Renewable energy annual output time sequence generation method

Abstract

The invention discloses a method for generating an annual output time sequence of renewable energy, which belongs to the technical field of power system control, and comprises the following steps of: modeling a current residual component and a current trend component passing stability verification in a historical output time sequence by using an Irai random process; obtaining an initial simulation trend component and an initial simulation residual component; performing inverse difference processing on the periodic component and the target simulation trend component to obtain a target simulation trend component and a target simulation residual component, and finally reconstructing a target output time sequence by utilizing the target simulation trend component, the target simulation residual component and the periodic component. According to the scheme, mathematical modeling is carried out on the aperiodic component through the Irai random process, massive simulation curves can be generated, random features and time sequence correlation of the aperiodic component can be accurately reflected at the same time, key features of renewable energy under a long time scale can be accurately quantified, and the new energy consumption capability of a power grid can be accurately grasped according to the key features.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system operation planning, and more specifically, relates to a method for generating a year-round output time series of renewable energy. Background Art

[0002] As the country puts forward requirements for energy conservation, carbon reduction and carbon reduction, the proportion of renewable energy in new power systems is gradually increasing, and the significance of simulating the year-round time operation of renewable energy is gradually becoming apparent.

[0003] Most existing renewable energy scenario generation methods consider power forecasts on short time scales, day-ahead or intraday, and lack appropriate modeling methods on medium and long-term time scales, making it impossible to quantify the seasonal characteristics of renewable energy. In addition, most existing time series operation simulation methods are based on typical days, and the power system planning based on this can only ensure that the situation on typical days is feasible, and may not be able to consider some extreme scenarios. In fact, as the proportion of renewable energy increases, typical days are also constantly changing. In addition, the traditional 8760-hour time series operation simulation method needs to add time constraints throughout the year, which will result in excessive computation and low computational efficiency.

[0004] Therefore, existing modeling methods are difficult to accurately quantify the key characteristics of renewable energy over long time scales, which makes the power system's power balance face the risk of imbalance, is not conducive to the correct grasp of the grid's new energy absorption capacity, and cannot support the planning and development of large-scale power systems. Summary of the invention

[0005] In response to the above defects or improvement needs of the prior art, the present invention provides a method for generating a time series of renewable energy output throughout the year, which aims to solve the technical problem that existing modeling methods are difficult to accurately quantify the key characteristics of renewable energy on a long time scale.

[0006] To achieve the above object, according to one aspect of the present invention, a method for generating a renewable energy output time series throughout the year is provided, comprising:

[0007] S1: Decompose the historical output time series of renewable energy to obtain trend component, residual component and periodic component;

[0008] S2: Perform a stationarity check on the current residual component and the current trend component;

[0009] S3: If the stationarity check fails, the current trend component and the current residual component are differentiated to update the current trend component and the current residual component, and the process returns to S2 until the stationarity check is finally passed and S4 is entered;

[0010] S4: Obtaining the probability distribution of the current residual component and the current trend component;

[0011] S5: applying the probability distribution to the integral Fokker-Planck equation of the Ito random process, and using the Ito random process to model the current residual component and the current trend component to obtain an initial simulated trend component and an initial simulated residual component;

[0012] S6: performing reverse differentiation processing on the initial simulation trend component and the initial simulation residual component to obtain a target simulation trend component and a target simulation residual component;

[0013] S7: Reconstruct the target output time series using the target simulation trend component, the target simulation residual component and the periodic component.

[0014] In one embodiment, the S1 includes:

[0015] The historical output time series of renewable energy is decomposed by using an additive model or a multiplicative model to obtain the trend component, the residual component and a plurality of periodic components.

[0016] In one embodiment, the S1 includes:

[0017] S11: Finding potential periodic components from the historical output time series;

[0018] S12: Decoupling the potential periodic component according to the time series decomposition algorithm based on LOESS regression to obtain a corresponding periodic component, and removing the periodic component from the current historical output time series to update the historical output time series;

[0019] S13: Check whether there are other potential periodic components in the current historical output time series. If so, return to S12; otherwise, enter S14;

[0020] S14: Decomposing the trend component and the residual component from the current historical output time series.

