to the output distribution curve based on cloud transformation to obtain a distribution function of the output distribution curve; An output unit, configured to use the distribution function as a curve fitting equation for fitting conventional power supply output data; The cloud-based transformation, according to the output distribution curve, fits the conventional power output data to obtain a distribution function of the distribution curve, including: Based on the cloud transformation, a cloud model fitting the output distribution curve is generated according to the output distribution curve, and based on the cloud model, a residual data distribution is determined, and based on the residual data distribution, a cloud model distribution function is repeatedly constructed until all peaks of the output distribution curve are fitted, and a distribution function of the distribution curve is generated; The cloud-based transformation generates a cloud model distribution function that fits the output distribution curve according to the output distribution curve, including: Based on cloud transformation, the output distribution curve is identified to determine the peak position of each conventional power output data, the peak position is used as the center of gravity position of the cloud, the center of gravity position is used as a reference to calculate the entropy of the cloud model, and based on the center of gravity position and entropy, a normal cloud model formula is applied to generate a cloud model; The determining of the residual data distribution based on the cloud model distribution function includes: Eliminate the data values ​​of the output distribution curve that are consistent with the cloud model fitting part, and use the data values ​​of the remaining part of the output distribution curve as the residual distribution; The method of repeatedly constructing the cloud model number based on the residual data distribution until all peaks of the output distribution curve are fitted includes: Repeating the peak identification on the output distribution curve for the residual data distribution to construct a cloud model until all peaks are fitted; After all peaks are fitted, multiple corresponding cloud model distribution functions are obtained.

6. The system according to claim 5, characterized in that The distribution function for generating the distribution curve includes: The distribution functions of the distribution curve are generated by superimposing and fitting multiple corresponding cloud model distribution functions.

7. The system according to claim 5, characterized in that When generating a cloud model for fitting the output distribution curve, the expected curve of the cloud model is used for fitting.

8. The system according to claim 7, characterized in that When fitting the expected curve of the cloud model, the peak position is determined. With the peak position as the center, the first trough and the first peak are screened out on the left and right sides of the center respectively. The difference between the first trough and the first peak is calculated, and the difference is compared with the preset threshold. If the difference is greater than the preset threshold, the value of the first peak is recorded. If the difference is less than the preset threshold, the next peak is searched until the difference is greater than the preset threshold. The value range of the data point is constructed based on the recorded value of the first peak, and the data points within the value range are used as cloud droplets. Based on the cloud droplets, the initial value of entropy is calculated using the inverse cloud algorithm.

9. A computer device, characterized in that: include: one or more processors; a processor for executing one or more programs; When the one or more programs are executed by the one or more processors, the method according to any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed, the method according to any one of claims 1 to 4 is implemented.

Images