Method and device for generating fault data of excitation system of hydropower station

Abstract

The invention discloses a hydropower station excitation system fault data generation method and device, and belongs to the technical field of hydropower station excitation system faults. The method comprises the following steps: building a model for connecting a single machine of a hydroelectric generating set of a hydropower station into a power grid; the hydropower station water turbine generator set single machine merging power grid model comprises an excitation system model, a speed regulator model, a generator model, a water turbine model and a power grid model; under different input state working conditions, an improved large field search algorithm is used for searching adjustable excitation parameters in an excitation system model in each iteration process, and then the adjustable excitation parameters are merged into a power grid model for simulation through a single hydroelectric generating set of a hydropower station so as to obtain fault data. According to the method, firstly, a hydropower station water-turbine generator set single machine merging power grid model is built, and then variable parameter, variable working condition and variable excitation parameter experiments are carried out based on the model, so that fault data can be found out.

Description

Technical Field

[0001] The invention relates to a method and a device for generating fault data of an excitation system of a hydropower station, belonging to the technical field of faults of excitation systems of hydropower stations. Background Art

[0002] Improper setting of excitation system parameters of hydropower stations can easily lead to abnormal output of regulators. In the actual production and operation of hydropower stations, if the abnormal regulation of the excitation system is not discovered in time, it may cause problems such as voltage imbalance, which in turn leads to instability of the generator output power, thereby affecting the operation stability of the unit or the power grid, and even causing equipment damage in severe cases. For example, on July 10, 2016, the No. 1 unit of the Longma Hydropower Station of Lixianjiang Datang Group in Yunnan Province caused grid frequency oscillation during operation. The reason was caused by the forced oscillation inside the excitation system. The forced oscillation was mainly caused by the output of PSS (Power System Stabilizer). According to the analysis report of a research unit, due to the superposition of low excitation output PSS output, the amplitude-frequency and phase-frequency characteristics of the AVR (Automatic Voltage Regulator) link changed. Therefore, the original PSS parameters for phase compensation may not be completely suitable. The greater the low excitation output, the greater the impact of PSS output. The above problem is only a typical example of improper parameter setting, and fails to fully reflect the adjustment problems caused by improper parameter setting. Therefore, how to avoid improper excitation parameter setting and ensure the safety of the power system and related electrical equipment has become a technical problem that needs to be solved urgently.

[0003] The acquisition and application of fault data generated during the actual operation of the current hydropower station excitation system mainly depends on the integrity, timeliness, closure and storage of data records. For example, one method can directly extract data from the power station monitoring device. Since most hydropower station excitation systems do not have independent fault recording devices, operators usually retrieve data from the monitoring system when analyzing data, and the monitoring data generally does not contain excitation intermediate scalar data, such as AVR regulation output, PSS output and other key data, resulting in the lack of key features in fault analysis and diagnosis, and quantitative analysis is difficult. At the same time, due to the privacy of these operating data, which usually involves sensitive information, the acquisition and sharing are subject to strict privacy and security regulations. Therefore, it is particularly difficult to collect data on different fault types under the operating conditions of the excitation system of each hydropower station. The other is the field test method, which simulates the influence of different adjustment parameters on voltage output under different working conditions to find the optimal adjustment parameters. However, due to the limitations of power generation conditions and safety factors in the actual simulation process, the test parameters are limited, which cannot reflect the operating status of the entire parameter domain of the excitation system, and it is not easy to find the parameter set that can induce adjustment faults. Therefore, based on the above problems, it has become a mainstream technical means to use professional simulation modeling technology to establish a detailed model of the excitation system of a hydropower station, and to generate fault data by simulating different fault scenarios. In recent years, researchers at home and abroad have proposed some methods for generating fault data of excitation systems. Existing research mainly focuses on faults caused by damage to the main device of the excitation system, but its method still has certain limitations and does not consider the regulation faults caused by improper parameter configuration. It can be seen that it is urgent to invent a data generation method for abnormal regulator output caused by improper configuration of hydropower station excitation parameters to make up for the shortcomings of existing excitation fault data. Summary of the invention

[0004] The present invention provides a method and device for generating fault data of an excitation system of a hydropower station. Firstly, a model of a single turbine generator set in a hydropower station being connected to a power grid is constructed. Then, based on the model, experiments of varying parameters, varying operating conditions (varying frequency, varying reactive power) and varying excitation parameters are carried out to find fault data.

[0005] The technical solution of the present invention is:

[0006] According to a first aspect of the present invention, a method for generating fault data of an excitation system of a hydropower station is provided, comprising the following steps: constructing a model of a single unit of a hydro-turbine generator set in a hydropower station being connected to a power grid; the model of a single unit of a hydro-turbine generator set in a hydropower station being connected to a power grid comprises an excitation system model, a governor model, a generator model, a turbine model and a power grid model; and obtaining fault data by searching the adjustable excitation parameters in the excitation system model in each iteration using an improved large domain search algorithm under different input state conditions and then simulating the model of a single unit of a hydro-turbine generator set in a hydropower station being connected to a power grid.

