New energy generating capacity evaluation and prediction method and system based on multi-time scale analysis

Abstract

The invention discloses a new energy generating capacity evaluation and prediction method and system based on multi-time scale analysis, and relates to the technical field of data analysis and prediction. Based on a multi-time scale analysis method, extracting change characteristics of the power generation data; constructing a time sequence prediction model; inputting the change characteristics of the power generation data into a time sequence prediction model; and providing a generating capacity evaluation report based on the time sequence prediction model. By introducing the multi-time scale analysis method, the characteristics of new energy power generation data in short-term fluctuation, middle-term periodic change and long-term trend are effectively extracted, so that the accuracy and stability of power generation capacity prediction are greatly improved. By integrating the multi-time scale features and the load demand data, a more accurate time sequence prediction model is constructed, and the defects that an existing prediction method cannot adapt to complex time-varying characteristics and loads are not matched are overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis and prediction, and specifically to a method and system for evaluating and predicting renewable energy power generation based on multi-time scale analysis. Background Art

[0002] With the transformation of the global energy structure and the rapid development of renewable energy, new energy power generation (such as solar energy, wind energy, etc.) has gradually become an important part of energy supply. However, the characteristics of new energy power generation make it face many challenges in practical application. Compared with traditional fossil energy power generation, new energy power generation has obvious intermittency and volatility, which is mainly due to the impact of natural factors such as weather and seasons on wind and photovoltaic energy output. Therefore, how to accurately evaluate and predict the performance of new energy power generation to improve the dispatch efficiency of the power grid has become an important issue in the research of smart grids and power systems.

[0003] At present, the research on the prediction of renewable energy power generation has made some progress. Traditional power generation prediction methods mainly rely on historical data and simple statistical models, such as time series analysis and linear regression. These methods can provide a certain degree of prediction, but often fail to fully consider the time-varying and complexity of renewable energy power generation. To solve this problem, research in recent years has begun to introduce advanced technologies such as machine learning and deep learning, and try to improve the prediction accuracy through multi-source data fusion and spatiotemporal feature extraction. In addition, data-driven multi-time scale analysis methods have gradually become a mainstream trend, especially feature extraction for different time scales (such as short-term, medium-term, and long-term), which has been proven to effectively improve the accuracy of renewable energy power generation prediction.

[0004] Although the existing technology has alleviated the problem of renewable energy power generation prediction to a certain extent, there are still many shortcomings. First, traditional methods mostly rely on a single time scale or a simple feature extraction method, and fail to fully capture the changing characteristics of power generation data at different time scales. This limits the accuracy of the prediction model, especially when modeling complex systems (such as variable weather conditions, power grid load, etc.), it is often unable to provide sufficiently refined predictions. Secondly, most of the existing methods only focus on the prediction of power generation itself, ignoring the impact of other factors related to power generation (such as load demand, power grid scheduling, weather changes, etc.) on system performance. Comprehensively evaluating these factors in multiple dimensions can more accurately reflect the operating status of the power grid and the characteristics of the power generation system. Therefore, how to construct a power generation evaluation and prediction method based on multi-time scale analysis and comprehensive feature vectors has become a problem that needs to be solved urgently in the current technology. It is in this context that the present invention is proposed. By introducing multi-time scale feature extraction, comprehensive feature vector fusion and time series prediction model, it can more accurately predict the power generation of renewable energy, and provide a basis for optimizing power grid scheduling and power generation management, which has significant technical advantages and application prospects. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problems solved by the present invention are: first, the existing new energy power generation prediction methods cannot effectively process data characteristics of multiple time scales, resulting in the failure to fully consider short-term fluctuations, medium-term cyclical changes and long-term trends, thereby affecting the accuracy of power generation prediction; second, when the power generation does not match the load demand, the existing methods fail to provide real-time and accurate load scheduling and emergency response strategies, and cannot timely schedule and optimize resource allocation; third, when facing sudden load changes and complex power generation environments, the existing technology lacks an intelligent decision-making mechanism based on the fusion of multi-data features, resulting in insufficient support for system stability and grid scheduling in the prediction results.

[0007] In order to solve the above technical problems, the present invention provides the following technical solution: a new energy power generation evaluation and prediction method based on multi-time scale analysis, which comprises the following steps:

[0008] Acquire power generation data; extract the changing characteristics of power generation data based on multi-time scale analysis methods; construct a time series prediction model; input the changing characteristics of power generation data into the time series prediction model; provide a power generation assessment report based on the time series prediction model.

