Network provincial scheduling method and device considering source load uncertainty, equipment and medium

Abstract

The invention discloses a network provincial scheduling method, device, equipment and medium considering source load uncertainty, and relates to the technical field of power systems, a two-stage robust optimization scheduling model comprises a robust optimization framework of a daily scheduling stage and a real-time scheduling stage, and is realized through a column and constraint generation algorithm. The model can effectively deal with the uncertainty of renewable energy output and load demand, effectively deal with the volatility of wind energy, photovoltaic energy and other renewable energy sources, ensure the stable operation of the power grid under various extreme conditions, and ensure the stable operation of the power grid under the condition of large fluctuation of wind energy and photovoltaic power generation by introducing uncertainty parameters into the scheduling model. And the power system can still operate stably. In addition, the residual load is managed and reduced through energy storage systems such as a pump storage power station, stable operation of the provincial power grid is ensured, collaborative optimization among the power grids of multiple provincials is realized through fine management of the residual load, and the impact of peak load on the power system is reduced.

Description

Technical Field

[0001] The present application relates to the technical field of power systems, and particularly to a network-provincial dispatching method, device, equipment and medium considering the uncertainty of power sources and loads. Background Art

[0002] With the development of power system technology, under the background of widespread access of new energy at present, considering the high integration of new energy in large-scale power grid dispatching, it is vulnerable to the influence of load and wind-solar power output fluctuations, and it is more likely to appear load peak situations, increasing power safety hazards. In particular, the formulation of dispatching strategies at the network level and provincial level should be considered to ensure power balance during load peak periods, and at the same time coordinate the network-level and provincial-level dispatching strategies to achieve the goal of ensuring stable power supply at the network level and promoting new energy consumption at the provincial level, and ensuring the power safety of large-scale power grids to the greatest extent.

[0003] In related technologies, at the regional power grid dispatching level, peak-valley filling and peak shaving operations are mainly carried out through energy storage systems to cope with the influence brought by load uncertainty and renewable energy volatility; at the provincial dispatching level, a two-stage robust optimization model is mainly adopted. On the basis of considering regional power grid dispatching, the in-province power generation dispatching and power trading arrangements are further optimized to minimize the economic cost. However, the applicant realizes that in regional power grid dispatching, the existing methods mainly focus on peak-valley filling and peak shaving operations and do not consider comprehensively enough, while in provincial dispatching, although the existing two-stage robust optimization model can cope with the uncertainty of the supply and demand sides, there is still room for further improvement in terms of coordination with regional power grid dispatching. Summary of the Invention

[0004] In view of this, the present application provides a network-provincial dispatching method, device, equipment and medium considering the uncertainty of power sources and loads, mainly aiming to solve the problems that the existing methods do not consider comprehensively enough in regional power grid dispatching and there is still room for further improvement in terms of coordination with regional power grid dispatching in provincial dispatching.

[0005] According to the first aspect of the present application, a network-provincial dispatching method considering the uncertainty of power sources and loads is provided, and the method includes:

[0006] Obtain the power data of conventional generating units and the power data of pumped-storage units, and construct a residual load fluctuation optimization model based on the power data of the conventional generating units and the power data of the pumped-storage units;

[0007] Obtain the new energy dispatching data and the load curve output by the residual load fluctuation optimization model, and construct a provincial dispatching model based on the new energy dispatching data and the load curve;

[0008] Construct a provincial two-stage robust optimization model based on the remaining load fluctuation optimization model and the provincial dispatching model, decompose the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtain the network-province coordinated dispatching result.

[0009] According to the second aspect of the present application, a network-province dispatching device considering source-load uncertainty is provided. The device includes:

[0010] A first construction module, configured to obtain the power data of conventional generating units and the power data of pumped-storage units, and construct a remaining load fluctuation optimization model based on the power data of the conventional generating units and the power data of the pumped-storage units;

[0011] A second construction module, configured to obtain new energy dispatching data and the load curve output by the remaining load fluctuation optimization model, and construct a provincial dispatching model based on the new energy dispatching data and the load curve;

[0012] A solving module, configured to construct a provincial two-stage robust optimization model based on the remaining load fluctuation optimization model and the provincial dispatching model, decompose the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtain the network-province coordinated dispatching result.

[0013] According to the third aspect of the present application, a device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method described in any one of the above first aspects are implemented.

[0014] According to the fourth aspect of the present application, a medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described in any one of the above first aspects are implemented.

[0015] With the above technical solution, a network-province dispatching method, device, equipment and medium considering the uncertainty of power sources and loads provided by this application. This application obtains the power data of conventional generating units and pumped-storage units, constructs a residual load fluctuation optimization model based on the power data of conventional generating units and pumped-storage units, obtains the new energy dispatching data and the load curve output by the residual load fluctuation optimization model, constructs a provincial dispatching model based on the new energy dispatching data and the load curve, constructs a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial dispatching model, decomposes the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtains the network-province coordinated dispatching result. The two-stage robust optimization dispatching model includes robust optimization frameworks for the daily dispatching stage and the real-time dispatching stage, which is implemented by the column and constraint generation algorithm. This model can effectively handle the uncertainty of renewable energy output and load demand, and ensure the stable operation of the power grid under various extreme conditions. Moreover, this application manages and reduces the residual load through energy storage systems such as pumped-storage power stations to ensure the stable operation of the provincial power grid. Through the refined management of the residual load, the coordinated optimization among multiple provincial power grids is realized, and the impact of peak load on the power system is reduced. In addition, this application effectively handles the volatility of renewable energy such as wind energy and photovoltaic by setting uncertain parameters and robust optimization strategies to ensure the reliable operation of the power system, and by introducing uncertain parameters into the dispatching model, it ensures that the power system can still operate stably under the condition of large fluctuations in wind energy and photovoltaic power generation.

[0016] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following specifically gives the specific implementation manners of this application. Brief Description of the Drawings

[0017] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of this application. And throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0018] Figure 1 It shows a schematic flow chart of a method for network-province dispatching considering the uncertainty of power sources and loads provided by an embodiment of this application;

[0019] Figure 2A It shows another schematic flow chart of a method for network-province dispatching considering the uncertainty of power sources and loads provided by an embodiment of this application;

[0020] Figure 2B It shows a logic diagram of a C&CG algorithm provided by an embodiment of this application;

[0021] Figure 3 shows a schematic structural diagram of a network-province dispatch considering source-load uncertainty provided by an embodiment of the present application;

[0022] Figure 4 shows a schematic device structure diagram of a device provided by an embodiment of the present application. Detailed implementation manners

[0023] Hereinafter, exemplary embodiments of the present application will be described in more detail with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present application can be more thoroughly understood and the scope of the present application can be fully conveyed to those skilled in the art.

[0024] In the prior art, some methods related to regional power grid dispatch and provincial dispatch have been proposed. At the regional power grid dispatch level, mainly through the energy storage system to carry out peak-valley filling and peak shaving operations to cope with the impact brought by the uncertainty of load and the volatility of renewable energy; at the provincial dispatch level, mainly adopt a two-stage robust optimization model, on the basis of considering the regional power grid dispatch, further optimize the in-province power generation dispatch and power trading arrangements to minimize the economic cost. However, the existing methods mainly focus on peak-valley filling and peak shaving operations, but the consideration of the load gap (i.e., the remaining load) that cannot be fully met by the regional power grid units is not comprehensive enough; in provincial dispatch, although the existing two-stage robust optimization model can cope with the uncertainty on the supply and demand sides, there is still room for further improvement in the coordination with the regional power grid dispatch.

[0025] To solve this problem, the present application proposes a network-province dispatch method considering source-load uncertainty. In the regional power grid dispatch, the energy storage system is used to effectively reduce the fluctuation of the remaining load and improve the dispatch flexibility; in provincial dispatch, on the basis of the regional power grid dispatch results, a robust optimization model is constructed to fully coordinate the dispatch strategies to minimize the overall economic cost. The execution subject of the present application can be a network-province collaborative dispatch system. The network-province collaborative dispatch system relies on the computing power of the server to provide services for users. The server can be an independent server or can provide basic cloud computing servers such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms, so as to realize the efficient operation of the power system and the optimal allocation of resources.

[0026] An embodiment of the present application provides a network-province dispatch method considering source-load uncertainty, asFigure 1 As shown, the method includes:

[0027] 101. Obtain the power data of the conventional generating units and the pumped-storage units, and construct a residual load fluctuation optimization model based on the power data of the conventional generating units and the pumped-storage units.