[0007] Furthermore, the improved large domain search algorithm is used to search the adjustable excitation parameters in the excitation system model in each iteration process, specifically:

[0008] S1: Select the adjustable excitation parameters in the excitation system model and determine the excitation system model parameter set X i =[x1 i ,x2 i ,x3 i …x j i …x m i ]Specific parameters; among them, x j i Indicated as in the input state δ i The jth adjustable excitation parameter under ;

[0009] S2: Construct the excitation system model parameter set X i =[x1 i ,x2 i ,x3 i …x j i …x m i ]’s discretization coding method;

[0010] S3: Set the first input state δ1, that is, determine (W G 1 ,V G 1 ), and determine other input states, a total of z input states are set; among them, (W G 1 ,V G 1 ) represents the given frequency per unit value of the hydro-generator set under the first input state W G And the given voltage per unit value V G ;

[0011] S4: Build X 1 The scale generation method determines X 1 The number of sets u that constitute G 1 (old) set, i.e. u1,u2∈u, represents the uth set of adjustable excitation parameters;

[0012] S5: To G 1 (old) Conduct parameter searches;

[0013] S6: Construct the evaluation function in the search process to evaluate the simulation result Y of each search1 Perform an evaluation, and if the evaluation value exceeds a preset threshold, it is determined to be fault data;

[0014] S7: After repeating steps S5-S6, select G 1 The adjustable excitation parameter set corresponding to the fault data in the (old) set constitutes the set G 1 (new), i.e., the set of adjustable excitation parameters corresponding to the fault data under the first input state;

[0015] S8: Change the input state, change δ1 in S3 to δ2, and continue to execute S3-S7 to select G 2 (new), repeat z times until the set G(total) = {G 1 (new),G 2 (new)…G Z (new)}, the algorithm stops and records the value of each element of G(total) Output the result; Indicated as in the input state δ i Next by G i The output results generated by each set of adjustable excitation parameter sets in (new).

[0016] Furthermore, the discretization encoding method of the adjustable parameters of the excitation model is specifically as follows: for X i Each parameter in is randomly drawn from an optional set containing D elements. j i The value of can be expressed as x j i ∈W j ,W j ={w j(1) i ,w j(2) i ,w j(3) i …w j(d) i …w j(D) i}; where W j is the adjustable excitation parameter x j i An optional set of values, w j(d) i It is the dth value result of the jth parameter in the ith input state.

[0017] Furthermore, the dth value w of the jth parameter of the ith input state j(d) i The calculation formula is:

[0018]

[0019] Among them, min(x j i )、max(x j i ) represent x j i The minimum and maximum values ​​of .

[0020] Furthermore, the input state δ i The setting range is as follows: V G i ∈[0.8,1.2]; W G i ∈[1-0.004,1+0.004]; where V G i ,W G i It represents the given voltage per unit value at the generator end and the given frequency per unit value of the hydro-generator set under the i-th input state.

[0021] Furthermore, for G i (old) Conduct a parameter search as follows:

[0022] The specific operation strategy of the first local search strategy is: in the i-th input state, from G i The first set of excitation parameters randomly selected in (old) Any j-th parameter x in j i In {w j(1) i ,w j(2) i ,W j(3) i …w j(d) i …w j(D) i} A new set of excitation parameters of the first group is generated by exchanging with other values ​​in the set range;

[0023] The specific operation strategy of the second local search strategy is: in the i-th input state, from G i Any g-th parameter x of the randomly selected l1-th set of excitation parameters in (old) p i The current value of is the same as any pth parameter x of the randomly selected l2th group of excitation parameters p i The current values ​​are mapped in their respective sets, g≠p.

[0024] Furthermore, the specific process of repeatedly executing steps S5-S6 is as follows:

[0025] S51, set at G i (old) Maximum number of iterations of the search, there are a total of z states; initial state i = 1, number of iterations gen = 1;

[0026] S52, executing the first local search strategy, using the evaluation function to evaluate the excitation parameter set participating in the first local search strategy If any one of the two conditions in the first evaluation rule is met, Deposit G i (new) set, and G i (old) Remove from collection If both conditions are not met, execute S53;

[0027] S53, executing the second local search strategy, and using the evaluation function to evaluate the two sets of excitation parameter sets participating in the second local search strategy If any one of the two conditions in the second evaluation rule or the third evaluation rule is met, the excitation parameter set that meets the condition is stored in G i (new) set, and G i (old) Remove from collection or If any one of the two conditions in the second evaluation rule is satisfied and any one of the two conditions in the third evaluation rule is satisfied, then both sets of excitation parameter sets that satisfy the conditions are stored in G i (new) set, and G i (old) Remove from collection and Otherwise, execute S54;

[0028] S54, repeatedly execute S52-S53 until the termination condition of the current state is reached and then enter the next state until all z states are executed.

[0029] Furthermore, the first evaluation rule is specifically:

[0030]

[0031] in, VG is the new set of excitation parameters generated according to the first local search strategy; i is the per unit value of the given voltage at the generator terminal under the i-th input state; The adjustable excitation parameter combination under the i-th input state is Fault data evaluation function during configuration; data_fault is the generation condition of excitation fault data: if If any one of the two conditions is met, Deposit G i (new) set, and G i (old) Remove from collection

[0032] Furthermore, the second evaluation rule is specifically:

[0033]

[0034] in, is the new set of excitation parameters l1 generated according to the second local search strategy; V G i is the per unit value of the given voltage at the generator terminal under the i-th input state; The adjustable excitation parameter combination under the i-th input state is Fault data evaluation function during configuration; data_fault is the generation condition of excitation fault data: if If any one of the two conditions is met, Deposit G i (new) set, and G i (old) Remove from collection