[0009] As a preferred solution of the method for evaluating and predicting the power generation of new energy based on multi-time scale analysis described in the present invention, it is characterized in that: the power generation data is preprocessed after being acquired;

[0010] The preprocessing includes denoising, filling missing values ​​and processing outliers on the power generation data.

[0011] As a preferred solution of the renewable energy power generation evaluation and prediction method based on multi-time scale analysis described in the present invention, it is characterized in that: the multi-time scale analysis method includes defining the time scale, and based on different time scales, adopting different feature extraction methods to extract the changing characteristics of power generation data.

[0012] The different time scales include a short-term time scale, a medium-term time scale, and a long-term time scale.

[0013] The different feature extraction methods include, based on a short-term time scale, converting the power generation data into the frequency domain, analyzing different frequency components, identifying the periodic changes in the short term, and obtaining the short-term fluctuation characteristics.

[0014] Based on the medium-term time scale, the power generation data is decomposed by wavelet transform to obtain the medium-term periodic characteristics.

[0015] Based on the long-term time scale, long-term trend characteristics are obtained through seasonal decomposition.

[0016] The short-term fluctuation characteristics, medium-term cyclical characteristics and long-term trend characteristics are combined into a comprehensive feature vector, and a time series prediction model is constructed based on the comprehensive feature vector.

[0017] As a preferred solution of the method for evaluating and predicting renewable energy power generation based on multi-time scale analysis described in the present invention, it is characterized in that: the power generation data is converted into the frequency domain to obtain a spectrum, which is expressed as:

[0018] X(f)=FFT{x(t)}

[0019] Where f represents frequency, X(f) represents a complex value in the frequency domain, and x(t) represents power generation data.

[0020] The short-term fluctuation feature vector is obtained by extracting the peak frequency and peak amplitude in the spectrum.

[0021] The peak frequency is expressed as,

[0022] f peak = argmax f |X(f)|

[0023] The peak amplitude is expressed as,

[0024] A peak =max f |X(f)|

[0025] The short-term fluctuation characteristic vector is expressed as,

[0026] S short =[f peak,short ,A peak ,PeriodicityIndex]

[0027] Among them, f peak,short represents the peak frequency in the short-term fluctuation feature vector;

[0028] The mid-term periodic feature vector is obtained by decomposing the power generation data through the wavelet transform, wherein the coefficients obtained after the power generation data is transformed by the wavelet transform are expressed as:

[0029]

[0030] Among them, a represents the scale factor, b represents the translation factor, and ψ represents the wavelet basis function.

[0031] The mid-term periodic eigenvector is expressed as,

[0032] S medium =[fpeak,medium ,E wavelet ,Scale low ]

[0033] Among them, Scale low represents the low-frequency scale after wavelet transformation, E wavelet Represents the energy of the wavelet transform coefficients, by calculating ∑|W a,b (t)| 2 Get, f peak,medium Represents the peak frequency of the medium-term periodic eigenvector.

[0034] The power generation data is decomposed into trend, seasonal component and residual by seasonal decomposition to obtain the long-term trend feature vector, which is expressed as:

[0035] x(t)=T(t)+S(t)+R(t)

[0036] Among them, T(t) represents the trend, S(t) represents the seasonal component, and R(t) represents the residual.

[0037] The trend is obtained by sliding average or weighted smoothing method and is expressed as,

[0038] T(t)=Smoothing(x(t))

[0039] The long-term trend feature vector is expressed as,

[0040] S long =[T slope ,T amplitude ,S variation ]

[0041] Among them, T slope Indicates the rate of change of trend, T amplitude Indicates the magnitude of the trend, S variation Indicates the magnitude of seasonal fluctuations.

[0042] The principal component analysis method is used to reduce the dimension and fuse the features of each time scale to obtain a comprehensive feature vector, which is expressed as:

[0043] S combined = PCA(S short ,S medium ,S long ).

[0044] As a preferred solution of the method for evaluating and predicting renewable energy power generation based on multi-time scale analysis described in the present invention, an autoregressive moving average model is introduced to construct the time series prediction model, and the time series prediction model is evaluated and optimized.

