Online Decision-making Method and System for Emergency Generator Tripping Based on Steady-state Control Effect Evaluation

Abstract

The present invention proposes an online decision-making method and system for emergency generator tripping based on the evaluation of the stability control effect. The method includes: collecting the inertia parameters of generators and recording the operating state data of generators after a fault occurs and after tripping; performing unit grouping, aggregation, and equivalence to calculate the convexity index of the phase trajectory, determining whether the system is unstable according to the convexity index, and when the system is unstable, forming an equivalent system power angle curve by using the curve fitting method based on the measurement data from the time after the fault to the time when the stability is judged, calculating the difference between the cumulative unbalanced energy and the remaining decelerable area to calculate the total amount of generator tripping, comparing the output of the leading generator group with the total amount of generator tripping to form a generator tripping combination that meets the requirement of the total amount of generator tripping to construct a set of potential generator tripping schemes, and performing parallel evaluation of the stability control effect for any scheme in the scheme set, and selecting the scheme with the smallest amount of generator tripping as the optimal generator tripping scheme. Based on this method, an online decision-making system for emergency generator tripping is also proposed. The present invention can realize online generation of the optimal generator tripping scheme by simultaneously giving the generator tripping position and the amount of generator tripping.

Description

Technical Field

[0001] The present invention belongs to the technical field of emergency control of new power systems, and particularly relates to an online decision-making method and system for emergency generator tripping based on the evaluation of the stability control effect. Background Art

[0002] At present, the large-scale access of renewable energy and the large-scale interconnection of AC / DC power grids have led to significant changes in the structural form and stability form of the power system, and the dynamic characteristics of the power grid have become increasingly complex. Emergency control is an important means to prevent the destruction of system stability and avoid large-scale power outages. The traditional emergency control decision-making method of "offline decision-making and online matching" designs pre-planned control strategies based on limited operating modes, and the mismatch risk increases significantly in modern power systems with increasingly complex and variable operating modes and fault conditions. Online stability control quickly determines whether the system will become unstable based on the real-time state or response information of the system, and decides control measures according to the real-time operating state of the system. In theory, it has stronger adaptability and can reduce the mismatch risk, and has gradually become the development direction of stability control.

[0003] The main task of online stability control decision-making is to judge the stability according to the real-time operating information of the power system, and select the optimal one from various stability control schemes when the system becomes unstable, so as to form a control scheme that can not only ensure the stability of the power system but also have a smaller control cost. Currently, two main methods are adopted for online stability control decision-making: one is to calculate the total amount of generator tripping by using the energy function method or the EEAC theory, and then determine the generator tripping location according to engineering experience based on the electrical distance or the power angle; the other is to use sensitivity to describe the difference in the stability control effect of different generator tripping locations, linearize the stability constraints in the emergency control decision-making model, and directly determine the generator tripping location and the amount of generator tripping by solving the optimal control model. The sensitivity can be obtained by online calculation based on super-real-time simulation or online mapping using a neural network trained offline.

[0004] The first method adopts the idea of first considering the total amount and then decomposing it, and can calculate the total amount of generator tripping without simulation, meeting the requirements of rapidity for on-line stability control. However, when calculating the remaining decelerable area, only the change in the mechanical power of the equivalent system is considered, and the influence of the change in parameters such as the equivalent electromagnetic power of the critical generator group and the system inertia time constant after generator tripping is not taken into account. Moreover, determining the generator tripping position based on the generator output, electrical distance, or power angle leading degree lacks theoretical support, and it is usually difficult to obtain the optimal generator tripping scheme. The second method is theoretically more complete in predicting the generator tripping effect based on sensitivity and then simultaneously determining the amount and position of generator tripping. However, calculating sensitivity based on super-real-time simulation has high requirements for simulation calculation performance, and it is difficult to balance the model detail level and sensitivity accuracy. In addition, the data-driven method for obtaining sensitivity has problems with generalization, and the sensitivity calculation error may increase due to changes in the operation mode or instability mode, thereby affecting the generator tripping control effect. Therefore, how to quickly and accurately evaluate the stability control effect of different generator tripping schemes and select the optimal one from multiple generator tripping schemes is an urgent problem to be solved in on-line stability control decision-making. Summary of the Invention

[0005] To solve the above technical problems, the present invention proposes an on-line decision-making method and system for emergency generator tripping based on stability control effect evaluation, which can simultaneously give the generator tripping position and the amount of generator tripping and generate an optimal generator tripping scheme online.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] The on-line decision-making method for emergency generator tripping based on stability control effect evaluation includes the following steps:

[0008] Collect the generator inertia parameters, record the generator operating state data after the fault occurs and the generator operating state data after the fault is cleared; based on the generator operating state data matrix after the fault is cleared, use the clustering algorithm to cluster the power angle curves to divide the generators into an advanced generator group and a lagging generator group; equalize the advanced generator group and the lagging generator group to a single-machine infinite system, calculate the concavity and convexity index of the angular velocity-power angle phase trajectory of the equivalent system, and judge whether the system is unstable according to whether the concavity and convexity index is greater than zero;

[0009] If the system is unstable, calculate the accelerating area during the generator fault, then calculate the decelerated area before the stability control using the equivalent system power angle curve from after the fault is cleared to the moment of stability judgment, and calculate the remaining decelerable area by curve fitting extrapolation to form an equivalent system power angle characteristic prediction curve; calculate the theoretical amount of generator tripping according to the principle that the accelerating area is not less than the sum of the total decelerated areas, and finally form a set of potential generator tripping schemes based on the principle that the sum of the real-time outputs of the tripped generators is greater than the theoretical amount of generator tripping;

[0010] Calculate the change in unbalanced energy for the selected generator tripping scheme, then calculate the remaining decelerable area of the equivalent system of the remaining units after generator tripping and the change in the remaining decelerable area caused by generator tripping, further evaluate the overall stability control effect of generator tripping to determine whether the generator tripping scheme can stabilize the system, and finally compare all generator tripping schemes that can stabilize the system, and select the scheme with the smallest amount of tripped generators as the optimal solution for the generator tripping scheme.

[0011] An online decision-making system for emergency generator tripping based on the evaluation of stability control effect, including a judgment module, a calculation module and a decision-making module;

[0012] The judgment module is used to collect generator inertia parameters, record the working state data of the generator after the fault occurs and the working state data of the generator after the fault is removed; based on the working state data matrix of the generator after the fault is removed, the clustering algorithm is used to cluster the power angle curves to divide the units into an advanced machine group and a lagging machine group; the advanced machine group and the lagging machine group are equivalent to a single-machine infinite bus system, calculate the concavity and convexity index of the angular velocity-power angle phase trajectory of the equivalent system, and judge whether the system is unstable according to whether the concavity and convexity index is greater than zero;

[0013] The calculation module is used to calculate the accelerating area during the generator fault if the system is unstable, then calculate the decelerated area before the stability control using the power angle curve of the equivalent system from the time when the fault is removed to the time when the stability is judged, and calculate the remaining decelerable area by extrapolating through curve fitting to form a predicted curve of the power angle characteristic of the equivalent system; calculate the theoretical amount of tripped generators according to the principle that the accelerating area is not less than the sum of the total decelerated areas, and finally form a set of potential generator tripping schemes based on the principle that the sum of the real-time output of the tripped units is greater than the theoretical amount of tripped generators;

[0014] The decision-making module is used to calculate the change in unbalanced energy for the selected generator tripping scheme, then calculate the remaining decelerable area of the equivalent system of the remaining units after generator tripping and the change in the remaining decelerable area caused by generator tripping, further evaluate the overall stability control effect of generator tripping to determine whether the generator tripping scheme can stabilize the system, and finally compare all generator tripping schemes that can stabilize the system, and select the scheme with the smallest amount of tripped generators as the optimal solution for the generator tripping scheme.