[0035] Furthermore, the third evaluation rule is specifically:

[0036]

[0037] in, is a new set of excitation parameters l2 generated according to the second local search strategy; V G i is the per unit value of the given voltage at the generator terminal under the i-th input state; The adjustable excitation parameter combination under the i-th input state is Fault data evaluation function during configuration; data_fault is the generation condition of excitation fault data: if If any one of the two conditions is met, Deposit G i (new) set, and G i (old) Remove from collection

[0038] According to a second aspect of the present invention, a device for generating fault data of an excitation system of a hydropower station is provided, comprising: a model building module for building a model of a single unit of a hydro-turbine generator set in a hydropower station being connected to the grid; the model of a single unit of a hydro-turbine generator set in a hydropower station being connected to the grid comprises an excitation system model, a governor model, a generator model, a turbine model and a grid model; an acquisition module for obtaining fault data by searching for adjustable excitation parameters in the excitation system model in each iteration process using an improved large field search algorithm under different input state conditions, and then simulating the model of a single unit of a hydro-turbine generator set in a hydropower station being connected to the grid.

[0039] The beneficial effects of the present invention are:

[0040] First, the present invention solves the problem of fault identification caused by multi-parameter adjustment of the excitation system and improves the efficiency of fault data generation of the excitation system. Specifically: Due to the complexity of the excitation system parameters of the hydropower unit and the large number of parameters, traditional manual methods or exhaustive methods are difficult to cope with the complex fault waveform identification caused by parameter changes; and the present invention, by designing the ILDSA algorithm, can efficiently cope with changes in up to more than ten excitation parameters, and quickly locate fault data through a domain search strategy, providing a powerful tool for system fault analysis and optimization. In addition, by designing an improved domain search algorithm (ILDSA), fault data caused by improper configuration of the excitation system parameters of the hydropower unit can be automatically and efficiently identified and searched; compared with traditional enumeration methods or exhaustive algorithms, the ILDSA algorithm can quickly and accurately discover fault waveforms, greatly improving the efficiency and accuracy of the data generation process, and avoiding the tedious work of manual verification.

[0041] Second, the present invention constructs a model of a single hydro-turbine generator set in a hydropower station connected to the grid, which includes an excitation system, a speed governor, a turbine, a generator, and a power grid. The model of a single hydro-turbine generator set in a hydropower station connected to the grid covers multiple complex system components. Based on this model, various operating conditions and parameter change experiments are carried out, so that the fault data of the excitation system can be comprehensively and efficiently identified, thereby improving the accuracy and applicability of fault diagnosis.

[0042] Third, the present invention provides a reasonable reference basis for the parameter configuration of the excitation system of the hydropower station through the generation and analysis of fault data; by reasonably configuring the excitation system parameters, it can effectively avoid faults caused by improper parameters, improve the stability and reliability of the hydropower unit, and thus improve the overall operating efficiency and safety of the hydropower station. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is the operation model of a single turbine generator set in a hydropower station connected to the power grid;

[0044] Figure 2 It is a fault waveform curve diagram. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. It should be noted that the embodiments in this application and the features in the embodiments can be combined with each other arbitrarily without conflict.

[0046] Since there are more than ten excitation parameters, it is impossible to complete the search using an exhaustive method and manually verify them. Therefore, the present invention provides a method and device for generating fault data of a hydropower station excitation system, so that the algorithm can efficiently search for the fault data of the hydropower station excitation system and provide a reasonable reference basis for the excitation parameter configuration.

[0047] like Figure 1-2 According to a first aspect of an embodiment of the present invention, a method for generating fault data of an excitation system of a hydropower station is provided, comprising: constructing Figure 1 The single-machine grid-connected model of a hydro-turbine generator set of a hydropower station shown in the figure comprises an excitation system model, a governor model, a generator model, a turbine model and a grid model. Under different input state conditions, the improved large domain search algorithm ILDSA is used to search for the adjustable excitation parameters in the excitation system model in each iteration process, and then the single-machine grid-connected model of a hydro-turbine generator set of a hydropower station is simulated to obtain fault data.

[0048] The speed regulator model is as follows:

[0049] C1(s)=D1(s)*Y1(s);

[0050]

[0051] W ERR (s) = W G (s)-W1(s);

[0052] Y g (s) = W ERR (s)*C1(s);

[0053] Where C1(s) is the transfer function of the speed regulator model, D1(s) is the isolated network model of the speed regulation system of a single unit in the frequency dead zone, and Y1(s) is the actuator model of the speed regulation system of a single unit; T y is the stationary time constant of the relay, s is the pull factor; K P is the controller PID proportional link gain, KI is the controller PID integral link gain, K D is the controller PID differential link gain; b p is the controller comprehensive output feedback coefficient, T 1v is the differential time constant; e fH is the upper limit of the isolated network frequency zone, e fL is the lower limit of the isolated network frequency dead zone; W ERR (s) is the frequency error, W G (s) is the given frequency of the hydro-generator set, W1(s) is the generator terminal frequency, Y g (s) is the speed regulator output.

[0054] The turbine model is as follows:

[0055]

[0056] P m (s)=P1(s)*Y g (s);

[0057] Where P1(s) is the transfer function of the turbine model, T w is the water flow inertia time constant, a and b are turbine model coefficients, P m (s) is the mechanical power of the turbine.