[0045] As a preferred solution of the new energy power generation evaluation and prediction method based on multi-time scale analysis described in the present invention, wherein: the power generation evaluation report provided based on the time series prediction model includes incorporating the comprehensive feature vector as an exogenous variable into the time series prediction model, expressed as,

[0046]

[0047] Among them, y t represents the power generation at time t, α is a constant term, φ i represents the autoregressive coefficient, θ j represents the sliding average coefficient, ε t-j represents the error term at the past j time points, γ k represents the coefficient of the exogenous variable, S combined，t-k represents the comprehensive characteristic vector at time tk, and k represents the lag period of the exogenous variable.

[0048] Power generation evaluation is performed based on power generation.

[0049] As a preferred scheme of the new energy power generation assessment and prediction method based on multi-time scale analysis described in the present invention, wherein: the power generation assessment based on power generation includes setting a normal range of power generation and making a judgment based on the power generation and the normal range of power generation.

[0050] The normal range of power generation is expressed as,

[0051] μ t -δ·σ t ≤y t ≤μ t +δ·σ t

[0052] Among them, μ t represents the mean value of power generation, δ is the coefficient, σ t Represents standard deviation.

[0053] If the power generation y t If the power generation is within the normal range, regular inspections will be carried out to optimize grid dispatch and resource allocation based on power generation capacity.

[0054] If the power generation y t Greater than μ t +δ·σ t, it means that the power generation system is in overload operation. When the power generation system is in overload operation, the load demand at the power grid dispatching system is obtained for secondary judgment. When the relative gap between the power generation and the load demand exceeds the specified threshold, it means that the power grid system is in a clear overload state. The excess power is then transmitted to other areas through the cross-regional interconnection of the power grid. If there is no power demand in other areas, the excess power is stored in the energy storage device.

[0055] If the power generation y t Less than μ t -δ·σ t If the power generation is greater than the load demand and less than μ, the power generation in other regions will be dispatched through the cross-regional interconnection of the power grid. If the load demand cannot be met, the backup generator or temporary power generation facilities will be activated. If the power generation is greater than the load demand and less than μ t -δ·σ t When the power grid is running, smart grid technology is used to optimize power distribution and scheduling based on real-time data, and the output of different generator sets is adjusted in real time by dynamically monitoring the relationship between load and power generation.

[0056] The relative gap is expressed as,

[0057]

[0058] Among them, L t Indicates load demand.

[0059] Another object of the present invention is to provide a new energy power generation assessment and prediction system based on multi-time scale analysis, which can extract the short-term fluctuation characteristics, medium-term periodic characteristics and long-term trend characteristics of power generation data through multi-time scale analysis methods, combine load demand with exogenous variables, and optimize the time series prediction model, thereby improving the accuracy and stability of power generation prediction, solving the shortcomings of existing new energy power generation prediction systems in processing complex time-varying data, capturing multi-time scale characteristics and responding to changes in load demand, and filling the problems of existing technologies in intelligent load scheduling and power generation assessment.

[0060] To solve the above technical problems, the present invention provides the following technical solutions: a new energy power generation evaluation and prediction system based on multi-time scale analysis, including a data acquisition module, a data preprocessing module, a multi-time scale analysis module, a feature fusion module and a power generation evaluation module.

[0061] The data acquisition module is responsible for collecting real-time power generation data from sensors, smart meters and external sources.

[0062] The data preprocessing module is responsible for data cleaning and preprocessing.

[0063] The multi-time scale analysis module extracts the variation characteristics of power generation data according to different time scales through frequency domain conversion, wavelet transform and seasonal decomposition methods.

[0064] The feature fusion module fuses the short-term fluctuation features, medium-term periodic features and long-term trend features extracted at different time scales to obtain a comprehensive feature vector.

[0065] The power generation evaluation module evaluates the power generation based on the output of the time series prediction model and the set threshold.

[0066] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method for evaluating and predicting new energy power generation based on multi-time scale analysis as described above are implemented.

[0067] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for evaluating and predicting new energy power generation based on multi-time scale analysis as described above.