[0015] The effects provided in the invention content are only the effects of the embodiments, rather than all the effects of the invention. One of the above technical solutions has the following advantages or beneficial effects:

[0016] The present invention proposes an online decision-making method and system for emergency generator tripping based on the evaluation of the stability control effect. The method includes the following steps: collecting the generator inertia parameters, recording the generator operating state data after the fault occurs and the generator operating state data after the fault is cleared; based on the generator operating state data matrix after the fault is cleared, using a clustering algorithm to cluster the power angle curves to divide the units into an advanced machine group and a lagging machine group; equalizing the advanced machine group and the lagging machine group to a single-machine infinite bus system, calculating the concavity and convexity index of the angular velocity-power angle phase trajectory of the equivalent system, and judging whether the system is unstable according to whether the concavity and convexity index is greater than zero; if the system is unstable, calculate the accelerating area during the generator fault, then calculate the decelerating area before the stability control using the power angle curve of the equivalent system from the time after the fault is cleared to the time when the stability is judged, and calculate the remaining decelerating area by extrapolating through curve fitting to form the power angle characteristic prediction curve of the equivalent system; calculate the theoretical generator tripping amount according to the principle that the accelerating area is not less than the sum of the total decelerating areas, and finally form a set of potential generator tripping schemes based on the principle that the sum of the real-time output of the tripped units is greater than the theoretical generator tripping amount; calculate the change in the unbalanced energy of the selected generator tripping scheme, then calculate the remaining decelerating area of the equivalent system of the remaining units after the generator tripping and the change in the remaining decelerating area caused by the generator tripping, further evaluate the overall stability control effect of the generator tripping to judge whether the generator tripping scheme can stabilize the system, and finally compare all the generator tripping schemes that can stabilize the system, and select the scheme with the smallest generator tripping amount as the optimal solution for the generator tripping scheme. Based on the online decision-making method for emergency generator tripping based on the evaluation of the stability control effect, an online decision-making system for emergency generator tripping based on the evaluation of the stability control effect is also proposed. The present invention can accurately and quickly evaluate the stability control effects of different generator tripping schemes based on the real-time state information of the power grid, further optimize the comprehensive optimal scheme by comprehensively comparing the control effects and control costs of the generator tripping schemes, and at the same time give the generator tripping location and the generator tripping amount to generate the optimal generator tripping scheme online. Description of the Drawings

[0017] Figure 1 It is the flow chart of the online decision-making method for emergency generator tripping based on the evaluation of the stability control effect in Embodiment 1 of the present invention;

[0018] Figure 2 It is the flow chart of the fast stability judgment based on the concavity and convexity of the phase trajectory in Embodiment 1 of the present invention;

[0019] Figure 3 It is the flow chart of the unit grouping based on the K-means clustering method in Embodiment 1 of the present invention;

[0020] Figure 4 It is the flow chart of generating a set of potential generator tripping schemes based on the equality of the accelerating and decelerating areas in Embodiment 1 of the present invention;

[0021] Figure 5 It is the flow chart of optimizing the generator tripping scheme based on the fast evaluation of the stability control effect of the generator tripping in Embodiment 1 of the present invention;

[0022] Figure 6 It is the topology structure diagram of the example system;

[0023] Figure 7 It is the continuous calculation result of the convexity and concavity index of the phase trajectory of the application example;

[0024] Figure 8 It is the fitting result of the power angle characteristic curve of the system after the fault is removed in the application example;

[0025] Figure 9 It is the system diagram of the online decision-making method for emergency generator tripping based on the evaluation of the stability control effect proposed in Embodiment 2 of the present invention. Specific implementation manners

[0026] To clearly illustrate the technical features of the present solution, the present invention will be described in detail below through specific implementation manners and in conjunction with its accompanying drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the present invention.

[0027] Embodiment 1

[0028] Embodiment 1 of the present invention proposes a flowchart of an online decision-making method for emergency generator tripping based on the evaluation of the stability control effect, which is used to solve the problem of how to select the optimal generator tripping scheme from multiple generator tripping schemes in the prior art.

[0029] In step 1: Collect the generator inertia parameters, record the generator operating state data after the fault occurs and the generator operating state data after the fault is removed; Based on the generator operating state data matrix after the fault is removed, use the clustering algorithm to cluster the power angle curves to divide the units into an advanced machine group and a lagging machine group; Equalize the advanced machine group and the lagging machine group to a single-machine infinite bus system, calculate the convexity and concavity index of the angular velocity-power angle phase trajectory of the equivalent system, and determine whether the system is unstable according to whether the convexity and concavity index is greater than zero.

[0030] Taking the generator inertia parameters and the power angle, angular velocity (the difference between the actual angular velocity per unit value and the synchronous angular velocity per unit value), electromagnetic power, and mechanical power measurement data during and after the fault as inputs, accumulate the measurement data, and based on the accumulated measurement data, perform unit grouping, equalize to a single-machine infinite bus system, calculate the convexity and concavity index of the angular velocity-power angle phase trajectory of the equivalent system, and finally determine whether the system is unstable according to whether the convexity and concavity index is greater than zero. Figure 2 It is the fast stability judgment flowchart based on the convexity and concavity of the phase trajectory proposed in Embodiment 1 of the present invention.

[0031] In step 1.1, based on the measurement information of devices such as PMU and WAMS, record the generator parameters and the operating state data during the fault, including: Collect the generator inertia parameter data in operation to form a generator inertia vector:

[0032] M0 = [M1 M2 … Mi ; (1)

[0033] Among them, M0 is the moment of inertia of all operating generators, and M i is the moment of inertia of the i-th generator;

[0034] After detecting the occurrence of a fault, record the working state data of each generator collected by devices such as PMU and WAMS at the current moment, including the generator power angle, angular velocity, mechanical power, and electromagnetic power, and store them in the variable matrices of the fault data buffer:

[0035]

[0036] Among them, δ bf is the recording matrix of the generator power angle operation data value; represents the power angle of the i-th generator at the k0-th moment after the fault; ω bf is the recording matrix of the generator angular velocity operation data value; represents the generator angular velocity of the i-th generator at the k0-th moment after the fault; is the recording matrix of the generator electromagnetic power operation data value; represents the generator electromagnetic power of the i-th generator at the k0-th moment after the fault; is the recording matrix of the generator mechanical power operation data value; represents the generator mechanical power of the i-th generator at the k0-th moment after the fault; i = 1, 2,..., n, k0 = 1, 2,...t bf , n represents the number of generators, and t bf represents the current sampling time number of the fault data record. Sampling is performed cyclically, and each sampled data is stored in a new row added to the matrix of the fault data buffer.

[0037] Step 1.2: Record the working state data of the generator after the fault is removed based on the measurement information of devices such as PMU and WAMS, including: After detecting the removal of the fault, record the working state data of each generator collected by devices such as PMU and WAMS at the current moment, including the generator power angle, angular velocity, mechanical power, and electromagnetic power, and store them in the variable matrices of the post-fault-removal data buffer:

[0038]

[0039] Among them, δ af represents the recording matrix of the generator power angle operation data value after the fault is removed; represents the power angle of the i-th generator at the k-th moment after the fault is removed; ω af represents the recording matrix of the generator angular velocity operation data value after the fault is removed; It represents the angular velocity of the ith generator at the kth moment after the fault is cleared; It represents the matrix for recording the operating data values of the generator's electromagnetic power after the fault is cleared; It represents the electromagnetic power of the ith generator at the kth moment after the fault is cleared; It represents the matrix for recording the operating data values of the generator's mechanical power after the fault is cleared; It represents the mechanical power of the ith generator at the kth moment after the fault is cleared; k = 1, 2,... t af , t af It represents the current sampling time count of the data record after the fault is cleared.

[0040] Perform cyclic sampling, and store each sampled data into a new row added to the matrix of the data buffer area after the fault is cleared;

[0041] Detect the number of stored data t in the matrix of the data buffer area after the fault is cleared af Whether it is greater than the minimum number of information N for calculation min (N min Generally take 10), if so, enter step 1.3, if not, continue sampling.

