CN114725926A

CN114725926A - Toughness-improvement-oriented black start strategy for distributed resource-assisted main network key nodes

Info

Publication number: CN114725926A
Application number: CN202210269343.2A
Authority: CN
Inventors: 刘方; 张�浩; 徐耀杰; 杨秀; 李承泽; 蒋家富; 汤金璋; 仇志鑫; 焦楷丹; 刘欣雨; 蒋倩; 张倩倩; 周从亨; 高凌霄
Original assignee: Shanghai University of Electric Power
Current assignee: Shanghai University of Electric Power
Priority date: 2022-03-18
Filing date: 2022-03-18
Publication date: 2022-07-08

Abstract

The invention discloses a toughness-improvement-oriented black start strategy for key nodes of a distributed resource-assisted main network, which comprises the following steps: step 1, analyzing output characteristics and regulation strategies of each DER in a power distribution network, forming a method for evaluating the rapid recovery feasibility of DER supporting a main network key node, establishing black start performability evaluation indexes, and selecting a DER control strategy; step 2, establishing a power distribution network DER starting and grid frame reconstruction combined optimization model based on DER control strategies, solving by adopting a method combining a genetic algorithm and a Floeider algorithm to obtain an optimal starting sequence of each DER, stably starting and connecting each DER according to the optimal starting sequence, and polymerizing to form a regional power supply; and 3, starting the thermal power plant auxiliary machines in batches by the regional power supply, starting the thermal power generating units after the thermal power plant auxiliary machines completely and normally operate, and continuously accessing the recovered thermal power generating units to the load of the peripheral power distribution network and providing starting power supply support for other thermal power plants, so that the recovery range is gradually expanded until the whole system is recovered.

Description

Toughness-improvement-oriented black start strategy for distributed resource-assisted main network key nodes

Technical Field

The invention relates to the field of power grids, in particular to a toughness-improvement-oriented black start strategy for key nodes of a distributed resource auxiliary main network.

Background

The effective implementation of the strategic target of 'double carbon' and the continuous promotion of energy transformation accelerate the construction process of 'a novel power system taking new energy as a main body' in China. High-proportion new energy and power electronic equipment in the power system are connected in a large scale, alternating current and direct current series-parallel connection power distribution is developed rapidly, the characteristic of load diversification is enhanced, the power distribution network gradually presents the characteristics of source generation, flexibility and multiple randomness, and the situation and the role of the power distribution network in the novel power system are more complicated. Meanwhile, internal and external security threats of the power system, such as cascading failures, broadband oscillation, network attack, artificial operation problems, extreme natural disasters and the like, increase year by year, and provide new challenges for the safe operation of the power system.

As an important development direction of a future power system, the tough power grid has 6 key features in a broad sense: sensory, synergistic, strain, defense, restorative and learning forces. The restoring force is used as an important function for coping with extreme weather, natural disasters and major power failure accidents caused by physical and network attacks, loss is reduced to the maximum extent, and multi-link potential needs to be excavated to support rapid restoration after power grid failure. However, as the scale of the power grid is gradually enlarged, the number and the layout of the traditional black start power supply are difficult to meet the requirements of rapidity and reliability of restoration, the traditional black start power supply points are expensive to manufacture and limited by the availability of energy resources, and the safe operation economy of the power grid is restricted to a certain extent. Therefore, how to realize the rapid restoration of the power grid infrastructure facing internal and external risks based on the existing resources and how to make a reasonable and effective black start scheme by aiming at solving the problem of fault restoration becomes an urgent need, which has important significance for improving the toughness level of the power grid and guaranteeing the national energy safety.

The power grid black start scheme can be divided into three layers of a micro-power grid, a power distribution network and a power transmission network according to the scale and the voltage grade. The conventional black start scheme is more oriented to the power transmission network layer, mainly takes a hydroelectric generating set with strong self-starting capability or a large thermal generating set with black start capability as a black start power supply, gradually drives other generating sets without black start capability to start, and finally recovers the whole power network; the recovery of the power distribution network usually needs to wait for the recovery of the superior power grid and then supply power to the critical loads of the superior power grid, so that the recovery time is long and the recovery force is insufficient. Along with the gradual tightness of the degree of association coupling between a distribution network and a main network, a large-scale distributed power supply (DG), an electric automobile, energy storage, a micro-grid and other diversified 'source-load-storage' distributed resources (DER) are dispersedly accessed, based on the advantages of small DER starting power, high speed, flexible reaction and the like, the important load power supply is ensured by aggregating multiple DER auxiliary distribution network fault recovery or local self-healing, and the DER is utilized to rapidly start key nodes such as a main network power supply, a transformer substation and the like to provide an auxiliary power supply and flexible resource support, so that the whole network recovery process is accelerated, the toughness level of the power network is improved in the aspect of recovery power, the problems of insufficient black-start power supply and high cost of the power network are solved, and the method becomes a new idea and a new direction for making a black-start scheme under the construction background of a 'novel power system'.

For the black start of DER auxiliary micro-grid and distribution network, the students at home and abroad have developed preliminary research. Research is carried out on a DG starting sequence in the micro-grid black start, a multi-objective optimization model considering DG power generation and load importance is established at the same time, and coordinated distribution of DG starting and load recovery time sequences is achieved. Researches focus on distributed resources such as photovoltaic power, wind power, energy storage, electric vehicles and micro-grids as black-start power sources to participate in recovery of a power distribution network system, the uncontrollable DER is started by the DER with black-start capability, a recovery path and a target grid frame are coordinated and optimized, the recovery range is gradually expanded, and meanwhile, important load power supply is guaranteed; the method comprises the following steps that a network frame reconstruction double-layer optimization model considering an electric vehicle charging station is established by taking the maximum available generated energy of a system and the minimum charging capacitance of a recovery path as targets; the method has the advantages of researching and analyzing the fault recovery of the microgrid serving as a black-start power supply support distribution network, establishing a grid reconstruction model for the recovery of the microgrid support distribution network by taking the maximum power generation amount and the minimum load loss of the system recovery as targets, and verifying the advantages of the recovery scheme in the aspects of economy and reliability based on a 39-node system of a new England 10 machine. However, the above research only aims at the black start of the microgrid or the power distribution network layer, the supporting effect of the power distribution network resources on the black start of the power transmission network layer is not considered, and the improvement of the overall resilience of the system is limited.

Black start schemes for the DER auxiliary grid level of the distribution grid typically include three phases: a power supply black start stage, a net rack reconstruction stage and a load recovery stage. The existing research on the black start of the DER auxiliary power transmission network mainly focuses on the black start stage of a power supply, the DER with the black start capability is started preferentially to provide starting power for a large thermal power generating unit which cannot be started automatically in the power transmission network, and time is saved by the parallel recovery of partitions of the power transmission network. The scholars drive the factory asynchronous motor by taking the optical storage combined power generation system as a black start power supply and simulate the whole black start process of DER supporting the key nodes of the main network, and the prepared optical storage system coordination control strategy has strong engineering applicability; research proposes configuring an energy storage system for a wind power plant to enable the wind power plant to have black start capability, formulating a strategy for rapidly starting a thermoelectric unit by the wind storage system, and analyzing the starting processes of an empty charging transmission line, a transformer and an auxiliary machine in detail through simulation; a black start scheme of a power transmission network system is researched and optimized, a micro gas turbine near a power plant is used as a novel standby power supply of a large-scale unit, the problems that a standby diesel generator set is poor in economy and insufficient in a black start power supply in an urban power grid are solved, simulation verification is carried out through the starting of an air charging transformer and an asynchronous motor of the micro gas turbine, and a black start strategy is feasible. The research effectively shortens the fault recovery time of the power transmission network by utilizing the DER of the power distribution network, but mainly focuses on the analysis of a single problem that the DER of the power distribution network supports the large-scale main network unit to start in the recovery process, only relates to the safety check and scheme evaluation of the processes of self-starting DER, an empty charging line and a transformer, starting auxiliary machines of a power plant and the like, does not consider the processes of DER starting sequence optimization, recovery path formulation, target grid frame reconstruction and the like in the recovery process, three aspects in the black start of the actual power grid are not mutually independent, but are mutually driven and mutually supported, and joint optimization needs to be considered comprehensively to construct a black start dynamic overall process scheme under the conditions of a whole period and multiple time scales.

Disclosure of Invention

The present invention is made to solve the above problems, and an object of the present invention is to provide a distributed resource assisted main network key node black start strategy for toughness improvement.

The invention provides a toughness improvement-oriented distributed resource assisted main network key node black start strategy, which is characterized by comprising the following steps of: step 1, analyzing output characteristics and regulation strategies of each DER in a power distribution network, forming a method for evaluating the rapid recovery feasibility of the DER supporting a key node of a main network, establishing black start performability evaluation indexes, evaluating black start executable time periods in one day, and selecting a DER control strategy meeting power supply characteristics and system recovery requirements for the black start executable time periods based on evaluation results; step 2, establishing a power distribution network DER starting and grid frame reconstruction combined optimization model with the minimum recovery time, the maximum available power generation capacity and the minimum recovery path weight as targets based on DER control strategies, solving by adopting a method of combining a genetic algorithm and a Floiede algorithm to obtain an optimal starting sequence of each DER, stably starting and connecting the DERs according to the optimal starting sequence, and aggregating to form a regional power supply; and 3, starting the thermal power plant auxiliary machines in batches by the regional power supply, starting the thermal power generating units after the thermal power plant auxiliary machines completely and normally operate, and continuously accessing the recovered thermal power generating units to the load of the peripheral power distribution network and providing starting power supply support for other thermal power plants, so that the recovery range is gradually expanded until the whole system is recovered.

In the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the invention, the method can also have the following characteristics: in the step 2, in the step 1, the DER comprises a fan, a photovoltaic, an energy storage and a micro gas turbine. The step 1 comprises the following steps: step 1-1, predicting the available power generation capacity of the DER in a future period of time in the black start process aiming at the output characteristics and the uncertainty of time-space distribution of the multi-type DER, and processing prediction error deviation by using a confidence interval method to obtain a DER output prediction result of a power distribution network system; step 1-2, based on a DER output prediction result of a power distribution network system, considering two aspects of minimum power requirement of starting of an auxiliary machine of a main network large-scale unit and continuous execution inclination of black start, establishing a black start executable evaluation index, and evaluating the executable time period of the black start in one day to obtain an evaluation result; and 1-3, selecting a DER control strategy which meets the power supply characteristics and system recovery requirements for the power failure time of the executable black start scheme based on the evaluation result, and ensuring the successful start and stable operation of the fault system.

In the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the invention, the method can also have the following characteristics: in the step 1-1, the multi-type DER output characteristics comprise wind power output characteristics, photovoltaic unit output characteristics and energy storage operation characteristics. The energy storage operation characteristic considering energy storage charging and discharging power constraint and energy storage charge state constraint is as follows:

-P_ES,N≤P_ES(t)≤P_ES,N (1)

SOC_ES,min≤SOC_ES(t)≤SOC_ES,max (2)

in the formula, P_ES,NRated power for energy storage; SOC_ES,max、SOC_ES,minAnd SOC_ES(t) respectively representing the upper limit and the lower limit of the charge state of the energy storage device and the charge state of the energy storage device in the black starting process; SOC_ES(t) the calculation formula can be expressed as:

in the formula, delta is the energy storage self-discharge rate; eta_c、η_dRespectively charge and discharge efficiency;

the charging and discharging power for energy storage at the time t; s_sIs the capacity of the energy storage device.