[0058] The excitation system model is as follows:

[0059] [V ERR (s)+P2(s)]*E(s)=E f (s);

[0060] V G (s) = V t (s) = V ERR (s);

[0061]

[0062]

[0063] Where V ERR (s) is the difference between the given voltage and the generator terminal voltage, P2(s) is the transfer function of PSS, E(s) is the overall transfer function of the excitation system, E f (s) is the excitation voltage input to the generator by the excitation system, V G (s) is the given voltage at the generator end, V t (t) The transfer function of the generator output terminal voltage, V ERR (s) is the transfer function of another input of the excitation system besides PSS; E AVR(s), E OL (s), E LL (s), E1(s) are the transfer functions of the AVR part of the excitation system, the overexcitation limit, the underexcitation limit part, and the output after the low-value gate, respectively. V L1 is the output of the low-value gate, V H is the low excitation limit output voltage, V AVR is the output voltage of the AVR part, V L2 is the output voltage of overexcitation limit; K, K A , K F , K H1 , K L , K S , K r , K p is the gain of each link of the transfer function; K V is the model selection factor, which takes the value of 0 or 1; T1, T2...T 25 、T H1 、T H2 、T L1 、T L2 is the time constant of each link of the transfer function; G U (s), G L (s), G B (s) are the transfer functions of different branches of PSS; ω n is the system frequency. Among the above parameters, the decision parameters are: T H1 、T H2 、T L1 、T L2 , K, K H1 , K L , K V Etc., the optimal algorithm can be determined by optimizing the decision parameters.

[0064] The generator model is as follows:

[0065]

[0066]

[0067] Where G(s) is the transfer function of the generator model, P e (s) is the transfer function of the generator output active power, Q e (s) is the transfer function of the generator output reactive power, E f (s) and P m (s) are the excitation voltage input to the generator by the excitation system and the mechanical power of the turbine, respectively; M is the moment of inertia, ζ is the damping coefficient, R s is the stator resistance, X sis the synchronous reactance, L is the stator inductance, φ is the phase difference between the terminal voltage and the stator current; j represents an imaginary number; V t (s) is the transfer function of the generator output terminal voltage, I a (s) is the transfer function of the generator stator current; δ(s) is the Laplace transform of the generator rotor phase angle δ(t); W1(s) is the Laplace transform of the generator rotor angular velocity ω1(t), that is, the generator terminal frequency.

[0068] The power grid model is as follows:

[0069]

[0070] Where H(s) is the transfer function of the power grid model, W1(s) is the generator terminal frequency, P m (s) is the mechanical power of the turbine, M is the moment of inertia, X s is the synchronous reactance and ζ is the damping coefficient.

[0071] According to the above model transfer function, by changing the given voltage per unit value V G And the given frequency per unit value W of the hydro-generator set G Constitute different input states δ i =(W G i ,V G i ),i∈{1,2,3…z}, where δ i Represented as the i-th input state, (W G i ,V G i ) represents the given voltage per unit value V at the generator terminal under the i-th input state G And the given frequency per unit value W of the hydro-generator set G (Here expressed as W G 、V G It is due to W G 、V G It is actually a constant, which is different from W expressed in the above theoretical model. G (s), V G (s)). z represents the total number of states generated in the fault generation experiment, and the input state set can be expressed as δ = {δ1, δ2, δ3…δ z}={(W G 1 ,V G 1 ),(W G 2 ,V G 2 ),…(W Gz ,V G z}.

[0072] Set a certain input state δ i , in δ i By changing the excitation model parameter set X i =[x1 i ,x2 i ,x3 i …x j i …x m i ], j∈{1,2,3…m}, after the single unit of the hydropower station turbine generator set is connected to the grid model simulation, different output results Y are generated i =[y1 i ,y2 i ,y3 i …y q i …y k i ], q∈{1,2,3…k}(such as Figure 2 is the output result corresponding to the excitation model parameter set, and the characteristics of the output result include active power, reactive power, generator terminal frequency, and generator terminal voltage); where X i Indicated as in the input state δ i The following adjustable excitation parameter set, x j i Indicated as in the input state δ i The jth adjustable excitation parameter under i Indicated as in the input state δ i The following is by X i The i-th output result, y q i Indicated as in the input state δ i The following constitutes Y i The qth output feature of .

[0073] Excitation model adjustable excitation parameter X i =[x1 i ,x2 i ,x3 i …x j i …x m i The main operating steps of the improved large domain search algorithm ILDSA include:

[0074] S1. Select the adjustable excitation parameters of the excitation model and determine the excitation model parameter set X i =[x1i ,x2 i ,x3 i …x j i …x m i ]Specific parameters; among them, x j i Indicated as in the input state δ i The jth adjustable excitation parameter under ;

[0075] S2. Construct excitation model parameter set X i =[x1 i ,x2 i ,x3 i …x j i …x m i ]’s discretization coding method;

[0076] S3. Set the first input state δ1, that is, determine (W G 1 ,V G 1 ), and determine other input states, a total of z input states are set; among them, (W G 1 ,V G 1 ) represents the given voltage per unit value V at the generator terminal under the first input state G And the given frequency per unit value W of the hydro-generator set G ;