[0028] At the network-level dispatching level, due to the integration of the loads of multiple provinces, the impact of new energy is greatly enhanced, and load peaks are extremely likely to occur during daily operation. Taking the East China Power Grid as an example, the main power supply provinces of the East China Power Grid are Shanghai, Jiangsu, Zhejiang, Fujian, and Anhui. The daily power supply pressure is relatively high, and it often participates in the mutual power dispatching with other provinces. Considering the line capacity limit and the power adjustment cost of conventional units, power cross-time-scale dispatching through an energy storage system is a relatively effective solution, which has the characteristics of flexibility, high efficiency, adjustable, fast response speed, and environmental friendliness. At present, there has been relatively in-depth research on energy storage peak shaving, but there is little research on the network-level dispatching situation. Therefore, in the embodiments of the present application, considering the common problem of insufficient unit output in the network-level dispatching situation and the resulting residual load situation, the network-province collaborative dispatching system constructs a residual load fluctuation optimization model based on the power data of the conventional generating units and the pumped-storage units, and uses the energy storage system to smooth the load of multiple provinces, and extends the goal of load leveling to reducing the degree of residual load fluctuation.

[0029] 102. Obtain the new energy dispatching data and the load curve output by the residual load fluctuation optimization model, and construct a provincial-level dispatching model based on the new energy dispatching data and the load curve.

[0030] At the provincial-level dispatching level, it is mainly responsible for the accommodation of new energy after the residual load peak shaving at the network level. However, in the face of the increasingly complex operating environment and various uncertain factors, traditional optimization methods often fail to meet the actual operating requirements. As an emerging optimization technology, the two-stage robust optimization method has attracted much attention in recent years. The two-stage robust optimization method has wide applicability in the application fields of power systems. Whether it is power market design, generation dispatching or power grid planning, the two-stage robust optimization method can be introduced to solve complex problems in actual operation. Especially in aspects such as new energy grid connection and cross-regional power trading, this method can effectively overcome the challenges brought by uncertainty and achieve the efficient operation of the power system and the optimal allocation of resources. Therefore, in the embodiments of the present application, the network-province collaborative dispatching system constructs a provincial-level dispatching model based on the new energy dispatching data and the load curve.

[0031] 103. Construct a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial-level dispatching model, decompose the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtain the network-province collaborative dispatching result.

[0032] In the embodiment of the present application, the provincial two-stage robust optimization model is constructed based on the residual load fluctuation optimization model and the provincial dispatching model in the network-province collaborative dispatching system. The provincial two-stage robust optimization model is decomposed into a master problem and a sub-problem for iterative solution to obtain the network-province collaborative dispatching result. Through two-stage robust optimization dispatching and combining the collaborative operations of daily dispatching and real-time dispatching, in the face of high uncertainty, the system can maintain the stability and economy of the power grid at an appropriate increase in cost.

[0033] The method provided by the embodiment of the present application obtains the power data of conventional generating units and pumped-storage generating units, constructs a residual load fluctuation optimization model based on the power data of conventional generating units and pumped-storage generating units, obtains the new energy dispatching data and the load curve output by the residual load fluctuation optimization model, constructs a provincial dispatching model based on the new energy dispatching data and the load curve, constructs a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial dispatching model, decomposes the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution to obtain the network-province collaborative dispatching result. The two-stage robust optimization dispatching model includes the robust optimization frameworks in the daily dispatching stage and the real-time dispatching stage, and is implemented through the column and constraint generation algorithm. This model can effectively handle the uncertainty of renewable energy output and load demand, and ensure the stable operation of the power grid in various extreme situations. Moreover, in the present application, energy storage systems such as pumped-storage power stations are used to manage and reduce the residual load, ensuring the stable operation of the provincial power grid. Through the refined management of the residual load, the collaborative optimization between multiple provincial power grids is realized, and the impact of peak load on the power system is reduced. In addition, in the present application, by setting uncertain parameters and robust optimization strategies, the volatility of renewable energy such as wind energy and photovoltaic is effectively handled to ensure the reliable operation of the power system, and by introducing uncertain parameters into the dispatching model, it is ensured that the power system can still operate stably when the fluctuations of wind power and photovoltaic power generation are large.

[0034] Further, as a refinement and extension of the specific implementation manner of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, the embodiment of the present application provides another network-province dispatching method considering source-load uncertainty, as Figure 2A shown, this method includes:

[0035] 201. Obtain the power data of conventional generating units and pumped-storage generating units, and determine the first decision variable based on the power data of conventional generating units and pumped-storage generating units.

[0036] At the regional power grid level, an optimization model aiming to minimize the residual load fluctuation is constructed in the embodiment of the present application. This model smooths the residual load curves of each provincial power grid by dispatching conventional generating units and pumped storage power stations, ensuring the stable operation of the provincial power grid. Through the refined management of the residual load, this method can achieve collaborative optimization among multiple provincial power grids and reduce the impact of peak load on the system. Specifically, the network-province collaborative dispatching system extracts the charge-discharge power and charge-discharge status of the output of conventional generating units from the power data of conventional generating units, extracts the charge-discharge power and charge-discharge status of pumped storage units from the power data of pumped storage units, and takes the charge-discharge power and charge-discharge status of the output of conventional generating units and the charge-discharge power and charge-discharge status of pumped storage units as the first decision variables.

[0037] 202. Establish the first constraint condition by using the power data of conventional generating units and the power data of pumped storage units.

[0038] In the embodiment of the present application, the types of units involved in the dispatching optimization problem at the network level can generally be divided into two types, namely conventional generating units and pumped storage power stations. The network-province collaborative dispatching system extracts the power generation power data of conventional generating units from the power data of conventional generating units, and generates the unit ramp constraint and the unit output boundary constraint by using the power generation power data of conventional generating units, as shown in Formula 1 below:

[0039] Formula 1:

[0040] Among them, P g,p,t+1 is the power generation power of the g-th conventional generating unit at the (t + 1)-th moment in the p-th province, P g,p,t is the power generation power of the g-th conventional generating unit at the t-th moment in the p-th province, RU g is the upper limit of the ramp rate of the g-th conventional generating unit. The upper limit of the ramp rate refers to the output increase or decrease per unit time. is the maximum output of the conventional generating unit, and the constraint is an empty lower bound, which is applicable to the power generation stage of all units.

[0041] Then, the network-province collaborative dispatching system generates the pumped storage unit constraint and the pumped storage unit boundary constraint by using the power data of pumped storage units. Among them, the pumped storage unit constraint includes the water level conversion energy equation constraint and the next-time period energy equation constraint, and the pumped storage unit boundary constraint includes the energy boundary constraint and the pumping and generating power constraint, as shown in Formula 2 below:

[0042] Formula 2: η charge ΔE = P charge t = η charge mgΔh

[0043]

[0044] Among them, E t is the equivalent stored energy of the pumped-storage unit at time t, ΔE is the change in equivalent energy when the water level changes, m is the mass of the water volume in the reservoir, g is the acceleration due to gravity of the water volume in the reservoir, h is the water level height of the water volume in the reservoir, ρ is the density of the water volume in the reservoir, V is the volume of the water volume in the reservoir, S is the bottom area of the reservoir, Δh is the change in water level height, Δt is the unit time period, P charge is the first intermediate variable, P discharge is the second intermediate variable, is the pumping power of the pumped-storage unit at time t, is the generating power of the pumped-storage unit at time t, η charge is the pumping and energy storage efficiency of the pumped-storage unit, η discharge is the discharging and generating efficiency of the pumped-storage unit, E min is the minimum value of the stored energy of the pumped-storage unit, E max is the maximum value of the stored energy of the pumped-storage unit, is the upper limit of the pumping and energy storage output of the pumped-storage unit, is the upper limit of the discharging and generating output of the pumped-storage unit, is the pumping and charging state of the pumped-storage unit at time t, is the discharging and generating state of the pumped-storage unit at time t, and is a Boolean variable. By using the pumped-storage power station for energy storage during the low-load period and releasing electric energy during the high-load period, the operating cost of the system can be effectively reduced.

[0045] Subsequently, the network-province coordinated dispatching system extracts the output power data from the power data of the conventional generating units and the pumped-storage units, and generates the output power ratio constraints by using the output power data. The output of all types of units needs to be distributed to each provincial power grid according to the agreed ratio. At the same time, the grid electric energy is consumed according to the corresponding ratio during the pumping and energy storage period, as shown in the following formula 3:

[0046] Formula 3:

[0047] Among them, P g,p,t ′ is the output power of the g-th unit in the p-th province at time t, P g,t is the output power of the g-th unit to the p-th province at time t, R g,p is the power transmission ratio of the g-th unit to the p-th province. The unit is a conventional generating unit or a pumped-storage unit.