[0021] In one embodiment, the S2 includes:

[0022] Performing unit root tests in various ways on the current trend component and the current residual component;

[0023] The unit root tests of the various modes include: ADF test and KPSS test. If the unit root tests of the various modes are all passed, it is considered that the stationarity test has been passed.

[0024] In one embodiment, the S5 includes:

[0025] S51: Modeling the current trend component using an Ito random process to obtain a first mapping that characterizes the relationship between a drift function, a diffusion function, and a probability distribution of the current trend component, bringing the probability distribution of the current trend component into the first mapping to obtain a drift function of the current trend component and a diffusion function of the current residual component, and using the two to reconstruct an initial simulated trend component;

[0026] S52: Use the Ito random process to model the current residual component to obtain a second mapping that characterizes the relationship between the drift function, diffusion function and probability distribution of the current residual component, and bring the probability distribution of the current residual component into the second mapping to obtain the drift function of the current residual component and the diffusion function of the current residual component, and use the two to reconstruct the initial simulated residual component.

[0027] In one embodiment, the S51 includes:

[0028] Using T(t)=T(t-1)+U T (t)*Δt+V T (t)*ΔW(t) models the current trend component T(t) to obtain a drift function U that characterizes the current trend component T (t), diffusion function V T (t) and probability distribution p T The first mapping of the relationship between (x, t) ΔW(t) is a random factor;

[0029] The probability distribution p of the current trend component T (x, t) is brought into the first mapping to obtain the drift function of the current trend component and the diffusion function of the current residual component, and the initial simulated trend component is reconstructed using the two.

[0030] In one embodiment, the S52 includes:

[0031] Using the formula R(t)=R(t-1)+U R (t)*Δt+V R (t)*ΔW(t) models the current residual component R(t) to obtain a drift function U that characterizes the current residual component R (t), diffusion function V R (t) and probability distribution p R The second mapping of the relationship between (x, t)

[0032] The probability distribution p of the current residual component R(x, t) is brought into the second mapping to obtain the drift function of the current residual component and the diffusion function of the current residual component, and the initial simulated residual component is reconstructed using the two.

[0033] In one embodiment, the step S1 further includes:

[0034] If the trend component and residual component obtained by S1 pass the stationarity check, S4 and S5 are executed, and the initial simulated trend component and the initial simulated residual component obtained by S5 are first differentiated and then back-differentiated to obtain the target simulated trend component and the target simulated residual component; finally, the target simulated trend component, the target simulated residual component and the periodic component are used to reconstruct the target output time series.

[0035] According to another aspect of the present invention, there is provided a device for generating a time series of annual output of renewable energy, comprising:

[0036] A decomposition module is used to decompose the historical output time series of renewable energy to obtain trend components, residual components and periodic components;

[0037] A verification module is used to verify the stability of the current residual component and the current trend component;

[0038] A loop module, for performing a difference processing on the current trend component and the current residual component to update the current trend component and the current residual component if the stationarity check fails, and returning to the check module until the stationarity check is finally passed and the acquisition module is entered;

[0039] An acquisition module, used for acquiring the respective probability distributions of the current residual component and the current trend component;

[0040] A modeling module, used for applying the probability distribution to the integral Fokker-Planck equation of the Ito random process, and using the Ito random process to model the current residual component and the current trend component to obtain an initial simulated trend component and an initial simulated residual component;

[0041] A calculation module, used for performing reverse differentiation processing on the initial simulation trend component and the initial simulation residual component to obtain a target simulation trend component and a target simulation residual component;

[0042] A reconstruction module is used to reconstruct a target output time series using the target simulation trend component, the target simulation residual component and the periodic component.

