Day-ahead dispatching method for hydro-wind hybrid power generation system based on dynamic frequency constraints
By establishing a dynamic frequency constraint in the water-wind complementary power generation system, the working status of the hydropower unit is optimized, and the impact of wind power generation power fluctuations on the grid frequency is solved, and the stability and reliability of the power system are improved.
Patent Information
- Application Number
- CN202110932099.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-08-13
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2041-08-13
AI Technical Summary
The prior art is difficult to effectively predict the power fluctuations of wind power generation, resulting in unstable grid frequency and affecting the safety and reliability of the power system.
Using the water-wind complementary power generation system recently dispatching method based on dynamic frequency constraints, a minimum water consumption model containing dynamic frequency response constraints is established by establishing a wind power uncertainty model and building linearized and nonlinearized constraints of output-related variables of hydropower units. Determine the working state of the water-power unit, optimize the switching state and output magnitude of the water-power unit to compensate for the uncertainty of wind power.
Accurate prediction of wind power generation power is achieved, the continuity and reliability of water and wind complementary power generation system is ensured, the burden of grid frequency adjustment is reduced, and the stability and power supply reliability of the power system are improved.
Smart Images

Figure CN113659620B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of hydropower and new energy complementary power generation, in particular to a day-ahead dispatching method for a hydro-wind complementary power generation system based on dynamic frequency constraints. Background Art
[0002] With the development of my country's economy and the improvement of people's living standards, energy consumption is increasing. In order to alleviate the increasingly serious energy and environmental problems, the country has vigorously developed clean energy such as hydropower and wind power. Energy and environmental issues are the main obstacles to my country's economic and social development, and the comprehensive utilization of renewable energy will play an increasingly important role.
[0003] At present, wind energy and hydropower are ideal and technologically mature renewable energy sources. However, the utilization of wind energy and hydropower is greatly affected by seasonality and regionality. For example, wind energy has strong intermittency and volatility. When large-scale wind power energy is connected, due to the decoupling of wind turbines and the power grid, the frequency regulation capability of the system decreases. Traditional dispatching methods cannot guarantee the complete absorption of wind power energy and the stability of system frequency. Therefore, the reliability of power supply is not high, which also affects the safe, stable and reliable operation of the power system.
[0004] Hydropower can effectively make up for the shortcomings of wind power generation. On the one hand, hydropower stations use their peak-to-valley performance to reduce the peak-to-valley difference of the system, optimize the working positions of other power sources, and improve the economic efficiency of the overall operation of the system. On the other hand, hydropower units are flexible to start and stop and respond quickly, and can achieve conversion between different working conditions in a short time, meet the requirements of rapid regulation of the power system, and effectively improve the power supply quality and safety and stability of the power system.
[0005] If we fully utilize the complementarity of wind and water, and make probabilistic, interval-based or scenario-based predictions on the randomness of wind power when considering the scheduling problem of wind power access, and work in different modes according to load demand, wind energy and hydropower can be combined to achieve a balance between electricity consumption, power supply and energy storage, greatly improve the continuity and reliability of power supply, and significantly expand the scope of economical energy utilization, it will have a high promotion value.
[0006] At present, in the wind and water complementary power generation, many scholars focus on the complementary output of water conservancy and wind power generation to achieve the actual load demand of the power system and ensure the feasibility of the power system operation. However, due to the uncertainty of resources, there will be an imbalance between the power generation and power load of the wind power system. This imbalance can be compensated by hydropower generation, but it does not take into account the fluctuation of power generation frequency caused by the instability of the wind power system. Frequency fluctuation is one of the important indicators to measure the quality of electric energy. Small frequency fluctuations may affect the normal use of electrical appliances, and large frequency fluctuations will endanger the normal operation of the generator. Therefore, it is very important to maintain frequency stability in the power system.
[0007] Therefore, how to make a more accurate prediction of wind power generation and use hydropower for compensation and regulation is an urgent problem that technicians in this field need to solve. Summary of the invention
[0008] In view of the above problems, the present invention provides a day-ahead scheduling method for a hydro-wind complementary power generation system based on dynamic frequency constraints, which at least solves some of the above technical problems. The method can accurately predict the power of wind power generation, ensure the continuity and reliability of the hydro-wind complementary power generation system when it is connected to the power grid, and reduce the frequency regulation burden of the power grid.
[0009] The embodiment of the present invention provides a method for day-ahead scheduling of a hydro-wind complementary power generation system based on dynamic frequency constraints, including:
[0010] S1. Establish a model to describe wind power uncertainty: collect historical data, obtain the probability density function of the prediction error through probability prediction, and then obtain the confidence interval of wind power output under various confidence levels through interval prediction;
[0011] S2. Construct linear and nonlinear constraints on the output-related variables of the hydropower units;
[0012] S3, establishing a water consumption minimization model with dynamic frequency response constraints;
[0013] S4. Determine the working state of the hydropower unit according to the wind power output confidence interval, the linear constraints and nonlinear constraints of the hydropower unit output-related variables, and the minimum water consumption model.