[0077] S4. Build X 1 The scale generation method determines X 1 The number of sets u that constitute G 1 (old) set, i.e. u1,u2∈u, represents the uth set of adjustable excitation parameters;

[0078] S5. vs. G 1 (old) Conduct parameter searches;

[0079] S6. Construct the evaluation function CFEI in the search process for each search simulation result Y 1 Perform an evaluation, and if the evaluation value exceeds a preset threshold, it is determined to be fault data;

[0080] S7. Repeat steps S5-S6 and select G 1 The parameter set corresponding to the fault data in the (old) set constitutes the set G 1(new), that is, the parameter set corresponding to the fault data under the first input state;

[0081] S8. Change the input state, change δ1 in S3 to δ2, and continue to execute S3-S7 to select G 2 (new), repeat z times until the set G(total) = {G 1 (new),G 2 (new)…G Z (new)}, the algorithm stops and records the value of each element of G(total) Output the result; Indicated as in the input state δ i Next by G i The output results generated by each set of adjustable excitation parameter sets in (new).

[0082] Exemplarily, the excitation model can adjust the excitation parameter X i =[x1 i ,x2 i ,x3 i …x j i …x m i ]for:

[0083] Table 1 Adjustable excitation parameter set

[0084] Parameter number Parameter name symbol Value range <![CDATA[x1 i ]]> Excitation AVR part time constant <![CDATA[T 19 ]]> [0.01,100] <![CDATA[x2 i ]]> Excitation AVR part time constant <![CDATA[T 20 ]]> [0.01,100] <![CDATA[x3 i ]]> Excitation AVR part time constant <![CDATA[T 21 ]]> [0.01,100] <![CDATA[x4 i ]]> Excitation AVR part time constant <![CDATA[T 22 ]]> [0.01,100] <![CDATA[x5 i ]]> Excitation AVR gain K (0,100] <![CDATA[x6 i ]]> Excitation AVR Part Model Selection Factor <![CDATA[K V ]]> 0 or 1 <![CDATA[x7 i ]]> Excitation low excitation limit part time constant <![CDATA[T H1 ]]> [0.01,100] <![CDATA[x8 i ]]> Excitation low excitation limit part time constant <![CDATA[T H2 ]]> [0.01,100] <![CDATA[x9 i ]]> Excitation low excitation limit part gain <![CDATA[K H1 ]]> [-50,50] but not 0 <![CDATA[x 10 i ]]> Excitation overexcitation limit part time constant <![CDATA[T L1 ]]> [0.01,100] <![CDATA[x 11 i ]]> Excitation overexcitation limit part time constant <![CDATA[T L2 ]]> [0.01,100] <![CDATA[x 12 i ]]> Excitation overexcitation limit part gain <![CDATA[K L ]]> (0,50] <![CDATA[x 13 i ]]> PSS Partial Gain <![CDATA[K S ]]> (0,100] <![CDATA[x 14 i ]]> PSS Partial Gain <![CDATA[K r ]]> (0,100] <![CDATA[x 15 i ]]> PSS Partial Gain <![CDATA[K p ]]> (0,100]

[0085] Furthermore, the discretization encoding method of the adjustable parameters of the excitation model is specifically as follows:

[0086] X i The selection rule for each parameter in is: randomly selected from an optional set containing D elements, for example, any parameter x j i The value of can be expressed as x j i ∈W j ,W j ={w j(1) i ,w j(2) i ,w j(3) i …w j(d) i …w j(D) i}, where W j is the adjustable excitation parameter x j i An optional set of values, w j(d)i It is the dth value result of the jth parameter in the ith input state.

[0087] The calculation formula for the dth value of the jth parameter of the i-th input state in the above formula is:

[0088]

[0089] Among them, min(x j i )、max(x j i ) represent x j i The minimum and maximum values ​​of are shown in Table 1. j i ) cannot be 0, then the value is 0.01, the smallest two-digit number close to 0. For example, the excitation adjustable parameter corresponds to x5 in Table 1 i , if the current state is the first input state, then min(x5 1 )=0.01,max(x5 1 )=100,D=10,d=5,then w 5(5) 1 =0.01+5*[(100-0.01) / 10]=49.96.

[0090] Furthermore, the input state δ i The setting range of VG is as follows: i ∈[0.8,1.2]; W G i ∈[1-0.004,1+0.004].

[0091] Furthermore, the large-scale generation method is specifically as follows:

[0092] To construct the uth set of adjustable excitation parameters To explain: Any j-th parameter x of j i In the optional {w j(1) i ,w j(2) i ,w j(3) i …w j(d) i …w j(D) i}The collection is randomly generated.

[0093] Furthermore, for G i (old) Conduct a parameter search as follows:

[0094] The first local search strategy is defined as swap_local. The specific operation strategy is: in the i-th input state, the search is performed from G i The first set of excitation parameters randomly selected in (old) Any j-th parameter x in j i In {w j(1) i ,w j(2) i ,w j(3) i …w j(d) i …w j(D) i}Exchange with other values ​​in the collection range, for example, the current parameter x j i The value of (old) is w j(3) i , let r = rand(1,D), rand(1,D) represents a random integer in [1,D], then the new parameter x j i (new) = w j(r) i , r≠3;

[0095] The second local search strategy is defined as map_global. The specific operation strategy is: in the i-th input state, the local search strategy is to search from G i Any j-th parameter x of the randomly selected l1-th set of excitation parameters in (old) j i The current value of is the same as any pth parameter x of the randomly selected l2th group of excitation parameters p i The current value is mapped in the respective set (j≠p), and the mapping rule is to swap the subscripts of the elements in the set. For example, x j i Currently in the collection {w j(1) i ,w j(2) i ,w j(3) i …w j(k1) i …w j(D) i The value in} is w j(k1) i , x p i Currently in the collection {w p(1) i ,w p(2) i ,wp(3) i …w p(k2) i …w p(D) i The value in} is w p(k2) i , swap k1, k2, then x j i The value of is changed to w j(k2) i , x p i The value of is changed to w p(k1) i , forming a new parameter x j i (new), x p i (new) Constructs a set of excitation parameters.