[0043] According to another aspect of the present invention, a renewable energy management system is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0044] According to another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0045] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0046] (1) This scheme provides a method for generating a renewable energy output time series throughout the year. The method uses the Ito random process to model the current residual component and the current trend component that have passed the stationarity check in the historical output time series to obtain an initial simulated trend component and an initial simulated residual component. The two are then subjected to reverse difference differentiation processing to obtain a target simulated trend component and a target simulated residual component. Finally, the target simulated trend component, the target simulated residual component and the periodic component are used to reconstruct the target output time series. This scheme uses the Ito random process to mathematically model the non-periodic component, which can generate a large number of simulation curves and accurately reflect its random characteristics and time series correlation. It can accurately quantify the key characteristics of renewable energy on a long time scale. Based on the above key characteristics, the new energy absorption capacity of the power grid can be accurately grasped, which is convenient for supporting the planning and development of large-scale power systems.

[0047] (2) This solution uses an additive model or a multiplicative model to decompose the historical output time series, and correspondingly uses an additive model or a multiplicative model to reconstruct the target output time series, which has low computational complexity and can improve the execution efficiency of the entire algorithm. Furthermore, this method can also generate multiple periodic components, such as annual components, quarterly components, monthly components, daily components, etc., to accurately characterize the key characteristics of renewable energy on a long time scale.

[0048] (3) This scheme first eliminates some easily identifiable periodic components to reduce the data size and computational complexity of the current historical output time series, and then further determines whether there are other potential periodic components. If so, the potential periodic components are decoupled according to the time series decomposition algorithm based on LOESS regression. This method can first find as many periodic components as possible, and secondly, the computational complexity is low and the algorithm execution efficiency is high.

[0049] (4) This scheme performs unit root tests on the current trend component and the current residual component in various ways, which can improve the accuracy of the stationarity test.

[0050] (5) This scheme adopts the Ito random process to model the current trend component and the current residual component respectively, obtains the corresponding first mapping and the second mapping, and brings the corresponding probability distribution into the two mappings, and finally reconstructs the initial simulated trend component and the initial simulated residual component; this method can accurately characterize the characteristics of the residual component and the trend component.

[0051] (6) This scheme uses T(t) = T(t-1) + U T (t)*Δt+V T (t)*ΔW(t) models the current trend component T(t) and also takes into account random factors, so that the trend component finally obtained is more accurate and closer to the actual application scenario.

[0052] (7) This scheme uses R(t)=R(t-1)+U R (t)*Δt+V R (t)*ΔW(t) models the current residual component R(t) and also takes into account random factors, so that the residual component finally obtained is more accurate and close to the actual application scenario.

[0053] (8) This scheme takes into account a special case. The trend component and residual component obtained by S1 can directly pass the stationarity check, and then execute the initial simulated trend component and the initial simulated residual component obtained by S4 and S5. However, the initial simulated trend component and the initial simulated residual component have not been differentiated. Therefore, they are first differentiated and then reversely differentiated to restore them to the time series of the simulation state, so as to reconstruct the target output time series with the periodic component. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a flow chart of a method for generating a renewable energy annual output time series provided by Example 1 of the present invention;

[0055] Figure 2 is a flow chart of another method for generating a renewable energy annual output time series provided in Example 1 of the present invention;

[0056] Figure 3 Schematic diagram of the probability distribution function of the mixed Gaussian distribution model provided in Example 1 of the present invention;

[0057] Figure 4 It is a comparison chart of the differential time series generated by the Ito process provided in Example 1 of the present invention and the differential time series obtained by differentiating the original time series;

[0058] Figure 5 It is a comparison diagram of the trend component generated by the Ito process provided in Example 1 of the present invention and the trend component in the original time series;

[0059] Figure 6 It is a schematic diagram of different components obtained by decomposing a wind speed time series according to the method for generating a renewable energy annual output time series provided in Example 1 of the present invention;

[0060] Figure 7 It is a schematic diagram of different components obtained by decomposing the solar radiation time series according to the method for generating the annual output time series of renewable energy provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0062] Example 1

[0063] like Figure 1 As shown, this embodiment provides a method for generating a renewable energy output time series throughout the year, including: S1-S7.