[0068] Beneficial effects of the present invention: Our invention effectively extracts the characteristics of new energy power generation data in short-term fluctuations, medium-term cyclical changes and long-term trends by introducing a multi-time scale analysis method, thereby greatly improving the accuracy and stability of power generation forecasts. By integrating multi-time scale characteristics and load demand data, a more accurate time series prediction model is constructed, which solves the shortcomings of existing prediction methods that cannot adapt to complex time variability and load mismatch. In addition, the present invention optimizes the load scheduling strategy by evaluating power generation in real time and combining it with the load demand of the power grid dispatching system. When the difference between power generation and load demand is large, load scheduling and resource optimization can be carried out in a timely manner, avoiding the situation of excess or insufficient power generation, improving the emergency response capability and resource utilization efficiency of the power grid system, and thus improving the overall reliability and intelligence level of the new energy power generation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0070] Figure 1 An overall flow chart of a new energy power generation assessment and prediction method based on multi-time scale analysis provided for the first embodiment of the present invention.

[0071] Figure 2An overall framework diagram of a renewable energy power generation assessment and prediction system based on multi-time scale analysis provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0073] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides a new energy power generation evaluation and prediction method based on multi-time scale analysis, characterized in that:

[0074] S1: Obtain power generation data;

[0075] After obtaining the power generation data, preprocessing is performed.

[0076] Preprocessing includes denoising, filling missing values ​​and handling outliers for power generation data.

[0077] It is necessary to collect data from different energy generation equipment (such as wind farms, photovoltaic power stations, etc.) and environmental sensors. The data includes but is not limited to real-time power generation, temperature, humidity, wind speed, radiation intensity, air pressure and other parameters.

[0078] S2: Extract the variation characteristics of power generation data based on multi-time scale analysis method.

[0079] The multi-time scale analysis method includes defining the time scale and taking different feature extraction methods to extract the changing characteristics of power generation data based on different time scales.

[0080] Different time scales include short term time scale, medium term time scale and long term time scale.

[0081] Different feature extraction methods include, based on the short-term time scale, by converting the power generation data into the frequency domain, and by analyzing different frequency components, identifying the periodic changes in the short term to obtain the short-term fluctuation characteristics.

[0082] Based on the medium-term time scale, the power generation data is decomposed by wavelet transform to obtain the medium-term periodic characteristics.

[0083] Based on the long-term time scale, long-term trend characteristics are obtained through seasonal decomposition.

[0084] The short-term fluctuation characteristics, medium-term cyclical characteristics and long-term trend characteristics are combined into a comprehensive feature vector, and a time series prediction model is constructed based on the comprehensive feature vector.

[0085] The power generation data is converted into the frequency domain to obtain the spectrum, which is expressed as,

[0086] X(f)=FFT{x(t)}

[0087] Where f represents frequency, X(f) represents a complex value in the frequency domain, and x(t) represents power generation data.

[0088] The short-term fluctuation feature vector is obtained by extracting the peak frequency and peak amplitude in the spectrum.

[0089] The peak frequency is expressed as,

[0090] f peak = argmax f |X(f)|

[0091] The peak amplitude is expressed as,

[0092] A peak =max f |X(f)|

[0093] The short-term volatility eigenvector is expressed as,

[0094] S short =[f peak,short ,A peak ,PeriodicityIndex]

[0095] Among them, f peak,short Represents the peak frequency in the short-term fluctuation eigenvector.

[0096] The mid-term periodic feature vector is obtained by decomposing the power generation data through wavelet transform, where the coefficients obtained by wavelet transform of the power generation data are expressed as:

[0097]

[0098] Among them, a represents the scale factor, b represents the translation factor, ψ represents the wavelet basis function,

[0099] The medium-term periodic eigenvector is expressed as,

[0100] S medium =[f peak,medium ,E wavelet ,Scale low ]

[0101] Among them, Scale low represents the low-frequency scale after wavelet transformation, Ewavelet Represents the energy of the wavelet transform coefficients, by calculating ∑|W a,b (t)| 2 Get, f peak,medium Represents the peak frequency of the medium-term periodic eigenvector.

[0102] The power generation data is decomposed into trend, seasonal component and residual by seasonal decomposition, and the long-term trend feature vector is obtained, which is expressed as:

[0103] x(t)=T(t)+S(t)+R(t)

[0104] Among them, T(t) represents the trend, S(t) represents the seasonal component, and R(t) represents the residual.

[0105] The trend is obtained by moving average or weighted smoothing method and is expressed as,

[0106] T(t)=Smoothing(x(t))

[0107] The long-term trend eigenvector is expressed as,

[0108] S long =[T slope ,T amplitude ,S variation ]

[0109] Among them, T slope Indicates the rate of change of trend, T amplitude Indicates the magnitude of the trend, S variation Indicates the magnitude of seasonal fluctuations.