[0014] Furthermore, the S1 specifically includes:
[0015] S11, obtaining historical wind power actual values and corresponding wind power forecast values;
[0016] S12, dividing the actual wind power value and the corresponding predicted wind power value according to different time periods, so as to obtain average actual wind power values and corresponding average predicted wind power values in multiple different time periods, and calculating the prediction error for each time period;
[0017] S13, extracting the maximum value and the minimum value of the average wind power prediction value to obtain a wind power prediction range;
[0018] S14, dividing the wind power prediction range into equal intervals to obtain n wind power prediction intervals;
[0019] S15, grouping the average wind power prediction value of each time period and the prediction error of the corresponding time period; and placing each group of data into the corresponding wind power prediction interval according to the size of the average wind power prediction value of the group; each wind power prediction interval constitutes a data set;
[0020] S16, using a kernel density estimation algorithm to calculate the probability density function of the prediction error in each of the data sets;
[0021] S17. Calculate confidence intervals of different confidence levels in the corresponding wind power prediction interval using the probability density function.
[0022] Furthermore, in S2, the linearization constraints of the hydropower unit output-related variables include: water balance constraints, reservoir flow constraints, reservoir water level constraints, reservoir storage capacity constraints, power balance constraints, hydropower unit output constraints, hydropower unit flow constraints, minimum startup and shutdown time constraints, maximum startup times constraints, unit generating head constraints, and spare capacity constraints;
[0023] In S2, the nonlinear constraints on the output-related variables of the hydropower unit include: water level-reservoir capacity constraint, tailwater level-discharge flow constraint, head loss constraint, hydropower unit power generation characteristic constraint, and hydropower unit vibration zone limit constraint.
[0024] Furthermore, in the linearization constraint of the hydropower unit output related variables, the hydropower unit output constraint equation is:
[0025]
[0026]
[0027] in, is the upper limit of the output of hydropower unit i; P i is the lower limit of the output of the i-th hydropower unit, in MW; u i,t is the operating status of hydropower unit i during period t; P i,t is the output of the i hydropower unit in period t; 0 means the unit is in shutdown state, and 1 means it is in startup state.
[0028] Furthermore, in the nonlinear constraints of the output-related variables of the hydropower unit, the relationship function between the reservoir water level and the flow in the water level-reservoir capacity constraint, the relationship function between the tail water level and the downstream flow in the tail water level-discharge flow constraint, the relationship function between the head loss and the downstream flow in the head loss constraint, and the output characteristic function of the hydropower unit in the unit power generation characteristics are piecewise linearized.
[0029] Furthermore, the S3 specifically includes:
[0030] S31, constructing a frequency response constraint equation;
[0031] S32, constructing a minimum objective function for water consumption in power generation;
[0032]
[0033]
[0034] Among them, Q i,t is the flow rate of the i-th unit in period t, in m 3 / s;Q sum t is the total water consumption of the units during period t; Q s t is the power station water discharge flow rate during period t, in m 3 / s; I is the set of hydropower units; T is the set of time periods; Δt is the duration of each time period, in h;
[0035] S33. Constructing a dynamic frequency response constraint according to the output constraint equation of the hydropower unit.
[0036] Furthermore, a piecewise multivariate least squares fitting method is used to linearize formula (14) in the output constraint equation of the hydropower unit;
[0037] The specifics of the piecewise multivariate least squares fitting method are:
[0038] Assume that the left side of formula (14) Let f(X) be
[0039] The f(X) is segmented to obtain multiple initial segmentation intervals, that is,
[0040] Set the upper limit of the number of iterations l max ;
[0041] Determine whether each initial segment interval satisfies Where j represents the segment; the optimal segment interval is obtained;
[0042] Combining the minimum startup and shutdown time constraints in the linearization constraints of the output-related variables of the hydropower unit and the reserve capacity constraints, formula (14) is transformed to obtain a new frequency constraint condition:
[0043] π 1 ≤t 1 ≤π 1 +v 1 M
[0044] π 2 ≤t 1 ≤π2 +(1-v 1 )M
[0045] t 1 ≤t 2 ≤t 1 +v 2 M
[0046] π 3 ≤t 2 ≤π 3 +(1-v 2 )M
[0047] L
[0048] t k-2 ≤t k-1 ≤t k-2 +v k-1 M
[0049] π k ≤t k-1 ≤π k +(1-v k-1 )M
[0050] t k-1 ≤f ΔUFLS (47)
[0051] Among them, M is a large constant; v is an auxiliary binary variable of M; π and t are both intermediate variables.
[0052] Compared with the prior art, the method for dispatching the hydro-wind complementary power generation system based on dynamic frequency constraints recorded in the present invention fully considers the impact of wind power output uncertainty on the frequency of the power system, and at the same time satisfies physical constraints or restrictions such as reservoir operating conditions, power station operating conditions, and power grid operating constraints. It can optimize the switch state and output size of each hydropower unit under different wind power prediction confidence levels to achieve the goal of minimizing water consumption. The present invention ensures the continuity and reliability of the hydro-wind complementary power generation system when it is incorporated into the power grid, and reduces the burden of power grid frequency regulation.
