A short-term robust optimization scheduling method for small hydropower groups considering prediction errors

Abstract

The present invention provides a short-term robust optimization scheduling method for a small hydropower group considering prediction errors, including: using a large amount of historical data to obtain samples of uncertain factors after considering prediction errors; the samples are processed by numerical optimization to construct an output set of run-of-river small hydropower; a short-term robust optimization scheduling model is constructed, and the output set is brought into the short-term robust optimization scheduling model for solution to obtain the final output of the small hydropower station. The optimization scheduling method of the present invention considers that samples of prediction errors of run-of-river small hydropower are included in the uncertainty set, and realizes the short-term optimization scheduling of the small hydropower group while considering the prediction error, reduces the conservatism of the model solution, and improves the robustness of the uncertainty set and the power generation efficiency of the power station in the region.

Description

Technical Field

[0001] The present invention relates to the field of short-term robust optimization of small power groups, and in particular to a short-term robust optimization scheduling method for small hydropower groups taking prediction errors into consideration. Background Art

[0002] In areas rich in hydropower resources, how to significantly reduce water abandonment, achieve optimal economic dispatch of small hydropower station groups, increase power generation, and solve "blind action" and "blind adjustment" have become key issues in the coordinated operation of multiple hydropower stations. Since hydropower is an intermittent power source, its power generation capacity has strong randomness and uncertainty, especially run-of-river small hydropower, which is limited by the reservoir regulation capacity, has strong seasonality, is greatly affected by climate, and has obvious volatility. Although many studies have conducted short-term power forecasts for small hydropower, there are still errors between actual power and predicted values. Studies have shown that smaller forecast errors will deteriorate the dispatch optimization results. Therefore, short-term power adjustment is the focus of small hydropower optimization dispatch.

[0003] Common methods for dealing with uncertain output are stochastic optimization and robust optimization. The characteristic of stochastic optimization is to construct a probability model of uncertain output, and then generate multiple scenarios and reduce the scenarios to simulate the scenarios of uncertain output. However, most uncertain outputs do not have accurate probability distribution models or the probability distribution models are extremely complex, and the generated probability scenarios cannot fully cover the scenarios that occur. Robust optimization includes the output of uncertain factors in the uncertain set and finds the optimal solution under the worst conditions. It can fully consider the occurrence of uncertain factors, but the construction of the uncertain set is the key to this method. The worse the robustness of the set, the more conservative the optimization result. When calculating large-scale systems, this method will occupy more computing resources and consume a lot of running time. Therefore, the robustness of the uncertain set affects the efficiency of model solving. Summary of the invention

[0004] The purpose of the present invention is to optimize the short-term dispatch of small hydropower groups while taking into account the prediction error, reduce the conservatism of the model solution, and improve the robustness of the uncertainty set and the power generation efficiency of the power stations in the region.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] The present invention provides a short-term robust optimization scheduling method for a small hydropower group taking into account prediction errors, comprising: using a large amount of historical output data to obtain samples of uncertain factors after taking into account prediction errors; using numerical optimization processing on the samples to construct an output set of run-of-river small hydropower; building a short-term robust optimization scheduling model, bringing the output set into the short-term robust optimization scheduling model for solution, and obtaining the final output of the small hydropower station.

[0007] Optionally, the historical output data is obtained by obtaining predicted output information of run-of-river small hydropower according to weather forecast information, and predicting the load in the same scheduling period.

[0008] Optionally, the numerical optimization process includes: optimizing a minimum volume ellipsoid set to construct an uncertain set of run-of-river small hydropower output sets.

[0009] Optionally, the short-term robust optimization scheduling model solution process includes: using K-means++ to cluster historical data into 24 categories, comparing the difference between each cluster center and the predicted value, and when the deviation between the two is less than 0.05% of the predicted value, outputting the cluster center of this clustering and s samples of the cluster center.

[0010] Optionally, in the process of constructing the uncertain set, first construct the minimum volume ellipsoid set as shown in the following formula:

[0011] ε knn ={u|||Pu+ρ||2≤1}

[0012] In the formula, P and ρ are the matrix variables to be optimized, and u is a column vector composed of s samples clustered at the cluster center. P and ρ are obtained by the following model:

[0013] min log det P -1

[0014] st||Pu+ρ||2≤1

[0015] Once P and ρ are solved, some parameters R and V of the proposed uncertainty set can be obtained by the following formula:

[0016] R=P -1

[0017] V=-Rρ

[0018] Finally, the uncertainty set of this method is expressed as follows:

[0019] ε′ knn ={v+RW|||W||2≤1}.