[0042] step 1.3: Based on the matrix δ of the data values after excision af , use the K-means clustering method to cluster the power angle curves of all generators, and divide the units into the leading machine group Ω S and the lagging machine group Ω A .

[0043] Figure 3 This is the flow chart of unit grouping based on the K-means clustering method proposed in Embodiment 1 of the present invention.

[0044] Taking a column of data af in the matrix δ of the data values after the fault is cleared (k = 1, 2,... t af ) as a power angle curve δ i Substitute it into the following formula (4) to calculate the Euclidean distance between any two power angle curves:

[0045]

[0046] where, D ij is the Euclidean distance between the power angle curve of the ith generator and the power angle curve of the jth generator;

[0047] Select the two curves with the largest Euclidean distance as the initial clustering centers C1 and C2 of the first machine group and the second machine group, and write the corresponding generator numbers into the index sets Ω1 and Ω2 of the first machine group and the second machine group;

[0048] From δ af Arbitrarily select one δ from the remaining curves m and calculate the Euclidean distances between δ m and the cluster centers C1 and C2:

[0049]

[0050] where D m,1 , D m,2 are the Euclidean distances between the m-th power angle curve and the cluster centers C1 and C2 respectively; C 1,k , C 2,k are the values of the cluster centers C1 and C2 at the k-th moment respectively

[0051] Compare the magnitudes of D m,1 and D m,2 If D m,1 > D m,2 , then δ m is classified into the second machine group and the curve column index is merged into the index set Ω2; otherwise, δ m is classified into the first machine group and the curve column index is merged into the index set Ω1; update the aggregation centers of the corresponding machine groups:

[0052]

[0053] where are the numbers of generators in the first machine group and the second machine group respectively;

[0054] If the calculation of the power angle curves in δ af is completed, then calculate the mean difference between the cluster centers C1 and C2 of the machine groups, and distinguish the leading machine group Ω S and the lagging machine group Ω A according to the sign of the difference:

[0055]

[0056] δ C1 , δ C2 are the average power angles of the first machine group and the second machine group respectively;

[0057] If δ C1 > δ C2 , then record the first machine group and the second machine group as the leading machine group and the lagging machine group respectively, i.e., Ω S = Ω1, Ω A = Ω2; otherwise, Ω S = Ω2, Ω A = Ω1.

[0058] Step 1.4: According to the clustering result, the leading machine group Ω Sand the lagging generator group Ω A Respectively adopt the equivalent values of their respective inertia centers, and further use the relative power angle difference to describe the system swing process to form a single-machine infinite system, and record the matrix δ based on the data values after fault removal af , ω af , Calculate the power angle, angular velocity, electromagnetic power, and mechanical power trajectories of the equivalent system. Including:

[0059] Use the weighted average of the power angles and angular velocities of the generators within the group to represent the power angle and angular velocity of the inertia center, use the sum of the electromagnetic powers and the sum of the mechanical powers of each generator to represent the total electromagnetic power and mechanical power of the inertia center, and use formula (8) to calculate the values of the power angle, angular velocity, electromagnetic power, and mechanical power trajectories of the inertia center at each moment:

[0060]

[0061] Among them, M S is the equivalent moment of inertia of the leading generator group Ω S ; M A is the equivalent moment of inertia of the lagging generator group Ω A ; δ S,k , ω S,k , P eS,k , P mS,k are respectively the value of the equivalent power angle of Ω S at the k-th moment, the value of the angular velocity at the k-th moment, the value of the electromagnetic power at the k-th moment, and the value of the mechanical power at the k-th moment; δ A,k , ω A,k , P eA,k , P mA,k are respectively the value of the equivalent power angle of Ω A at the k-th moment, the value of the angular velocity at the k-th moment, the value of the electromagnetic power at the k-th moment, and the value of the mechanical power at the k-th moment;

[0062] Use the difference of the equivalent parameters of the two groups to characterize the system state and equivalent it to a single-machine infinite system, and use formula (9) to calculate the values of the power angle, angular velocity, electromagnetic power, and mechanical power trajectories of the equivalent system at each moment;

[0063]

[0064] M T =M S +M A , k = 1, 2...t af ; (9)

[0065] Among them, M T is the moment of inertia of the equivalent system; δ k is the value of the power angle of the equivalent system at the k-th moment, ωk is the value of the angular velocity at the k-th moment, and P e,k is the value of the electromagnetic power at the k-th moment, and P m,k is the value of the mechanical power at the k-th moment;

[0066] After the fault is cleared, the generator operating state data δ af , ω af , are respectively converted into the form represented by the equivalent system state trajectory after the fault is cleared by using equations (8) - (9) to form;

[0067]

[0068] In step 1.5, according to the equivalent single-machine infinite-bus system state trajectory, calculate the concavity and convexity of the current power angle - angular velocity phase trajectory, and judge whether the system has power angle instability through the concavity and convexity of this phase trajectory, including:

[0069] Take the angular velocity and power at the last two moments in the state trajectory, and substitute them into formula (11) to calculate the first derivative of the power angle - angular velocity phase trajectory at the current moment and the previous moment:

[0070]

[0071] where, d k is the first derivative of the power angle - angular velocity phase trajectory at the k-th moment;

[0072] Based on the first derivative of the power angle - angular velocity phase trajectory and the values of the power angle of the equivalent system at the last two moments, substitute them into formula (12) to calculate the second derivative of the power angle - angular velocity phase trajectory of the equivalent system at the current moment:

[0073]

[0074] Based on the second derivative of the power angle - angular velocity phase trajectory at the current moment and the angular velocity, calculate the trajectory concavity and convexity index according to formula (13):

[0075]

[0076] where, μ taf is the concavity and convexity of the power angle - angular velocity phase trajectory at the current t af moment;

[0077] Judge whether the system has power angle instability. If μ taf ≥0, it is determined that the system enters the power angle instability state, and enter step2 for power angle stability emergency control.

[0078] Calculate If The system is determined to be in a stable state all the time, clear all cached data and return to step 1.1, waiting for the next fault; if Only judge the stability of this fault, return to step 1.2 to collect data at the next moment, and start the next round of stability judgment.

[0079] In step 2: If the system is unstable, calculate the accelerating area during the generator fault, then calculate the decelerating area before stable control using the equivalent system power angle curve from the fault removal to the stability judgment moment, and calculate the remaining decelerating area by extrapolating through curve fitting to form an equivalent system power angle characteristic prediction curve; calculate the theoretical generator cutting amount according to the principle that the sum of the accelerating area and the total decelerating area is equal, and finally form a set of potential generator cutting schemes based on the principle that the sum of the real-time output of the cut-off units is greater than the theoretical generator cutting amount.

[0080] First, calculate the accelerating area during the fault using the recorded fault data based on the grouping scheme given by the stability judgment, then calculate the decelerating area before stable control using the equivalent system power angle curve from the fault removal to the stability judgment moment, and calculate the remaining decelerating area by extrapolating through curve fitting to form an equivalent system power angle characteristic prediction curve. Further, calculate the theoretical generator cutting amount according to the principle that the sum of the accelerating area and the total decelerating area is equal, and finally form a set of potential generator cutting schemes based on the principle that the sum of the real-time output of the cut-off units is greater than the theoretical generator cutting amount. Figure 4 It is the flowchart for generating a set of potential generator cutting schemes based on the equality of accelerating and decelerating areas proposed in Embodiment 1 of the present invention;

[0081] step 2.1: Based on the leading machine group set Ω S and the lagging machine group set Ω A determined by the stability judgment grouping, convert the variable matrices δ af , ω af , recorded in the fault data buffer into the form represented by the equivalent system state trajectory after the fault removal respectively using Equations (8)-(9).