[0053] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.
[0054] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0055] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:
[0056] Figure 1 A flow chart of a method for day-ahead scheduling of a hydro-wind complementary power generation system based on dynamic frequency constraints provided in an embodiment of the present invention.
[0057] Figure 2 A wind power prediction interval diagram provided by an embodiment of the present invention.
[0058] Figure 3 The probability density function of the prediction error in each interval provided by the embodiment of the present invention.
[0059] Figure 4 Confidence intervals of different confidence levels in the wind power prediction interval provided by an embodiment of the present invention.
[0060] FIG5( a ) is a comparison diagram of optimization scheduling results without considering dynamic frequency constraints provided by an embodiment of the present invention.
[0061] FIG5( b ) is a comparison diagram of optimization scheduling results considering dynamic frequency constraints provided by an embodiment of the present invention.
[0062] Figure 6 A diagram of changes in the limit value of the system frequency in each scheduling period provided by an embodiment of the present invention.
[0063] Figure 7 A diagram of frequency fluctuations in a power system provided by an embodiment of the present invention.
[0064] Figure 8 A frequency drop diagram provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0065] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.
[0066] See also Figure 1 As shown, the embodiment of the present invention provides a method for day-ahead scheduling of a hydro-wind complementary power generation system based on dynamic frequency constraints, which specifically includes the following steps:
[0067] S1. Establish a model to describe wind power uncertainty: collect historical data, obtain the probability density function of the prediction error through probability prediction, and then obtain the confidence interval of wind power output under various confidence levels through interval prediction;
[0068] S2. Construct linear and nonlinear constraints on the output-related variables of the hydropower units;
[0069] S3, establishing a water consumption minimization model with dynamic frequency response constraints;
[0070] S4. Determine the working state of the hydropower unit according to the wind power output confidence interval, the linear constraints and nonlinear constraints of the hydropower unit output-related variables, and the minimum water consumption model.
[0071] The present invention aims to use hydropower to compensate and adjust wind power after successfully describing the uncertainty of wind power during the day-ahead scheduling of Feng Shui, so as to meet the load demand of the power system, take the minimum water consumption as the goal, and introduce dynamic frequency response constraints at the same time, adjust the frequency fluctuations caused by the uncertainty of wind power, thereby reducing the harm of frequency fluctuations to the power system and ensuring its smooth operation. In order to describe this scheme more clearly below, it is introduced from two stages: the first stage is to establish a model to describe the uncertainty of wind power, and determine the output and fluctuation of wind power in the day-ahead scheduling; the second stage is to establish a minimum water consumption model with dynamic frequency response constraints. The two stages are linked by the working status of the hydropower unit, and the working status of the hydropower unit is finally determined.
[0072] Next, we will introduce the first and second stages in detail:
[0073] (1) Phase I: Wind power uncertainty description method - wind power range estimation
[0074] The uncertainty of wind power output is one of the main reasons for load imbalance and further frequency fluctuation. It is particularly important to accurately describe this uncertainty in day-ahead dispatching. Many scholars have proposed different methods for describing wind power uncertainty, such as analytical methods based on point prediction, fitting methods based on interval prediction, and simulation methods based on scenario prediction. The planning and operation of power systems involve a large number of uncertain factors, but how to deal with these factors in the analysis and control of power systems still needs in-depth research. Therefore, when describing the uncertainty of wind power, it is inevitable that there will be deviations from the actual situation. At present, the prediction methods for uncertain energy mainly include interval prediction, probability prediction and scenario prediction. Scenario prediction outputs multiple possible value scenarios of the predicted object in the future, reflecting the uncertainty of the predicted object value. However, the number of scenarios is limited, and most of them are general typical scenarios after scenario reduction, and some extreme scenarios will be filtered out. Therefore, this paper combines interval prediction and probability prediction, obtains the probability density function of the prediction error by statistically analyzing historical data, and then obtains the confidence interval of wind power output at each confidence level through interval prediction.