[0064] Specifically, S1: Decompose the historical output time series of renewable energy to obtain trend component, residual component and periodic component. S2: Perform stationarity check on the current residual component and the current trend component. S3: If the stationarity check fails, perform difference differentiation on the current trend component and the current residual component to update the current trend component and the current residual component, and return to S2 until the stationarity check is finally passed and enter S4. S4: Obtain the probability distribution of the current residual component and the current trend component. S5: Apply the probability distribution to the integral Focke-Planck equation of the Ito random process, and use the Ito random process to model the current residual component and the current trend component to obtain the initial simulated trend component and the initial simulated residual component. S6: Perform reverse differentiation on the initial simulated trend component and the initial simulated residual component to obtain the target simulated trend component and the target simulated residual component. S7: Reconstruct the target output time series using the target simulated trend component, the target simulated residual component and the periodic component.

[0065] As an optional implementation, after S1, it also includes: if the trend component and residual component obtained by S1 pass the stationarity check, then S4 and S5 are executed, and the initial simulated trend component and the initial simulated residual component obtained by S5 are first subjected to difference differentiation processing, and then subjected to reverse difference differentiation processing to obtain the target simulated trend component and the target simulated residual component. Finally, the target simulated trend component, the target simulated residual component and the periodic component are used to reconstruct the target output time series.

[0066] As an optional implementation, S1 includes: using an additive model or a multiplicative model to decompose the historical output time series of renewable energy to obtain a trend component, a residual component and multiple periodic components.

[0067] Among them, the MSTL method (Multiple-STL) is improved from the STL method (time series decomposition algorithm based on LOESS regression). After STL decomposition, the time series can be decomposed into trend component, periodic component, and residual component. After MSTL decomposition, the time series can be decomposed into trend component, multiple periodic components, and residual component. Among them, the multiple periodic components can include annual component, quarterly component, monthly component, and daily component. The trend component and residual component are non-periodic components. The STL method can only obtain one periodic component, while the MSTL method can obtain multiple.

[0068] It should be noted that for the MSTL method, the additive model and the multiplicative model are as follows:

[0069] X(t)=T(t)+S 1 (t)+...+S n (t)+R(t);

[0070] X(t)=T(t)×S 1 (t)×...×S n (t)×R(t).

[0071] If the additive model is used for decomposition, S7 is the accumulation of all components to form a complete renewable energy time series throughout the year to complete the modeling. The accumulation formula is as follows: X(t) = T(t) + S 1 (t)+S 2 (t)+R(t); By repeating the whole process continuously, a large number of time series curves can be obtained.

[0072] As an optional implementation, S1 includes: S11: Find the potential periodic component from the historical output time series. S12: Decouple the potential periodic component according to the time series decomposition algorithm based on LOESS regression to obtain the corresponding periodic component, and remove the periodic component from the current historical output time series to update the historical output time series. S13: Detect whether there are other potential periodic components in the current historical output time series, and if so, return to S12. Otherwise, enter S14. S14: Decompose the trend component and the residual component from the current historical output time series.

[0073] It should be noted that the STL method is divided into an inner loop and an outer loop. The steps of the inner loop are as follows:

[0074] (1) Set the initial value of the trend component and the initial value of the robust weight factor.

[0075] (2) Subtract the trend component from the complete time series to obtain the detrended time series X′(t), X′(t) = X(t) - T(t).

[0076] (3) Under the initial value of the robust weight factor, LOESS regression is used to smooth all periodic sequences in X′(t) to obtain the temporary periodic component X T (t).

[0077] (4) For the temporary periodic component X T (t) Perform low-pass filtering to remove the periodic features and obtain the low-pass component X L (t).

[0078] (5) Perform residual processing on the smoothed periodic sequence to obtain the periodic component S(t), S(t) = X T (t)-X L (t).

[0079] (6) Subtract the periodic component from the original time series to obtain the time series T without the periodic component. o (t), namely the initial trend component T o (t) = X(t) - S(t).

[0080] (7) T o (t) is smoothed by LOESS to obtain the trend component T(t). The inner loop ends here, and the trend component and the period component are obtained.

[0081] The steps of the outer loop are as follows: subtract the trend component and the periodic component from the original time series to obtain the residual component R(t)=X(t)-T(t)-S(t), and test its size.