[0048] Then, the network-province coordinated dispatching system takes the unit ramp rate constraint, the unit output boundary constraint, the pumped-storage unit constraint, the pumped-storage unit boundary constraint, and the unit output ratio constraint as the first constraint condition.

[0049] Based on the first decision variable and the first constraint condition, a first optimization objective is established, and a residual load fluctuation optimization model is constructed using the first decision variable, the first optimization objective, and the first constraint condition.

[0050] An optimization model is established for the above variables and constraints. The main objective at the network level is to use the unit output and pumped-storage power stations to smooth the residual load curves of each province, thereby reducing the impact of load peaks on provincial dispatching. To achieve this goal, referring to the generation dispatching function of the Chinese power system, the minimum residual load variance is selected as the objective function. Therefore, the objective function of this optimization problem is the following formula 4:

[0051] Formula 4:

[0052] Since there is a corresponding load variance for each province, this model is a multi-objective optimization model, and the weighted sum method is prepared to combine multiple objectives into an equivalent scalar objective.

[0053] To avoid unreasonable results caused by large differences in load magnitudes among multiple provincial power grids, the total accepted unit output is normalized, and then the objective functions of different provinces are integrated into a single objective with reasonable weighting coefficients. Therefore, the composite objective function can be expressed as the following formula 5:

[0054] Formula 5:

[0055] where obj′ p is the normalized form of the objective function for province p, that is, the following formula 6:

[0056] Formula 6:

[0057] where weight p is the weight of the p-th province in the objective function, and max(Load p ) is the maximum value in the load curve of the p-th province.

[0058] Therefore, substituting the total unit output gives the following formula 7:

[0059] Formula 7:

[0060] Because the variances of peak and valley periods are considered for separate optimization, the multi-objective problem is transformed into a single-objective problem again by weighting the two optimization objectives, that is, the following formula 8:

[0061] Formula 8: minobj = weight peak ×obj peak + weiht valley ×objvalley

[0062] wherein, weight peak、 weight valley are the corresponding weights in the peak period and the valley period respectively, and obj peak , obj valley are the corresponding objective functions in the peak period and the valley period respectively.

[0063] So far, using the above formula transformation, the multi-objective problem is transformed into a single-objective problem. It should be noted that in this model, the objective function is a quadratic function of the decision variables.

[0064] However, the power balance constraint in this model cannot be satisfied. After research and testing, simply deleting the power balance constraint will lead to the negative output of the conventional generating units and the non-uniqueness of the solution to the problem. The main reason is that the degree of relaxation of the model constraints is too high, and additional constraints need to be set or the objective function needs to be changed.

[0065] Therefore, the embodiment of this application adopts the penalty function method, adding a penalty term to the basis of the objective function, as shown in Formula 9 below:

[0066] Formula 9:

[0067] The construction idea of this penalty function comes from the weighted method. Among them, the max function is used to judge whether the maximum output of the unit in period t can meet the load demand. If it cannot be met, there will be a remaining load. The difference is used as the weight and multiplied by the difference between the actual unit output and the load. Adding this penalty term to the original objective function can solve the problem of negative output of conventional generating units. Therefore, the objective function after adding the penalty term is as shown in Formula 10 below:

[0068] Formula 10: minobj = weight peak ×obj peak +weight valley ×obj valley +φ

[0069] where:

[0070] [[ID=4i5]]

[0071] wherein, P′ p,t represents the total unit output received by the power grid p in period t, P g,p,t is the transmission power of unit g to power grid p in period t, and T is the total number of periods.

[0072] To sum up, the network-province collaborative dispatching system establishes the first optimization objective based on the first decision variable and the first constraint condition, as shown in Formula 11 below:

[0073] Formula 11: minobj = weight peak × obj peak + weight valley × obj valley + φ

[0074]

[0075]

[0076] Wherein, minobj is the first optimization objective, weight peak is the peak period weight, weight valley is the valley period weight, obj peak is the objective function of the peak period, obj valley is the objective function of the valley period, φ is the penalty term, obj′ p is the first objective function, and the first objective function is the objective function of the peak period or the objective function of the valley period, P′ p,t is the total unit output received by the power grid at time t, P′ p,i is the unit output of province p at the i-th moment, P g,p,t ” is the transmission power of the g-th unit at time t, T is the total number of periods, max(Load p ) is the maximum value in the load curve of province p, Load p,t is the load curve of province p at time t, is the maximum output of the conventional generator set. By using the pumped-storage power station to cut peaks and fill valleys for the remaining load of each province, the impact of peak load on the provincial power grid can be reduced, thereby stabilizing the overall operation of the power grid. By smoothing the load of each province, the load fluctuation during the peak period can be reduced, and the start-stop frequency of the standby generator set can be lowered, thereby improving the stability of the system.

[0077] The highest degree term of the objective function of this model is quadratic, so this optimization problem is a quadratic programming problem. To prove that this model can converge to the global optimum, the compact form of this model is as follows in Formula 12:

[0078] Formula 12:

[0079] In the compact form model, K eq , D are the coefficient matrices of the constraints in the model, and k, d are the constant term matrices of the corresponding constraints. Since the coefficient matrix a of the quadratic term of the objective function is a symmetric matrix, its eigenvalues are real numbers. The eigenvalue solution equation is as follows in Formula 13:

[0080] Formula 13: |λE - a| = 0

[0081] All the eigenvalues can be obtained as non - negative numbers, so \(a\) is a positive semi - definite matrix. Therefore, this problem is a convex quadratic optimization problem. For convex optimization, any local optimal solution is a global optimal solution under satisfied conditions, and this problem is solvable within polynomial time.

[0082] 204. Calculate the cost functions of controllable generating units, energy storage elements, demand response loads, and power purchase and sale costs using new - energy dispatching data and load curves, and respectively determine the constraint conditions corresponding to the cost function of controllable generating units, the cost function of energy storage elements, the cost function of demand response loads, and the cost function of power purchase and sale.

[0083] The provincial two - stage robust optimization model is based on the load curve after load shedding provided by the network - level dispatching strategy to model the source - load uncertainty problem under new - energy access.

[0084] Specifically, in the embodiment of this application, the network - province collaborative dispatching system extracts the power generation power data of controllable generating units from new - energy dispatching data and load curves, and calculates the cost function of controllable generating units and the corresponding constraint conditions of the cost function of controllable generating units using the power generation power data of controllable generating units, as shown in Formula 14 below:

[0085] Formula 14:

[0086]

[0087] Where \(C\) g is the cost function of controllable generating units, \(P\) h,t is the power generation power of controllable generating units at time \(t\), \(a\) is the first cost coefficient, \(b\) is the second cost coefficient, \(\Delta t\) is the unit time period, is the minimum power generation power of controllable generating units, is the maximum power generation power of controllable generating units. Controllable generating units include adjustable gas turbines, diesel - generating sets, etc.

[0088] Next, the network - province collaborative dispatching system extracts the charging power data and discharging power data of energy storage elements from new - energy dispatching data and load curves, and calculates the cost function of energy storage elements and the corresponding constraint conditions of the cost function of energy storage elements using the charging power data and discharging power data of energy storage elements, as shown in Formula 15 below:

[0089] Formula 15:

[0090] Where \(C\) s is the cost function of energy storage elements, is the charging power of energy storage elements at time \(t\), is the discharge power of the energy storage element at time t, η is the charge-discharge efficiency, and K s is the unit charge-discharge cost, and P max is the maximum allowable output of the energy storage element. is a Boolean variable representing the charge-discharge state of the energy storage element, and E min is the lower bound of the energy storage capacity, and E max is the upper bound of the energy storage capacity, E0 is the initial energy of the energy storage. is the charging power of the energy storage element at the i-th moment. is the discharge power of the energy storage element at the i-th moment.

[0091] Subsequently, the provincial and grid collaborative dispatching system considers the flexible dispatching process of the demand response load. Under the condition of meeting the electricity consumption characteristics of providing demand response services, the power grid can adjust the electricity consumption plan of users and needs to make appropriate compensation to users. Thus, the adjustment cost of the demand response load is as shown in Equation 16 below:

[0092] Equation 16:

[0093] where is the actual dispatching power of the demand response load at time t, and D DR is the total electricity consumption demand for all periods. are the minimum and maximum values of the demand response load in the t-th period, respectively.