[0110] The principal component analysis method is used to reduce the dimension and fuse the features of each time scale to obtain a comprehensive feature vector, which is expressed as:

[0111] S combined = PCA(S short ,S medium ,S long ).

[0112] Furthermore, the short-term scale is specific, such as minute and hour levels, reflecting instantaneous changes and short-term fluctuations (for example, fluctuations caused by sudden weather changes and equipment failures).

[0113] The medium-term scale is specific, such as the daily and monthly levels, and mainly reflects changes in daily and weekly cycles (for example, changes in day and night, changes in sunshine, changes in wind speed, etc.).

[0114] Long-term scales, such as seasonal or annual scales, reflect long-term trends and cyclical changes (e.g., seasonal climate change, average annual wind speed, solar radiation, etc.).

[0115] Furthermore, by extracting features at different time scales, the present invention can fully capture the multi-level variation patterns of power generation data. Frequency domain analysis at a short-term time scale can identify periodic changes, wavelet transform at a medium-term time scale helps extract periodic features, and seasonal decomposition at a long-term time scale reveals long-term trends. This multi-dimensional feature extraction method helps to improve the accuracy and stability of power generation forecasts, especially in the face of complex climate change and load fluctuations, and can ensure the accuracy of power generation forecasts.

[0116] S3: Build a time series forecasting model.

[0117] The autoregressive moving average model is introduced to construct a time series prediction model, and the time series prediction model is evaluated and optimized.

[0118] The present invention introduces an autoregressive moving average (ARMA) model to construct a time series forecasting model, and models the model by integrating the feature vectors extracted from each time scale. This method not only takes into account the historical information of power generation data, but also incorporates the changing factors of the external environment (such as meteorological conditions) into the forecasting model, which significantly improves the flexibility and accuracy of the forecast. The autoregressive moving average model has good adaptability to the dynamic changes of time series and can effectively capture the short-term fluctuations and trend changes in power generation.

[0119] S4: Input the changing characteristics of power generation data into the time series prediction model.

[0120] Providing a power generation assessment report based on a time series forecasting model includes incorporating a comprehensive feature vector as an exogenous variable into the time series forecasting model, expressed as,

[0121]

[0122] Among them, y t represents the power generation at time t, α is a constant term, φ i represents the autoregressive coefficient, θ j represents the sliding average partial coefficient, ∈ t-j represents the error term at the past j time points, γ k represents the coefficient of the exogenous variable, S combined，t-k represents the comprehensive characteristic vector at time tk, and k represents the lag period of the exogenous variable.

[0123] Power generation evaluation is performed based on power generation.

[0124] S5: Provide power generation assessment report based on time series prediction model.

[0125] The power generation evaluation based on the power generation includes setting a normal range of the power generation and making a judgment based on the power generation and the normal range of the power generation.

[0126] The normal range of power generation is expressed as,

[0127] μ t -δ·σ t ≤y t ≤μ t +δ·σ t

[0128] Among them, μ t represents the mean value of power generation, δ is the coefficient, σ t Represents standard deviation.

[0129] If the power generation y t If the power generation is within the normal range, regular inspections will be carried out to optimize grid dispatch and resource allocation based on power generation capacity.

[0130] If the power generation y t Greater than μ t +δ·σ t , it means that the power generation system is in overload operation. When the power generation system is in overload operation, the load demand at the power grid dispatching system is obtained for secondary judgment. When the relative gap between the power generation and the load demand exceeds the specified threshold, it means that the power grid system is in a clear overload state. The excess power is then transmitted to other areas through the cross-regional interconnection of the power grid. If there is no power demand in other areas, the excess power is stored in the energy storage device.

[0131] If the power generation y t Less than μ t -δ·σ t If the power generation is greater than the load demand and less than μ, the power generation in other regions will be dispatched through the cross-regional interconnection of the power grid. If the load demand cannot be met, the backup generator or temporary power generation facilities will be activated. If the power generation is greater than the load demand and less than μ t -δ·σ t When the power grid is running, smart grid technology is used to optimize power distribution and scheduling based on real-time data, and the output of different generator sets is adjusted in real time by dynamically monitoring the relationship between load and power generation.

[0132] The relative gap is expressed as,

[0133]

[0134] Among them, L t Indicates load demand.