[0075] Since the uncertainty of wind power is related to its output power, the higher the output power, the greater the prediction error; therefore, in order to accurately describe the confidence interval, it is necessary to divide the wind power prediction data into intervals, and use the kernel density estimation method to calculate the confidence intervals under different confidence levels for each interval. The specific steps are as follows:
[0076] S11, obtaining historical wind power actual values and corresponding wind power forecast values;
[0077] S12, dividing the actual wind power value and the corresponding predicted wind power value according to different time periods, so as to obtain average actual wind power values and corresponding average predicted wind power values in multiple different time periods, and calculating the prediction error for each time period;
[0078] For the above S11-S12, in this embodiment, the actual wind power value and the corresponding wind power forecast value are taken for each day in 2017, and divided every 15 minutes, so that the average actual wind power value and the corresponding average wind power forecast value of 96 time periods can be obtained every day, and the average actual wind power value and the corresponding average wind power forecast value of 35045 time periods can be obtained in the whole year of 2017; then the prediction error of each time period is calculated;
[0079] S13, extracting the maximum value and the minimum value of the average wind power prediction value to obtain a wind power prediction range;
[0080] S14, dividing the wind power prediction range into equal intervals to obtain n wind power prediction intervals, as shown in the accompanying drawings. Figure 2 ;
[0081] S15, grouping the average wind power prediction value of each time period and the prediction error of the corresponding time period; and placing each group of data into a corresponding wind power prediction interval according to the size of the average wind power prediction value of the group; each wind power prediction interval constitutes a data set;
[0082] For the above S13-S15, in this embodiment, the average wind power prediction value of each time period and the prediction error of the corresponding time period are grouped together; thereby 35045 groups are obtained; these 35045 groups of data are all placed into corresponding wind power prediction intervals according to the size of the average wind power prediction value in the group; based on this, each wind power prediction interval constitutes a data set;
[0083] S16, using a kernel density estimation algorithm to calculate the probability density function of the prediction error in each of the data sets, see the attached drawings for details Figure 3 ;
[0084] Among them, the probability density function is recorded as F(ξ), where ξ is the prediction error variable, and the kernel density function selects the Gaussian kernel function;
[0085] S17, calculating the confidence intervals of different confidence levels in the corresponding wind power prediction interval by using the probability density function, with specific reference to the accompanying drawings Figure 4 ;
[0086] In this embodiment, the confidence level is set to β. When the confidence level is 1-β, the prediction interval is [P pred +F -1 (β 1 ),P pred +F -1 (β 2 )], where P pred is the predicted value; F -1 (ξ) is the inverse function of the probability density distribution function F(ξ); F -1 (β 1 ) is the minimum value of the prediction error variable; F -1 (β 2 ) is the maximum value of the prediction error variable; β 1 =β / 2,β 2 =1-β / 2;
[0087] Then, the uncertain model describing wind power is verified, and the optimization dispatch results considering and not considering frequency constraints are compared:
[0088] When the wind power output level is high and the residual load is small, due to the minimum output limit of hydropower, multiple units cannot be operated simultaneously. In order to ensure frequency safety, part of the wind power output is cut off and compensated by hydropower. For details, please refer to Figure 5, where Figure 5 (a) shows the case without considering the dynamic frequency constraint; Figure 5 (b) shows the case with considering the dynamic frequency constraint;
[0089] After considering the frequency constraint, the limit value of the system frequency in each scheduling cycle is reduced. Please refer to the attached figure for details. Figure 6 .
[0090] (2) Second stage: Day-ahead optimization model
[0091] Derivation and linearization of dynamic frequency response model:
[0092] The fluctuation of power system frequency is caused by the deviation between power generation and load due to load fluctuation or random energy output power changes such as wind and solar. Figure 7The primary frequency adjustment process is divided into an inertial response stage and a primary frequency regulation recovery stage. For example, when the output of the power generation side suddenly exceeds the load at a certain moment, the frequency deviation f' value gradually increases from 0, and a part of the excess electric energy is converted into rotor mechanical energy. This stage is the inertial response stage. When f' exceeds the boundary of the primary frequency regulation dead zone, the frequency regulation unit operates to reduce the unit output by adjusting the guide vane opening, so that the mechanical torque gradually equals the electromagnetic torque, suppressing the frequency from further increasing, and finally reaching a new balance. This stage is the primary frequency regulation stage. The inertial response stage is generally 2-3s, and the primary frequency regulation recovery time is generally 1-2min.
[0093] The frequency response is given by the swing equation:
[0094]
[0095] In formula (1), Δf(t) is the frequency change; t is the time period; ΔP L is the power imbalance; ΔP i is the power change of the i-th unit; H is the grid inertia; N is the number of units participating in the primary frequency regulation; D is the grid damping coefficient;
[0096] Linearize the curve process of primary frequency modulation. When the system load suddenly increases, equation (1) can be expressed as:
[0097]
[0098] In formula (2), Δf(t) is the frequency change; t is the time period; ΔP L is the power imbalance; H is the grid inertia; D is the grid damping coefficient; R is the frequency modulation rate.
[0099] Usually, the speed regulator of the hydropower unit will set a frequency regulation dead zone. When the frequency change exceeds the set value, the speed regulator will start to act. Then the frequency reduction process can be divided into two parts, namely the frequency regulation dead zone and the primary frequency regulation process. Please refer to the attached figure for details. Figure 8 , before t=0 is the frequency regulation dead zone, and at t=0, the hydropower unit begins to participate in the primary frequency regulation.
[0100] In the frequency modulation dead zone Δf DB From formula (2), we can know that ΔP i is 0, the frequency decreases at a speed of S 0 Affected only by the inertia of the system, we have:
[0101]
[0102]
[0103] When the frequency changes to outside the frequency dead zone, the speed governor of the hydropower unit starts to increase the unit output; at this time, solving equation (2) can obtain the first frequency fluctuation expression:
[0104]
[0105] Among them, c is an intermediate variable and has no actual meaning;
[0106] From Δf(t)=Δf DB Inference:
[0107]
[0108] When the derivative of Δf(t) is zero, the lowest point of the frequency drop can be obtained.