In the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the invention, the method also has the following characteristics: wherein, the step 1-2 comprises the following steps: step 1-2-1, confidence interval estimation: the actual output of the generator set is distributed around the predicted value, the actual output of the wind power generator set and the photovoltaic generator set in each time period in the black start process is expressed as the sum of the predicted value and the prediction error, and the description is as follows:

in the formula, P_WT(t)、P_PV(t) respectively representing the actual power generation power of the wind power generation unit and the photovoltaic unit at the moment t,

respectively represents the predicted power of the wind power generator and the photovoltaic generator at the moment t,

and respectively representing the output prediction errors of the wind power generator and the photovoltaic generator at the moment t. Uncertainty handlingThe method adopts a confidence interval method to ensure that a scheme meets a certain confidence level, firstly, based on the output predicted values of the wind power and photovoltaic units, an uncertain model is adopted to fit the predicted error distribution in different power sections to obtain a probability density function, secondly, based on the probability density function, an interval set with the accumulated probability more than or equal to the confidence level 1-alpha is extracted, and finally, the interval with the shortest length under the same confidence level 1-alpha is selected from the interval set to serve as the confidence interval under the predicted power. The confidence interval under the confidence 1-alpha calculated based on the principle of the shortest confidence interval is as follows

In the formula, Δ P is an error corresponding to the power prediction value P, and conforms to the probability distribution F (Δ P); and 1-2-2, evaluating the black start performability by adopting two evaluation indexes of the minimum power requirement of the black start and the continuous execution inclination of the black start. Minimum power requirement for black start is P_DER,minComprises the following steps:

P_DER,min＝(1+α)P_M (6)

in the formula, P_MRated power of an auxiliary engine of the thermal power generating unit to be started; the coefficient alpha comprises the per se power consumption rate, the line loss rate and a certain reserved allowance, and the uncertainty influence on the output of the wind power generation unit and the photovoltaic unit is dealt with by reserving the rotary reserve. DER power generation amount Q in black start continuous process_DERMinimum required electric quantity Q of black start_DER,minAnd the black start continuous execution inclination γ is expressed as:

in the formula, P_DER(t) is DER's force at time t, where γ is greater than or equal to 1 indicates a higher sustainable execution inclination, and where γ is less than 1 indicates that the black start scenario is more prone to be non-executable.

In the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the invention, the method can also have the following characteristics: in the step 2, a power distribution network DER starting and grid frame reconstruction combined optimization model is established according to the characteristic that the DER starting sequence and the recovery path are mutually driven and mutually supported. The process of solving for the optimal start-up sequence includes the following sub-steps: step 2-1, selecting an energy storage and micro gas turbine with self-starting capability and stronger adjusting performance as a black-start power supply for self-starting, and optimizing DER model selection and DER starting sequence through a genetic algorithm by taking shortest system recovery time, maximum available power generation capacity and minimum recovery path weight as targets; and 2-2, based on the line importance evaluation, carrying out path optimization on the DER starting sequence by adopting a shortest path algorithm to obtain a shortest path, transmitting the shortest path back to the DER starting sequence optimization model to obtain the optimal starting sequence of each DER, and realizing global optimization.

In the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the invention, the method can also have the following characteristics: in step 2-1, the minimum time, the maximum available power generation capacity and the minimum requirement of the DER recovery path weight to be started are taken as DER starting sequence optimization targets, and the optimization targets are as follows:

in the formula, Ψ_BGAnd Ψ_NBGSet of respectively starting black start power supply and non-black start power supply in system network reconstruction stage_LA set of all lines l in the restoration path; z_lIs the recovery weight, P, of the selected line l_BG,nAnd P_NBG,kThe power generation output recovered by the nth black start unit and the kth non-black start unit at the stage is respectively provided.And establishing a DER starting sequence optimization model by taking the DER available power generation amount at the power failure time as a reference. The constraint conditions in the step 2-1 comprise DER output constraint, node voltage constraint, line thermal stability limit constraint, power balance constraint, thermal power unit hot start time constraint, distribution network topology radial structure constraint and auxiliary engine motor self-starting verification. DER output constraints are:

in the formula, P_GnActive output, Q, of the activated DER_GnPsi reactive output of activated DER_GA set of all DER in the system. The node voltage constraint is:

V_i ^min≤V_i≤V_i ^max i∈ψ_N (16)

in the formula, V_iIs the voltage magnitude, Ψ, of node i in the system_NA set of all nodes in the system. The line thermal stability limit constraints are:

in the formula, P_LlIs the active power transmitted in line i. The power balance constraint is:

in the formula, P_MFor thermal power plant auxiliary machinery power, P_GnFor the started DER output, the coefficient alpha comprises plant power rate, line loss rate and certain margin reserved for output prediction error of wind power and photovoltaic units, and the coefficient alpha is (1+ alpha) P_MTo meet the minimum requirement of black startThe power requirements. The thermal power generating unit hot start time constraint is as follows:

0≤T_L≤T_max (19)

in the formula, T_LThe time consumed for starting the self-black start power supply and transmitting power to the auxiliary engine of the thermal power generating unit through the recovery path T_maxThe maximum time limit for the thermal power generating unit to be started in a hot mode. The constraint of the topological radiation structure of the power distribution network is as follows:

in the formula, z_lSwitching state binary variables for branch l: 0 denotes open branch, 1 denotes closed branch, n_b、n_sThe total number of the nodes in the power distribution network system and the number of the root nodes are respectively. The auxiliary motor performs self-starting verification, namely voltage verification and capacity verification:

in the formula of U_*1The factory high-voltage bus voltage per unit value is obtained; u shape_*2The voltage per unit value of the factory low-voltage bus is obtained.

In the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the invention, the method can also have the following characteristics: in step 2-2, the indicators of the line importance evaluation include a line operation time weight setting, a line restoration cost weight setting, and a line weight setting. In the setting of the weight of the line operation time, the time consumed by each line operation is divided into optimistic estimation time A, pessimistic estimation time B and most probable estimation time M respectively, and the operation time t required for recovering a certain line l_lIts expected value E (t)_l) Sum variance σ_lRespectively as follows:

where l ∈ Ψ_L，Ψ_LFor a set formed by all lines in the system, assuming that a unit starting path at a certain node i is composed of L lines in total, the mean and variance of the operation time required by the recovery path are as follows:

if T_iAnd if the probability of falling within the starting time limit of the thermal power generating unit is higher, the thermal power generating unit is considered to be listed in a hot starting plan. In the setting of the line restoration cost weight, the total charging capacitance on the restoration path for starting a certain unit i is represented as:

setting the line recovery cost weight as:

W_l＝C_l+S_l (27)

in the formula, S_lThe operation cost of switching on and switching off a certain line l is high. The setting of the line weight mainly considers 2 evaluation indexes: time operation weight t and recovery cost weight W, and weight Z of line l_lThe following settings are set:

Z_l＝k₁W_l+k₂t_l (28)

in the formula, k₁And k₂To assign a coefficient, k₁+k₂When the line time operation weight and the recovery cost weight are smaller, the line weight of the line is smaller, the importance of the line is higher, and priority is used in the recovery path selection processThe line with smaller weight of the recovery line, i.e. higher importance of the line, is selected. In step 2-2, the minimum sum of the line weights in the restoration path is used as a target for optimization, and an objective function is expressed as:

in the dynamic optimization of the recovery path, when the next section of line is recovered based on the current state, a system network is divided into a charged part and a dead part, the charged part is aggregated into a power supply point, the power supply point calculates the optimal path for starting the next target point, the line with higher importance is recovered, and the steps are repeated until the thermal power plant is started.

In the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the invention, the method can also have the following characteristics: in the step 2, the algorithm implementation process of model solution by adopting a method of combining a genetic algorithm and a Floeider algorithm is as follows: and calling a Floiede algorithm to search a recovery path for each DER starting sequence based on DER starting sequences randomly generated by the genetic algorithm, and calling the genetic algorithm to select an optimal DER starting sequence and a recovery path in the sequence, thereby obtaining a global optimal solution of the target grid. The Floeid algorithm is sequentially provided with S_DERMiddle S₁～S_NThe nodes are used as the initial and final points of the search to search for the recovery path, and the basic thought is as follows: let L (S)_i,S_k) Represents a point S_iTo point S_kShortest path of (1), L (S)_k,S_j) Represents a point S_kTo point S_jShortest path of (1), L (S)_i,S_j) Represents a point S_iTo point S_jIs a path of (S) is a point S_iTo point S_jIs expressed as min (L (S)_i,S_k)+L(S_k,S_j),L(S_i,S_j)). The basic idea of the genetic algorithm is as follows: assuming that N DER nodes and 1 main power plant node Q exist in the system, a binary coding mode is adopted for chromosomes in a genetic algorithm, and DER starting sequence is represented by chromosomes with the length of k (N +1)I.e. Chrom ═ a₁,A₂,…,A_N|A_N+1]Wherein A is_N+1Is a chromosome fragment of length k, A₁～A_NRepresents the starting moment of each DER, A_N+1Setting the initial chromosomes as the final start of the thermal power plant, randomly generating a batch of initial chromosomes by a genetic algorithm, converting an objective function of a DER start sequence optimization model into a fitness function of the chromosomes, randomly selecting through roulette based on the fitness function values of the chromosomes, and carrying out operation steps of crossing, mutation, recombination and the like to generate offspring chromosomes, recording and reserving parent excellent chromosomes, and repeating the steps until set conditions are met.

In the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the invention, the method can also have the following characteristics: wherein, in step 3, after all kinds of DER start polymerization formation regional black start power in the distribution network, begin to start and close on main network in the thermal power plant, for thermal power unit auxiliary engine provides the power, the power balance expression is:

P_W(t)+P_PV(t)+P_ES(t)+P_MT(t)-P_load(t)＝0 (10)

in the formula: p_W(t) represents the fan output, P_PV(t) represents the photovoltaic contribution, P_ES(t) represents the stored energy output power, P_MT(t) represents the fan output, P_loadAnd (t) represents the power of the auxiliary engine of the thermal power generating unit. The output of the wind power and photovoltaic generator set is respectively as follows compared with the excess and shortage of the auxiliary machine power:

ΔP₊(t)＝|P_W(t)+P_PV(t)-P_load(t)| (11)

ΔP_-(t)＝|P_W(t)+P_PV(t)-P_load(t)| (12)

the DER starts the coordinated control process of the thermal power generating unit as follows: step 3-1, at the moment t, DER starts polymerization to form a regional power supply; step 3-2, sequentially putting the regional power supplies into auxiliary machines of the thermal power plant; 3-3, in the initial stage of black start, when the sum of the output of the wind power generator and the output of the photovoltaic generator set is smaller than the power of the auxiliary machine, the wind power generator and the photovoltaic generator set both adopt an MPPT control strategy, and the rest isThe power shortage deltap- (t) is taken over by the energy storage and micro gas turbine. When the sum of the output of the wind power generator and the output of the photovoltaic generator set is larger than the power of the auxiliary machine and the excess quantity delta P₊(t) less than 10% P_PVAnd (t) in the process, the photovoltaic unit is controlled by MPPT, and the wind turbine unit is controlled by a load tracking control strategy to track the power of the auxiliary machine and the photovoltaic output deficit change output. When the sum of the output of the wind power and the photovoltaic set is greater than the power of the auxiliary machine and the excess delta P₊(t) greater than 10% P_PVDuring (t), the photovoltaic unit adopts voltage load reduction control deviating from the maximum power point, and the load reduction amount is 10% P_PV(t), the wind turbine generator adopts a load tracking control strategy to track the auxiliary engine power and photovoltaic output deficit change output; and 3-4, if the thermal power generating unit does not recover to generate power, repeating the steps 3-2-3 until the thermal power generating unit recovers to generate power.