[0094] Since the above formula is a non-linear function and variables need to be introduced to linearize it, the provincial and grid collaborative dispatching system extracts the actual dispatching power data of the demand response load and the expected electricity consumption power data of the demand response load from the new energy dispatching data and the load curve, and calculates the demand response load cost function and the corresponding constraint conditions of the demand response load cost function using the actual dispatching power data of the demand response load and the expected electricity consumption power data of the demand response load, as shown in Equation 17 below:

[0095] Equation 17:

[0096] where C DR is the demand response load cost function. is the first linearization auxiliary variable. is the second linearization auxiliary variable, and K DR is the demand response cost-related coefficient. is the actual dispatching power of the demand response load at time t. is the expected electricity consumption power of the demand response load at time t.

[0097] When the unit output cannot meet the actual load demand, it is necessary to purchase electricity from other departments. At the same time, when there is surplus electricity, it can be sold. Therefore, the power purchase and sale process must meet the basic power balance, as shown in the following formula 18:

[0098] Formula 18:

[0099] in, is the purchased power at time t, is the power sold at time t, P g,t is the output power of the g-th unit at time t, is the charging power of the energy storage element at time t, is the discharge power of the energy storage element at time t, is the actual dispatch power of the demand response load at time t, is the load data at time t in the power system operation data, is the photovoltaic power generation data at time t in the power system operation data.

[0100] Therefore, the grid-provincial coordinated dispatching system extracts power system operation data, day-ahead transaction prices, and unit output data from the renewable energy dispatching data and load curves. It uses the power system operation data, day-ahead transaction prices, unit output data, charging power data and discharging power data of energy storage elements, and actual dispatching power data of demand response loads to calculate the power purchase and sales cost function, as well as the constraints corresponding to the power purchase and sales cost function, as shown in the following formula 19:

[0101] Formula 19:

[0102] Among them, C m is the cost function for purchasing and selling electricity, is the purchased power at time t, is the power sold at time t, P t is the day-ahead electricity price, P is a Boolean variable representing the buying and selling status. max is the maximum switching power value.

[0103] 205. Generate a second objective function using the controllable generator set cost function, the energy storage element cost function, the demand response load cost function, and the power purchase and sale cost function.

[0104] In the embodiment of the present application, the grid-province coordinated dispatching system generates a second objective function using the controllable generator set cost function, the energy storage element cost function, the demand response load cost function, and the power purchase and sale cost function, as shown in the following formula 20:

[0105] Formula 20: minobj = C g +Cs +C DR +C m

[0106] where minobj is the second objective function, and C g is the cost function of the controllable generating set, C s is the cost function of the energy storage element, C DR is the cost function of the demand response load, C m is the cost function of power purchase and sale.

[0107] 206. Construct the uncertain parameters and determine the boundary constraints of the uncertain parameters.

[0108] In the embodiment of the present application, considering the conservatism of the robust optimization result, the provincial and network collaborative dispatching system adds an uncertain parameter variable and the maximum value of the uncertain parameter to the model to limit the number of times the photovoltaic and load data take the worst-case values, so that the result of the entire model will not be too conservative. Among them, the uncertain parameters and the boundary constraints of the uncertain parameters are as shown in Formula 21 below:

[0109] Formula 21:

[0110] where is the uncertain parameter of the photovoltaic output, is the uncertain parameter of the load demand, is the upper limit value of the uncertain parameter of the photovoltaic output, is the upper limit value of the uncertain parameter of the load demand. The uncertain parameters include the uncertain parameter of the photovoltaic output and the uncertain parameter of the load demand. The uncertain parameter is a Boolean variable indicating whether the worst-case value is taken, which is used to control the actual values of the photovoltaic and load data in the optimization model. Generally, the worst-case scenario is that the photovoltaic power generation is insufficient while the load demand is large. Therefore, consider the lower bound deviation of the photovoltaic and the upper bound deviation of the load, as shown in Formula 22 below:

[0111] Formula 22:

[0112] where is the predicted value of the photovoltaic data, is the predicted value of the load data, is the deviation value of the photovoltaic and load powers at time t obtained by fuzzy clustering. By introducing uncertainty parameters into the dispatching model, it can ensure that the system can still operate stably under the condition of large fluctuations in wind energy and photovoltaic power generation.

[0113] 207. A provincial dispatching model is constructed by using the second objective function, the constraint conditions corresponding to the controllable generator set cost function, the constraint conditions corresponding to the energy storage element cost function, the constraint conditions corresponding to the demand response load cost function, the constraint conditions corresponding to the power purchase and sale cost function, the uncertain parameters, and the boundary constraints of the uncertain parameters.

[0114] In the embodiment of the present application, the network-province collaborative dispatching system constructs a provincial dispatching model by using the second objective function, the constraint conditions corresponding to the controllable generator set cost function, the constraint conditions corresponding to the energy storage element cost function, the constraint conditions corresponding to the demand response load cost function, the constraint conditions corresponding to the power purchase and sale cost function, the uncertain parameters, and the boundary constraints of the uncertain parameters.

[0115] 208. A provincial two-stage robust optimization model is constructed based on the residual load fluctuation optimization model and the provincial dispatching model, and the provincial two-stage robust optimization model is decomposed into a master problem and a sub-problem for iterative solution to obtain the network-province collaborative dispatching result.

[0116] In the embodiment of the present application, the network-province collaborative dispatching system constructs a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial dispatching model. The compact form of the provincial dispatching problem is as shown in Formula 23 below:

[0117] Formula 23:

[0118] Among them, the variables represented by x are:

[0119]

[0120] The variables represented by y are:

[0121]

[0122] u represents the number of times of taking the worst-case scenario:

[0123]

[0124] c is the correlation constraint coefficient matrix of the y variable, G is the first coefficient matrix corresponding to the inequality constraint, h is the second coefficient matrix corresponding to the inequality constraint, E is the third coefficient matrix corresponding to the inequality constraint, M is the fourth coefficient matrix corresponding to the inequality constraint, G eq is the first coefficient matrix corresponding to the equality constraint, h eq is the second coefficient matrix corresponding to the equality constraint, E eq is the third coefficient matrix corresponding to the equality constraint, M eq is the fourth coefficient matrix corresponding to the equality constraint, is a Boolean variable representing the charge and discharge state of the energy storage element, is a Boolean variable representing the purchase and sale status, P g,t is the output power of the g-th unit at time t, is the charging power of the energy storage element at time t, is the discharging power of the energy storage element at time t, is the actual scheduling power of the demand response load at time t, is the first linearization auxiliary variable, is the second linearization auxiliary variable, is the purchase power at time t, is the sale power at time t, is the load data at time t, is the photovoltaic power generation data at time t, is the uncertain parameter of photovoltaic output, is the uncertain parameter of load demand.

[0125] For the objective function, it is divided into two layers, the outer layer represents solving the minimum value of the inner layer function and making a decision on the variable x, so that x is brought into the inner layer solution as a constant term; the inner layer function represents the cost minimization result in the worst case, y ∈ F(x,u) represents the feasible region of y under the values of the variable x and the uncertain parameter u, represents the optimization result under the feasible region at this time, represents that among all the cost minimization results, the worst case of photovoltaic and load is taken.

[0126] From the above analysis, the main problem and sub-problem of this problem can be sorted out:

[0127] The main problem is to make decisions on the state variables of energy storage and power purchase and sale in advance, select the predicted values of photovoltaic and load data, optimize, and obtain the lower bound of the final solution, as shown in Equation 24 below:

[0128] Equation 24:

[0129]

[0130] The sub-problem is to find the decision with the minimum cost in the worst case based on the determined values during the day, as shown in Equation 25 below:

[0131] Equation 25:

[0132]

[0133] where, u kis the uncertain parameter selected in the k-th iteration, α is the lower bound of the final solution, and x value is a constant.

[0134] Since the sub-problem is a bilevel problem, by using the KKT (Karush-Kuhn-Tucker) algorithm for constrained optimization, the minimization of the internal cost is reformulated into KKT conditions, and the bilevel problem is transformed into a single-level problem, as shown in Equation 26 below:

[0135] Equation 26:

[0136]

[0137] Since the problem contains non-linear constraints, the constraints are linearized by the Big M algorithm, and finally the sub-problem becomes a single-level optimization problem, that is, the target sub-problem, as shown in Equation 27 below:

[0138] Equation 27:

[0139]

[0140] where v is the first auxiliary operator of the Big M algorithm, l is the second auxiliary operator of the Big M algorithm, w is the third auxiliary operator of the Big M algorithm, and v, l, w are boolean variables, and M ∞ is a constant, π1 is the first Lagrange multiplier of the constrained optimization condition algorithm, and π2 is the second Lagrange multiplier of the constrained optimization condition algorithm.