[0082] If it is too large, it is adjusted and then returns to the inner loop. At this time, the trend component and robust weight factor have been updated, and the inner loop steps are repeated to obtain new trend components and residual components. If the size meets the requirements, the trend component, period component, and residual component can be output to complete the decomposition.

[0083] As an optional implementation, S2 includes: performing unit root tests in various ways on the current trend component and the current residual component. Among them, the various unit root tests include: ADF test (extended Dickey-Fuller test) and KPSS test (Yatkowski-Phillips-Schmidt-Shin test). If the unit root tests in various ways are all passed, it is considered to have passed the stationarity test.

[0084] Among them, before using the Ito random process to process the trend component and the residual component, it is necessary to perform a stationarity check on the time series. If the unit root test of multiple methods is not satisfied, it means that it is not stationary. It is necessary to differentiate the time series to obtain the differentiated time series. If it is still not satisfied, the difference is continued until the stationarity check is satisfied. The difference formula is as follows: T′(t)=T(t)-T(t-1). If the original time series is differentiated during the stationarity check, the inverse differentiation process of the same order is applied to the differentiated time series in S6 to obtain the simulated non-periodic components, that is, the original time series of the trend component and the residual component.

[0085] As an optional implementation, S5 includes: S51: using the Ito random process to model the current trend component to obtain a first mapping that characterizes the relationship between the drift function, diffusion function and probability distribution of the current trend component, bringing the probability distribution of the current trend component into the first mapping, obtaining the drift function of the current trend component and the diffusion function of the current residual component, and using the two to reconstruct the initial simulated trend component. S52: using the Ito random process to model the current residual component to obtain a second mapping that characterizes the relationship between the drift function, diffusion function and probability distribution of the current residual component, bringing the probability distribution of the current residual component into the second mapping, obtaining the drift function of the current residual component and the diffusion function of the current residual component, and using the two to reconstruct the initial simulated residual component.

[0086] Since the probability distribution of random variables needs to be obtained first when using the Ito random process to generate simulated time series, the probability distribution of trend components and residual components is obtained using the mixed Gaussian distribution model. The general formula for the probability density of the mixed Gaussian distribution model is as follows: By adjusting the parameter a j ,b j ,c j , which can be used to fit different data.

[0087] Among them, the periodic component is relatively easy to model, so only the Ito random process is used to model the trend component and the residual component, which can simulate the random changes of the trend component and the residual component and retain the time series correlation. The initial description of the Ito random process is as follows.

[0088] The form of a typical Ito random process is as follows: It can be described as the sum of an integral over time and an integral over Brownian motion. Among them, X(0) is the initial value, and the drift function U(t) and the diffusion function V(t) are both time-varying functions. The drift function U(t) reflects the drift of the random variable value around the expected value, and the diffusion function V(t) reflects the influence of the random factor W(t) on the random variable, reflecting the net change in the value of the random variable.

[0089] The general form of the Ito random process is rewritten into a differential form as follows: dX(t) = U(t)dt + V(t)dW. In fact, it is generally impossible to obtain a continuous curve for the output of renewable energy, so the differential form is rewritten into a discrete form as follows: X(t)-X(t-1) = U(t)*Δt + V(t)*ΔW(t); and after determining the probability distribution of the random variable p(x, t), the drift function U(t) and the diffusion function V(t) satisfy the Fokker-Planck equation as follows:

[0090]

[0091] By integrating the above formula, we can get the relationship between the drift function, diffusion function, and probability distribution of random variables as follows:

[0092] As an optional implementation, S51 includes: using T(t)=T(t-1)+U T (t)*Δt+V T (t)*ΔW(t) models the current trend component T(t) and obtains the drift function U that represents the current trend component T (t), diffusion function V T (t) and probability distribution p T The first mapping of the relationship between (x, t) ΔW(t) is a random factor. The probability distribution of the current trend component p T (x, t) is brought into the first mapping to obtain the drift function of the current trend component and the diffusion function of the current residual component, and the two are used to reconstruct the initial simulation trend component.