Action and Effect of the invention

According to the toughness improvement oriented distributed resource auxiliary main network key node black start strategy, the start steps are as follows: step 1, analyzing output characteristics and regulation strategies of each DER in a power distribution network, forming a method for evaluating the rapid recovery feasibility of the DER supporting a key node of a main network, establishing black start performability evaluation indexes, evaluating black start executable time periods in one day, and selecting a DER control strategy meeting power supply characteristics and system recovery requirements for the black start executable time periods based on evaluation results; step 2, based on DER control strategies, establishing a power distribution network DER starting and grid frame reconstruction combined optimization model with the minimum recovery time, the maximum available power generation capacity and the minimum recovery path weight as targets, solving by adopting a method of combining a genetic algorithm and a Floerd algorithm to obtain an optimal starting sequence of each DER, stably starting and connecting each DER according to the optimal starting sequence, and aggregating to form a regional power supply; and 3, starting the thermal power plant auxiliary machines in batches by the regional power supply, starting the thermal power generating units after the thermal power plant auxiliary machines completely and normally operate, and continuously accessing the recovered thermal power generating units to the load of the peripheral power distribution network and providing starting power supply support for other thermal power plants, so that the recovery range is gradually expanded until the whole system is recovered.

The black start performability in one day is evaluated based on DER output prediction, a grid frame reconstruction model for joint optimization of DER start sequence and recovery path is established, an optimal recovery scheme for starting a main grid thermal power plant by the DER of a power distribution network system is formed, the thermal power generating unit is guaranteed to be started safely as soon as possible within a hot start time limit, and the toughness of a power grid is improved.

In addition, the toughness-improvement-oriented distributed resource-assisted main network key node black-start strategy provided by the invention obviously reduces the main network power supply node recovery time, and effectively improves the toughness of the power grid and the DER utilization efficiency.

Drawings

FIG. 1 is a diagram of a black start process of a distribution grid DER supporting a main thermal power plant in accordance with an embodiment of the present invention;

FIG. 2 is a flow diagram of path restoration in an embodiment of the present invention;

FIG. 3 is a flow chart of an algorithm solution in an embodiment of the present invention;

FIG. 4 is a flow chart of coordinated control of DER starting a thermal power generating unit according to an embodiment of the invention;

FIG. 5 is a schematic diagram of a 35-node distribution network system in a certain area according to an embodiment of the present invention;

FIG. 6 is a typical solar photovoltaic, wind power effort diagram in an embodiment of the present invention;

FIG. 7 is a plot of the total power of the DER versus a day for an embodiment of the present invention;

FIG. 8 is a block diagram illustrating the continuous black start performance slope in an embodiment of the present invention;

FIG. 9 is a diagram of a scene I grid reconstruction scheme in an embodiment of the present invention;

FIG. 10 is a graph of DER and auxiliary power changes in scenario I in an embodiment of the present invention;

FIG. 11 is a diagram of a scenario II rack reconstruction scheme in an embodiment of the present invention;

FIG. 12 is a graph of DER and auxiliary power changes in scenario II in an embodiment of the present invention;

FIG. 13 is a diagram of a scene III lattice reconstruction scenario in an embodiment of the present disclosure; and

fig. 14 is a graph showing DER and auxiliary power changes in scene iii in the embodiment of the present invention.

Detailed Description

In order to make the technical means, the creation features, the achievement purposes and the effects of the present invention easy to understand, the following embodiments specifically describe a toughness-enhancement-oriented black start strategy for key nodes of a distributed resource-assisted main network according to the present invention with reference to the accompanying drawings.

In this embodiment, a distributed resource assisted main network key node black start strategy for toughness improvement is provided.

Fig. 1 is a diagram of a black start process of the distribution network DER supporting the main power plant in the embodiment.

As shown in fig. 1, the toughness-promotion-oriented distributed resource assisted host network key node black start policy according to this embodiment includes the following steps:

and S1, analyzing the output characteristics and the regulation and control strategies of a fan, a photovoltaic, an energy storage and a micro gas turbine in the power distribution network, forming a distributed resource support main network key node rapid recovery feasibility evaluation method, establishing a black start performability evaluation index, evaluating the black start executable time period in one day, and selecting a distributed resource control strategy meeting the power supply characteristics and the system recovery requirements for the black start executable time period based on the evaluation result.

The uncertainty of the multi-type DER output time-space distribution, the difference of the control strategy, the inconsistency of the output prediction deviation and the like greatly increase the difficulty and complexity of the DER participating in black start. In contrast, in the embodiment, starting from three aspects of DER output fluctuation, prediction error uncertainty and coordination control strategy, deep research is performed on the feasibility of DER supporting black start of the main network. DER resources in the power distribution network mainly comprise a fan, photovoltaic, energy storage and a micro gas turbine. The energy storage and micro gas turbine has the characteristics of rapid self-starting, continuous and stable output power, strong voltage and frequency regulation capacity and the like, and can be used as a black-start power supply to provide stable reactive voltage and active frequency support for a system. The output of the wind power and photovoltaic units is limited by natural conditions of wind speed and illumination intensity, has the characteristics of randomness, volatility and uncontrollable property, is difficult to meet the requirements of frequency modulation and voltage regulation, is used as a non-black starting power supply, and needs to be provided with starting power by an energy storage or a micro gas turbine during starting.

Step S1 specifically includes the following steps:

and step S1-1, predicting the available power generation capacity of the distributed resources in a future period of time in the black start process aiming at the output characteristics and the uncertainty of the time-space distribution of the various types of distributed resources, and processing the prediction error deviation by using a confidence interval method to obtain the output prediction result of the distributed resources of the power distribution network system.

The multi-type distributed resource output characteristics comprise wind power output characteristics, photovoltaic unit output characteristics and energy storage operation characteristics.

A large amount of research and statistical data show that the average wind speed probability density follows Weibull distribution, the illumination intensity can be similar to Beta distribution, and the wind speed and illumination intensity prediction model, the probability density distribution of the output of the distributed wind power and photovoltaic units and the cumulative distribution function in the embodiment refer to the prior art.

The energy storage starting state time in the power distribution network is short, the SOC of the energy storage starting state is basically unchanged, the stability of charge and discharge power, voltage and frequency can be maintained in the black starting process, and voltage and frequency support is provided for other power sources. However, when a blackout accident occurs, the state of charge SOC of the stored energy at the time t_ES(t) has randomness, and obviously influences the energy storage and power supply capacity in the black start process, and the energy storage operation characteristic considers the energy storage charge-discharge power constraint and the energy storage charge state constraint as follows:

-P_ES,N≤P_ES(t)≤P_ES,N (1)

SOC_ES,min≤SOC_ES(t)≤SOC_ES,max (2)

in the formula, P_ES,NRated power for stored energy; SOC_ES,max、SOC_ES,minAnd SOC_ES(t) the upper limit and the lower limit of the charge state of the energy storage device and the charge state of the energy storage device in the black starting process are respectively. SOC (system on chip)_ES(t) the calculation formula can be expressed as:

the charging and discharging power for energy storage at the moment t; s. the_sIs the capacity of the energy storage device.

And step S1-2, based on the power distribution network system distributed resource output prediction result, considering two aspects of the minimum power requirement for starting the main network large-scale unit auxiliary machine and the black start continuous execution inclination, establishing a black start performability evaluation index, and evaluating the black start executable time period in one day to obtain an evaluation result.

The output time-varying rule and randomness of the wind power and the photovoltaic units cause that the difference of available capacities of starting power supplies in systems at different time intervals is obvious, and black start performability when major power failure accidents occur at different moments in the day needs to be judged. The DER supports the judgement basis of the black start feasibility of major network thermal power plant and mainly includes two aspects: 1) judging whether the DER generating power of the power distribution network system meets the minimum black start power requirement or not by predicting the output of the wind power generator and the photovoltaic generator set in the black start process; 2) and evaluating the DER continuous effective output in the black start time period, and judging the performability of the black start scheme. At present, output prediction methods for wind and light intermittent power sources exist, but prediction accuracy is still unsatisfactory, wherein the average relative error of short-term power generation prediction of a wind turbine generator is up to 15% -30%, and the power generation prediction error of a photovoltaic generator is generally 10% -20%.

Step S1-2 includes the steps of:

step S1-2-1, confidence interval estimation: analysis of a large amount of research data shows that output prediction errors of the wind power and photovoltaic units conform to a mathematical distribution rule to a certain extent, characteristics such as thick tails, peaks and asymmetry of output prediction error probability density distribution can be well described by using a Gaussian mixture probability distribution model, actual output of the units is distributed around a predicted value, actual output of the wind power and photovoltaic units in each time period in a black start process is expressed as the sum of the predicted value and the prediction error, and the description is as follows:

and respectively representing the output prediction errors of the wind power generator and the photovoltaic generator at the moment t.

The uncertainty processing method adopts a confidence interval method to ensure that a scheme meets a certain confidence level, firstly, based on output predicted values of wind power and photovoltaic units, an uncertainty model is adopted to fit prediction error distribution in different power sections to obtain a probability density function, secondly, based on the probability density function, an interval set with the accumulated probability more than or equal to confidence 1-alpha is extracted, and finally, an interval with the shortest length under the same confidence 1-alpha is selected from the interval set to serve as a confidence interval under the prediction power. The confidence interval under the confidence 1-alpha calculated based on the principle of the shortest confidence interval is as follows:

in the formula, Δ P is an error corresponding to the power prediction value P, and corresponds to the probability distribution F (Δ P).

In the embodiment, in order to avoid over conservative prediction of DER total output, the probability that the output of all wind power and photovoltaic units is lower than the lower prediction power limit is assumed to be extremely small, in order to simplify the calculation complexity, the lower prediction power limit is taken as the actual output of the units, meanwhile, a certain power margin is considered for uncertainty factors, and black start performability is evaluated.

Step S1-2-2, the black start of the distributed resource support main grid power supply is not completed instantaneously, but needs to provide an uninterruptible power supply support in the entire continuous process of recovering the power supply of the thermal power generating unit. For the performability evaluation of the black start scheme of the DER supporting main power plant in one day, a certain time margin is considered fully, and 60min is taken as a black start continuous time period, wherein the black start continuous time period comprises a series of steps of DER start, path recovery, starting of auxiliary machines of thermal power units and the like in a power distribution network system. Because the resilience directly influences the toughness level after the power grid fails, the toughness level of the power transmission and distribution network is judged by analyzing the performability period of the black start scheme. The embodiment adopts two evaluation indexes of the minimum power requirement of the black start and the continuous execution inclination of the black start to evaluate the performability of the black start.