[0082] Taking the power angle as the independent variable and the mechanical power and electromagnetic power as the dependent variables, substitute them into the following formula and use the rectangle method to calculate the accelerating area during the fault:

[0083]

[0084] In the formula, S A is the accelerating area of the equivalent system during the fault;

[0085] step 2.2: Calculate the decelerating area before stable control using the equivalent system power angle curve from the fault removal to the stability judgment moment, and form an equivalent system power angle characteristic prediction curve by extrapolating through curve fitting. Calculate the remaining decelerating area based on the extrapolated curve, including:

[0086] The deceleration area that has occurred from the time of fault clearing to the time of stability judgment:

[0087]

[0088] S D1 is the deceleration area of the equivalent system from the time of fault clearing to the time of stability judgment;

[0089] The deceleration area of the equivalent system from the time of fault clearing to the time of stability judgment

[0090] ΔS0 = S A -S D1 ; (16)

[0091] Assume that the power angle curve and mechanical power curve of the equivalent system conform to the following form:

[0092] P e = α + βsin(γδ + C)

[0093] P m = a + bδ; (17)

[0094] Respectively, taking P e = P e af , δ = δ af and as the mapping sequence, using the non - linear and least - squares parameter identification method to calculate the first parameter to be identified α, the second parameter to be identified β, the third parameter to be identified γ, the fourth parameter to be identified C, as well as the fifth parameter to be identified a and the sixth parameter to be identified b;

[0095] In interval, equally divide it into a preset number of power angle points to form a power angle prediction sequence In this application, the preset number is 100. Substitute it into formula (16) to calculate the corresponding electromagnetic power and mechanical power prediction sequences Search forward from the front to find the sequence elements that meet the conditions to form the electromagnetic and mechanical power prediction sequences corresponding to the decelerable area and its corresponding power angle prediction sequence

[0096] Taking the predicted power angle sequence as the independent variable, the selected predicted mechanical power and electromagnetic power as the dependent variables, substitute them into formula (18) and use the rectangle method to calculate the remaining decelerable area:

[0097]

[0098] Among them, S D2It is the remaining deceleratable area of the equivalent system from the stability judgment moment to the non-return point.

[0099] Determine the shortage of the remaining deceleratable area based on the principle that the accelerating area is not less than the sum of the total decelerating areas:

[0100] S D3 =λ D (S A -S D1 -S D2 ) ; (19)

[0101] In the formula, S D3 is the shortage of the decelerating area required to maintain system stability, and λ D is the reliability coefficient that can still ensure the stability control effect considering calculation errors; usually take 1.05 - 1.1;

[0102] Assume that the potential increase in the decelerating area caused by the generator tripping control is rectangular, and calculate the equivalent generator tripping amount ΔP m :

[0103]

[0104] Among them, are the power angles of the equivalent system at the stability judgment moment and the power angle of the farthest point respectively;

[0105] Calculate the theoretical generator tripping amount ΔP M :

[0106]

[0107] The process of forming a set of potential generator tripping schemes based on the principle that the sum of the real-time outputs of the generators to be tripped is greater than the theoretical generator tripping amount is as follows: Read the set of generator indices Ω S , set the maximum allowable number of generators to be tripped N cmax ; Set the current allowable number of generators to be tripped as N c to 1; Form all subsets of the set Ω c containing only N S elements; Calculate the sum of the real-time outputs of the generators in each subset. If the sum of the real-time outputs is greater than the theoretical generator tripping amount, add this subset to the set of potential generator tripping schemes; Until the number of generators to be tripped N c increases to the maximum allowable number of generators to be tripped N cmax , complete the set of generator tripping schemes

[0108] In Step 3: Calculate the change in unbalanced energy of the selected generator tripping scheme, then calculate the remaining decelerating area of the equivalent system of the remaining units after generator tripping and the change in the remaining decelerating area caused by generator tripping, further evaluate the overall stability control effect of generator tripping to determine whether the generator tripping scheme can stabilize the system, and finally compare all generator tripping schemes that can stabilize the system and select the scheme with the smallest amount of tripped generators as the optimal solution for the generator tripping scheme. Figure 5 It is the flowchart for optimizing the generator tripping scheme based on the rapid evaluation of the stability control effect proposed in Embodiment 1 of the present invention;

[0109] In Step 3.1, arbitrarily select a generator tripping scheme Ω from the set of generator tripping schemes c , and calculate the change in unbalanced energy caused by the generator tripping scheme, including:

[0110] In the leading generator group index set Ω S , remove the generator numbers corresponding to the generator tripping scheme Ω c to form the remaining leading generator group index set Ω s '

[0111] Determine the remaining leading generator group index set Ω' s ; According to Ω' s and Ω A calculate the state trajectories δ c , ω bf' of the equivalent system during the fault period after removing the unit set Ω bf' ,

[0112] Substitute the power angle as the independent variable, and the mechanical power and electromagnetic power as the dependent variables into formula (22) to calculate the accelerating area of the equivalent system after removing the tripped units during the fault period:

[0113]

[0114] where S' A is the accelerating area of the equivalent system after removing the tripped units during the fault period;

[0115] Determine the state trajectories δ c , ω af' of the equivalent system during the fault period after removing the unit set Ω af' ,

[0116] Calculate the decelerating area that has occurred in the equivalent system from the time of fault removal to the time of stability judgment:

[0117]

[0118] where S' D1 is the decelerating area that has occurred in the equivalent system from the time of fault removal to the time of stability judgment;

[0119] Compare the difference between the equivalent system's accelerating area and the decelerated area before and after the generator set is removed, and use equation (24) to calculate the change in unbalanced energy caused by generator tripping control:

[0120] ΔS1 = (S A - S D1 ) - (S' A - S' D1 ) ; (24)

[0121] In the formula, ΔS1 is the change in the cumulative unbalanced energy before and after generator tripping control;

[0122] Step3.2: Based on the state trajectories δ c , ω af' , af' , of the equivalent system during the fault period after removing the generator set from the set Ω, perform power angle characteristic fitting and extrapolation calculations to obtain the remaining decelerating area;

[0123] Assume that the power angle curve and mechanical power curve of the equivalent system after removing the tripped generator are as shown in equation (17). Using and as the mapping sequences, use the non - linear and least - squares parameter identification method to calculate the first parameter to be identified α, the second parameter to be identified β, the third parameter to be identified γ, the fourth parameter to be identified C, the fifth parameter to be identified a, and the sixth parameter to be identified b;

[0124] In the interval, equally divide it into a preset number of power angle points to form a power angle prediction sequence Substitute it into formula (23) to calculate the corresponding electromagnetic power and mechanical power prediction sequences In this application, the preset number is 100. Search forward to find the sequence elements that meet the condition to form the electromagnetic and mechanical power prediction sequences corresponding to the decelerating area and its corresponding power angle prediction sequence

[0125] Using the predicted power angle sequence as the independent variable, and the selected predicted mechanical power and electromagnetic power as the dependent variables, substitute them into formula (25) and use the rectangle method to calculate the remaining decelerating area:

[0126]

[0127] In the formula, S' D2 is the remaining decelerating area of the equivalent system after removing the tripped generator from the critical stability moment to the non - return point.