[0093] Here, the probability distribution of the trend component and the residual component has been obtained using the mixed Gaussian distribution model, and the sequence after differentiation has satisfied the stationarity check, so the Ito random process can be used to model the trend component and the residual component. The differential equation is defined as follows:

[0094] Specifically, the discretized Ito random process is applied to the trend component, and we get: T(t) = T(t-1) + U T (t)*Δt+V T (t)*ΔW(t); the drift function of the trend component is expressed as: U T (t) = -θ T (T(t)-T avg );Use the mixed Gaussian distribution model to fit the probability distribution of the trend component:

[0095] The first mapping relationship between the drift function, diffusion function, and probability distribution of the trend component is as follows:

[0096] As an optional implementation, S52 includes: using the formula R(t)=R(t-1)+U R (t)*Δt+V R (t)*ΔW(t) models the current residual component R(t) to obtain the drift function U that characterizes the current residual component R (t), diffusion function V R (t) and probability distribution p R The second mapping of the relationship between (x, t) The probability distribution p of the current residual component R (x, t) is brought into the second mapping to obtain the drift function of the current residual component and the diffusion function of the current residual component, and the initial simulated residual component is reconstructed using the two.

[0097] Specifically, the discretized Ito random process is applied to the residual component, and we get: R(t)=R(t-1)+U R (t)*Δt+V R (t)*ΔW(t); the drift function of the residual component is expressed as: U R (t) = -θ R (R(t)-R avg );Use the mixed Gaussian distribution model to fit the probability distribution of the residual component and get: The second mapping relationship between the drift function, diffusion function and probability distribution of the residual component is:

[0098] By solving the above differential equations, the differentiated simulated trend component and residual component can be generated.

[0099] The following is a method for generating a renewable energy output time series throughout the year provided by the present invention, and a decomposition test is performed on the wind speed time series and photovoltaic radiation time series in a certain area. Two cases are set to illustrate the effectiveness of the present invention:

[0100] First case: historical time series;

[0101] Second case: 1000 simulated time series are generated according to the method for generating the annual output time series of renewable energy proposed by the present invention.

[0102] like Figure 2 The process shown in the figure decomposes the wind speed time series and photovoltaic radiation time series of a certain area. The method for generating the annual output time series of renewable energy proposed in the second case specifically includes the following steps:

[0103] S1. Using the MSTL method, the historical time series of renewable energy is decomposed to obtain trend component, annual component, periodic component and residual component;

[0104] S2. Perform stationarity test on historical time series, namely ADF test and KPSS test.

[0105] S3. If these two tests are not satisfied, it means that the system is not stable. It is necessary to differentiate the time series of the trend component and the residual component to obtain the differentiated time series. If it is still not satisfied, continue to differentiate until the stability test is satisfied. The differential formula is as follows:

[0106] T′(t)=T(t)-T(t-1);

[0107] S4. The probability distribution of the trend component and the residual component is obtained by fitting the mixed Gaussian distribution model. The general formula of the probability density of the mixed Gaussian distribution model is as follows: By adjusting the parameter a j ,b j ,c j , which can be used to fit different data. Figure 3 It is a schematic diagram of the probability distribution function of the mixed Gaussian distribution model.

[0108] S5. Use Ito random process to model the trend component and residual component to generate differential trend time series and residual time series; Figure 4 It is a comparison chart of the difference time series generated by the Ito process and the difference time series obtained by differentiating the original time series; Figure 5 It is a comparison chart of the trend component generated by the adopted Ito process and the trend component in the original time series.

[0109] S6. If the original time series is differentiated during the stationarity check, then the inverse differentiation process of the same order is applied to the differentiated time series here to obtain the original time series of the simulated trend component and the residual component.

[0110] S7. Accumulate all components to form a complete renewable energy time series throughout the year and complete the modeling. The accumulation formula is as follows: X(t) = T(t) + S 1 (t)+S 2 (t)+R(t), and the whole process is repeated continuously to obtain a large number of time series curves.