[0096] Furthermore, the evaluation function CFEI evaluation method is as follows:

[0097]

[0098] N=(t end i -t start i ) / f sample i

[0099] Among them, F CFEI (X i ) is the adjustable excitation parameter combination under the i-th input state is X i Fault data evaluation function during configuration, V t i (t) is the unit value of the terminal voltage output at time t under the i-th input state, t start i is the disturbance start time under the i-th input state, t end i The end time of disturbance under the i-th input state, f sample i The simulation sampling frequency under the i-th input state, V G i The per unit value of the given voltage at the generator terminal under the i-th input state.

[0100] Furthermore, the search strategy of ILDSA and the evaluation rules of CFEI are as follows:

[0101] S51, set at G i(old) The total number of iterations of the search is gen, and there are z states in total; the initial state i = 1, and the initial number of iterations gen = 1;

[0102] S52, execute the first local search strategy, using F CFEI Evaluate the set of excitation parameters participating in the first local search strategy

[0103]

[0104] in, is a new set of excitation parameters l generated according to the first local search strategy; is the generation condition of excitation fault data; if If any one of the two conditions is met, Deposit G i (new) set, and G i (old) Remove from collection If both conditions are not met, execute S53;

[0105] S53, execute the second local search strategy, using f CFEI Evaluation of two sets of excitation parameters participating in the second local search strategy

[0106]

[0107] like or If any one of the conditions is met, the excitation parameter set that meets the condition is stored in G i (new) set, and G i (old) Remove from collection or like and If any one of the conditions is met, the two sets of excitation parameter sets that meet the conditions are stored in G i (new) set, and G i (old) Remove from collection and Otherwise, execute S54;

[0108] S54, at the end of each iteration, the number of iterations is increased by 1, and S52-S53 are repeated until the termination condition of the current state is reached and then the next state is entered until all z states are executed; the termination condition of the current state is: reaching the maximum number of iterations; or G i All the data in (old) are eliminated (that is, if any of the above termination conditions is met, the search for the next state will be directly entered).

[0109] In order to verify the effectiveness of the ILDSA proposed in this invention and the accuracy of the model. The specific experimental parameter settings are as follows: the algorithm is programmed in Matlab 2022a, and the hardware environment is tested on a PC with Intel core i72.80GHz and 16G memory. The simulation parameters are set as follows: Set the adjustable parameters of the excitation system as shown in Table 1, set the inherent parameters or identification parameters T of the speed regulator and turbine y =0.10s,T w =1.6s, a=1, b=0.5, K P =4,K I =3,K D =1,b P =4%, e fl =e fH =0.05Hz Set the inherent parameters of the excitation system as shown in Table 2 below.

[0110] Table 2 Excitation fixed parameter set

[0111] <![CDATA[T1]]> 36 <![CDATA[T2]]> 0.00746 <![CDATA[T3]]> 0.2587 <![CDATA[T4]]> 3.448 <![CDATA[T5]]> 21.517 <![CDATA[T6]]> 23.548 <![CDATA[T7]]> 54.3264 <![CDATA[T8]]> 13.48 <![CDATA[T9]]> 6 <![CDATA[T 10 ]]> 60 <![CDATA[T 11 ]]> 76 <![CDATA[T 12 ]]> 17 <![CDATA[T 13 ]]> 0.046 <![CDATA[T 14 ]]> 0.476 <![CDATA[T 15 ]]> 1.43 <![CDATA[T 16 ]]> 0.00036 <![CDATA[T 17 ]]> 0.0306 <![CDATA[T 18 ]]> 0.65 <![CDATA[T 23 ]]> 0.004 <![CDATA[T 24 ]]> 0.04 <![CDATA[T 25 ]]> 0.02 7 1

[0112] Set the inherent parameters of the generator and its power grid system M = 12969.11 kg m 2 , ζ=0pu,R s =2.8544×10 -3 pu,X s =1.305pu, L = 0pu. The ILDSA algorithm is set as follows: gen = 100, u = 100, z = 1, D = 5, δ1 = (1.004, 1.1).