[0109]
[0110]
[0111] In formula (2-6), t d is the moment when the frequency drops to the lowest point; Δf nadir is the lowest frequency point; set the preset frequency lower limit to Δf UFLS , then the lowest frequency point Δf nadir Need to be higher than the frequency lower limit Δf UFLS , therefore, the first frequency highest point constraint is expressed as:
[0112]
[0113] In formula (2-7), H and R are determined by the number of hydropower units and the output power of the hydropower units;
[0114] Similarly, during a frequency regulation process, the load of the hydropower unit suddenly drops, ΔP L and R is less than 0, then formula (2) can be transformed into:
[0115]
[0116] The second frequency fluctuation expression and the second frequency highest point constraint can be expressed as:
[0117]
[0118]
[0119]
[0120] Among them, c′ represents an intermediate variable and has no actual meaning;
[0121] Combining formulas (2-7) and (2-10), when the load damping coefficient D = 1, the frequency response constraint should be:
[0122]
[0123]
[0124]
[0125]
[0126] Among them, h i The system inertia provided for hydropower unit i; u i,t is the operating status of hydropower unit i during period t; Δf DB is the frequency dead zone; I is the set of hydropower units; r i The frequency modulation rate provided for hydropower unit i; p t,s W is the random output of the wind turbine in time period t under scenario s; P i,t is the output of hydropower unit i in period t;
[0127] Objective function: The objective function is to minimize the water consumption for power generation:
[0128]
[0129]
[0130] Among them, Q i,t is the flow rate of the i-th unit in period t, in m 3 / s;Q sum t is the total water consumption of the units during period t; Q s t is the power station water discharge flow rate during period t, in m 3 / s; I is the set of hydropower units; T is the set of time periods; Δt is the duration of each time period, in h;
[0131] In the current day-ahead dispatch model, at the hour / 15-minute time scale, the power imbalance caused by the uncertainty of wind and solar output can be handled by setting the backup capacity, but at the second scale, the frequency fluctuation of the system caused by the instantaneous power imbalance caused by the random output of wind and solar power is not considered; at the same time, when large-scale wind and solar energy participate in the grid connection, the inertia of the grid will be reduced, and the frequency stability problem will be more prominent. When the overall frequency regulation capability of the grid-connected units is not high, excessive prediction errors will lead to frequency over-limit, low-frequency load reduction or high-frequency machine-cutting relay operation, resulting in load shedding or load shedding of the grid-connected units, threatening the reliability of grid operation; in addition, although compared with thermal power units, hydropower units have stronger climbing capabilities, and without considering capacity limitations, the number of hydropower units required for the same backup capacity is less, but this will lead to a further reduction in the level of grid inertia, especially as thermal power units are gradually withdrawn in the future, this impact will be more prominent; therefore, it is necessary to linearize and nonlinearly constrain the relevant variables based on the minimum water consumption objective function;
[0132] Among them, the linearization constraints include water balance constraints, reservoir flow constraints, reservoir water level constraints, reservoir storage capacity constraints, power balance constraints; hydropower unit output constraints, hydropower unit flow constraints, minimum startup and shutdown time constraints, maximum startup times constraints, unit generating head constraints and spare capacity constraints; the details are as follows:
[0133] (a) Water balance constraints:
[0134]
[0135] In formula (9), V t is the reservoir capacity in time period t, in m 3 ;Q IN t is the runoff flow in period t, in m 3 / s;Q sum t is the discharge flow in period t, in m 3 / s;
[0136] (b) Reservoir flow constraints:
[0137]
[0138] In formula (10), is the upper limit of reservoir discharge, in m 3 / s; Q is the lower limit of reservoir discharge, in m 3 / s;
[0139] (c) Reservoir water level constraints:
[0140]
[0141] In formula (11), is the upper limit of the reservoir water level, in m, Z is the lower limit of the reservoir water level, in m; Z t is the reservoir water level;
[0142] (d) Reservoir capacity constraints:
[0143]
[0144] In formula (12), is the upper limit of reservoir capacity, in m 3 ; V is the lower limit of reservoir capacity, in m 3 ; V t is the reservoir capacity;
[0145] (e) Power balance constraints:
[0146]
[0147] In formula (13), P t load is the load demand during period t; P i,t is the output of hydropower unit i in period t; P t wind is the wind power output in period t, in MW;
[0148] (f) Output constraints of hydropower units:
[0149]
[0150]
[0151] In formulas (14)-(15), is the upper limit of the output of hydropower unit i; P i is the lower limit of the output of the i-th hydropower unit, in MW; u i,t is the operating status of the hydropower unit i in period t; 0 represents that the unit is in the shutdown state, and 1 represents that it is in the startup state; P i,t is the output of hydropower unit i in period t;
[0152] (g) Flow constraints of hydropower units:
[0153]
[0154] is the upper limit of flow rate of hydropower unit i, in m 3 / s; Q iis the lower limit of flow rate of hydropower unit i, in m 3 / s;Q i,t is the flow rate of the i-th unit in period t, in m 3 / s;