[0141] Then, the provincial and grid collaborative scheduling system obtains the Constraint-and-Column Generation (C&CG) algorithm, and uses the C&CG algorithm to perform iterative calculations on the master problem and the target sub-problem to obtain the provincial and grid collaborative scheduling results. By setting uncertain parameters and robust optimization strategies, the volatility of renewable energy such as wind energy and photovoltaic energy can be effectively handled to ensure the reliable operation of the system.

[0142] Specifically, the provincial and network collaborative scheduling system obtains the initial iteration number, the initial upper bound value, and the initial lower bound value of the column constraint generation algorithm, solves the master problem, obtains the optimal decision result of the current iteration number, updates the initial lower bound value with the optimal decision result, and obtains the lower bound value of the current iteration number. Next, the provincial and network collaborative scheduling system uses the optimal decision result of the current iteration number to solve the objective sub-problem, obtains the optimal objective value and scenario probability of the current iteration number, updates the initial upper bound value with the optimal objective value and scenario probability of the current iteration number, and obtains the upper bound value of the current iteration number. Obtain the maximum gap between the upper and lower bounds of the column constraint generation algorithm, and compare the difference between the lower bound value and the upper bound value with the maximum gap between the upper and lower bounds; if the difference between the lower bound value and the upper bound value is less than the maximum gap between the upper and lower bounds, then use the optimal decision result of the current iteration number as the provincial and network collaborative scheduling result; if the difference between the lower bound value and the upper bound value is greater than or equal to the maximum gap between the upper and lower bounds, then use the column constraint generation algorithm to perform the next iteration calculation on the master problem and the objective sub-problem. Through two-stage robust optimization scheduling, combined with the collaborative operation of daily scheduling and real-time scheduling, in the face of high uncertainty, the system can maintain the stability and economy of the power grid while appropriately increasing the cost.

[0143] As Figure 2B shown in the C&CG algorithm logic diagram, set the initial lower bound value LB = -∞, the initial upper bound value UB = +∞, the initial iteration number n = 0, solve the master problem to obtain the optimal decision result, and update the lower bound value LB. Fix the optimal decision result, solve the sub-problem to obtain the optimal objective value and scenario probability, and update the upper bound value UB. If LB - UB < △, where △ is the maximum gap between the upper and lower bounds, then output the decision result; otherwise, update the variables and constraints in the master problem, n = n + 1, and re-solve the master problem. Using the KKT condition and the C&CG algorithm can effectively decompose complex optimization problems into master problems and sub-problems, improving the accuracy and execution efficiency of scheduling decisions. By iteratively solving the master problem and the sub-problem, the system can obtain the global optimal solution in a short time, thereby improving the response speed of scheduling.

[0144] The method provided by the embodiments of this application obtains the power data of conventional generating units and pumped-storage units, constructs a residual load fluctuation optimization model based on the power data of conventional generating units and pumped-storage units, obtains new energy scheduling data and the load curve output by the residual load fluctuation optimization model, constructs a provincial scheduling model based on the new energy scheduling data and the load curve, constructs a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial scheduling model, decomposes the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtains the network-province coordinated scheduling result. The two-stage robust optimization scheduling model includes robust optimization frameworks for the daily scheduling stage and the real-time scheduling stage, and is implemented through the column and constraint generation algorithm. This model can effectively handle the uncertainties of renewable energy output and load demand, and ensure the stable operation of the power grid under various extreme conditions. Moreover, this application manages and reduces the residual load through energy storage systems such as pumped-storage power stations to ensure the stable operation of the provincial power grid. Through the refined management of the residual load, the coordinated optimization among multiple provincial power grids is realized, and the impact of peak load on the power system is reduced. In addition, this application effectively handles the volatility of renewable energy such as wind energy and photovoltaic energy by setting uncertain parameters and robust optimization strategies to ensure the reliable operation of the power system, and ensures the stable operation of the power system even when the fluctuations of wind power and photovoltaic power generation are large by introducing uncertain parameters into the scheduling model.

[0145] Further, as Figure 1 a specific implementation of the method, the embodiments of this application provide a network-province scheduling device considering source-load uncertainty, as Figure 3 shown. The device includes: a first construction module 301, a second construction module 302, and a solution module 303.

[0146] The first construction module 301 is used to obtain the power data of conventional generating units and pumped-storage units, and construct a residual load fluctuation optimization model based on the power data of the conventional generating units and the pumped-storage units;

[0147] The second construction module 302 is used to obtain new energy scheduling data and the load curve output by the residual load fluctuation optimization model, and construct a provincial scheduling model based on the new energy scheduling data and the load curve;

[0148] The solution module 303 is used to construct a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial scheduling model, decompose the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtain the network-province coordinated scheduling result.

[0149] In a specific application scenario, the first construction module 301 is used to extract the charge-discharge power and charge-discharge status of the output of the conventional generating unit from the power data of the conventional generating unit, and extract the charge-discharge power and charge-discharge status of the pumped-storage unit from the power data of the pumped-storage unit; take the charge-discharge power and charge-discharge status of the output of the conventional generating unit and the charge-discharge power and charge-discharge status of the pumped-storage unit as the first decision variables; establish the first constraint condition by using the power data of the conventional generating unit and the power data of the pumped-storage unit, and establish the first optimization objective based on the first decision variables and the first constraint condition; construct the remaining load fluctuation optimization model by using the first decision variables, the first optimization objective, and the first constraint condition.

[0150] In a specific application scenario, the first construction module 301 is used to extract the power generation power data of the conventional generating unit from the power data of the conventional generating unit, and generate the unit ramp constraint and the unit output boundary constraint by using the power generation power data of the conventional generating unit.

[0151]

[0152] where P g,p,t+1 is the power generation power of the g-th conventional generating unit at the (t + 1)-th moment in the p-th province, and P g,p,t is the power generation power of the g-th conventional generating unit at the t-th moment in the p-th province, and RU g is the upper limit of the ramp rate of the g-th conventional generating unit, and the upper limit of the ramp rate refers to the output increase or decrease per unit time. is the maximum output of the conventional generating unit; generate the pumped-storage unit constraint and the pumped-storage unit boundary constraint by using the power data of the pumped-storage unit.

[0153] η charge ΔE = P charge t = η charge mgΔh

[0154]

[0155] where E t is the equivalent stored energy of the pumped-storage unit at the t-th moment, ΔE is the equivalent energy change when the water level changes, m is the mass of the water volume in the reservoir, g is the acceleration due to gravity of the water volume in the reservoir, h is the water level height of the water volume in the reservoir, ρ is the density of the water volume in the reservoir, V is the volume of the water volume in the reservoir, S is the bottom area of the reservoir, Δh is the water level height change amount, Δt is the unit time period, and P charge is the first intermediate variable, and P discharge is the second intermediate variable. is the pumping power of the pumped-storage unit at the t-th moment. is the power generation power of the pumped - storage unit at time t, η char ge is the pumping - storage efficiency of the pumped - storage unit, η discharge is the water - discharging power generation efficiency of the pumped - storage unit, E min is the minimum value of the stored energy of the pumped - storage unit, E max is the maximum value of the stored energy of the pumped - storage unit, is the upper limit of the pumping - storage output of the pumped - storage unit, is the upper limit of the water - discharging power generation output of the pumped - storage unit, is the pumping - charging state of the pumped - storage unit at time t, is the water - discharging power generation state of the pumped - storage unit at time t, and is a Boolean variable; extract the output power data from the power data of the conventional generator set and the power data of the pumped - storage unit, and generate the unit output ratio constraint using the output power data,

[0156]

[0157] where, P g,p,t ′ is the output power of the g - th unit in the p - th province at time t, P g,t is the output power of the g - th unit to the p - th province at time t, R g,p is the power transmission ratio of the g - th unit to transmit power to the p - th province, and the unit is the conventional generator set or the pumped - storage unit; take the unit ramp - up constraint, the unit output boundary constraint, the pumped - storage unit constraint, the pumped - storage unit boundary constraint, and the unit output ratio constraint as the first constraint condition; establish the first optimization objective based on the first decision variable and the first constraint condition,

[0158] minobj = weight peak ×obj peak +weight valley ×obj valley +φ

[0159]

[0160] where, minobj is the first optimization objective, weight peak is the peak - period weight, weight valley is the valley - period weight, obj peak is the objective function in the peak period, obj valley is the objective function in the valley period, φ is the penalty term, obj′ p is the first objective function, and the first objective function is the objective function in the peak period or the objective function in the valley period, P′ p,tis the total unit output received by the power grid during period t, P′ p,i is the unit output of the i-th moment in province p, P g,p,t ” is the transmission power of the g-th unit during period t, T is the total number of periods, max(Load p ) is the maximum value in the load curve of province p, Load p,t is the load curve of province p at period t, is the maximum output of conventional generating units.