Evaluation index 1: black starts minimum power requirements. The main load object in the black start process is the thermal power unit auxiliary engine, the output power of the started DER in the power distribution network is larger than the load power, and meanwhile, the uncertainty of the self power consumption and line loss of the DER and the output power of the wind power and photovoltaic units is required to be met. Minimum power requirement for black start is P_{Distributed resources, min}Comprises the following steps:

P_DER,min＝(1+α)P_M (6)

in the formula, P_MRated power of an auxiliary engine of the thermal power generating unit to be started; the coefficient alpha comprises the self power consumption rate and the line loss rate of the distributed resources and a certain reserved allowance, and the uncertainty influence on the output of the wind power generator and the photovoltaic generator set is dealt with by reserving the rotary reserve.

Evaluation index 2: the black start continues to perform the inclination. In order to ensure the smooth starting of the auxiliary machine of the fire-electric generating set in the black starting process, the DER needs to continuously and effectively provide starting power within a certain time period, and accordingly an evaluation index of the continuous execution inclination of the black starting is established, namely the DER continuous power supply quantity in 60min is larger than the total load requirement of the auxiliary machine in the starting process of the main grid set. Distributed resource power generation capacity Q in black start continuous process_{Distributed resources}Minimum required electric quantity Q in black start_{Distributed resources, min}And black start continuous execution inclinationγ is represented as:

in the formula, P_{Distributed resources}And (t) the distributed resources exert force at the time t, wherein when gamma is greater than or equal to 1, the sustainable execution inclination is higher, and when gamma is less than 1, the black start scheme is more prone to be unexecutable.

And step S1-3, based on the evaluation result, selecting a distributed resource control strategy meeting the power supply characteristics and the system recovery requirements for the power failure time of the executable black start scheme, and ensuring the successful start and stable operation of the failed system.

Step S2, based on the distributed resource control strategy, aiming at the characteristics that the distributed resource starting sequence and the recovery path are mutually driven and mutually supported, a power distribution network distributed resource starting and grid frame reconstruction joint optimization model with the minimum recovery time, the maximum available power generation capacity and the minimum recovery path weight as targets is established, the optimal starting sequence of each distributed resource is obtained by solving by adopting a method of combining a genetic algorithm and a Floeud algorithm, each distributed resource is stably started and grid-connected according to the optimal starting sequence, and the distributed resources are aggregated to form a regional power supply.

The main task of grid reconstruction is to restore the main grid power plant as soon as possible, initially establish a relatively stable grid structure and lay a foundation for the comprehensive restoration of the system at the next stage. The present embodiment provides a DER start sequence and restoration path sequence joint optimization method for a system restoration strategy in a rack reconstruction stage, and avoids the problem that 2 optimization links, namely determining a target rack and searching for a restoration path, are separated from each other to a certain extent. Firstly, determining the starting sequence of each DER by taking the minimization of the maximum available power generation capacity of the recovery system and the DER recovery path weight to be started as soon as possible as a target; meanwhile, based on a power supply starting sequence generated by a DER optimization model, determining a recovery path for starting the DER and supplying power to the auxiliary engine of the thermal power generating unit by taking the minimum sum of path line weights as a target; and finally, forming a system recovery strategy of the net rack reconstruction stage.

In step S2, the process of solving the optimal start-up sequence includes the following sub-steps:

and step S2-1, selecting an energy storage and micro gas turbine with self-starting capability and stronger regulation performance as a black-start power supply for self-starting, and optimizing the distributed resource selection and the distributed resource starting sequence by a genetic algorithm with the aims of shortest system recovery time, maximum available power generation capacity and minimum recovery path weight.

Arrange reasonable DER start-up sequence in black start-up process, coordinate through different grade type power, be favorable to improving available generating capacity, promote the reliability and the persistence of starting the mains operated to accelerate system recovery process, start earlier by the power that the adjustable ability is strong simultaneously and can improve black start success rate, guarantee to restore system safety and stability operation. Compared with the traditional black start unit, the DER has the characteristics of short start time, quick response, flexible control and the like, and can meet the following principle according to the optimal DER start sequence required by DER self-start capability and output characteristic:

1) DER start with strong black start capability is prioritized. The DER can be divided into a black start power supply and a non-black start power supply according to the black start capability, the DER with strong black start capability is arranged to be started preferentially, and the non-black start power supply needs to be supported by the power provided by the black start power supply to start grid connection in sequence.

2) DER with large available capacity is prioritized for startup. The goal in the early recovery phase of the system is to obtain as much of the available generating capacity as possible, thereby providing more starting power to the system.

3) And preferably arranging the path weight small DER to be started. When the non-black start power supply start sequence is arranged, factors influencing the path weight, such as a line charging capacitor, the switching operation times, the path recovery time and the like, need to be considered.

4) DER start-up with FM and voltage regulation capability is prioritized. DER needs to have certain frequency modulation and voltage regulation capability to maintain stable operation of the system, and in the black start process, different DER adopts different control strategies according to local conditions so as to ensure the stability of the voltage and the frequency of the system.

The DER starting sequence affects the final path recovery scheme, and if only a single type of DER starting sequence is considered, the obtained optimal path recovery scheme only represents local optimization rather than global optimization under the DER starting sequence, so that the DER starting sequence and the recovery path need to be considered in a coordinated manner and optimized in a combined manner. The aim of the initial stage of the black start is to obtain available power generation capacity as much as possible, provide larger starting power for the system, comprehensively consider the coupling relation between the DER starting sequence and the recovery path, and optimize the distributed resource starting sequence by taking the minimum time, the maximum available power generation capacity and the minimum weight of the recovery path of the distributed resources to be started as follows:

in the formula, Ψ_BGAnd Ψ_NBGSet of start black start power supply and non-black start power supply for system network reconfiguration stage_LA set of all lines l in the restoration path; z_lIs the recovered weight, P, of the selected line l_BG,nAnd P_NBG,kThe power generation output recovered by the nth black start unit and the kth non-black start unit at the stage is respectively provided.

Due to uncertainty of output of the wind power generator and the photovoltaic generator, DER starting sequence selection in the black starting process can be influenced. If the black start period is [0, T ]]In the presence of t₁∈[0,T]，t₂∈[0,T]And t is₁≠t₂If at t₁And t₂At any moment, the output fluctuation of the wind power generation unit and the photovoltaic unit is large, different recovery strategies occur at the two moments, and finally the difference of the built grid structures is obvious. Thus, when DER is at [0, T]When the output fluctuation is large in a time period, the DER starting sequence and the target grid structure are easy to change frequently,the line switch is frequently operated, and the auxiliary machine of the thermal power generating unit cannot obtain continuous and stable power supply. In order to ensure smooth propulsion of the black start process and improve the stability of the recovery system, the method establishes a DER start sequence optimization model by taking the DER available power generation amount at the power failure time as a reference.

The constraint conditions comprise distributed resource output constraint, node voltage constraint, line thermal stability limit constraint, power balance constraint, thermal power generating unit hot start time constraint, distribution network topological radiation structure constraint and auxiliary engine motor self-starting verification.

The distributed resource contribution constraints are:

in the formula, P_GnActive power output, Q, for the activated distributed resource_GnReactive power out of the activated distributed resource, Ψ_GA set of all distributed resources in the system.

The node voltage constraint is:

V_i ^min≤V_i≤V_i ^max i∈ψ_N (16)

in the formula, V_iIs the voltage magnitude, Ψ, of node i in the system_NA set of all nodes in the system.

The line thermal stability limit constraints are:

in the formula, P_LlIs the real power transmitted in line i.

In the black start scheme, the main load is the thermal power unit auxiliary engine, the thermal power unit can be started only when the DER output meets the power requirement of the auxiliary engine, but the wind and light output is random, and the photovoltaic start can be executed only in the daytime. Therefore, DER and auxiliary power in black start should satisfy the following constraints:

in the formula, P_MFor thermal power plant auxiliary machinery power, P_GnFor the output of the started distributed resources, the coefficient alpha comprises the plant power rate, the line loss rate and a certain margin reserved for the output prediction error of the wind power and photovoltaic units, and the (1+ alpha) P_MTo meet the minimum power requirements of the black start requirement.

The thermal power generating unit hot start time constraint is as follows:

0≤T_L≤T_max (19)

in the formula, T_LThe time consumed for starting the self-black start power supply and transmitting power to the auxiliary engine of the thermal power generating unit through the recovery path T_maxThe maximum time limit for the thermal power generating unit to be started in a hot mode is shown.

Aiming at the requirements of the protection setting consideration and the small short-circuit current of the power distribution network, the power distribution network is generally required to run radially^[30]Namely, the system has no 'loop' and 'island' topological structures, and the topological radiation structure constraint of the power distribution network is expressed as follows:

in the formula, z_lSwitching state binary variables for branch l: 0 denotes open branch, 1 denotes closed branch, n_b、n_sThe total number of the nodes in the power distribution network system and the number of the root nodes are respectively. The radial network can be abstracted to a tree structure in a graph theory, each feeder line is represented as a tree, a root node corresponds to the root of the tree, a node corresponds to a node of the tree, a branch corresponds to an edge of the tree, and the requirement that the total edge number is equal to the node number-root number is met.

The self-starting verification is performed on the auxiliary motor, and can be divided into voltage verification and capacity verification. Because the starting current can reach 4-7 times of rated current when the auxiliary engine is started, the auxiliary engine is equivalent to a short circuit at the moment of starting, and the residual voltage of the bus is used by a checking calculation factory. The voltage check requirement for the self-starting bus of the motor is as follows: the lowest voltage of the high-voltage bus which is self-started by the service high-voltage bus power supply motor is 60 percent of rated voltage; the lowest voltage of the low-voltage bus for self-starting of the motor is 55% of the rated voltage by connecting the factory low-voltage bus and the high-voltage bus in series.

And step S2-2, based on the line importance evaluation, performing path optimization on the distributed resource starting sequence by adopting a shortest path algorithm to obtain a shortest path, transmitting the shortest path back to the distributed resource starting sequence optimization model to obtain the optimal starting sequence of each distributed resource, and realizing global optimization.

Aiming at a given DER starting sequence, a plurality of different system recovery paths can be obtained, the recovery net racks established by the method are different, so that the line operation time, the charging capacitor and the switch-on operation times of the switch are different in the path recovery process, the reasonable recovery path is selected, the system recovery time can be effectively shortened, and the safe and stable execution of a black starting process is guaranteed. In the embodiment, when the target network frame is established, the influence of the line operation time, the charging capacitor and the switch-on operation times in 3 aspects on the line weight is comprehensively considered, the line importance in the system is evaluated by setting the line weight, and an optimization model is established to obtain a recovery path formed by lines with relatively higher importance.

The indexes of the line importance evaluation comprise line operation time weight setting, line recovery cost weight setting and line weight setting.