[0155] (h) Minimum startup and shutdown time constraints:
[0156]
[0157]
[0158]
[0159]
[0160]
[0161] In formulas (17)-(21), Indicates the start-up action of the hydropower unit. When it is 1, it means that the hydropower unit i has performed the start-up action in the t period, and it is in the shutdown state in the t-1 period. When it is 0, it is the opposite; Indicates the shutdown action of the hydropower unit. When it is 1, it means that the i hydropower unit has executed the shutdown action at time t, and it is in the start-up state during the t-1 period. When it is 0, it is the opposite; T i ON is the minimum startup time of the i hydropower unit, that is, the minimum interval between the startup operation and the next shutdown operation of the unit; T i OFF is the minimum shutdown time of the i hydropower unit, that is, the minimum interval between the shutdown operation of the unit and the next startup operation. The minimum startup and shutdown time constraint is to limit the frequent start and stop of the unit; m represents the next moment of t; and Respectively represent the on and off operation variables;
[0162] (i) Maximum boot times constraint:
[0163]
[0164] In formula (22), is the maximum number of startup times of the i-th hydropower unit in a scheduling cycle;
[0165] (j) Unit generating head constraints:
[0166]
[0167] In formula (23), H i,t is the net water head of the i-th hydropower unit in the t-th period, in m; Z dt is the tailwater level, in m; H t loss is the head loss of the i-th hydropower unit in the t-th time period, in m.
[0168] (k) Reserve capacity constraints:
[0169]
[0170] In formula (24), is the random output of the wind turbine in time period t under scenario s; R Di is the downhill rate of the hydropower unit, in MW / 15min; R Ui is the ramp-up rate of the hydropower unit, in MW / 15min; is the lower limit of output; P t W Provide output for wind power forecasting;
[0171] The nonlinear constraints include water level-storage capacity constraints, tailwater level-discharge flow constraints, head loss constraints, unit power generation characteristics constraints and unit vibration zone restriction processing; the details are as follows:
[0172] 2. Nonlinear constraints:
[0173] (l) Water level-reservoir capacity constraints:
[0174]
[0175] In formula (25), f Z-V is the relationship function between reservoir water level and flow;
[0176] (m) Tailwater level-discharge flow constraint:
[0177]
[0178] In formula (26), Z d t is the tailwater level; is the relationship function between tailwater level and discharge flow; Q sum t is the sum of the water consumption of the units during period t;
[0179] (n) Head loss constraint:
[0180]
[0181] In formula (27), H t loss is the head loss of the i-th hydropower unit in the t-th period, in m; is the relationship function between head loss and downstream flow rate; Q it is the flow rate of the i-th hydropower unit in period t;
[0182] (o) Constraints on unit power generation characteristics:
[0183]
[0184] In formula (28), f P-Q-H is the output characteristic function of the hydropower unit; H i,t is the net water head of the i-th hydropower unit in the t-th period;
[0185] (p) Unit vibration zone limitation treatment:
[0186]
[0187] In formula (29), are the upper and lower limits of the kth vibration zone of unit n, in MW; P n,t is the water head output of hydropower unit n at time t;
[0188] Considering the upper and lower limits of the unit output, the k vibration zones divide the unit into k+1 safe operation zones, and equation (27) can be transformed as follows:
[0189]
[0190] In formula (30), is the indicator variable, Indicates that the output of unit n in period t is in the kth safe operation zone: are the upper and lower limits of the kth safe operating area of unit n, respectively, and The P with f as the subscript represents the safe operating area as a whole;
[0191] The relationship function between reservoir water level and flow rate can be calculated by the above formulas (25)-(28): Z-V , the relationship function between tailwater level and discharge flow Relationship function between head loss and downstream flow And the output characteristic function f of the hydropower unit P-Q-H ; Linearize these four relationship functions respectively. The specific steps are as follows:
[0192] f Z-V , and It is obtained by fitting a fourth-order polynomial and then performing piecewise linearization.
[0193]
[0194]
[0195]
[0196] Among them, a, b, c are polynomial coefficients; f P-Q-H There are two independent variables, flow rate Q and water head H, which are three-dimensional surfaces. The linearization process is divided into two steps. First, the net water head of each hydropower unit is discretized into n intervals, and a curve is selected from each interval to replace the unit operation characteristics in the water head interval. By introducing auxiliary variables Determine the range of the unit's operating net water head. is 1, it means that the net water head H of the i-th hydropower unit in period t i,t In [H n lower ,H n upper ] range.
[0197]
[0198]
[0199]
[0200] Select an output characteristic curve within the net head range To approximate the relationship between the unit output and flow rate within the net head range, the output of the hydropower unit can be expressed as
[0201]
[0202] in, is an auxiliary variable; The processing method is the same as the relationship functions between water level-reservoir capacity, tailwater level-discharge flow and head loss.