[0161] In a specific application scenario, the second construction module 302 is used to calculate the controllable generating unit cost function, the energy storage element cost function, the demand response load cost function, and the power purchase and sale cost function by using the new energy scheduling data and the load curve, and respectively determine the constraint conditions corresponding to the controllable generating unit cost function, the constraint conditions corresponding to the energy storage element cost function, the constraint conditions corresponding to the demand response load cost function, and the constraint conditions corresponding to the power purchase and sale cost function; generate a second objective function by using the controllable generating unit cost function, the energy storage element cost function, the demand response load cost function, and the power purchase and sale cost function,

[0162] minobj = C g + C s + C DR + C m

[0163] where minobj is the second objective function, C g is the controllable generating unit cost function, C s is the energy storage element cost function, C DR is the demand response load cost function, C m is the power purchase and sale cost function; construct uncertain parameters, determine the boundary constraints of the uncertain parameters, and the uncertain parameters include photovoltaic output uncertain parameters and load demand uncertain parameters,

[0164]

[0165]

[0166] where, is the photovoltaic output uncertain parameter, is the load demand uncertain parameter, is the upper limit value of the photovoltaic output uncertain parameter, is the upper limit value of the uncertain parameter of the load demand; the provincial dispatch model is constructed by using the second objective function, the constraint conditions corresponding to the controllable generator set cost function, the constraint conditions corresponding to the energy storage element cost function, the constraint conditions corresponding to the demand response load cost function, the constraint conditions corresponding to the power purchase and sale cost function, the uncertain parameter, and the boundary constraints of the uncertain parameter.

[0167] In a specific application scenario, the second construction module 302 is configured to extract the power generation power data of the controllable generator set from the new energy dispatch data and the load curve, and calculate the controllable generator set cost function and the constraint conditions corresponding to the controllable generator set cost function by using the power generation power data of the controllable generator set.

[0168]

[0169] where C g is the controllable generator set cost function, P h,t is the power generation power of the controllable generator set at time t, a is the first cost coefficient, b is the second cost coefficient, Δt is the unit time period, is the minimum power generation power of the controllable generator set, is the maximum power generation power of the controllable generator set; extract the charging power data and discharging power data of the energy storage element from the new energy dispatch data and the load curve, and calculate the energy storage element cost function and the constraint conditions corresponding to the energy storage element cost function by using the charging power data and discharging power data of the energy storage element.

[0170]

[0171] where C s is the energy storage element cost function, is the charging power of the energy storage element at time t, is the discharging power of the energy storage element at time t, η is the charge-discharge efficiency, K s is the unit charge-discharge cost, P max is the maximum allowable output of the energy storage element, is a Boolean variable representing the charge-discharge state of the energy storage element, E min is the lower bound of the energy storage capacity, E max is the upper bound of the energy storage capacity, E0 is the initial energy of the energy storage, is the charging power of the energy storage element at the i-th moment, is the discharge power of the energy storage element at the i-th moment; extracting the actual scheduling power data of the demand response load and the expected power consumption data of the demand response load from the new energy scheduling data and the load curve, and calculating the demand response load cost function and the corresponding constraint conditions of the demand response load cost function by using the actual scheduling power data of the demand response load and the expected power consumption data of the demand response load,

[0172]

[0173] where C DR is the demand response load cost function, is the first linearization auxiliary variable, is the second linearization auxiliary variable, K DR is the demand response cost correlation coefficient, is the actual scheduling power of the demand response load at time t, is the expected power consumption of the demand response load at time t; extracting the power system operation data, the day-ahead trading electricity price and the unit output data from the new energy scheduling data and the load curve, and calculating the power purchase and sale cost function and the corresponding constraint conditions of the power purchase and sale cost function by using the power system operation data, the day-ahead trading electricity price, the unit output data, the charging power data and the discharge power data of the energy storage element, and the actual scheduling power data of the demand response load,

[0174]

[0175] where C m is the power purchase and sale cost function, is the purchased power at time t, is the sold power at time t, P g,t is the output power of the g-th unit at time t, is the charging power of the energy storage element at time t, is the discharge power of the energy storage element at time t, is the actual scheduling power of the demand response load at time t, is the load data at time t in the power system operation data, is the photovoltaic power generation data at time t in the power system operation data, P t is the day-ahead trading electricity price, is the boolean variable representing the purchase and sale status, P max is the maximum exchange power value.

[0176] In a specific application scenario, the solution module 303 is configured to construct a provincial two-stage robust optimization model based on the remaining load fluctuation optimization model and the provincial dispatching model.

[0177]

[0178] Among them, the variables represented by x are:

[0179]

[0180] The variables represented by y are:

[0181]

[0182] u represents the number of times the worst-case scenario is reached:

[0183]

[0184] c is the correlation constraint coefficient matrix of the y variable, G is the first coefficient matrix corresponding in the inequality constraint, h is the second coefficient matrix corresponding in the inequality constraint, E is the third coefficient matrix corresponding in the inequality constraint, M is the fourth coefficient matrix corresponding in the inequality constraint, G eq is the first coefficient matrix corresponding in the equality constraint, h eq is the second coefficient matrix corresponding in the equality constraint, E eq is the third coefficient matrix corresponding in the equality constraint, M eq is the fourth coefficient matrix corresponding in the equality constraint, is a Boolean variable representing the charge and discharge state of the energy storage element, is a Boolean variable representing the purchase and sale state, P g,t is the output power of the g-th unit at time t, is the charging power of the energy storage element at time t, is the discharging power of the energy storage element at time t, is the actual dispatching power of the demand response load at time t, is the first linearization auxiliary variable, is the second linearization auxiliary variable, is the purchase power at time t, is the sale power at time t, is the load data at time t, is the photovoltaic power generation data at time t, is the photovoltaic output uncertainty parameter, is the load demand uncertainty parameter; the provincial two-stage robust optimization model is decomposed into a master problem and a sub-problem, where the master problem is:

[0185]

[0186] The sub - problems are as follows:

[0187]

[0188] Among them, u k is the uncertain parameter selected in the k - th iteration, α is the lower bound of the final solution, and x value is a constant; Obtain the constrained optimization condition algorithm and the big M algorithm, and transform the sub - problem by using the constrained optimization condition algorithm and the big M algorithm to obtain the target sub - problem.

[0189]

[0190] Among them, v is the first auxiliary operator of the big M algorithm, l is the second auxiliary operator of the big M algorithm, w is the third auxiliary operator of the big M algorithm, and v, l, w are Boolean variables, and M ∞ is a constant, π1 is the first Lagrange multiplier of the constrained optimization condition algorithm, and π2 is the second Lagrange multiplier of the constrained optimization condition algorithm; Obtain the column - constraint generation algorithm, and perform iterative calculations on the main problem and the target sub - problem by using the column - constraint generation algorithm to obtain the network - province collaborative scheduling result.

[0191] In a specific application scenario, the solving module 303 is used to obtain the initial iteration number, the initial upper - bound value, and the initial lower - bound value of the column - constraint generation algorithm, solve the main problem to obtain the optimal decision result of the current iteration number, update the initial lower - bound value by using the optimal decision result to obtain the lower - bound value of the current iteration number; Solve the target sub - problem by using the optimal decision result of the current iteration number to obtain the optimal objective value and the scenario probability of the current iteration number, update the initial upper - bound value by using the optimal objective value and the scenario probability of the current iteration number to obtain the upper - bound value of the current iteration number; Obtain the maximum gap between the upper and lower bounds of the column - constraint generation algorithm, and compare the difference between the lower - bound value and the upper - bound value with the maximum gap between the upper and lower bounds; If the difference between the lower - bound value and the upper - bound value is less than the maximum gap between the upper and lower bounds, then use the optimal decision result of the current iteration number as the network - province collaborative scheduling result; If the difference between the lower - bound value and the upper - bound value is greater than or equal to the maximum gap between the upper and lower bounds, then perform the next - iteration calculation on the main problem and the target sub - problem by using the column - constraint generation algorithm.