[0020] Optionally, the uncertainty set has an optimal uncertainty factor fluctuation range, and the calculation process of the optimal uncertainty factor fluctuation range includes:

[0021] The optimal uncertainty factor fluctuation range is obtained through a two-term exponential formula:

[0022] W=1 / 2·(a exp(bs)+c exp(ds))

[0023] In the formula, W is the parameter for determining the optimal uncertainty factor fluctuation range, s is the number of samples gathered in the cluster center, and ad is the coefficient to be optimized, which is solved by the following coefficient optimization model:

[0024] min a exp(bs)+c exp(ds)

[0025]

[0026] Optionally, constructing a short-term robust optimization scheduling model includes:

[0027] The goal of small hydropower group coordination is to maximize run-of-river power generation while minimizing the total depreciation cost of regional power stations.

[0028]

[0029] Where E is the coordination target of small hydropower groups, M is the total number of run-of-river small hydropower stations, N is the total number of reservoir-type hydropower stations, T is the dispatching period, and P is the total number of small hydropower stations. i (t) is the power generation of the i-th small hydropower station at time t, c i is the depreciation cost per unit of electricity generated;

[0030] Run-of-river small hydropower has no regulation capability, which is mainly related to the water inflow. The main constraints considered in the model solution are the operation constraints of reservoir-capacity small hydropower and the coordinated control constraints of reservoir-capacity and run-of-river;

[0031] Water balance constraints:

[0032]

[0033] Where n is the number of the reservoir-type hydropower, V n (t+1) is the reservoir capacity of hydropower station n at the time t+1, V n (t) is the reservoir capacity of hydropower station n at time t, is the flow into hydropower station n at time t, is the flow out of hydropower station n at time t;

[0034] Outbound traffic restrictions:

[0035]

[0036] In the formula They are the upper and lower limits of the outflow of hydropower station n at time t;

[0037] Reservoir water level constraints

[0038]

[0039] In the formula is the upper and lower limits of the water level of the reservoir-type hydropower station at time t, Zn (t) is the water level at time t in the reservoir-type hydropower station;

[0040] Power station output upper and lower limit constraints:

[0041]

[0042]

[0043] In the formula It is the upper and lower limits of the output of the reservoir-type hydropower station. are the upper and lower limits of run-of-river small hydropower, which are determined by the proposed uncertainty set, P n (t), P m (t) is the output of run-of-river small hydropower station and reservoir capacity hydropower station at time t;

[0044] Power station output ramp constraints:

[0045] |P n (t)-P n (t-1)|≤ΔP n (t)

[0046] ΔP n (t) is the maximum climbing capacity of the nth reservoir-type hydropower station;

[0047] Power balance constraints:

[0048]

[0049] Where P l (t) is the load forecast value at time t, and ΔP(t) is the power interacting with the power grid at time t.

[0050] The beneficial effects of the present invention are:

[0051] The advantages of the short-term robust optimization scheduling method for a small hydropower group considering prediction errors of the present invention are: (1) a framework combining machine learning and numerical optimization is proposed to take prediction errors into account in the short-term robust optimization scheduling model, thereby reducing the impact of prediction errors on degrading scheduling results and improving the power generation efficiency of the small hydropower group in the region; (2) machine learning is used to analyze historical data, and samples that may appear after considering prediction errors are obtained by learning historical output characteristics; (3) the constructed optimal uncertainty set can contain possible hydropower output values ​​and the interval of the uncertainty set will not be too large. The robust model solution based on the uncertainty set is more conservative, which reduces the scheduling model solution time and computing resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0053] Figure 1 It is a flow chart of the overall solution of an embodiment of the present invention;

[0054] Figure 2 A schematic diagram of a short-term optimization scheduling result of a small hydropower group according to an embodiment of the present invention;

[0055] Figure 3 A schematic diagram of sample distribution of each power station close to the predicted value at each moment in an embodiment of the present invention;

[0056] Figure 4 A schematic diagram of the collective conservatism of each power station at each time in an embodiment of the present invention;

[0057] Among them, P1-P4 represent run-of-river small hydropower, P5 represents reservoir capacity power station, and Load represents load; Figure 2 The dispatching results have satisfied the power balance between supply and demand, and each small hydropower station has produced power according to the maximum benefit; Figure 4 The dotted line area is the distribution area of ​​samples close to the predicted value, and the solid line area with stars is the interval formed by the uncertain set. Figure 4 It can be seen that the possible output of each run-of-river power station is included in the interval composed of the uncertainty set. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] like Figure 1-4As shown, the present invention provides a short-term robust optimization scheduling method for a small hydropower group taking into account prediction errors, including: using a large amount of historical output data to obtain samples of uncertain factors after taking into account prediction errors; the samples are processed using numerical optimization to construct an output set of run-of-river small hydropower; a short-term robust optimization scheduling model is constructed, and the output set is brought into the short-term robust optimization scheduling model for solution to obtain the final output of the small hydropower station.