Setting a line operation time weight: during the black start, there is uncertainty about the operation time required for line recovery. For each lineThe operation time evaluation standard is set as a time operation weight t, and t is used for measuring the time spent by the line from the beginning of charging to the end of charging. In an actual system, generally, based on experience of an operator, time consumed by operation of each line is divided into optimistic estimated time a, pessimistic estimated time B and most probable estimated time M, and then the time actually required for restoring each line is distributed between a and B in a beta manner. The operating time t required for restoring a certain line l_lIts expected value E (t)_l) Sum variance σ_lRespectively as follows:

if T_iAnd if the probability of falling within the starting time limit of the thermal power generating unit is higher, the thermal power generating unit is considered to be listed in a hot starting plan.

In the network frame reconstruction stage, the safety of a system is affected by the overlarge charging capacitor in a recovery path, and the line overvoltage is caused by reactive power generated by no-load or light load of a line, so that the charging capacitor is one of important factors affecting the recovery cost weight of the line, and the charging capacitor of the line needs to be contained in a consideration range. When a power transmission path for starting the thermal power unit auxiliary engine is selected, a scheme that a line charging capacitor is relatively small is selected, so that reactive power generated on the power transmission path is reduced. The total charge capacitance on the recovery path to start a unit i can be expressed as:

recovering the line cost weight W_lAs an index for measuring the difficulty of a certain line l in the recovery process, if the recovery cost weight W of each line is set only by considering an influence factor of the line charging capacitance, it is easy to cause that a recovery path composed of a plurality of lines with relatively small charging capacitances is better than a recovery path composed of a single line. In order to avoid excessive circuit switch operation times caused by preferentially selecting a path consisting of excessive circuits, a circuit recovery cost weight value is set as:

W_l＝C_l+S_l (27)

in the formula, S_lThe operation cost of switching on and switching off for a certain line l is usually the additional circuit breakers at the two ends of the line, and the switching-on operation is needed when the line is recovered, so the S of each line is assumed_lSame by changing S_lThe influence of the charging capacitor of the circuit and the operation times of the switch on the optimization of the recovery path can be adjusted by the numerical value.

Setting a line weight: in this embodiment, 2 evaluation indexes are mainly considered for setting the line weight Z: time operation weight t and recovery cost weight W, and weight Z of line l_lThe following settings are set:

Z_l＝k₁W_l+k₂t_l (28)

in the formula, k₁And k₂To assign a coefficient, k₁+k₂When the line time operation weight and the recovery cost weight are smaller, the line weight of the line will be smaller, and the line importance will be higher, and in the recovery path selection process, the line with the smaller weight, that is, the higher importance of the line, is recovered by using priority.

When a line to be recovered is determined, the importance of line operation time, a charging capacitor and switch-on operation times is comprehensively considered, namely, the two aspects of efficiency and risk are considered when a skeleton network is required to be established. The larger the scale of the reconstructed grid frame is, the more important lines and power supplies are recovered, but the longer the final system recovery time is, and meanwhile, the more recovery lines increase, so that the charging capacitor and the switching operation frequency also increase, and the black start failure risk is higher. Therefore, in this embodiment, the minimum sum of the line weights in the restoration path is used as the target for optimization, and the objective function is expressed as:

the model constraint is the same as the DER start sequence optimization model constraint.

Fig. 2 is a flow chart of path restoration in the present embodiment.

As shown in fig. 2, in the rack reconfiguration process, as part of the lines are charged and restored, the importance of the remaining power-losing lines changes, and the importance of the power-losing lines needs to be determined again, so that the final restoration path optimization failure caused by repeated calculation of the restored lines is avoided. In this respect, the present embodiment provides a dynamic rating index of the line importance — the dynamic line importance. In the path recovery process, when the next section of line is recovered based on the current state, the system network is divided into a charged part and a dead part, the charged part is aggregated into a power supply point, the power supply point calculates the optimal path for starting the next target point, the line with higher importance is recovered, and the steps are repeated until the thermal power plant is started.

The grid reconstruction problem in this embodiment is a joint optimization problem in which the mutual influence relationship between the DER starting sequence and the restoration path is considered, and the model can be solved by adopting a method of combining a genetic algorithm and a shortest path algorithm. The algorithm is realized as follows:

in the embodiment, a genetic algorithm and a Floyd algorithm are combined to optimize a grid reconstruction model, based on DER starting sequences randomly generated by the genetic algorithm, the Floyd algorithm is called to search for recovery paths for each DER starting sequence, and meanwhile, the genetic algorithm is called to select the optimal DER starting sequence and the recovery paths in the sequence, so that the global optimal solution of the target grid is obtained.

1) Local optimization algorithm based on Floyd algorithm

Assuming that the starting sequence of each DER in the known power distribution network is S_DER＝{S₁,S₂,…,S_NS is adopted in the system in sequence by adopting a shortest path method₁～S_NAnd searching a recovery path, then aggregating the started DER into a power supply point and searching a power supply path for the auxiliary machine, and finally obtaining an optimal target grid frame. Because an optimal recovery path exists in different DER starting sequences, only the recovery path optimization is considered, and only a local optimal solution can be obtained by neglecting the DER starting sequence optimization.

The embodiment calls a classical shortest path algorithm, namely Floyd algorithm when the recovery path is optimized, and sequentially uses S_DERMiddle S₁～S_NAnd the nodes are used as the starting and ending points of the search to perform the recovery path search. The core thought of the algorithm is as follows: let L (S)_i,S_k) Represents a point S_iTo a point S_kShortest path of (1), L (S)_k,S_j) Represents a point S_kTo point S_jShortest path of (1), L (S)_i,S_j) Represents point S_iTo point S_jIs a path of (S) is a point S_iTo point S_jCan be expressed as min (L (S)_i,S_k)+L(S_k,S_j),L(S_i,S_j))。

2) Global optimization algorithm based on genetic algorithm

The DER starting sequence is known to influence the finally established recovery net rack, and the DER starting sequence needs to be optimized while the recovery path is optimized, so that the optimal recovery path under the optimal DER starting sequence is solved, and the global optimal solution of the target net rack is obtained.

In this embodiment, a genetic algorithm is invoked to perform global optimization on the DER start sequence, and the basic idea of the algorithm is as follows: assume that there are N DER nodes and 1 main thermal power plant node Q in the system. For the chromosome in the genetic algorithm, the DER is started by adopting a binary coding modeThe kinetic sequence can be represented by a chromosome of length k x (N +1), i.e. Chrom ═ a₁,A₂,…,A_N|A_N+1]Wherein A is_N+1Is a chromosome fragment of length k, A₁～A_NRepresents the starting time of each DER, A_N+1The setting is the final start of the thermal power plant. The algorithm firstly generates a batch of initial chromosomes at random, then converts an objective function of a DER starting sequence optimization model into a fitness function of the chromosomes, generates offspring chromosomes through operation steps of roulette random selection, crossing, mutation, recombination and the like based on the fitness function values of the chromosomes, records and retains superior chromosomes of parent generations, and repeats the steps until set conditions are met.

Fig. 3 is a flowchart of the algorithm solving in the present embodiment.

And step S3, starting the thermal power plant auxiliary machines in batches by the regional power supply, starting the thermal power generating units after the thermal power plant auxiliary machines completely and normally operate, and enabling the recovered thermal power generating units to be continuously connected to the loads of the peripheral power distribution network and provide starting power supply support for other thermal power plants, so that the recovery range is gradually expanded until the whole system is recovered.

And in the process of black start of the DER support main network key node, selecting stored energy as a main power supply to provide voltage and frequency support for the system. The over-charging and over-discharging of the stored energy will influence the DER to support the execution of the black start process of the thermal power plant, and the black start can be smoothly performed by limiting the charging and discharging depth of the stored energy in a reasonable interval from the perspective of coordination control. Therefore, a reasonable DER control strategy needs to be selected according to various power supply characteristics and system recovery requirements, the required energy storage capacity is reduced by controlling the power generation output of the power supply at each black start time period, and black start failure caused by over-charge and over-discharge of the stored energy is avoided. Therefore, a reasonable DER control strategy is a necessary condition for a failed system to restart successfully and run stably.

After various DER start polymerization formation region black start power in the distribution network, begin to start and close to main network medium thermal power plant, for thermal power unit auxiliary engine provides the power, the power balance expression is:

P_W(t)+P_PV(t)+P_ES(t)+P_MT(t)-P_load(t)＝0 (10)

in the formula, P_W(t) represents the fan output, P_PV(t) represents the photovoltaic contribution, P_ES(t) represents the stored energy output power, P_MT(t) represents the fan output, P_load(t) represents the power of the auxiliary engine of the thermal power generating unit,

the output of the wind power and photovoltaic generator set is respectively as follows compared with the excess and shortage of the auxiliary machine power:

ΔP₊(t)＝|P_W(t)+P_PV(t)-P_load(t)| (11)

ΔP_-(t)＝|P_W(t)+P_PV(t)-P_load(t)| (12)

load tracking and MPPT combined control: the photovoltaic unit adopts MPPT control and load shedding control, and the wind generation unit adopts MPPT control and load tracking control to combine^[28]The specific control flow principle is shown in figure 4. In the initial stage of black start, when the sum of the output of the wind power generation unit and the output of the photovoltaic generation unit is smaller than the power of the auxiliary engine, the wind power generation unit and the photovoltaic generation unit both adopt an MPPT control strategy to ensure that the output as much as possible is provided, and the rest power shortage delta P- (t) is borne by the energy storage and the micro gas turbine so as to improve the utilization rate of the wind power generation unit and the photovoltaic generation unit, reduce the required energy storage capacity and avoid the black start failure caused by the insufficient residual discharge capacity in the energy storage device. When the sum of the output of the wind power and the photovoltaic set is greater than the power of the auxiliary machine and the excess delta P₊(t) less than 10% P_PVAnd (t) in the process, the photovoltaic unit is controlled by MPPT, and the wind turbine unit is controlled by a load tracking control strategy to track the power of the auxiliary machine and the photovoltaic output deficit change output. When the sum of the output of the wind power generator and the output of the photovoltaic generator set is larger than the power of the auxiliary machine and the excess quantity delta P₊(t) greater than 10% P_PVDuring (t), the photovoltaic unit adopts voltage load reduction control deviating from the maximum power point, and the load reduction amount is 10% P_PVAnd (t) tracking the auxiliary engine power and the photovoltaic output deficit change output by the wind turbine generator by adopting a load tracking control strategy.

Fig. 4 is a flow chart of coordinated control of DER starting thermal power generating unit in this embodiment.

As shown in fig. 4, the process of starting the coordinated control of the thermal power generating unit by using the distributed resources includes:

step S3-1, at time t, distributed resources are started to aggregate to form a regional power supply;

step S3-2, sequentially putting the regional power supplies into the auxiliary machines of the thermal power plant;

step S3-3, in the initial stage of black start, when the sum of the output of the wind power set and the output of the photovoltaic set is smaller than the power of the auxiliary machine, the wind power set and the photovoltaic set both adopt an MPPT control strategy, the rest power shortage Delta P- (t) is borne by the energy storage and the micro gas turbine,

when the sum of the output of the wind power and the photovoltaic set is greater than the power of the auxiliary machine and the excess delta P₊(t) less than 10% P_PV(t), the photovoltaic set adopts MPPT control, the wind turbine set adopts a load tracking control strategy to track the auxiliary engine power and photovoltaic output deficit change output,

when the sum of the output of the wind power generator and the output of the photovoltaic generator set is larger than the power of the auxiliary machine and the excess quantity delta P₊(t) greater than 10% P_PVDuring (t), the photovoltaic unit adopts voltage load reduction control deviating from the maximum power point, and the load reduction amount is 10% P_PV(t), the wind turbine generator adopts a load tracking control strategy to track the auxiliary engine power and photovoltaic output deficit change output;

and S3-4, if the thermal power generating unit does not recover to generate power, repeating the steps 3-2-3 until the thermal power generating unit recovers to generate power.