[0192] The device provided by the embodiment of the present application acquires the power data of conventional generating units and pumped-storage units, constructs a residual load fluctuation optimization model based on the power data of conventional generating units and pumped-storage units, acquires the new energy scheduling data and the load curve output by the residual load fluctuation optimization model, constructs a provincial scheduling model based on the new energy scheduling data and the load curve, constructs a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial scheduling model, decomposes the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtains the network-province collaborative scheduling result. The two-stage robust optimization scheduling model includes the robust optimization frameworks in the daily scheduling stage and the real-time scheduling stage, and is implemented through the column and constraint generation algorithm. This model can effectively handle the uncertainties of renewable energy output and load demand, and ensure the stable operation of the power grid under various extreme conditions. Moreover, in the present application, energy storage systems such as pumped-storage power stations are used to manage and reduce the residual load, ensuring the stable operation of the provincial power grid. Through the refined management of the residual load, the collaborative optimization among multiple provincial power grids is achieved, and the impact of peak load on the power system is reduced. In addition, by setting uncertain parameters and robust optimization strategies, the present application effectively handles the volatility of renewable energy such as wind energy and photovoltaic energy, ensuring the reliable operation of the power system, and by introducing uncertain parameters into the scheduling model, ensuring that the power system can still operate stably under the condition of large fluctuations in wind energy and photovoltaic power generation.

[0193] It should be noted that for other corresponding descriptions of each functional unit involved in the network-province scheduling device considering source-load uncertainty provided by the embodiment of the present application, reference can be made to Figure 1 and Figures 2A to 2B the corresponding descriptions therein, which will not be elaborated here. The user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties.

[0194] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered that the scope described in this specification is covered. The above embodiments only represent several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several deformations and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

[0195] In an exemplary embodiment, refer to Figure 4, a device is also provided. The device includes a bus, a processor, a memory, and a communication interface, and may further include an input / output interface and a display device. Among them, each functional unit can complete mutual communication through the bus. The memory stores a computer program, and the processor is configured to execute the program stored on the memory to execute the provincial power grid dispatching method considering source-load uncertainty in the above embodiments.

[0196] A medium stores a computer program, and when the computer program is executed by a processor, the steps of the provincial power grid dispatching method considering source-load uncertainty are implemented.

[0197] Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented through hardware or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of the present application.

[0198] Those skilled in the art can understand that the drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the drawings are not necessarily essential for implementing the present application. Those skilled in the art can understand that the modules in the device in the implementation scenario can be distributed in the device in the implementation scenario according to the description of the implementation scenario, or can be correspondingly changed and located in one or more devices different from this implementation scenario. The modules in the above implementation scenario can be combined into one module, or can be further split into multiple sub-modules.

[0199] The above serial numbers of the present application are only for description and do not represent the advantages or disadvantages of the implementation scenarios.

[0200] The above discloses only several specific implementation scenarios of the present application. However, the present application is not limited thereto, and any changes that can be thought of by those skilled in the art should fall within the protection scope of the present application.

Claims

1. A coordinated dispatching method for provincial power grids considering the uncertainties of power sources and loads, characterized in that Including: Obtain the power data of conventional generator sets and pumped-storage power generation units, and construct a residual load fluctuation optimization model based on the power data of the conventional generator sets and the pumped-storage power generation units; Obtain the new energy dispatching data and the load curve output by the residual load fluctuation optimization model, and construct a provincial dispatching model based on the new energy dispatching data and the load curve; Construct a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial dispatching model, decompose the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtain the network-province coordinated dispatching result.

2. The method according to claim 1, wherein The constructing of the residual load fluctuation optimization model based on the power data of the conventional generator sets and the pumped-storage power generation units includes: Extract the charge-discharge power and charge-discharge state of the output of the conventional generator sets from the power data of the conventional generator sets, and extract the charge-discharge power and charge-discharge state of the pumped-storage power generation units from the power data of the pumped-storage power generation units; Take the charge-discharge power and charge-discharge state of the output of the conventional generator sets and the charge-discharge power and charge-discharge state of the pumped-storage power generation units as the first decision variables; Use the power data of the conventional generator sets and the pumped-storage power generation units to establish the first constraint conditions, and establish the first optimization objective based on the first decision variables and the first constraint conditions; Construct the residual load fluctuation optimization model by using the first decision variables, the first optimization objective, and the first constraint conditions.

3. The method according to claim 2, wherein The establishing of the first constraint conditions by using the power data of the conventional generator sets and the pumped-storage power generation units and the establishing of the first optimization objective based on the first decision variables and the first constraint conditions include: Extract the power generation power data of the conventional generator sets from the power data of the conventional generator sets, and generate the unit ramp constraint and the unit output boundary constraint by using the power generation power data of the conventional generator sets; -RU g ≤P g,p,t+1 -P g,p,t ≤RU g Among them, P g,p,t+1 is the power generation power of the g-th conventional generating unit in the p-th province at the (t + 1)-th moment, and P g,p,t is the power generation power of the g-th conventional generating unit in the p-th province at the t-th moment, and RU g is the upper limit of the ramp rate of the h-th conventional generating unit. The upper limit of the ramp rate refers to the output increase or decrease per unit time, is the maximum output of the conventional generating unit; Generate the pumped-storage power generation unit constraint and the pumped-storage power generation unit boundary constraint by using the power data of the pumped-storage power generation units; Among them, E t is the equivalent stored energy of the pumped-storage unit at time t, ΔE is the equivalent energy change when the water level changes, m is the mass of the water volume in the reservoir, g is the acceleration due to gravity of the water volume in the reservoir, h is the water level height of the water volume in the reservoir, ρ is the density of the water volume in the reservoir, V is the volume of the water volume in the reservoir, S is the bottom area of the reservoir, Δh is the change in water level height, Δt is the unit time period, P charge is the first intermediate variable, P discharge is the second intermediate variable, is the pumping power of the pumped-storage unit at time t, is the power generation power of the pumped-storage unit at time t, η charge is the pumping and energy storage efficiency of the pumped-storage unit, η discharge is the water release and power generation efficiency of the pumped-storage unit, E min is the minimum value of the stored energy of the pumped-storage unit, E max is the maximum value of the stored energy of the pumped-storage unit, is the upper limit of the pumping and energy storage output of the pumped-storage unit, is the upper limit of the water release and power generation output of the pumped-storage unit, is the pumping and charging state of the pumped-storage unit at time t, is the water release and power generation state of the pumped-storage unit at time t, and is a Boolean variable; Extract the output power data from the power data of the conventional generator sets and the pumped-storage power generation units, and generate the unit output ratio constraint by using the output power data; Among them, P g,p,t ′ is the output power of the g-th unit in the p-th province at time t, and P g,t is the output power of the g-th unit to the p-th province at time t, and R g,p is the power transmission ratio of the g-th unit to the p-th province. The unit is the conventional generating unit or the pumped storage unit; Take the unit ramp constraint, the unit output boundary constraint, the pumped-storage power generation unit constraint, the pumped-storage power generation unit boundary constraint, and the unit output ratio constraint as the first constraint conditions; Establish the first optimization objective based on the first decision variables and the first constraint conditions. Among them, minobj is the first optimization objective, weight peak is the peak period weight, weight valley is the valley period weight, obj peak is the objective function for the peak period, obj valley is the objective function for the valley period, φ is the penalty term, obj′ p is the first objective function, and the first objective function is the objective function for the peak period or the objective function for the valley period, P′ p,t is the total unit output received by the power grid at time t, P′ p,i is the unit output of the i-th moment in province p, P g,p,t ” is the transmission power of the g-th unit at time t, T is the total number of periods, max(Loda p ) is the maximum value in the load curve of province p, Loda p,t is the load curve of province p at time t, is the maximum output of the conventional generator set.