[0113] Supplementary excitation fault waveform curve and parameter list 3:

[0114] Table 3 Adjustable parameter values ​​for excitation fault

[0115] Parameter number Parameter name symbol Value <![CDATA[x1 1 ]]> Excitation AVR part time constant <![CDATA[T 19 ]]> 100 <![CDATA[x2 1 ]]> Excitation AVR part time constant <![CDATA[T 20 ]]> 100 <![CDATA[x3 1 ]]> Excitation AVR part time constant <![CDATA[T 21 ]]> 10 <![CDATA[x4 1 ]]> Excitation AVR part time constant <![CDATA[T 22 ]]> 1 <![CDATA[x5 1 ]]> Excitation AVR gain K 1 <![CDATA[x6 1 ]]> Excitation AVR Part Model Selection Factor <![CDATA[K V ]]> 1 <![CDATA[x7 1 ]]> Excitation low excitation limit part time constant <![CDATA[T H1 ]]> 1 <![CDATA[x8 1 ]]> Excitation low excitation limit part time constant <![CDATA[T H2 ]]> 10 <![CDATA[x9 1 ]]> Excitation low excitation limit part gain <![CDATA[K H1 ]]> -10 <![CDATA[x 10 1 ]]> Excitation overexcitation limit part time constant <![CDATA[T L1 ]]> 9 <![CDATA[x 11 1 ]]> Excitation overexcitation limit part time constant <![CDATA[T L2 ]]> 6 <![CDATA[x 12 1 ]]> Excitation overexcitation limit part gain <![CDATA[K L ]]> 1.5 <![CDATA[x 13 1 ]]> PSS Partial Gain <![CDATA[K S ]]> 1 <![CDATA[x 14 1 ]]> PSS Partial Gain <![CDATA[K r ]]> 1 <![CDATA[x 15 1 ]]> PSS Partial Gain <![CDATA[K p ]]> 80

[0116] By applying the above technical solution, it can be known that, on the one hand, the present invention constructs an overall mathematical model of a hydro-turbine generator set, and the overall mathematical model of the hydro-turbine generator set includes a governor model (including mechanical and electrical models), an excitation system model (including an underexcitation limit model, a PSS model, and an overexcitation limit model), a turbine model, a generator model, and a power grid model; by searching and simulating the excitation parameters in each iteration process using ILDSA under different model input conditions, possible faults in the excitation system are found, that is, the excitation system parameters that may have abnormal regulation are continuously searched, including the time constant of the excitation AVR part, the gain of the excitation AVR part, the excitation underexcitation and overexcitation limit and other related parameters (see Table 1 for details of the excitation adjustable parameter set), and the values ​​are changed within the adjustable range and then simulated, and the result simulation is judged by the designed Comprehensive Fault Evaluation Index (CFEI) in each iteration of ILDSA searching for unreasonable excitation parameters, and the fault data is screened out and marked, and finally the parameter configuration data with fault marks is generated.

[0117] According to the second aspect of an embodiment of the present invention, a device for generating fault data of an excitation system of a hydropower station is provided, comprising: a model building module for building a model of a single unit of a hydro-turbine generator set in a hydropower station being connected to the power grid; the model of a single unit of a hydro-turbine generator set in a hydropower station being connected to the power grid comprises an excitation system model, a governor model, a generator model, a turbine model and a power grid model; an acquisition module for simulating the model of a single unit of a hydro-turbine generator set in a hydropower station being connected to the power grid by searching the adjustable excitation parameters in the excitation system model in each iteration process using an improved large domain search algorithm under different input state conditions, so as to obtain fault data. For the parts not described in detail in the above for each module, please refer to the relevant description of this embodiment.

[0118] The specific implementation modes of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above implementation modes, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A method for generating fault data of a hydropower station excitation system, characterized in that: The following steps are involved: Build a model for connecting a single hydro-turbine generator set of a hydropower station to the power grid; the model for connecting a single hydro-turbine generator set of a hydropower station to the power grid includes an excitation system model, a governor model, a generator model, a turbine model and a power grid model; By using the improved large-area search algorithm to search the adjustable excitation parameters in the excitation system model in each iteration under different input state conditions, a single unit of the hydropower station's turbine generator set is connected to the grid model for simulation to obtain fault data.

2. The method for generating fault data of a hydropower station excitation system according to claim 1, characterized in that: The improved large domain search algorithm is used to search the adjustable excitation parameters in the excitation system model in each iteration process, specifically: S1: Select the adjustable excitation parameters in the excitation system model and determine the excitation system model parameter set X i =[x1 i ,x2 i ,x3 i …x j i …x m i ]Specific parameters; among them, x j i Indicated as in the input state δ i The jth adjustable excitation parameter under ; S2: Construct the excitation system model parameter set X i =[x1 i ,x2 i ,x3 i …x j i …x m i ]’s discretization coding method; S3: Set the first input state δ1, that is, determine (W G 1 ,V G 1 ), and determine other input states, a total of z input states are set; among them, (W G 1 ,V G 1 ) represents the given frequency per unit value of the hydro-generator set under the first input state W G And the given voltage per unit value V G ; S4: Build X 1 The scale generation method determines X 1 The number of sets u that constitute G 1 (old) set, i.e. u>2,u1,u2∈u, represents the uth set of adjustable excitation parameters; S5: To G 1 (old) Conduct parameter searches; S6: Construct the evaluation function in the search process to evaluate the simulation result Y of each search 1 Perform an evaluation, and if the evaluation value exceeds a preset threshold, it is determined to be fault data; S7: After repeating steps S5-S6, select G 1 The adjustable excitation parameter set corresponding to the fault data in the (old) set constitutes the set G 1 (new), i.e., the set of adjustable excitation parameters corresponding to the fault data under the first input state; S8: Change the input state, change δ1 in S3 to δ2, and continue to execute S3-S7 to select G 2 (new), repeat z times until the set G(total) = {G 1 (new),G 2 (new)…G Z (new)}, the algorithm stops and records the value of each element of G(total) Output the result; Indicated as in the input state δ i Next by G i The output results generated by each set of adjustable excitation parameter sets in (new).