[0061] Optionally, the historical output data is obtained by obtaining predicted output information of run-of-river small hydropower according to weather forecast information, and predicting the load in the same scheduling period.

[0062] Optionally, the numerical optimization process includes: optimizing a minimum volume ellipsoid set to construct an uncertain set of run-of-river small hydropower output sets.

[0063] Optionally, the short-term robust optimization scheduling model solution process includes: using K-means++ to cluster historical data into 24 categories, comparing the difference between each cluster center and the predicted value, and when the deviation between the two is less than 0.05% of the predicted value, outputting the cluster center of this clustering and s samples of the cluster center.

[0064] Optionally, in the process of constructing the uncertain set, first construct the minimum volume ellipsoid set as shown in the following formula:

[0065] ε knn ={u|||Pu+p||2≤1}

[0066] In the formula, P and ρ are the matrix variables to be optimized, and u is a column vector composed of s samples clustered at the cluster center. P and ρ are obtained by the following model:

[0067] min log det P -1

[0068] st||Pu+ρ||2≤1

[0069] Once P and ρ are solved, some parameters R and V of the proposed uncertainty set can be obtained by the following formula:

[0070] R=P -1

[0071] V=-Rρ

[0072] Finally, the uncertainty set of this method is expressed as follows:

[0073] ε′ knn ={v+RW|||W||2≤1}.

[0074] Optionally, the uncertainty set has an optimal uncertainty factor fluctuation range, and the calculation process of the optimal uncertainty factor fluctuation range includes:

[0075] The optimal uncertainty factor fluctuation range is obtained through a two-term exponential formula:

[0076] W=1 / 2·(a exp(bs)+c exp(ds))

[0077] In the formula, W is the parameter for determining the optimal uncertainty factor fluctuation range, S is the number of samples gathered in the cluster center, and ad is the coefficient to be optimized, which is solved by the following coefficient optimization model:

[0078] min a exp(bs)+c exp(ds)

[0079]

[0080] Optionally, constructing a short-term robust optimization scheduling model includes:

[0081] The goal of small hydropower group coordination is to maximize run-of-river power generation while minimizing the total depreciation cost of regional power stations.

[0082]

[0083] Where E is the coordination target of small hydropower groups, M is the total number of run-of-river small hydropower stations, N is the total number of reservoir-type hydropower stations, T is the dispatching period, and P is the total number of small hydropower stations. i (t) is the power generation of the i-th small hydropower station at time t, c i is the depreciation cost per unit of electricity generated;

[0084] Run-of-river small hydropower has no regulation capability, which is mainly related to the water inflow. The main constraints considered in the model solution are the operation constraints of reservoir-capacity small hydropower and the coordinated control constraints of reservoir-capacity and run-of-river;

[0085] Water balance constraints:

[0086]

[0087] Where n is the number of the reservoir-type hydropower, V n (t+1) is the reservoir capacity of hydropower station n at the time t+1, V n (t) is the reservoir capacity of hydropower station n at time t, is the flow into hydropower station n at time t, is the flow out of hydropower station n at time t;

[0088] Outbound traffic restrictions:

[0089]

[0090] In the formula They are the upper and lower limits of the outflow of hydropower station n at time t;

[0091] Reservoir water level constraints

[0092]

[0093] In the formula is the upper and lower limits of the water level of the reservoir-type hydropower station at time t, Z n (t) is the water level at time t in the reservoir-type hydropower station;

[0094] Power station output upper and lower limit constraints:

[0095]

[0096]

[0097] In the formula It is the upper and lower limits of the output of the reservoir-type hydropower station. are the upper and lower limits of run-of-river small hydropower, which are determined by the proposed uncertainty set, P n (t), P m (t) is the output of run-of-river small hydropower station and reservoir capacity hydropower station at time t;

[0098] Power station output ramp constraints:

[0099] |P n (t)-P n (t-1)|≤ΔP n (t)

[0100] ΔP n (t) is the maximum climbing capacity of the nth reservoir-type hydropower station;

[0101] Power balance constraints:

[0102]

[0103] Where P l (t) is the load forecast value at time t, and ΔP(t) is the power interacting with the power grid at time t.