[0111] The two statistics of mean and standard deviation are selected to verify the accurate extraction of component characteristics of the renewable energy simulation time series curve generated by the method for generating the renewable energy output time series throughout the year proposed in the present invention. Through the above steps, a total of 1000 simulated time series are generated in the second case. From the calculation results in the following table, it can be seen that the mean and standard deviation of the original time series in the first case are 5.79 and 0.29 respectively, and the mean and standard deviation of the simulated time series in the second case are 5.84 and 0.23 respectively. It can be seen that the mean and variance obtained in the first case and the second case are almost the same, which proves the accuracy and effectiveness of the full-year time series decomposition modeling method proposed in the present invention, and verifies its accurate characterization of the characteristic components of the renewable energy time series.

[0112] Statistics First case Second case average value 5.79 5.84 Standard Deviation 0.29 0.23

[0113] in, Figure 6 This is a schematic diagram of different components obtained by decomposing the wind speed time series based on the generation method of the annual output time series of renewable energy; Figure 7 The schematic diagram of different components obtained by decomposing the solar radiation time series based on the generation method of the renewable energy annual output time series also further verifies that the proposed method for generating the renewable energy annual output time series can accurately capture the characteristics of renewable energy and generate a long-time scale simulation time series similar to the original historical time series.

[0114] Example 2

[0115] This embodiment provides a device for generating a renewable energy output time series throughout the year, including: a decomposition module, a verification module, a circulation module, an acquisition module, a modeling module, a calculation module and a reconstruction module.

[0116] The decomposition module is used to decompose the historical output time series of renewable energy to obtain trend components, residual components and periodic components.

[0117] The verification module is used to perform a stationarity check on the current residual component and the current trend component.

[0118] The loop module is used to perform differentiation processing on the current trend component and the current residual component to update the current trend component and the current residual component if the stationarity check fails, and return to the check module until the stationarity check is finally passed and the acquisition module is entered.

[0119] The acquisition module is used to obtain the probability distribution of the current residual component and the current trend component.

[0120] The modeling module is used to apply the probability distribution to the integral Fokker-Planck equation of the Ito random process, and use the Ito random process to model the current residual component and the current trend component to obtain the initial simulated trend component and the initial simulated residual component.

[0121] The calculation module is used for performing reverse differentiation processing on the initial simulation trend component and the initial simulation residual component to obtain the target simulation trend component and the target simulation residual component.

[0122] The reconstruction module is used to reconstruct the target output time series using the target simulation trend component, the target simulation residual component and the periodic component.

[0123] Example 3

[0124] This embodiment provides a renewable energy management system, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0125] Example 4

[0126] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are implemented.

[0127] It will be appreciated by those skilled in the art that 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. Moreover, 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 codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0128] 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 flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, 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 flowchart and / or block diagram. 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.

[0129] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate 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 A function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0131] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the present invention is intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0132] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the protection scope of the present invention, the present invention is also intended to include these modifications and variations.

Claims

1. A method for generating a time series of annual output of renewable energy, characterized in that: include: S1: Decompose the historical output time series of renewable energy to obtain trend component, residual component and periodic component; S2: Perform a stationarity check on the current residual component and the current trend component; S3: If the stationarity check fails, the current trend component and the current residual component are differentiated to update the current trend component and the current residual component, and the process returns to S2 until the stationarity check is finally passed and S4 is entered; S4: Obtaining the respective probability distributions of the current residual component and the current trend component; S5: applying the probability distribution to the integral Fokker-Planck equation of the Ito random process, and using the Ito random process to model the current residual component and the current trend component to obtain an initial simulated trend component and an initial simulated residual component; S6: performing reverse differentiation processing on the initial simulation trend component and the initial simulation residual component to obtain a target simulation trend component and a target simulation residual component; S7: Reconstruct the target output time series using the target simulation trend component, the target simulation residual component and the periodic component.

2. The method for generating a renewable energy annual output time series according to claim 1, characterized in that: The S1 includes: The historical output time series of renewable energy is decomposed by using an additive model or a multiplicative model to obtain the trend component, the residual component and a plurality of the periodic components.

3. The method for generating a renewable energy annual output time series according to claim 2, characterized in that: The S1 includes: S11: Finding potential periodic components from the historical output time series; S12: Decoupling the potential periodic component according to the time series decomposition algorithm based on LOESS regression to obtain the corresponding periodic component, and removing the periodic component from the current historical output time series to update the historical output time series; S13: Check whether there are other potential periodic components in the current historical output time series. If so, return to S12; otherwise, enter S14; S14: Decomposing the trend component and the residual component from the current historical output time series.