[0128] Step 3.3: Based on the fact that generator tripping leads to the transfer of electromagnetic power, use rectangular approximation to calculate the change in the remaining decelerable area, including: model the un-tripped generators using the voltage source model, model the tripped generators using the change in injected current, form a circuit model with the change in injected power as the input, and form a nodal admittance matrix according to the circuit interconnection relationship;

[0129] Use the nodal voltage method to calculate the change in the injected current of the un-tripped units caused by generator tripping

[0130]

[0131] are the injected currents at the r-th and s-th nodes of the tripped unit before tripping, respectively;

[0132] are the change in the injected current of the i-th generator and the change in the injected current of the j-th generator of the un-tripped units after the tripping of the tripped unit; y ir and y is are the associated admittance elements of the r-th and s-th nodes of the i-th generator of the tripped unit and the un-tripped units, respectively; y jr and y js are the associated admittance elements of the r-th and s-th nodes of the j-th generator of the tripped unit and the un-tripped units, respectively;

[0133] Multiply the change in the injected current of the un-tripped units by the corresponding nodal voltage to obtain the change in the injected power of the un-tripped units:

[0134]

[0135] In the formula, are the internal electromotive force, the change in the injected power, and the change in the injected current of the un-tripped units, respectively;

[0136] Use formula (28) to calculate the total change in the electromagnetic power of the equivalent system after generator tripping:

[0137]

[0138] In the formula, is the change in the electromagnetic power of the equivalent system caused by generator tripping, M C is the sum of the equivalent inertias of the tripped units;

[0139] Use formula (29) to calculate the change in the decelerable area caused by the change in the electromagnetic power of the equivalent system

[0140]

[0141] Among them, S' D3It is the change amount of the decelerable area caused by the change amount of the electromagnetic power of the equivalent system;

[0142] Sum the change amount of the unbalanced area, the remaining decelerable area, and the change value of the decelerable area to obtain the comprehensive decelerable area that the generator tripping control can provide:

[0143] S dec = ΔS1 + S' D2 + S' D3 ; (30)

[0144] Compare the comprehensive decelerable area S dec with the initial unbalanced energy ΔS0: If S dec > ΔS0, then generator tripping can stabilize the system. Add the generator set included in this generator tripping scheme to the stable generator tripping scheme set Ω stable and record the corresponding generator tripping amount;

[0145] Compare the total generator tripping power of all generator tripping schemes in the stable generator tripping scheme set Ω stable and select the scheme with the smallest total generator tripping amount as the optimal generator tripping scheme.

[0146] The emergency generator tripping online decision-making method proposed in Embodiment 1 of the present invention can accurately and quickly evaluate the stable control effects of different generator tripping schemes based on the real-time state information of the power grid, further optimize the comprehensive optimal scheme by comprehensively comparing the control effects and control costs of the generator tripping schemes, and simultaneously give the generator tripping location and generator tripping amount to online generate the optimal generator tripping scheme.

[0147] Taking the IEEE 39-bus system as an example to verify the effectiveness of the method proposed in this patent. The IEEE 39-bus system contains 39 buses, including 3 generator buses, 6 substation buses and 30 load buses, 10 generators, and a total of 46 branches. The main generators are distributed in different regions and are connected by 46 transmission lines to form a complex power network. Figure 6 It is the topological structure diagram of the example system.

[0148] The fault scenario is to set a three-phase short-circuit fault at the head of line 23-24 at 1 s, and the fault duration is 0.22 s. Then the relay protection operates to trip the fault line. The application steps of the patented method are as follows:

[0149] First, when the system detects a fault, first record the measurement data during the fault according to step 1 to form the variable matrices of each buffer area of the fault data, and the data time step is 0.01 s;

[0150] Further detect the fault clearing, record the measurement data after the fault clearing, and form the variable matrices of each buffer area of the data after the fault clearing. The data time step is still 0.01 s;

[0151] After the fault is cleared, when 10 pieces of data are accumulated in the data buffer, the system enters the stability judgment process. Based on the measured data, clustering and equivalent calculation are carried out according to Equations (8)-(9) to obtain the state trajectory of the single-machine infinite-bus system. According to this trajectory, the concavity and convexity of the current power angle-angular velocity phase trajectory are calculated;

[0152] As the accumulated data increases, the concavity and convexity index of the phase trajectory changes with time; Figure 7 For the application example, the continuous calculation results of the concavity and convexity index of the phase trajectory are shown;

[0153] At 0.13 s after the fault is cleared (the absolute time is 1.35 s, that is, after 13 data points are accumulated after the fault is cleared), the concavity and convexity curve of the trajectory crosses the 0 axis and is greater than zero. Then, it is determined that the system is unstable and the emergency control is started;

[0154] Then, according to the method in Step 2, the equivalent system state trajectory during the fault is formed, and the accelerating area of the system is calculated to be 4.05;

[0155] The equivalent system state trajectory after the fault is cleared is formed, and the decelerating area is calculated to be 0.89;

[0156] The accelerating area is subtracted by the decelerating area to obtain the initial unbalanced energy of 3.16;

[0157] Based on the equivalent system state trajectory after the fault is cleared, the power angle curve of the system is obtained by curve fitting, and the equivalent system power angle characteristic prediction curve after the stability judgment moment is extrapolated. Figure 8 For the application example, the curve fitting result of the power angle characteristic curve of the system after the fault is cleared is shown;

[0158] Based on the predicted results of the electromagnetic power and mechanical power from the power angle curve, the remaining decelerating area is calculated to be 0.72;

[0159] The decelerating area deficit is obtained by subtracting the remaining decelerating area from the unbalanced energy, which is 2.44. Substituting it into Equations (19)-(20), the theoretical generator tripping amount can be calculated to be 858 MW;

[0160] Based on the principle of being greater than 90% of the theoretical generator tripping amount, the maximum number of tripped generators is one. Combining with the real-time output of the leading group of units, the potential generator tripping set is formed as {{G33}, {G34}, {G35}, {G36}}, and the real-time power of each unit is 1052 MW, 853 MW, 786 MW, and 878 MW.

[0161] Finally, following the method in Step 3, after removing the variables related to G33, G34, G35, and G36 respectively, the data recorded during the fault and the data recorded after the fault removal are re-clustered, equivalent, and curve-fitted and predicted. The change in unbalanced energy, the remaining decelerable area, and the change in decelerable area after removing different units are calculated, and the sum of the three is obtained as the comprehensive decelerable area. The calculation results are shown in Table 1.

[0162] Table 1 Comprehensive decelerable area and other indicators for different generator tripping schemes

[0163]

[0164]

[0165] Compare the comprehensive decelerable area of each generator tripping scheme in the above potential generator tripping scheme set with the initial unbalanced energy, and select the generator tripping schemes with a comprehensive decelerable area greater than the initial unbalanced energy to form a stable generator tripping scheme set. Based on the calculation results in Table 1, only the {G36} generator tripping scheme enters the stable generator tripping scheme set; select the scheme with the smallest total amount of tripped generators in the stable generator tripping scheme set as the optimal generator tripping scheme. Since there is only one element in the scheme set, {G36} is the optimal generator tripping scheme.

[0166] Embodiment 2

[0167] Based on the emergency generator tripping online decision-making method for stability control effect evaluation, an emergency generator tripping online decision-making system based on stability control effect evaluation is also proposed. Figure 9 The system diagram of the emergency generator tripping online decision-making method proposed in Embodiment 2 of the present invention. The system includes: a judgment module, a calculation module, and a decision module.

[0168] The judgment module is used to collect the generator inertia parameters, record the generator operating state data after the fault occurs and the generator operating state data after the fault removal; based on the generator operating state data matrix after the fault removal, the clustering algorithm is used to cluster the power angle curves to divide the units into a leading machine group and a lagging machine group; the leading machine group and the lagging machine group are equivalent to a single-machine infinite bus system, and the convexity index of the angular velocity-power angle phase trajectory of the equivalent system is calculated. Whether the system is unstable is judged according to whether the convexity index is greater than zero.

[0169] The calculation module is used to calculate the accelerating area of the generator during the fault if the system is unstable, then calculate the decelerated area before the stability control using the power angle curve of the equivalent system from the fault removal to the moment of stability judgment, and calculate the remaining decelerable area by curve fitting extrapolation to form the power angle characteristic prediction curve of the equivalent system; calculate the theoretical amount of tripped generators according to the principle that the accelerating area is not less than the sum of the total decelerated areas, and finally form a potential generator tripping scheme set based on the principle that the sum of the real-time outputs of the tripped units is greater than the theoretical amount of tripped generators.