[0203] (o) Dynamic frequency response constraints
[0204] The dynamic frequency response constraint involves the multiplication of two decision variables and there are nonlinear terms. The piecewise multivariate least squares fitting method is used to linearize equation (14). Let the left side of equation (14) be f(X), where X includes two variables H and R, and it can be proved that f(X) is The inner part is convex. The piecewise fitting result of f(X) can be obtained by Indicates that:
[0205]
[0206] in and are two unknown parameters in each segment j. Assuming the number of fitting data points is m, the least square fitting expression of f(X) is:
[0207]
[0208] Where k is the total number of segments for segmented fitting, and each segment contains a data set P j Will affect a j and b j The calculation results will affect the overall fitting effect, so a heuristic algorithm is used to determine the optimal segmentation method.
[0209] set up is the jth data set in the lth iteration, then
[0210]
[0211] The parameter value at the lth iteration is and Then the parameter of the l+1 iteration is and pass The specific method is as follows:
[0212] For each segment (j=1,...,k), use the data set The data in the least squares method Approximation f(X i ), that is, solving
[0213]
[0214] get and In general, there is a unique solution that minimizes the above equation, namely
[0215]
[0216] When the matrix to be inverted in equation (33) is singular, equation (32) has multiple minima. In this case, we can add a regularization term Using the new coefficient and get If there is X i If equation (34) is satisfied, then X i Assign to middle.
[0217]
[0218] If the new partitions generated during the iteration are the same as those generated in the previous iteration, or the set number of iterations is reached, the algorithm converges.
[0219] The specific process of the piecewise least squares fitting algorithm is as follows:
[0220] First, the initial segmentation interval is given, that is, Set the upper limit of the number of iterations l max
[0221] ① Calculate by formula (33) and
[0222] ②Generate through formula (34)
[0223] ③ Determine whether each segment satisfies
[0224] After obtaining the optimal segmentation, set α 1 =H,α 2 =R, then f(X) can be expressed as
[0225]
[0226] in Parameter a j and b j Calculated by formula (30), the calculation result can be expressed as
[0227]
[0228] Introducing auxiliary variables v 1 ,...,v k-1 ∈{0,1} and transform equation (35), let
[0229] t 1 =max{π 1 ,π 2},t 2 =max{t 1 ,π 3},...,t k-1 =max{t k-2 ,π k} (46)
[0230] Then the frequency constraint equation (25) can be transformed into
[0231] π 1 ≤t 1 ≤π 1 +v 1 M
[0232] π 2 ≤t 1 ≤π 2+(1-v 1 )M
[0233] t 1 ≤t 2 ≤t 1 +v 2 M
[0234] π 3 ≤t 2 ≤π 3 +(1-v 2 )M
[0235] L
[0236] t k-2 ≤t k-1 ≤t k-2 +v k-1 M
[0237] π k ≤t k-1 ≤π k +(1-v k-1 )M
[0238] t k-1 ≤f ΔUFLS (47)
[0239] In summary, the embodiment of the present invention provides a method for scheduling a hydro-wind complementary power generation system based on dynamic frequency constraints. The invention aims to improve the scheduling scheme of the wind-hydro complementary power generation system, fully consider the physical constraints or restrictions such as reservoir operating conditions, power station operating conditions, and grid operating constraints, and introduce dynamic frequency response constraints, thereby calculating the combination of hydropower units and the on / off state, so as to achieve the goal of stable grid frequency and minimum water consumption; the invention ensures the continuity and reliability of wind-hydropower when it is integrated into the grid, and the specific benefits are as follows:
[0240] (1) The wind power stochastic model based on the prediction error distribution effectively describes the characteristics of wind power prediction errors in different intervals. The wind power output scenarios sampled and generated by this model can more accurately describe the randomness of wind power output, and will become more and more accurate as the amount of historical data increases.
[0241] (2) The dynamic frequency response model of the frequency under disturbance is derived, which can describe the dynamic change process of the frequency under disturbance in detail and calculate the amplitude of the frequency offset. It is suitable for frequency offset analysis in scenarios with large wind power fluctuations. At the same time, the inequality relationship can be added to convert it into a constraint condition that limits the frequency offset and added to the optimization scheduling model to help the model absorb wind power energy.
[0242] (3) A day-ahead optimal dispatch model considering frequency response was established, which has the effect of stabilizing the frequency of power systems containing a large amount of wind power. Especially when the wind power fluctuates greatly and the system load is small, the dispatch decision result can ensure that the system inertia and frequency regulation capacity of the starting units are sufficient to make up for the power shortage and stabilize the frequency deviation. At the same time, the model can also reduce the waste of wind power energy and avoid the losses caused to electricity users by load shedding to a certain extent.