[0104] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A short-term robust optimization scheduling method for small hydropower groups considering prediction errors, characterized in that: include: Use a large amount of historical output data to obtain samples that take into account the uncertainty factors after the forecast error; The sample adopts numerical optimization processing to construct the output set of run-of-river small hydropower; Construct a short-term robust optimization scheduling model, bring the output set into the short-term robust optimization scheduling model for solution, and obtain the final output of the small hydropower station; Acquiring the historical output data is to obtain the predicted output information of run-of-river small hydropower according to the weather forecast information, and predict the load in the same dispatching cycle; The numerical optimization processing includes: using the optimization of the minimum volume ellipsoid set to construct an uncertain set of the output set of run-of-river small hydropower; the short-term robust optimization scheduling model solution process includes: using K-means++ to cluster historical data into 24 categories, comparing the difference between each cluster center and the predicted value, and when the deviation between the two is less than 0.05% of the predicted value, outputting the cluster center of this clustering and s samples of the cluster center.

2. The short-term robust optimization scheduling method for small hydropower groups considering prediction errors according to claim 1 is characterized in that: In the process of constructing an uncertain set, first construct the minimum volume ellipsoid set as shown in the following formula e knn ={u|||Pu+ρ||2≤1} In the formula, P and ρ are the matrix variables to be optimized, and u is a column vector composed of s samples clustered at the cluster center. P and ρ are obtained by the following model: my log it P -1 st||Pu+ρ||2≤1 Once P and ρ are solved, some parameters R and V of the proposed uncertainty set can be obtained by the following formula: R=P -1 V=-Rρ Finally, the uncertainty set of this method is expressed as follows: ε′ knn (V+RW|||W||2≤1}.

3. The short-term robust optimization scheduling method for small hydropower groups considering prediction errors according to claim 2 is characterized in that: The uncertainty set has an optimal uncertainty factor fluctuation range, and the calculation process of the optimal uncertainty factor fluctuation range includes: The optimal uncertainty factor fluctuation range is obtained through a two-term exponential formula: W = 1 / 2 (aexp(bs) + cexp(ds)) In the formula, W is the parameter for determining the optimal uncertainty factor fluctuation range, s is the number of samples gathered in the cluster center, and ad is the coefficient to be optimized, which is solved by the following coefficient optimization model: min a exp(bs)+c exp(ds) 4. The short-term robust optimization scheduling method for small hydropower groups considering prediction errors according to claim 1 is characterized in that: Constructing a short-term robust optimization scheduling model includes: The goal of small hydropower group coordination is to maximize run-of-river power generation while minimizing the total depreciation cost of regional power stations. Where E is the coordination target of small hydropower groups, M is the total number of run-of-river small hydropower stations, N is the total number of reservoir-type hydropower stations, T is the dispatching period, and P is the total number of small hydropower stations. i (t) is the power generation of the i-th small hydropower station at time t, c i is the depreciation cost per unit of electricity generated; Run-of-river small hydropower has no regulation capability, which is mainly related to the water inflow. The main constraints considered in the model solution are the operation constraints of reservoir-capacity small hydropower and the coordinated control constraints of reservoir-capacity and run-of-river; Water balance constraints: Where n is the number of the reservoir-type hydropower, V n (t+1) is the reservoir capacity of hydropower station n at the time t+1, V n (t) is the storage capacity of hydropower station n at the time t, is the flow into hydropower station n at time t, is the flow out of hydropower station n at time t; Outbound traffic restrictions: In the formula They are the upper and lower limits of the outflow of hydropower station n at time t; Reservoir water level constraints In the formula is the upper and lower limits of the water level of the reservoir-type hydropower station at time t, Z n (t) is the water level at time t in the reservoir-type hydropower station; Power station output upper and lower limit constraints: In the formula It is the upper and lower limits of the output of the reservoir-type hydropower station. are the upper and lower limits of run-of-river small hydropower, which are determined by the proposed uncertainty set, P n (t), P m (t) is the output of run-of-river small hydropower station and reservoir capacity hydropower station at time t; Power station output ramp constraints: |P n (t)-P n (t-1)|≤ΔP n (t) ΔP n (t) is the maximum climbing capacity of the nth reservoir-type hydropower station; Power balance constraints: Where P l (t) is the load forecast value at time t, and ΔP(t) is the power interacting with the grid at time t.

Images