[0135] Based on the power generation assessment report provided by the time series prediction model, it is possible to timely assess whether the power generation is within the normal range. If the power generation fluctuates abnormally, the present invention adjusts the output of the generator set or allocates power between regions in real time through the power grid dispatching system. If the power generation is overloaded or insufficient, the present invention can also perform real-time optimization through smart grid technology to ensure the balance of power supply and demand. In addition, the present invention can also transfer power to other areas when the power generation system is overloaded, or store the remaining power in energy storage equipment, thereby improving the flexibility and emergency response capabilities of the power grid.

[0136] Example 2, reference Figure 2 , which is an embodiment of the present invention, provides a system for evaluating and predicting new energy power generation based on multi-time scale analysis, characterized in that it includes a data acquisition module 100, a data preprocessing module 200, a multi-time scale analysis module 300, a feature fusion module 400 and a power generation evaluation module 500.

[0137] The data acquisition module 100 is responsible for collecting real-time power generation data from sensors, smart meters and external sources.

[0138] The data preprocessing module 200 is responsible for data cleaning and preprocessing.

[0139] The multi-time scale analysis module 300 extracts the variation characteristics of power generation data according to different time scales through frequency domain conversion, wavelet transform and seasonal decomposition methods.

[0140] The feature fusion module 400 fuses the short-term fluctuation features, medium-term periodic features and long-term trend features extracted at different time scales to obtain a comprehensive feature vector.

[0141] The power generation evaluation module 500 evaluates the power generation based on the output of the time series prediction model and the set threshold.

[0142] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0143] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0144] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0145] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0146] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention 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 invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A new energy power generation evaluation and prediction method based on multi-time scale analysis, characterized in that: include: Obtain power generation data; Extract the variation characteristics of power generation data based on multi-time scale analysis method; Build time series forecasting models; Input the changing characteristics of power generation data into the time series forecasting model; Provide power generation assessment report based on time series forecasting model.

2. The method for evaluating and predicting renewable energy power generation based on multi-time scale analysis according to claim 1, characterized in that: After the power generation data is acquired, preprocessing is performed; The preprocessing includes denoising, filling missing values ​​and processing outliers on the power generation data.

3. The method for evaluating and predicting renewable energy power generation based on multi-time scale analysis according to claim 2, characterized in that: The multi-time scale analysis method includes defining a time scale, and taking different feature extraction methods based on different time scales to extract the variation characteristics of power generation data; The different time scales include short-term time scale, medium-term time scale and long-term time scale; The different feature extraction methods include, based on the short-term time scale, converting the power generation data into the frequency domain, identifying the periodic changes in the short term by analyzing different frequency components, and obtaining the short-term fluctuation characteristics; Based on the medium-term time scale, the power generation data is decomposed by wavelet transform to obtain the medium-term periodic characteristics; Based on the long-term time scale, long-term trend characteristics are obtained through seasonal decomposition; The short-term fluctuation characteristics, medium-term cyclical characteristics and long-term trend characteristics are combined into a comprehensive feature vector, and a time series prediction model is constructed based on the comprehensive feature vector.

4. The method for evaluating and predicting renewable energy power generation based on multi-time scale analysis according to claim 3, characterized in that: The power generation data is converted into the frequency domain to obtain the spectrum, which is expressed as: X(f)=FFT{x(t)} Where f represents frequency, X(f) represents the complex value in the frequency domain, and x(t) represents power generation data; After extracting the peak frequency and peak amplitude in the spectrum, the short-term fluctuation feature vector is obtained; The peak frequency is expressed as, f peak =argmax f |X(f)| The peak amplitude is expressed as, A peak =max f |X(f)| The short-term fluctuation characteristic vector is expressed as, S short =[f peak,short ,A peak ,PeriodicityIndex] Among them, f peak,short represents the peak frequency in the short-term fluctuation feature vector; The mid-term periodic feature vector is obtained by decomposing the power generation data through the wavelet transform, wherein the coefficients obtained after the power generation data is transformed by the wavelet transform are expressed as: Among them, a represents the scale factor, b represents the translation factor, and ψ represents the wavelet basis function; The mid-term periodic eigenvector is expressed as, S medium =[f peak,medium ,E wavelet ,Scale low ] Among them, Scale low represents the low-frequency scale after wavelet transformation, E wavelet Represents the energy of the wavelet transform coefficients, by calculating ∑|W a,b (t)| 2 Get, f peak,medium represents the peak frequency of the mid-term periodic eigenvector; The power generation data is decomposed into trend, seasonal component and residual by seasonal decomposition to obtain the long-term trend feature vector, which is expressed as: x(t)=T(t)+S(t)+R(t) Among them, T(t) represents the trend, S(t) represents the seasonal component, and R(t) represents the residual; The trend is obtained by sliding average or weighted smoothing method and is expressed as, T(t)=Smoothing(x(t)) The long-term trend feature vector is expressed as, S long =[T slope ,T amplitude ,S variation ] Among them, T slope Indicates the rate of change of trend, T amplitude Indicates the magnitude of the trend, S variation Indicates the magnitude of seasonal fluctuations; The principal component analysis method is used to reduce the dimension and fuse the features of each time scale to obtain a comprehensive feature vector, which is expressed as: S combined =PCA(S short ,S medium ,S long )。 5. The method for evaluating and predicting renewable energy power generation based on multi-time scale analysis according to claim 4, characterized in that: An autoregressive moving average model is introduced to construct the time series prediction model, and the time series prediction model is evaluated and optimized.