[0243] 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 fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A day-ahead dispatching method for a hydro-wind complementary power generation system based on dynamic frequency constraints, characterized in that: include: S1. Establish a model to describe wind power uncertainty: collect historical data, obtain the probability density function of the prediction error through probability prediction, and then obtain the confidence interval of wind power output under various confidence levels through interval prediction; S2. Construct linear and nonlinear constraints on the output-related variables of the hydropower units; S3, establishing a water consumption minimization model with dynamic frequency response constraints; S4, determining the working state of the hydropower unit according to the wind power output confidence interval, the linearization constraint and nonlinearization constraint conditions of the hydropower unit output-related variables, and the minimum water consumption model; The S1 specifically includes: S11, obtaining historical wind power actual values and corresponding wind power forecast values; S12, dividing the actual wind power value and the corresponding predicted wind power value according to different time periods, so as to obtain average actual wind power values and corresponding average predicted wind power values in multiple different time periods, and calculating the prediction error for each time period; S13, extracting the maximum value and the minimum value of the average wind power prediction value to obtain a wind power prediction range; S14, dividing the wind power prediction range into equal intervals to obtain n wind power prediction intervals; S15, grouping the average wind power prediction value of each time period and the prediction error of the corresponding time period; and placing each group of data into the corresponding wind power prediction interval according to the size of the average wind power prediction value of the group; each wind power prediction interval constitutes a data set; S16, using a kernel density estimation algorithm to calculate the probability density function of the prediction error in each of the data sets; S17. Calculate confidence intervals of different confidence levels in the corresponding wind power prediction interval using the probability density function.
2. The method for day-ahead dispatching of a hydro-wind complementary power generation system based on dynamic frequency constraints according to claim 1, characterized in that: In S2, the linearization constraints of the hydropower unit output-related variables include: water balance constraint, reservoir flow constraint, reservoir water level constraint, reservoir storage capacity constraint, power balance constraint, hydropower unit output constraint, hydropower unit flow constraint, minimum startup and shutdown time constraint, maximum startup number constraint, unit generating head constraint, and spare capacity constraint; In S2, the nonlinear constraints on the output-related variables of the hydropower unit include: water level-reservoir capacity constraint, tailwater level-discharge flow constraint, head loss constraint, hydropower unit power generation characteristic constraint, and hydropower unit vibration zone limit constraint.
3. The method for day-ahead dispatching of a hydro-wind hybrid power generation system based on dynamic frequency constraints according to claim 2, characterized in that: In the linearization constraint of the hydropower unit output related variables, the hydropower unit output constraint equation is: in, is the upper limit of the output of hydropower unit i; P i is the lower limit of the output of the i-th hydropower unit, in MW; u i,t is the operating status of hydropower unit i during period t; P i,t is the output of the i hydropower unit in period t; 0 means the unit is in shutdown state, and 1 means it is in startup state.
4. The method for day-ahead dispatching of a hydro-wind hybrid power generation system based on dynamic frequency constraints according to claim 2, characterized in that: In the nonlinear constraints of the output-related variables of the hydropower unit, the relationship function between the reservoir water level and the flow in the water level-reservoir capacity constraint, the relationship function between the tail water level and the downstream flow in the tail water level-discharge flow constraint, the relationship function between the head loss and the downstream flow in the head loss constraint, and the output characteristic function of the hydropower unit in the hydropower unit power generation characteristic constraint are subjected to piecewise linearization processing.
5. The method for day-ahead dispatching of a hydro-wind complementary power generation system based on dynamic frequency constraints according to claim 3, characterized in that: The S3 specifically includes: S31, constructing a frequency response constraint equation; S32, constructing an objective function for minimizing water consumption for power generation; Among them, Q i,t is the flow rate of the i-th unit in period t, in m 3 / s;Q sum t is the total water consumption of the units during period t; Q s t is the power station water discharge flow rate during period t, in m 3 / s; I is the set of hydropower units; T is the set of time periods; △t is the duration of each time period, in h; S33. Constructing a dynamic frequency response constraint according to the output constraint equation of the hydropower unit.
6. The method for day-ahead dispatching of a hydro-wind complementary power generation system based on dynamic frequency constraints according to claim 5, characterized in that: The segmented multivariate least squares fitting method is used to linearize the formula (14) in the output constraint equation of the hydropower unit; The specifics of the piecewise multivariate least squares fitting method are: Assume that the left side of formula (14) P i ×u i,t Let f(X) be The f(X) is segmented to obtain multiple initial segmentation intervals, that is, Set the upper limit of the number of iterations l max ; Determine whether each initial segment interval satisfies Where j represents the segment; the optimal segment interval is obtained; Combining the minimum startup and shutdown time constraints in the linearization constraints of the output-related variables of the hydropower unit and the reserve capacity constraints, formula (14) is transformed to obtain a new frequency constraint condition: π1≤t1≤π1+v1M π2≤t1≤π2+(1-v1)M t1≤t2≤t1+v2M π3≤t2≤π3+(1-v2)M L t k-2 ≤t k-1 ≤t k-2 +v k-1 M p k ≤t k-1 ≤π k +(1-v k-1 )M t k-1 ≤Δf UFLS (47) Among them, M is a large constant; v is an auxiliary binary variable of M; π and t are both intermediate variables; Δf UFLS is the preset frequency lower limit.