V/F control: the inverter in the V/F (constant voltage and constant frequency) control can establish stable reference voltage and frequency according to the control requirement, the energy storage in the system and the micro gas turbine power supply are controlled by selecting the V/F to realize self-starting, and the system can stabilize the power fluctuation of the system and provide voltage and frequency support in the processes of starting other power supplies and putting the power supplies into the auxiliary engine of the thermal power generating unit. When a plurality of DER with frequency modulation and voltage regulation capabilities such as energy storage and micro gas turbines exist in the black start process, a plurality of groups of inverter control strategies with droop characteristics can be selected, and a large amount of endless power adjustment operations caused by slight differences among different inverter controls are avoided.

In this embodiment, taking a 35-node main and distribution network system in a certain area as an example, a black start model and a black start method for key nodes of a DER auxiliary main network are expanded and analyzed.

Fig. 5 is a schematic diagram of a 35-node distribution network system in a certain area in the present embodiment.

As shown in fig. 5, in which the solid line and the broken line represent the lines of which the switches are closed and opened in the normal operation state, respectively, all the line parameters in the system are shown in table 1 (using the standard line in pandapawer v2.6.0).

Table 1 shows all line parameters in the system.

TABLE 1 line parameters

The auxiliary machine of the thermal power generating unit with the target load of 100MW in the system to be recovered has the total capacity of 3MW, the black start duration is set to be 60min, and transient problems such as excitation inrush current generated by a starting transformer substation in the black start process and impact current generated when an asynchronous motor is started are not considered in the embodiment. DER is connected to

nodes

4, 11, 14, 20, 29 and 31, namely DER1, DER2, DER3, DER4, DER5 and DER6 respectively, wherein the initial SOC of stored energy is set to be 50%, so that the stored energy can be continuously supplied with the maximum discharge power for 1h under the condition of complete discharge, a 0 node represents thermal power plant auxiliary equipment connected to a 10kV bus, and specific DER and auxiliary equipment detailed parameters are shown in tables 2 and 3.

Table 2 is a DER parameter table in the power distribution network.

Table 3 is a main auxiliary parameter table.

TABLE 2 DER parameters in distribution networks

TABLE 3 Primary and auxiliary parameters

Tab.2 main auxiliary machine parameters

And (5) processing the DER output prediction deviation by adopting the uncertain factor processing method in the step S1 to obtain a wind power and photovoltaic generator set power generation prediction output curve with a typical day time granularity of 5min in the area as shown in FIG. 6. Typical day black start executability was divided by time period for DER daily output curves, as shown in fig. 7.

Fig. 6 is a typical daily photovoltaic, wind power effort diagram in an embodiment of the present invention.

FIG. 7 is a plot of DER total power versus time in an embodiment of the present invention.

Based on the black start performability assessment method, line loss, self service power and uncertainty factor influence are comprehensively considered, P_DER,minThe value is 3.5MW, and the black start continuous execution inclination of each power failure time in one day is obtained as shown in FIG. 8.

Fig. 8 is a diagram of the black start continuous execution inclination in the present embodiment.

As shown in fig. 8, it can be known that the difference of the system restoring force is large at different time, wherein the higher the execution inclination represents that the system has higher probability of successfully realizing black start in the period, and the toughness is stronger. From the analysis of FIGS. 6-8, the black start non-executable time period can be estimated as 03: 30-07: 30 and 16: 00-19: 00, the proportion of the distribution network in the whole day is about 29%, and the proportion of the black start executable time period in the whole day can be improved by reasonably configuring distributed resources during planning of the distribution system, so that the toughness level of the transmission and distribution network is improved. Based on the analysis, the DER output level in the distribution network system has larger difference at different power failure moments, and can directly influence whether the thermal power plant can be started successfully or not. In addition, if the DER has large output fluctuation in the system black start execution period, the execution inclination is reduced, and different time domain scenes in one day need to be selected to carry out deep analysis on the DER assisted thermal power plant black start scheme.

Scene i: when the power failure time is 09:00, the power failure is in a proper wind/light environment, the photovoltaic available power generation capacity is about 39% (total 0.702MW), and the wind power available power generation capacity is about 41% (total 0.902 MW).

And scene II: when the power failure time is 12:00, the photovoltaic power generation system is in a weak wind and strong light environment, the photovoltaic available power generation capacity is up to 80% (total 1.44MW), and the wind power available power generation capacity is reduced to 22% (total 0.484 MW).

Scene III: when the power failure time is 23:00, the photovoltaic power generation system is in a strong wind and no light environment, the photovoltaic available power generation capacity is 0%, and the wind power available power generation capacity is as high as about 80% (1.76 MW in total).

Aiming at the grid reconstruction optimization model established in the embodiment, a genetic algorithm and a Floyd algorithm are called to be combined to solve the DER auxiliary thermal power plant starting scheme under different scenes, the population number of the genetic algorithm is set to be 50, the iteration frequency is 200, the minimum output of a black start power supply for supporting a safe and stable starting auxiliary machine is determined to be 3.5MW by considering the service power, the transmission line loss and the reserved margin, and the black start optimization scheme under 3 scenes is obtained and is shown in the figures 9, 11 and 13.

Fig. 9 is a view of the scene i grid reconstruction scheme in this embodiment.

Fig. 11 is a view of the scene ii grid reconstruction scheme in this embodiment.

Fig. 13 is a view of a scene iii grid reconstruction scheme in the present embodiment.

As shown in fig. 9, 11, 13, the green color in the figure represents the enabled DER and recovered line. Then, based on the control strategy of the DER in each black start stage, analyzing the output change of the DER when the DER starts the thermal power plant auxiliary machinery within the black start execution time period T being 60min, and obtaining DER and auxiliary machinery power change curves under different scenes as shown in fig. 10, 12 and 14.

Fig. 10 is a graph showing the power changes of DER and auxiliary machines in scene i in the present embodiment.

Fig. 12 is a graph showing the DER and auxiliary power changes in scene ii in the present embodiment.

Fig. 14 is a graph showing the DER and auxiliary power changes in scene iii in the present embodiment.

Comparing and analyzing the grid reconstruction scheme in the 3-class power failure scene, selecting DER4 at the node 20 to be self-started preferentially, and having the capabilities of energy storage, quick charge and discharge and reactive voltage regulation, providing stable voltage and frequency support for successfully starting and grid connection of other power supplies, and bearing larger charging power to prevent node voltage from exceeding the limit when the line is in a light load state and the path to be recovered is charged. According to DER characteristic evaluation analysis, the black start power supply is preferred due to the good black start supporting characteristic of the stored energy.

1) Scene I black start scheme

For the framework reconfiguration scheme in the scenario i, as shown in fig. 9 and table 4, first, DER4 (stored energy) at the node 20 is arranged to be self-started to support stable power output and frequency modulation and voltage regulation; then, other DER sequences to be started are determined based on the minimum black start power constraint, and are DER3, DER6, DER5 and DER1 in sequence. Since the blackout time occurred at 09 a.m.: 00, the output of photovoltaic and wind power is moderate, the difference of available capacity of photovoltaic and wind power sets in the system is not obvious, the starting sequence of the photovoltaic and wind power sets mainly considers the weight of a recovery path, the main aim of the black start initial stage is to provide sufficient power for the system by the minimum recovery path weight as soon as possible, based on the principle, the interconnection switch 20-16 is closed, the DER3 which is closest to the DER4 and has larger available power generation amount is recovered, the recovered power is aggregated into a power supply point, the current time that the DER6 is the minimum recovery path weight away from the power supply point is judged, therefore, the micro gas turbine DER6 which has stable output power and larger capacity is started through the optimal recovery path, the remaining photovoltaic sets DER1 and DER2 and the wind power set DER5 which have smaller output and larger recovery path weight are sequentially connected in parallel according to the starting optimization principle, the DER which is started successfully and aggregated is put into a line 3-4 through the node 4, 2-3, 1-2 and 0-1, supplying power to a 10kV bus of an auxiliary machine of the thermal power generating unit, gradually recovering the auxiliary machine, and further starting the thermal power generating unit. Finally, the total operation time for establishing the target net rack is about 19min, and the recovery system in the scene I obtains about 3.804MW of available power supply power.

Table 4 is a scene i grid reconstruction scheme table.

Table 4 scene i net rack reconstruction scheme

And starting each DER and recovering the required path in 0-19 min in the scene I, wherein the subsequent time period is the starting process of the auxiliary engine of the thermal power plant, and is shown in figure 10. Firstly, the electric water feeding pump is put into the photovoltaic wind turbine generator at the moment of 19min to be 1.4MW, the available output of the photovoltaic wind turbine generator at the moment is about 1.614MW to meet the requirement of the auxiliary engine power, the photovoltaic power is subjected to load shedding control by deviating from the maximum power point voltage, the fan is subjected to a load tracking control strategy, the auxiliary engine power and the photovoltaic output shortage are tracked to adjust the output, and the output of the photovoltaic wind turbine generator and the photovoltaic output shortage does not need to be called in the period of time because the output of the photovoltaic wind turbine generator and the photovoltaic power generation meets the load requirement. The water circulating pump is put into at 30min and is 0.5MW, the total load of the auxiliary machine is 1.9MW, and the available output of the photovoltaic generator and the wind turbine generator is about 1.656MW and is smaller than the power of the auxiliary machine, so that the photovoltaic generator and the wind turbine generator both adopt MPPT control strategies, wind and light resources are utilized to the maximum extent, the power difference is borne by the micro gas turbine, and the black start failure caused by insufficient charge in the subsequent period due to the calling of the stored energy is avoided. And (3) putting the coal mill and the induced draft fan at 39min for 0.9MW totally, wherein the micro gas turbine reaches the maximum output power due to the increase of the load, and the energy storage and discharge are required to be called to jointly supply power to the load together with other DER. And finally, a condensate pump and an injection pump are put in at the moment of 49min, so that the auxiliary machine of the thermal power plant is started successfully.

2) Scene II black start scheme

For the network frame reconfiguration scheme in the scene ii, as shown in fig. 11 and table 5, since the power failure occurs at 12:00 noon and is in a weak wind and strong light environment, the photovoltaic power generation power is far higher than that of the wind turbine generator, and the photovoltaic power in the system is higher than the starting priority level of the wind turbine generator. Different from a grid reconstruction scheme of a scene I, a wind turbine at a starting node 14 is not preferentially selected after energy storage self-starting at a node 20, a micro gas turbine DER6 with higher available capacity and more stable output is started by closing interconnection switches 20-16 and 15-32 through input lines 16-15 and 32-31, then the wind turbine DER3, a photovoltaic unit DER1 and a photovoltaic unit DER2 are sequentially started based on the condition that sufficient power is provided for a system by taking the minimum recovery path weight as an optimization target, the wind turbine DER5 which is far away from an auxiliary machine of a thermal power plant is abandoned as a power source to be started due to the fact that the recovery path weight is larger and the available power generation capacity is lower, and finally starting power is provided for the auxiliary machine of the thermal power plant through a recovery path 4-3-2-1-0. In the scene, the total operation time of the network frame reconstruction is about 15min, and the commonly obtained available power supply power in the recovery system is about 3.948 MW.