4. The method according to claim 1, characterized in that The constructing of the provincial dispatching model based on the new energy dispatching data and the load curve includes: Calculate the controllable generator set cost function, the energy storage element cost function, the demand response load cost function, and the power purchase and sale cost function by using the new energy dispatching data and the load curve, and respectively determine the constraint conditions corresponding to the controllable generator set cost function, the constraint conditions corresponding to the energy storage element cost function, the constraint conditions corresponding to the demand response load cost function, and the constraint conditions corresponding to the power purchase and sale cost function; Generate a second objective function by using the controllable generator set cost function, the energy storage element cost function, the demand response load cost function, and the power purchase and sale cost function. minobj = C g + C s + C DR + C m where minobj is the second objective function, C g is the controllable generating unit cost function, C s is the energy storage element cost function, C DR is the demand response load cost function, C m is the power purchase and sale cost function; Construct uncertain parameters, and determine the boundary constraints of the uncertain parameters. The uncertain parameters include photovoltaic output power uncertain parameters and load demand uncertain parameters. Among them, is the uncertain parameter of PV output, is the uncertain parameter of load demand, is the upper limit value of the uncertain parameter of PV output, is the upper limit value of the uncertain parameter of load demand; Construct the provincial dispatching model by using the second objective function, the constraint conditions corresponding to the controllable generator set cost function, the constraint conditions corresponding to the energy storage element cost function, the constraint conditions corresponding to the demand response load cost function, the constraint conditions corresponding to the power purchase and sale cost function, the uncertain parameters, and the boundary constraints of the uncertain parameters.

5. The method according to claim 4, characterized in that The steps of calculating the controllable generator set cost function, the energy storage element cost function, the demand response load cost function, and the power purchase and sale cost function by using the new energy dispatching data and the load curve, and respectively determining the constraint conditions corresponding to the controllable generator set cost function, the constraint conditions corresponding to the energy storage element cost function, the constraint conditions corresponding to the demand response load cost function, and the constraint conditions corresponding to the power purchase and sale cost function are as follows: Extract the power generation power data of the controllable generator set from the new energy dispatching data and the load curve, and calculate the controllable generator set cost function and the constraint conditions corresponding to the controllable generator set cost function by using the power generation power data of the controllable generator set. Among them, C g is the cost function of the controllable generating set, P h,t is the power generation power of the controllable generating set at time t, a is the first cost coefficient, b is the second cost coefficient, Δt is the unit time period, is the minimum power generation power of the controllable generating set, is the maximum power generation power of the controllable generating set; Extract the charging power data and discharging power data of the energy storage element from the new energy dispatching data and the load curve, and calculate the energy storage element cost function and the constraint conditions corresponding to the energy storage element cost function by using the charging power data and discharging power data of the energy storage element. Among them, C s is the cost function of the energy storage element, is the charging power of the energy storage element at time t, is the discharging power of the energy storage element at time t, η is the charge-discharge efficiency, K s is the unit charge-discharge cost, P max is the maximum allowable output of the energy storage element, is a Boolean variable representing the charge-discharge state of the energy storage element, E min is the lower bound of the energy storage capacity, E max is the upper bound of the energy storage capacity, E0 is the initial energy of the energy storage, is the charging power of the energy storage element at the i-th moment, is the discharging power of the energy storage element at the i-th moment; Extract the actual dispatching power data of the demand response load and the expected power consumption data of the demand response load from the new energy dispatching data and the load curve, and calculate the demand response load cost function and the constraint conditions corresponding to the demand response load cost function by using the actual dispatching power data of the demand response load and the expected power consumption data of the demand response load. Among them, C DR is the demand response load cost function, is the first linearization auxiliary variable, is the second linearization auxiliary variable, K DR is the demand response cost correlation coefficient, is the actual scheduling power of the demand response load at time t, is the expected power consumption of the demand response load at time t; Extract the power system operation data, the day-ahead trading electricity price, and the unit output data from the new energy dispatching data and the load curve, and calculate the power purchase and sale cost function and the constraint conditions corresponding to the power purchase and sale cost function by using the power system operation data, the day-ahead trading electricity price, the unit output data, the charging power data and discharging power data of the energy storage element, and the actual dispatching power data of the demand response load. Among them, C m is the cost function of power purchase and sale, is the purchased power at time t, is the sold power at time t, P g,t is the output power of the g-th unit at time t, is the charging power of the energy storage element at time t, is the discharging power of the energy storage element at time t, is the actual scheduling power of the demand response load at time t, is the load data at time t in the power system operation data, is the photovoltaic power generation data at time t in the power system operation data, P t is the day-ahead trading electricity price, is a Boolean variable representing the purchase and sale status, P max is the maximum exchange power value.

6. The method according to claim 1, wherein Construct a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial dispatching model, and decompose the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution to obtain the network-province coordinated dispatching result, including: Construct a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial dispatching model. Among them, the variables represented by x are: The variables represented by y are: u represents the number of times of taking the worst case: c is the correlation constraint coefficient matrix of the y variable, G is the first coefficient matrix corresponding to the inequality constraint, h is the second coefficient matrix corresponding to the inequality constraint, E is the third coefficient matrix corresponding to the inequality constraint, M is the fourth coefficient matrix corresponding to the inequality constraint, G eq is the first coefficient matrix corresponding to the equality constraint, h eq is the second coefficient matrix corresponding to the equality constraint, E eq is the third coefficient matrix corresponding to the equality constraint, M eq is the fourth coefficient matrix corresponding to the equality constraint, is a Boolean variable representing the charge and discharge state of the energy storage element, is a Boolean variable representing the purchase and sale state, P g,t is the output power of the g-th unit at time t, is the charging power of the energy storage element at time t, is the discharging power of the energy storage element at time t, is the actual scheduling power of the demand response load at time t, is the first linearization auxiliary variable, is the second linearization auxiliary variable, is the purchase power at time t, is the sale power at time t, is the load data at time t, is the photovoltaic power generation data at time t, is the uncertain parameter of photovoltaic output, is the uncertain parameter of load demand; Decompose the provincial two-stage robust optimization model into a master problem and a sub-problem, where the master problem is: The sub-problem is: where, u k is the uncertain parameter selected in the k-th iteration, α is the lower bound of the final solution, and x value is a constant; Obtain the constrained optimization condition algorithm and the big-M algorithm, and use the constrained optimization condition algorithm and the big-M algorithm to transform the sub-problem to obtain the target sub-problem. max cy wherein, v is the first auxiliary operator of the big M algorithm, l is the second auxiliary operator of the big M algorithm, w is the third auxiliary operator of the big M algorithm, and v, l, and w are Boolean variables, M ∞ is a constant, π1 is the first Lagrange multiplier of the constrained optimization condition algorithm, and π2 is the second Lagrange multiplier of the constrained optimization condition algorithm; Obtain the column constraint generation algorithm, and use the column constraint generation algorithm to perform iterative calculations on the master problem and the target sub-problem to obtain the network-province coordinated scheduling result.

7. The method according to claim 6, characterized in that, The step of using the column constraint generation algorithm to perform iterative calculations on the master problem and the target sub-problem to obtain the network-province coordinated scheduling result includes: Obtain the initial iteration number, the initial upper bound value, and the initial lower bound value of the column constraint generation algorithm, solve the master problem to obtain the optimal decision result of the current iteration number, and update the initial lower bound value with the optimal decision result to obtain the lower bound value of the current iteration number; Use the optimal decision result of the current iteration number to solve the target sub-problem to obtain the optimal objective value and the scenario probability of the current iteration number, and update the initial upper bound value with the optimal objective value and the scenario probability of the current iteration number to obtain the upper bound value of the current iteration number; Obtain the maximum gap between the upper and lower bounds of the column constraint generation algorithm, and compare the difference between the lower bound value and the upper bound value with the maximum gap between the upper and lower bounds; If the difference between the lower bound value and the upper bound value is less than the maximum gap between the upper and lower bounds, then use the optimal decision result of the current iteration number as the network-province coordinated scheduling result; If the difference between the lower bound value and the upper bound value is greater than or equal to the maximum gap between the upper and lower bounds, then use the column constraint generation algorithm to perform the next iteration calculation on the master problem and the target sub-problem.

8. A provincial grid dispatching device considering the uncertainties of power sources and loads, characterized in that, It includes: The first construction module is used to obtain the power data of conventional generating units and the power data of pumped storage units, and construct a residual load fluctuation optimization model based on the power data of conventional generating units and the power data of pumped storage units; The second construction module is used to obtain the new energy scheduling data and the load curve output by the residual load fluctuation optimization model, and construct a provincial scheduling model based on the new energy scheduling data and the load curve; The solution module is used to construct a provincial two-stage robust optimization model based on the residual load fluctuation optimization model and the provincial scheduling model, decompose the provincial two-stage robust optimization model into a master problem and a sub-problem for iterative solution, and obtain the network-province coordinated scheduling result.

9. An apparatus, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A medium on which a computer program is stored, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 7.

Images