[0170] The decision-making module is used to calculate the change in unbalanced energy of the selected generator tripping scheme, and then calculate the remaining available deceleration area of the equivalent system of the remaining units after generator tripping and the change in the remaining available deceleration area caused by generator tripping, further evaluate the overall stability control effect of generator tripping to determine whether the generator tripping scheme can stabilize the system, and finally compare all generator tripping schemes that can stabilize the system, and select the scheme with the smallest amount of generator tripping as the optimal solution for the generator tripping scheme.

[0171] In the judgment module, the specific process of collecting the generator inertia parameters is as follows: collect the inertia parameter data of the operating generators to form a generator inertia vector:

[0172] M0 = [M1 M2 … M i ; (1)

[0173] where M0 is the moment of inertia of all operating generators, and M i is the moment of inertia of the i-th generator;

[0174] Record the generator operating state data after the fault as:

[0175]

[0176] where δ bf is the recording matrix of the generator power angle operating data value; represents the power angle of the i-th generator at the k0-th moment after the fault; ω bf is the recording matrix of the generator angular velocity operating data value; represents the generator angular velocity of the i-th generator at the k0-th moment after the fault; is the recording matrix of the generator electromagnetic power operating data value; represents the generator electromagnetic power of the i-th generator at the k0-th moment after the fault; is the recording matrix of the generator mechanical power operating data value; represents the generator mechanical power of the i-th generator at the k0-th moment after the fault; i = 1, 2,..., n, k0 = 1, 2, … t bf , n represents the number of generators, and t bf represents the current sampling moment number of the fault data record;

[0177] The generator operating state data after the fault removal is:

[0178]

[0179] where δ af represents the recording matrix of the generator power angle operating data value after the fault removal; represents the power angle of the i-th generator at the k-th moment after the fault removal; ω afIt represents the record matrix of the generator angular velocity operation data values after fault removal; It represents the generator angular velocity of the i-th generator at the k-th moment after fault removal; It represents the record matrix of the generator electromagnetic power operation data values after fault removal; It represents the generator electromagnetic power of the i-th generator at the k-th moment after fault removal; It represents the record matrix of the generator mechanical power operation data values after fault removal; It represents the generator mechanical power of the i-th generator at the k-th moment after fault removal; k = 1, 2,..., t af , t af It represents the current sampling time number of the data record after fault clearing.

[0180] Using the clustering algorithm to cluster the power angle curves, the units are divided into the leading machine group and the lagging machine group, specifically including:

[0181] Taking a column of data in the data value record matrix δ af after fault removal (k = 1, 2,..., t af ) as a power angle curve δ i and substituting it into the following formula (4) to calculate the Euclidean distance between any two power angle curves:

[0182]

[0183] where D ij is the Euclidean distance between the power angle curve of the i-th generator and the power angle curve of the j-th generator;

[0184] Select the two curves with the largest Euclidean distance as the initial clustering centers C1 and C2 of the first machine group and the second machine group, and write the corresponding generator numbers into the index sets Ω1 and Ω2 of the first machine group and the second machine group;

[0185] Randomly select a δ af from the remaining curves of δ m , and calculate the Euclidean distances between δ m and the clustering centers C1 and C2:

[0186]

[0187] where D m,1 , D m,2 are respectively the Euclidean distances between the m-th power angle curve and the clustering centers C1 and C2; C 1,k , C 2,k are respectively the values of the clustering centers C1 and C2 at the k-th moment

[0188] Compare D m,1and D m,2 size relationship, if D m,1 > D m,2 , then δ m is classified into the second machine group and the curve column index is merged into the index set Ω2, otherwise δ m is classified into the first machine group and the curve column index is merged into the index set Ω1; update the aggregation center of the corresponding machine group:

[0189]

[0190] wherein, are the numbers of generator sets in the first machine group and the second machine group respectively;

[0191] If the power angle curves in δ af are calculated, then calculate the average value difference between the clustering centers C1 and C2 of the machine groups, and distinguish the leading machine group Ω S and the lagging machine group Ω A according to the positive and negative of the difference:

[0192]

[0193] δ C1 , δ C2 are the average power angles of the first machine group and the second machine group respectively;

[0194] If δ C1 > δ C2 , then record the first machine group and the second machine group as the leading machine group and the lagging machine group respectively, that is, Ω S = Ω1, Ω A = Ω2; otherwise Ω S = Ω2, Ω A = Ω1.

[0195] Equalize the leading machine group and the lagging machine group to the single-machine infinite system. The process of calculating the concavity and convexity index of the angular velocity-power angle phase trajectory of the equivalent system includes:

[0196] Equalize the leading machine group Ω S and the lagging machine group Ω A with their respective inertia centers respectively;

[0197]

[0198] wherein, M S is the equivalent moment of inertia of the leading machine group Ω S ; M A is the equivalent moment of inertia of the lagging machine group Ω A ; δ S,k , ω S,k , P eS,k , P mS,k are respectively for ΩS The values of the equivalent power angle, angular velocity, electromagnetic power, and mechanical power at the k-th moment; δ A,k , ω A,k , P eA,k , P mA,k are respectively the values of the equivalent power angle, angular velocity, electromagnetic power, and mechanical power at the k-th moment of Ω A ;

[0199] The two-group equivalent parameter difference is used to characterize the system state and is equivalent to a single-machine infinite-bus system. The values of the power angle, angular velocity, electromagnetic power, and mechanical power trajectories of the equivalent system at each moment are calculated using formula (9);

[0200]

[0201] where M T is the moment of inertia of the equivalent system; δ k is the value of the power angle of the equivalent system at the k-th moment, ω k is the value of the angular velocity at the k-th moment, P e,k is the value of the electromagnetic power at the k-th moment, P m,k is the value of the mechanical power at the k-th moment;

[0202] The generator operating state data δ af , ω af , after the fault is cleared are respectively converted into the form represented by the state trajectory of the equivalent system after the fault is cleared using equations (8) - (9) to form;

[0203]

[0204] The process of judging whether the system is unstable according to whether the concavity-convexity index is greater than zero includes:

[0205] Calculate the concavity-convexity of the current power angle - angular velocity phase trajectory according to the state trajectory of the equivalent single-machine infinite-bus system, and judge whether the system has power angle instability through the concavity-convexity of this phase trajectory;

[0206] Take the angular velocity and power at the last two moments in the state trajectory and substitute them into formula (11) to calculate the first derivative of the power angle - angular velocity phase trajectory at the current moment and the previous moment:

[0207]

[0208] where d k is the first derivative of the power angle - angular velocity phase trajectory at the k-th moment;

[0209] Based on the first derivative of the power angle - angular velocity phase trajectory and the values of the power angle of the equivalent system at the last two moments, substitute them into formula (12) to calculate the second derivative of the power angle - angular velocity phase trajectory of the equivalent system at the current moment:

[0210]

[0211] Based on the second derivative of the power angle - angular velocity phase trajectory and the angular velocity at the current moment, calculate the trajectory concavity index according to formula (13):

[0212]

[0213] where, is the concavity of the power angle - angular velocity phase trajectory at the current t af moment; if then it is determined that the system enters the power angle instability state, otherwise calculate If then it is determined that the system is always in a stable state, clear all cached data; if only judge the stability for this fault, continue to collect data at the next moment and start the next round of stability judgment.