And table 5 is a scene ii grid reconstruction scheme table.

Table 5 scene II grid reconstruction scheme

And in a scene II, carrying out net rack entering reconstruction based on the determined DER starting sequence and the preferred recovery path for 0-15 min. As shown in fig. 12, the remaining period is a thermal power plant auxiliary engine starting process, the electric water feeding pump is put into the thermal power plant auxiliary engine starting process at 15min, the power available to the photovoltaic unit is 1.458MW at this time, the power available to the wind turbine unit is less than 0.322MW, the auxiliary engine load is mainly borne by the photovoltaic unit, the photovoltaic unit adopts voltage load shedding control deviating from the maximum power point according to the control strategy, the wind turbine unit adopts a load tracking control strategy, the auxiliary engine power and the photovoltaic output deficit change output are tracked, and the photovoltaic and wind turbine output meet the load requirement, and the energy storage or micro gas turbine power increasing power is not required to be called in the period. The input of the circulating water pump is 0.5MW at the moment of 26min, the total load of the auxiliary machine is 1.9MW, and at the moment, the available output of the photovoltaic generator and the available output of the wind turbine are about smaller than the input of the auxiliary machine power, so the output of the photovoltaic generator and the output of the wind turbine both adopt an MPPT control strategy, and the residual power difference is borne by the micro gas turbine, so that the required energy storage capacity is reduced. And (3) respectively putting the coal mill, the draught fan, the condensate pump and the injection water pump at 37min and 48min, wherein the power supply control strategy is similar to that of the scene I, the photovoltaic power generation unit, the wind power generation unit and the micro gas turbine reach the maximum output power at the period along with the increase of the load, the energy storage and the discharge are needed to make up the power difference, and finally the auxiliary machine of the thermal power plant is started successfully.

3) Scene III black start scheme

As shown in fig. 13 and table 6, for the grid reconfiguration scheme in the scene iii, since the power failure time occurs at 23:00 night and is in a strong wind and dark environment, the available capacity of the photovoltaic unit is 0 and cannot participate in black start, and the available output of the wind turbine unit is higher by about 1.76 MW. And obtaining an optimal DER starting sequence and a recovery path based on the optimization goal of providing sufficient power for the system by using the minimum recovery path weight. The method comprises the steps of firstly starting an energy storage DER4 with frequency modulation and voltage regulation capabilities, then starting a wind turbine DER3 with higher available capacity through a recovery path 20-16-15-14, further starting a micro gas turbine DER6 which is closer to a recovered power supply point and has more stable output, wherein at the moment, the available power generation capacity of a system does not reach 3.5MW required by the minimum black start, the wind turbine DER5 which is further away from an auxiliary machine needs to be started to obtain more power supplies, and finally the determined available power supplies are DER4, DER3, DER6 and DER5 in sequence. The total operation time of the reconstruction of the net rack under the scene is about 17min, and the total output of the obtained available power supply is about 3.96 MW.

Table 6 shows a scene iii grid reconstruction scheme table.

Table 6 scene III net rack reconstruction scheme

And in the scene III, establishing a target grid through a grid reconstruction optimization model for 0-17 min. As shown in fig. 14, the remaining period is the starting process of the auxiliary engine of the thermal power plant, the electric feed pump is firstly put into the thermal power plant at 17min, the available output of the photovoltaic unit is 0MW at this time, and the available output of the wind turbine is more 1.76MW at this time, so the load of the auxiliary engine is mainly borne by the wind turbine, the wind turbine tracks the output of the auxiliary engine power change by adopting a load tracking control strategy, and since the output of the wind turbine meets the load requirement at this time, the energy storage or the power increase of the micro gas turbine is not required. The input of the circulating water pump is 0.5MW at the moment of 28min, the total load of the auxiliary machine is 1.9MW, and the available output of the wind turbine is slightly lower than the input of the auxiliary machine power at the moment, so that the wind turbine adopts an MPPT control strategy, wind power resources are fully utilized, the residual power difference is borne by the micro gas turbine, and the required energy storage capacity is reduced. And respectively putting the coal mill, the draught fan, the condensate pump and the injection water pump at 39min and 50min, wherein the wind turbine generator and the micro gas turbine reach the maximum output power along with the increase of the load, the energy storage and discharge are needed to make up the power difference, and finally the auxiliary machine of the thermal power plant is started successfully.

According to the results, obviously, the power failure occurs different moments, the scheme that the distributed power supply of the power distribution network assists the main network to be started black is different, the restoring force corresponding to the power transmission and distribution network is different, the restoring speed is higher when the available capacity of DER is more sufficient in the black starting period, the wind power, the photovoltaic generator set, the energy storage and the micro gas turbine are reasonably configured in the accessible, and the toughness level of the power grid is improved.

The net rack reconstruction strategies under the three different scenes are shown in fig. 9, 11 and 13, and it can be seen through comparative analysis that the output of the wind power generator and the photovoltaic generator set is greatly different due to the influence of weather changes, and the output of the stored energy and the available power generation of the micro gas turbine are relatively stable, so that in the recovery schemes formulated for the three scenes by using the model and the method provided by the embodiment, power supplies with relatively stable outputs, such as the stored energy and the micro gas turbine, are preferentially started, and whether intermittent power supplies, such as the wind power generator and the photovoltaic, are preferentially started or not is selected according to the weather changes. As can be seen from fig. 10, 12 and 14, for 3 scenes, the main grid thermal power plant black start scheme is supported by formulating different power distribution network DER to complete the starting of the thermal power plant auxiliary machine, in the process of starting the auxiliary machine, the uncertainty of the output of the wind power and the photovoltaic power generation set is comprehensively considered, the power generated by the wind power and the photovoltaic power supply is preferentially and fully utilized to supply power to the thermal power plant auxiliary machine, the times of energy storage charging and discharging switching and the required energy storage capacity are reduced, and the black start failure caused by insufficient energy storage residual charge is avoided.

The method is characterized in that the starting verification of the auxiliary engine of the thermal power plant is carried out aiming at the established black-start scheme, the starting current when the asynchronous motor is started is overlarge and can reach 4-7 times of rated current, so that the voltage of a power port is remarkably dropped when the asynchronous motor is started, and the auxiliary engine motor can be instantly started as a short circuit for calculation. The method comprises the steps that a power supply for supplying power to an auxiliary machine in a power distribution network system is treated as an infinite power supply, namely, the bus voltage is 1, a factory high-voltage bus is 0.682-0.60 and a factory low-voltage bus is 0.559-0.55 through calculation, the auxiliary machine can be successfully started in a black starting period (T60 min) under the three scenes, and the effectiveness of the model and the method is proved.

Effects and effects of the embodiments

According to the toughness-improvement-oriented distributed resource-assisted main network key node black start strategy provided by the embodiment, the start steps are as follows: step 1, analyzing output characteristics and regulation strategies of each DER in a power distribution network, forming a method for evaluating the rapid recovery feasibility of the DER supporting a key node of a main network, establishing black start performability evaluation indexes, evaluating black start executable time periods in one day, and selecting a DER control strategy meeting power supply characteristics and system recovery requirements for the black start executable time periods based on evaluation results; step 2, based on DER control strategies, establishing a power distribution network DER starting and grid frame reconstruction combined optimization model with the minimum recovery time, the maximum available power generation capacity and the minimum recovery path weight as targets, solving by adopting a method of combining a genetic algorithm and a Floerd algorithm to obtain an optimal starting sequence of each DER, stably starting and connecting each DER according to the optimal starting sequence, and aggregating to form a regional power supply; and 3, starting the thermal power plant auxiliary machines in batches by the regional power supply, starting the thermal power generating units after the thermal power plant auxiliary machines completely and normally operate, and continuously accessing the recovered thermal power generating units to the load of the peripheral power distribution network and providing starting power supply support for other thermal power plants, so that the recovery range is gradually expanded until the whole system is recovered.

In addition, the toughness-improvement-oriented black-start strategy for the key nodes of the distributed resource auxiliary main network obviously reduces the recovery time of the main network power source nodes, and effectively improves the toughness of the power grid and the DER utilization efficiency.

The above embodiments are preferred examples of the present invention, and are not intended to limit the scope of the present invention.

Claims

1. A toughness-promotion-oriented black start strategy for key nodes of a distributed resource-assisted main network is characterized by comprising the following steps:

step 1, analyzing the output characteristics and the regulation and control strategies of all distributed resources in a power distribution network, forming a distributed resource support main network key node rapid recovery feasibility evaluation method, establishing a black start executable evaluation index, evaluating a black start executable time period in one day, and selecting a distributed resource control strategy which meets the power supply characteristics and system recovery requirements for the black start executable time period based on an evaluation result;

step 2, based on the distributed resource control strategy, establishing a power distribution network distributed resource starting and grid frame reconstruction joint optimization model taking minimum recovery time, maximum available power generation capacity and minimum recovery path weight as targets, solving by adopting a method of combining a genetic algorithm and a Floerd algorithm to obtain an optimal starting sequence of each distributed resource, stably starting and connecting each distributed resource according to the optimal starting sequence, and aggregating to form a regional power supply;

and 3, starting thermal power plant auxiliary machines by the regional power supply in batches, starting thermal power generating units after the thermal power plant auxiliary machines completely and normally operate, and enabling the recovered thermal power generating units to be continuously connected to peripheral power distribution network loads and providing starting power supply support for other thermal power plants, so that the recovery range is gradually expanded until the whole system is recovered.

2. The toughness-enhancement-oriented distributed resource-assisted main network key node black-start strategy according to claim 1, characterized in that:

wherein, in the step 1, the distributed resources comprise a fan, a photovoltaic, an energy storage and a micro gas turbine,

the step 1 comprises the following steps:

step 1-1, predicting available power generation capacity of distributed resources in a future period of time in a black start process aiming at output characteristics and uncertainty of time-space distribution of various types of distributed resources, and processing prediction error deviation by using a confidence interval method to obtain a power distribution network system distributed resource output prediction result;

step 1-2, based on the power distribution network system distributed resource output prediction result, considering two aspects of minimum power requirement of starting of the main network large-scale unit auxiliary machine and continuous execution inclination of black start, establishing a black start performability evaluation index, and evaluating the black start executable time period in one day to obtain an evaluation result;

and 1-3, selecting a distributed resource control strategy which meets the power supply characteristics and system recovery requirements for the power failure time of the executable black start scheme based on the evaluation result, and ensuring the successful start and stable operation of the fault system.

3. The toughness-enhancement-oriented distributed resource-assisted main network key node black-start strategy according to claim 2, characterized in that:

wherein in the step 1-1, the output characteristics of the multi-type distributed resources comprise wind power output characteristics, photovoltaic unit output characteristics and energy storage operation characteristics,

the energy storage operation characteristic takes the energy storage charge-discharge power constraint and the energy storage charge state constraint into consideration as follows:

-P_ES,N≤P_ES(t)≤P_ES,N (1)

SOC_ES,min≤SOC_ES(t)≤SOC_ES,max (2)

the charging and discharging power for energy storage at the moment t; s_sIs the capacity of the energy storage device.