3. The method for generating fault data of a hydropower station excitation system according to claim 1, characterized in that: The discretization encoding method of the adjustable parameters of the excitation model is as follows: i Each parameter in is randomly drawn from an optional set containing D elements. j i The value of can be expressed as x j i ∈W j ,W j ={w j(1) i ,w j(2) i ,w j(3) i …w j(d) i …w j(D) i }; where W j is the adjustable excitation parameter x j i An optional set of values, w j(d) i It is the dth value result of the jth parameter in the ith input state.

4. The method for generating fault data of a hydropower station excitation system according to claim 3, characterized in that: The dth value w of the jth parameter of the i-th input state j(d) i The calculation formula is: Among them, min(x j i )、max(x j i ) represent x j i The minimum and maximum values ​​of .

5. The method for generating fault data of a hydropower station excitation system according to claim 2, characterized in that: The input state δ i The setting range is as follows: V G i ∈[0.8,1.2]; W G i ∈[1-0.004,1+0.004]; Among them, V G i ,W G i It represents the given voltage per unit value at the generator end and the given frequency per unit value of the hydro-generator set under the i-th input state.

6. The method for generating fault data of a hydropower station excitation system according to claim 2, characterized in that: For G i (old) Conduct a parameter search as follows: The specific operation strategy of the first local search strategy is: in the i-th input state, from G i The first set of excitation parameters randomly selected in (old) Any j-th parameter x in j i In {w j(1) i ,w j(2) i ,w j(3) i …w j(d) i …w j(D) i } A new set of excitation parameters of the first group is generated by exchanging other values ​​in the set range; The specific operation strategy of the second local search strategy is: in the i-th input state, from G i Any g-th parameter x of the randomly selected l1-th set of excitation parameters in (old) g i The current value of is the same as any pth parameter x of the randomly selected l2th group of excitation parameters p i The current values ​​are mapped in their respective sets, g≠p.

7. The method for generating fault data of a hydropower station excitation system according to claim 2, characterized in that: The specific process of repeatedly executing steps S5-S6 is as follows: S51, set at G i (old) Maximum number of iterations of the search, there are a total of z states; initial state i = 1, number of iterations gen = 1; S52, executing the first local search strategy, using the evaluation function to evaluate the excitation parameter set participating in the first local search strategy If any one of the two conditions in the first evaluation rule is met, Deposit G i (new) set, and G i (old) Remove from collection If both conditions are not met, execute S53; S53, executing the second local search strategy, and using the evaluation function to evaluate the two sets of excitation parameter sets participating in the second local search strategy If any one of the two conditions in the second evaluation rule or the third evaluation rule is met, the excitation parameter set that meets the condition is stored in G i (new) set, and G i (old) Remove from collection or If any one of the two conditions in the second evaluation rule is satisfied and any one of the two conditions in the third evaluation rule is satisfied, then both sets of excitation parameter sets that satisfy the conditions are stored in G i (new) set, and G i (old) Remove from collection and Otherwise, execute S54; S54, repeatedly execute S52-S53 until the termination condition of the current state is reached and then enter the next state until all z states are executed.

8. The method for generating fault data of a hydropower station excitation system according to claim 7, characterized in that: The first evaluation rule is specifically: in, is a new set of excitation parameters generated according to the first local search strategy; V G i is the per unit value of the given voltage at the generator terminal under the i-th input state; The adjustable excitation parameter combination under the i-th input state is Fault data evaluation function during configuration; data_fault is the generation condition of excitation fault data: if If any one of the two conditions is met, Deposit G i (new) set, and G i (old) Remove from collection 9. The method for generating fault data of a hydropower station excitation system according to claim 7, characterized in that: The second evaluation rule is specifically: in, is the new set of excitation parameters l1 generated according to the second local search strategy; V G i is the per unit value of the given voltage at the generator terminal under the i-th input state; The adjustable excitation parameter combination under the i-th input state is Fault data evaluation function during configuration; data_fault is the generation condition of excitation fault data: if If any one of the two conditions is met, Deposit G i (new) set, and G i (old) Remove from collection 10. The method for generating fault data of a hydropower station excitation system according to claim 7, characterized in that: The third evaluation rule is specifically: in, is a new set of excitation parameters l2 generated according to the second local search strategy; V G i is the per unit value of the given voltage at the generator terminal under the i-th input state; The adjustable excitation parameter combination under the i-th input state is Fault data evaluation function during configuration; data_fault is the generation condition of excitation fault data: if If any one of the two conditions is met, Deposit G i (new) set, and G i (old) Remove from collection 11. A device for generating fault data of an excitation system of a hydropower station, characterized in that: include: Model building module, used to build a model for connecting a single turbine generator set in a hydropower station to the power grid; The model of a single hydroelectric generator set in a hydropower station being connected to the power grid comprises an excitation system model, a speed regulator model, a generator model, a turbine model and a power grid model; The acquisition module is used to obtain fault data by searching the adjustable excitation parameters in the excitation system model in each iteration process using the improved large-domain search algorithm under different input state conditions, and then simulating the single machine of the hydropower station's turbine generator set connected to the power grid model.

Images