[0214] In the calculation module, if the system is unstable, calculate the accelerating area during the generator fault, then calculate the decelerating area that has occurred from after the fault removal to the moment of stability judgment using the power angle curve of the equivalent system, and calculate the remaining decelerating area through curve fitting extrapolation to form the predicted curve of the power angle characteristic of the equivalent system. The process includes: The accelerating area during the generator fault is:

[0215]

[0216] where, S A is the accelerating area of the equivalent system during the fault

[0217] The decelerating area that has occurred from after the fault removal to the moment of stability judgment:

[0218]

[0219] S D1 is the decelerating area of the equivalent system from after the fault removal to the moment of stability judgment;

[0220] The decelerating area of the equivalent system from after the fault removal to the moment of stability judgment

[0221] ΔS0 = S A - S D1 ; (16)

[0222] Assume that the power angle curve and the mechanical power curve of the equivalent system conform to the following form:

[0223] P e= α + βsin(γδ + C)

[0224] P m = a + bδ; (17)

[0225] Respectively with P e = P e af , δ = δ af and as the mapping sequences, the first parameter to be identified α, the second parameter to be identified β, the third parameter to be identified γ, the fourth parameter to be identified C, as well as the fifth parameter to be identified a and the sixth parameter to be identified b are calculated by using the non - linear and least - square parameter identification method;

[0226] At interval, it is equally divided into a preset number of power angle points to form a power angle prediction sequence Substitute into formula (16) to calculate the corresponding electromagnetic power and mechanical power prediction sequences Search forward from the front to find the sequence elements that meet the conditions to form the electromagnetic and mechanical power prediction sequences corresponding to the decelerating area and its corresponding power angle prediction sequence

[0227] Taking the predicted power angle sequence as the independent variable, the selected predicted mechanical power and electromagnetic power as the dependent variables, substitute into formula (18) and use the rectangle method to calculate the remaining decelerating area:

[0228]

[0229] where S D2 is the remaining decelerating area of the equivalent system from the stability judgment moment to the non - return point.

[0230] Calculate the theoretical amount of generator tripping according to the principle that the sum of the accelerating area and the total decelerating area is equal. Finally, form a set of potential generator tripping schemes based on the principle that the sum of the real - time output of the tripped generators is greater than the theoretical amount of generator tripping;

[0231] Determine the shortage of the remaining decelerating area according to the principle that the accelerating area is not less than the sum of the total decelerating areas:

[0232] S D3 = λ D (S A - S D1 - S D2 ) ; (19)

[0233] In the formula, S D3 is the shortage of the decelerating area required to maintain system stability, and λ DThe reliability coefficient to ensure the stability control effect even considering calculation errors;

[0234] Calculate the equivalent generator tripping amount ΔP m :

[0235]

[0236] Wherein, are the power angles of the equivalent system and the farthest point at the moment of stability judgment respectively;

[0237] Calculate the theoretical generator tripping amount ΔP M :

[0238]

[0239] The process of forming a potential generator tripping plan set based on the principle that the sum of the real-time outputs of the tripped units is greater than the theoretical generator tripping amount is as follows: Read the generator index set Ω of the leading generator group S , Set the maximum allowable number of tripped units N cmax ; Set the current allowable number of tripped units to 1; Form all subsets of the set Ω containing only N c elements; Calculate the sum of the real-time outputs of the generators in each subset. If the sum of the real-time outputs is greater than the theoretical generator tripping amount, add this subset to the potential generator tripping plan set; Until the number of tripped units reaches N c increases to the maximum allowable number of tripped units N S , and complete the generator tripping plan set. c cmax cmax cmax

[0240] In the decision-making module, the specific process of calculating the change in unbalanced energy of the selected generator tripping plan and then calculating the remaining decelerable area of the equivalent system of the remaining units after tripping and the change amount of the remaining decelerable area caused by tripping includes: Determine the remaining leading generator group index set Ω' s ; According to Ω' s and Ω A Calculate the state trajectories δ c of the equivalent system during the fault period after removing the tripped unit set Ω bf' , ω bf' ,

[0241] Calculate the accelerating area of the equivalent system after removing the tripped units during the fault period:

[0242]

[0243] Wherein, S' A is the accelerating area of the equivalent system after removing the tripped units during the fault period;

[0244] Determine the state trajectories δ c of the equivalent system during the fault period after removing the tripped unit set Ωaf' , ω af' ,

[0245] Calculate the deceleration area that has occurred in the equivalent system from after the fault is cleared to the moment of stability judgment:

[0246]

[0247] Among them, S' D1 is the deceleration area that has occurred in the equivalent system from after the fault is cleared to the moment of stability judgment;

[0248] Calculate the change in unbalanced energy caused by the generator tripping control:

[0249] ΔS1 = (S A - S D1 ) - (S' A - S' D1 ) ; (24)

[0250] In the formula, ΔS1 is the change in the cumulative unbalanced energy before and after the generator tripping control;

[0251] Based on the state trajectories δ c , ω af' , ω af' , of the equivalent system during the fault period after excluding the generator set Ω

[0252] Assume that the power angle curve and mechanical power curve of the equivalent system after excluding the tripped generators are as shown in Equation (17). Respectively, taking and as the mapping sequences, use the non - linear and least - squares parameter identification method to calculate the first parameter to be identified α, the second parameter to be identified β, the third parameter to be identified γ, the fourth parameter to be identified C, as well as the fifth parameter to be identified a and the sixth parameter to be identified b;

[0253] At interval, equally divide it into a preset number of power angle points to form a power angle prediction sequence Substitute it into Formula (23) to calculate the corresponding electromagnetic power and mechanical power prediction sequences Search forward to find the sequence elements that satisfy condition, and form the electromagnetic and mechanical power prediction sequences corresponding to the decelerable area and its corresponding power angle prediction sequence

[0254] Taking the predicted power angle sequence as the independent variable, the selected predicted mechanical power and electromagnetic power as the dependent variables, substitute them into Formula (25) and use the rectangle method to calculate the remaining decelerable area:

[0255]

[0256] Wherein, S' D2 is the remaining deceleration area of the equivalent system after removing the cut-off generator from the moment of stability judgment to the non-return point.

[0257] The process of evaluating the overall stability control effect of generator tripping and judging whether the generator tripping scheme can make the system stable, and finally comparing all generator tripping schemes that can make the system stable and selecting the scheme with the smallest amount of generator tripping as the optimal solution for the generator tripping scheme includes: modeling the uncut generators using the voltage source model, modeling the cut-off generators using the change in injected current, forming a circuit model with the change in injected power as the input, and forming a nodal incidence admittance matrix according to the circuit interconnection relationship; calculating the change in injected current of the uncut generators caused by generator tripping using the nodal voltage method

[0258]

[0259] are respectively the injected currents at the r-th and s-th nodes of the cut-off generator before removal;

[0260] are the change in injected current of the i-th generator and the change in injected current of the j-th generator of the uncut generators after the cut-off generator is removed; y ir and y is are respectively the incidence admittance elements at the r-th and s-th nodes of the i-th generator of the cut-off generator and the uncut generators; y jr and y js are respectively the incidence admittance elements at the r-th and s-th nodes of the j-th generator of the cut-off generator and the uncut generators;

[0261] Multiplying the change in injected current of the uncut generators by the corresponding nodal voltage to obtain the change in injected power of the uncut generators:

[0262]

[0263] Wherein, are respectively the internal electromotive force, the change in injected power and the change in injected current of the uncut generators;

[0264] Calculating the total change in electromagnetic power of the equivalent system after generator tripping using formula (28):

[0265]

[0266] Wherein, is the change in electromagnetic power of the equivalent system caused by generator tripping, M C is the sum of the equivalent inertias of the cut-off generators;

[0267] Calculate the change in the decelerable area caused by the change in the electromagnetic power of the equivalent system using Equation (29).

[0268]

[0269] where S' D3 is the change in the decelerable area caused by the change in the electromagnetic power of the equivalent system;

[0270] Sum the change in the unbalanced area, the remaining decelerable area, and the change value of the decelerable area to obtain the comprehensive decelerable area that the generator tripping control can provide:

[0271] S dec = ΔS1 + S' D2 + S' D3 ; (30)

[0272] Compare the comprehensive decelerable area S dec with the initial unbalanced energy ΔS0: If S dec > ΔS0, then generator tripping can stabilize the system. Add the set of generators included in this generator tripping scheme to the stable generator tripping scheme set Ω stable and record the corresponding generator tripping amount; Compare the total generator tripping power of all generator tripping schemes in the stable generator tripping scheme set Ω stable and select the scheme with the smallest total generator tripping amount as the optimal generator tripping scheme.