4. The toughness-enhancement-oriented distributed resource-assisted main network key node black-start strategy according to claim 2, characterized in that:

wherein, the step 1-2 comprises the following steps:

step 1-2-1, confidence interval estimation: the actual output of the generator set is distributed around the predicted value, the actual output of the wind power generator set and the photovoltaic generator set in each time period in the black start process is expressed as the sum of the predicted value and the prediction error, and the description is as follows:

respectively represents the prediction errors of the wind power and the photovoltaic set at the moment t,

the uncertainty processing method adopts a confidence interval method to ensure that a scheme meets a certain confidence level, firstly, based on the output predicted values of the wind power and the photovoltaic set, an uncertainty model is adopted to fit the predicted error distribution in different power sections to obtain a probability density function, secondly, based on the probability density function, an interval set with the accumulated probability more than or equal to the confidence level 1-alpha is extracted, and finally, the interval with the shortest length under the same confidence level 1-alpha is selected from the interval set as the confidence interval under the predicted power,

the confidence interval under the confidence 1-alpha calculated based on the principle of the shortest confidence interval is as follows

In the formula, Δ P is an error corresponding to the power prediction value P, and conforms to the probability distribution F (Δ P);

step 1-2-2, adopting two evaluation indexes of black start minimum power requirement and black start continuous execution inclination to evaluate the black start performability,

the black start minimum power requirement is P_{Distributed resources, min}Comprises the following steps:

P_DER,min＝(1+α)P_M (6)

in the formula, P_MRated power of an auxiliary engine of the thermal power generating unit to be started; the coefficient alpha comprises the self power consumption rate of the distributed resources, the line loss rate and a certain reserved allowance, the uncertain influence of the output of the wind power generation set and the photovoltaic generation set is responded by reserving the rotary reserve,

distributed resource power generation capacity Q in black start continuous process_{Distributed resources}Minimum required electric quantity Q of black start_{Distributed resources, min}And the black start continuous execution inclination γ is expressed as:

in the formula, P_{Distributed resources}(t) is the output of the distributed resources at the time t, and when gamma is more than or equal to 1, the distributed resources are sustainableThe execution continuation inclination is higher, and when γ is less than 1, it means that the black start scheme is more likely to be inoperable.

5. The toughness-enhancement-oriented distributed resource-assisted main network key node black-start strategy according to claim 1, characterized in that:

wherein, in the step 2, aiming at the characteristics that the distributed resource starting sequence and the recovery path are mutually driven and mutually supported, a joint optimization model of the distributed resource starting and the net rack reconstruction of the power distribution network is established,

the process of solving for the optimal start-up sequence comprises the sub-steps of:

step 2-1, selecting an energy storage and micro gas turbine with self-starting capability and stronger adjusting performance as a black-start power supply for self-starting, and optimizing distributed resource selection and a distributed resource starting sequence through the genetic algorithm by taking shortest system recovery time, maximum available power generation capacity and minimum recovery path weight as targets;

and 2-2, based on the line importance evaluation, carrying out path optimization on the distributed resource starting sequence by adopting a shortest path algorithm to obtain a shortest path, transmitting the shortest path back to a distributed resource starting sequence optimization model to obtain the optimal starting sequence of each distributed resource, and realizing global optimization.

6. The toughness-enhancement-oriented distributed resource-assisted main network key node black-start strategy according to claim 5, wherein:

in step 2-1, the minimum time, the maximum available power generation capacity and the minimum weight satisfying the recovery path of the distributed resources to be started are taken as the optimization target of the starting sequence of the distributed resources, as follows:

in the formula, Ψ_BGAnd Ψ_NBGStarting black start power supply for system network reconstruction stage respectivelyAnd a non-black start power supply, Ψ_LA set of all lines l in the restoration path; z_lIs the recovered weight, P, of the selected line l_BG,nAnd P_NBG,kThe power generation output recovered by the nth black start unit and the kth non-black start unit at the stage respectively,

by taking the available generated energy of the distributed resources at the moment of power failure as reference, a distributed resource starting sequence optimization model is established,

the constraint conditions in the step 2-1 comprise distributed resource output constraint, node voltage constraint, line thermal stability limit constraint, power balance constraint, thermal power unit hot start time constraint, distribution network topological radiation structure constraint and auxiliary engine motor self-starting verification,

the distributed resource contribution constraints are:

in the formula, P_GnActive power output, Q, for the initiated distributed resource_GnReactive power out of the activated distributed resource, Ψ_GIs a collection of all distributed resources in the system,

the node voltage constraint is:

V_i ^min≤V_i≤V_i ^max i∈ψ_N (16)

in the formula, V_iIs the voltage magnitude, Ψ, of node i in the system_NIs a set of all nodes in the system,

the line thermal stability limit constraints are:

in the formula, P_LlFor the active power transmitted in the line i,

the power balance constraint is:

in the formula, P_MFor thermal power plant auxiliary machinery power, P_GnFor the output of the started distributed resources, the coefficient alpha comprises the plant power rate, the line loss rate and a certain margin reserved for the output prediction error of the wind power and photovoltaic units, and the (1+ alpha) P_MTo meet the minimum power requirements of the black start requirement,

the thermal power generating unit hot start time constraint is as follows:

0≤T_L≤T_max (19)

in the formula, T_LThe time consumed for starting the self-black start power supply and transmitting power to the auxiliary engine of the thermal power generating unit through the recovery path T_maxIs the maximum time limit that the thermal power generating unit can be started up,

the constraint of the topological radiation structure of the power distribution network is as follows:

in the formula, z_lSwitching state binary variables for branch l: 0 denotes open branch, 1 denotes closed branch, n_b、n_sRespectively the total number of nodes and the number of root nodes in the power distribution network system,

the auxiliary motor carries out self-starting verification and is divided into voltage verification and capacity verification:

7. The toughness-boosting-oriented distributed resource assisted main network key node black start strategy according to claim 5, wherein:

wherein, in step 2-2, the indicators of the line importance evaluation include a line operation time weight setting, a line recovery cost weight setting and a line weight setting,

in the setting of the weight of the line operation time, the time consumed by each line operation is divided into optimistic estimation time A, pessimistic estimation time B and most probable estimation time M respectively, and the operation time t required for recovering a certain line l_lIts expected value E (t)_l) Sum variance σ_lRespectively as follows:

if T_iIf the probability of falling within the starting time limit of the thermal power generating unit is higher, the thermal power generating unit is considered to be listed in a hot start plan,

in the setting of the line restoration cost weight, the total charging capacitance on the restoration path for starting a certain unit i is represented as:

setting the line recovery cost weight as:

W_l＝C_l+S_l (27)

in the formula, S_lThe operating cost of switching on and off a certain line l,

the line weight setting mainly considers 2 evaluation indexes: time operation weight t and recovery cost weight W, and weight Z of line l_lThe following settings are set:

Z_l＝k₁W_l+k₂t_l (28)

in the formula, k₁And k₂To assign a coefficient, k₁+k₂When the line time operation weight and the recovery cost weight are smaller, the line weight of the line will be smaller, and the line importance will be higher, in the recovery path selection process, the line with smaller weight, i.e. higher importance, is recovered by priority,

in step 2-2, the minimum sum of the line weights in the restoration path is used as a target for optimization, and an objective function is expressed as:

8. The toughness-enhancement-oriented distributed resource-assisted main network key node black-start strategy according to claim 1, characterized in that:

in the step 2, the algorithm implementation process of model solution by adopting a method of combining a genetic algorithm and a Floeider algorithm is as follows:

based on the distributed resource starting sequence randomly generated by the genetic algorithm, calling the Flouard algorithm to search a recovery path for each distributed resource starting sequence, and simultaneously calling the genetic algorithm to select the optimal distributed resource starting sequence and the recovery path in the sequence, thereby obtaining the global optimal solution of the target net rack,

the Froude algorithm is sequentially represented by S_{Distributed resources}Middle S₁～S_NThe node is used as the starting point and the end point of the search to search the recovery path, and the basic thought is as follows: let L (S)_i,S_k) Represents point S_iTo point S_kShortest path of (1), L (S)_k,S_j) Represents a point S_kTo point S_jShortest path of (1), L (S)_i,S_j) Represents a point S_iTo point S_jIs a path of (S) is a point S_iTo point S_jIs expressed as min (L (S)_i,S_k)+L(S_k,S_j),L(S_i,S_j))，

The basic thought of the genetic algorithm is as follows: assuming that N distributed resource nodes and 1 main power plant node Q exist in the system, a binary coding mode is adopted for chromosomes in the genetic algorithm, and the distributed resource starting sequence is represented by chromosomes with the length of k x (N +1), namely Chrom [ [ A ] ]₁,A₂,…,A_N|A_N+1]Wherein A is_N+1Is a chromosome fragment of length k, A₁～A_NRepresenting the starting time of each distributed resource, A_N+1The method comprises the steps of setting the thermal power plant to be started finally, randomly generating a batch of initial chromosomes by the genetic algorithm, converting an objective function of a distributed resource starting sequence optimization model into a fitness function of the chromosomes, randomly selecting through roulette based on the fitness function values of the chromosomes, and carrying out operation steps of crossing, mutation, recombination and the like to generate offspring chromosomes, recording and reserving parent excellent chromosomes, and repeating the steps until set conditions are met.

9. The toughness-boosting-oriented distributed resource assisted main network key node black start strategy of claim 1, wherein:

in step 3, after various distributed resources in the power distribution network start and aggregate to form an area black start power supply, starting to close to the thermal power plant in the main network, and providing a power supply for the thermal power unit auxiliary machine, wherein the power balance expression is as follows:

P_W(t)+P_PV(t)+P_ES(t)+P_MT(t)-P_load(t)＝0 (10)

ΔP₊(t)＝|P_W(t)+P_PV(t)-P_load(t)| (11)

ΔP_-(t)＝|P_W(t)+P_PV(t)-P_load(t)| (12)

the distributed resource starting thermal power generating unit coordination control process comprises the following steps:

step 3-1, at the moment t, starting and aggregating distributed resources to form a regional power supply;

step 3-2, the regional power supplies are put into the auxiliary machines of the thermal power plant in sequence;

step 3-3, in the initial stage of black start, when the sum of the output of the wind power generator and the output of the photovoltaic generator is smaller than the power of the auxiliary machine, the wind power generator and the photovoltaic generator both adopt an MPPT control strategy, and the rest power shortage delta P_-(t) is borne by the energy storage and micro gas turbine,

when the sum of the output of the wind power generator and the output of the photovoltaic generator set is larger than the power of the auxiliary machine and the excess quantity delta P₊(t) less than 10% P_PV(t), the photovoltaic set adopts MPPT control, the wind turbine set adopts a load tracking control strategy to track the auxiliary engine power and photovoltaic output deficit change output,

and 3-4, if the thermal power generating unit does not recover to generate power, repeating the steps 3-2-3 until the thermal power generating unit recovers to generate power.