4. The method for generating a renewable energy annual output time series according to claim 1, characterized in that: The S2 includes: Performing unit root tests in various ways on the current trend component and the current residual component; Among them, the multiple unit root tests include: ADF test (extended Dickey-Fuller test) and KPSS test (Yatkowski-Phillips-Schmidt-Shin test). If the unit root tests of multiple methods are all passed, it is considered that the stationarity test has passed.

5. The method for generating a renewable energy annual output time series according to claim 1, characterized in that: The S5 includes: S51: Modeling the current trend component using an Ito random process to obtain a first mapping that characterizes the relationship between a drift function, a diffusion function, and a probability distribution of the current trend component, bringing the probability distribution of the current trend component into the first mapping to obtain a drift function of the current trend component and a diffusion function of the current residual component, and using the two to reconstruct an initial simulated trend component; S52: Use the Ito random process to model the current residual component to obtain a second mapping that characterizes the relationship between the drift function, diffusion function and probability distribution of the current residual component, and bring the probability distribution of the current residual component into the second mapping to obtain the drift function of the current residual component and the diffusion function of the current residual component, and use the two to reconstruct the initial simulated residual component.

6. The method for generating a renewable energy annual output time series according to claim 5, characterized in that: The S51 includes: Using T(t)=T(t-1)+U T (t)*Δt+V T (t)*ΔW(t) models the current trend component T(t) to obtain a drift function U that characterizes the current trend component T (t), diffusion function V T (t) and probability distribution p T The first mapping of the relationship between (x, t) ΔW(t) is a random factor; The probability distribution p of the current trend component T (x, t) is brought into the first mapping to obtain the drift function of the current trend component and the diffusion function of the current residual component, and the initial simulated trend component is reconstructed using the two.

7. The method for generating a renewable energy annual output time series according to claim 5, characterized in that: The S52 includes: Using the formula R(t)=R(t-1)+U R (t)*Δt+V R (t)*ΔW(t) models the current residual component R(t) to obtain a drift function U that characterizes the current residual component R (t), diffusion function V R (t) and probability distribution p R The second mapping of the relationship between (x, t) ΔW(t) is a random factor; The probability distribution p of the current residual component R (x, t) is brought into the second mapping to obtain the drift function of the current residual component and the diffusion function of the current residual component, and the initial simulated residual component is reconstructed using the two.

8. The method for generating a renewable energy annual output time series according to any one of claims 1 to 7, characterized in that: The S1 also includes: If the trend component and residual component obtained by S1 pass the stationarity check, S4 and S5 are executed, and the initial simulated trend component and the initial simulated residual component obtained by S5 are first differentiated and then back-differentiated to obtain the target simulated trend component and the target simulated residual component; finally, the target simulated trend component, the target simulated residual component and the periodic component are used to reconstruct the target output time series.

9. A device for generating a time series of annual output of renewable energy, characterized in that: include: A decomposition module is used to decompose the historical output time series of renewable energy to obtain trend components, residual components and periodic components; A verification module is used to verify the stability of the current residual component and the current trend component; A loop module, for performing a difference processing on the current trend component and the current residual component to update the current trend component and the current residual component if the stationarity check fails, and returning to the check module until the stationarity check is finally passed and the acquisition module is entered; An acquisition module, used for acquiring the respective probability distributions of the current residual component and the current trend component; A modeling module, used for applying the probability distribution to the integral Fokker-Planck equation of the Ito random process, and using the Ito random process to model the current residual component and the current trend component to obtain an initial simulated trend component and an initial simulated residual component; A calculation module, used for performing reverse differentiation processing on the initial simulation trend component and the initial simulation residual component to obtain a target simulation trend component and a target simulation residual component; A reconstruction module is used to reconstruct a target output time series using the target simulation trend component, the target simulation residual component and the periodic component.

10. A renewable energy management system, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

Images