6. The method for evaluating and predicting renewable energy power generation based on multi-time scale analysis according to claim 5, characterized in that: The provision of a power generation assessment report based on a time series prediction model includes integrating a comprehensive feature vector as an exogenous variable into the time series prediction model, expressed as: Among them, y t represents the power generation at time t, α is a constant term, φ i represents the autoregressive coefficient, θ j represents the sliding average partial coefficient, ∈ t-j represents the error term at the past j time points, γ k represents the coefficient of the exogenous variable, S combined，t-k represents the comprehensive characteristic vector at time tk, where k represents the lag period of the exogenous variable; Power generation evaluation is performed based on power generation.

7. The method for evaluating and predicting renewable energy power generation based on multi-time scale analysis according to claim 6, characterized in that: The power generation evaluation based on the power generation includes setting a normal range of the power generation and making a judgment based on the power generation and the normal range of the power generation; The normal range of power generation is expressed as, m t -d·s t ≤y t ≤μ t +d·s t Among them, μ t represents the mean value of power generation, δ is the coefficient, σ t represents standard deviation; If the power generation y t If the power generation is within the normal range, regular inspections are carried out to optimize grid dispatch and resource allocation based on power generation capacity; If the power generation y t Greater than μ t +δ·σ t , it means that the power generation system is in overload operation. When the power generation system is in overload operation, the load demand at the power grid dispatching system is obtained for secondary judgment. When the relative gap between the power generation and the load demand exceeds the specified threshold, it means that the power grid system is in a clear overload state. Then, the excess power is transmitted to other areas through the cross-regional interconnection of the power grid. If there is no power demand in other areas, the excess power is stored in the energy storage device. If the power generation y t Less than μ t -δ·σ t If the power generation is greater than the load demand and less than μ, the power generation in other regions will be dispatched through the cross-regional interconnection of the power grid. If the load demand cannot be met, the backup generator or temporary power generation facilities will be activated. If the power generation is greater than the load demand and less than μ t -δ·σ t When using smart grid technology, we can optimize power distribution and dispatch based on real-time data, dynamically monitor the relationship between load and power generation, and adjust the output of different generator sets in real time; The relative gap is expressed as, Among them, L t Indicates load demand.

8. A system using the method for evaluating and predicting renewable energy power generation based on multi-time scale analysis as claimed in any one of claims 1 to 7, characterized in that: It comprises a data acquisition module (100), a data preprocessing module (200), a multi-time scale analysis module (300), a feature fusion module (400) and a power generation evaluation module (500); The data acquisition module (100) is responsible for collecting real-time power generation data from sensors, smart meters and external sources; The data preprocessing module (200) is responsible for data cleaning and preprocessing; The multi-time scale analysis module (300) extracts the variation characteristics of power generation data through frequency domain conversion, wavelet transformation and seasonal decomposition methods according to different time scales; The feature fusion module (400) fuses the short-term fluctuation features, medium-term periodic features and long-term trend features extracted at different time scales to obtain a comprehensive feature vector; The power generation evaluation module (500) evaluates the power generation based on the output of the time series prediction model and in combination with a set threshold.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for evaluating and predicting new energy power generation based on multi-time scale analysis described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for evaluating and predicting new energy power generation based on multi-time scale analysis as described in any one of claims 1 to 7 are implemented.

Images