[0273] The online decision-making system for emergency generator tripping based on the evaluation of the stability control effect proposed in Embodiment 2 of the present invention can achieve the same technical effects as Embodiment 1. For the description of the relevant parts in the online decision-making system for emergency generator tripping based on the evaluation of the stability control effect provided in Embodiment 2 of this application, reference can be made to the detailed description of the corresponding parts in the online decision-making method for emergency generator tripping based on the evaluation of the stability control effect provided in Embodiment 1 of this application, which will not be elaborated here.

[0274] Although the specific implementation manners of the present invention are described above in conjunction with the accompanying drawings, it is not a limitation on the protection scope of the present invention. For those skilled in the art, other different forms of modifications or deformations can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. Based on the technical solutions of the present invention, various modifications or deformations that can be made by those skilled in the art without creative efforts are still within the protection scope of the present invention.

Claims

1. An online decision-making method for emergency machine switching based on stabilization effect evaluation is characterized by: The following steps are involved: Collect generator inertia parameters, record generator working status data after a fault occurs and after the fault is cleared; Based on the data matrix of the working state of the generator after the fault is removed, a clustering algorithm is used to cluster the power angle curve to divide the unit into a leading group and a lagging group; the leading group and the lagging group are equal to a single-machine infinite system, and the convexity index of the angular velocity-power angle phase trajectory of the equivalent system is calculated, and whether the system is unstable is judged according to whether the convexity index is greater than zero; If the system is unstable, the acceleration area during the generator fault period is calculated, and then the deceleration area before stable control is calculated using the equivalent system power angle curve from the time of fault removal to the time of stabilization judgment, and the equivalent system power angle characteristic prediction curve is formed by curve fitting and extrapolation to calculate the remaining deceleration area; the theoretical machine cutting amount is calculated based on the principle that the acceleration area is not less than the sum of the total deceleration area, and finally a potential machine cutting scheme set is formed based on the principle that the sum of the real-time output of the cut-off units is greater than the theoretical machine cutting amount; The unbalanced energy change of the selected machine cutting scheme is calculated, and then the remaining retarded area of ​​the equivalent system of the remaining units after the machine cutting and the change in the remaining retarded area caused by the machine cutting are calculated. The overall stabilization effect of the machine cutting is further evaluated to determine whether the machine cutting scheme can stabilize the system. Finally, all the machine cutting schemes that can stabilize the system are compared, and the scheme with the smallest machine cutting amount is selected as the optimal solution for the machine cutting scheme.

2. The method for online decision-making for emergency machine cutting based on stabilization effect evaluation according to claim 1 is characterized in that: The specific step of collecting the generator inertia parameters is to collect the inertia parameter data of the operating generator to form the generator inertia vector: M0=[M1 M2 … M i ]; (1) Among them, M0 is the moment of inertia of all operating generators, M i is the moment of inertia of the i-th generator; The data of the generator working status after the fault occurs are recorded as follows: Among them, δ bf A matrix for recording the operating data values ​​of the power angle of the generator; represents the power angle of the i-th generator at the k0th moment after the fault; ω bf Recording matrix of running data values ​​for generator angular velocity; represents the angular velocity of the generator of the i-th generator at the k0th moment after the fault; A matrix for recording the electromagnetic power operation data values ​​of the generator; represents the electromagnetic power of the i-th generator at the k0th moment after the fault; A matrix for recording the values ​​of the mechanical power operation data of the generator; represents the mechanical power of the i-th generator at the k0th moment after the fault; i = 1, 2, ..., n, k0 = 1, 2, ... t bf , n represents the number of generators, t bf Indicates the current sampling time of fault data record; The working status data of the generator after the fault is removed is: Among them, δ af Represents the generator power angle operation data value record matrix after the fault is removed; represents the power angle of the i-th generator at the k-th moment after the fault is removed; ω af Represents the generator angular velocity operation data value record matrix after the fault is removed; represents the angular velocity of the i-th generator at the k-th moment after the fault is removed; Represents the record matrix of the electromagnetic power operation data value of the generator after the fault is removed; It represents the electromagnetic power of the generator of the i-th generator at the k-th moment after the fault is removed; Represents the record matrix of the mechanical power operation data value of the generator after the fault is removed; represents the mechanical power of the i-th generator at the k-th moment after the fault is removed; k = 1, 2, ... t af , t af Indicates the current sampling time of data record after the fault is cleared.

3. The method for online decision-making for emergency machine cutting based on stabilization effect evaluation according to claim 2 is characterized in that: The clustering algorithm is used to cluster the power angle curves to divide the units into an advanced unit group and a lagging unit group; the advanced unit group and the lagging unit group are equalized to a single-machine infinite system, and the specific process of calculating the concavity and convexity index of the angular velocity-power angle phase trajectory of the equivalent system includes: Record the matrix δ with the data value after fault removal af A column of data (k=1,2,…t af ) as a power angle curve δ i Substitute the following formula (4) to calculate the Euclidean distance between any two power angle curves: Among them, D ij is the Euclidean distance between the power angle curve of the i-th generator and the power angle curve of the j-th generator; Select the two curves with the largest Euclidean distance as the initial cluster centers C1 and C2 of the first and second clusters, and write the corresponding generator numbers into the index sets Ω1 and Ω2 of the first and second clusters; From δ af Choose any one of the remaining curves m , calculate δ m The Euclidean distance between cluster centers C1 and C2: Among them, D m,1 ,D m,2 are the Euclidean distances between the mth power angle curve and the cluster centers C1 and C2 respectively; C 1,k , C 2,k are the values ​​of cluster centers C1 and C2 at the kth moment respectively. Contrast D m,1 and D m,2 Size relationship, if D m,1 >D m,2 , then δ m Classified as the second cluster and the curve column index is merged into the index set Ω2, otherwise δ m Classify it as the first cluster and merge the curve column index into the index set Ω1; update the aggregation center of the corresponding cluster: in, are the number of generators in the first and second machine groups respectively; If δ af After the power angle curve in is calculated, the average value difference between the cluster centers C1 and C2 is calculated, and the leading cluster Ω is distinguished according to the positive and negative difference. S and the lagging group Ω A : δ C1 , δ C2 are the average power angles of the first and second machine groups respectively; If δ C1 >δ C2 , then the first group and the second group are respectively the leading group and the lagging group, that is, Ω S =Ω1,Ω A =Ω2; otherwise Ω S =Ω2,Ω A =Ω1; The leading group Ω S and the lagging group Ω A Use the same value of their respective centers of inertia; Among them, M S For the leading group Ω S Equivalent moment of inertia; M A is the lagging group Ω A Equivalent moment of inertia; δ S,k ,ω S,k , P eS,k , P mS,k Ω S The value of the equivalent power angle at the kth moment, the value of the angular velocity at the kth moment, the value of the electromagnetic power at the kth moment, and the value of the mechanical power at the kth moment; δ A,k ,ω A,k , P eA,k , P mA,k Ω A The value of the equivalent power angle at the kth moment, the value of the angular velocity at the kth moment, the value of the electromagnetic power at the kth moment and the value of the mechanical power at the kth moment; The difference between two groups of equivalent parameters is used to characterize the system state, and the equivalent is a single-machine infinite system. Formula (9) is used to calculate the power angle, angular velocity, electromagnetic power and mechanical power trajectory of the equivalent system at each moment; M T =M S +M A ,k=1,2...t af ;(9) Among them, M T is the moment of inertia of the equivalent system; δ k is the value of the power angle of the equivalent system at the kth moment, ω k is the value of angular velocity at the kth moment, P e,k is the value of electromagnetic power at the kth moment, P m,k is the value of mechanical power at the kth moment; The generator working status data δ after the fault is removed af ,ω af , Equations (8) and (9) are used to convert the equivalent system state trajectory after fault removal to form; 4. The method for online decision-making for emergency machine cutting based on stabilization effect evaluation according to claim 3 is characterized in that: The process of judging whether the system is unstable according to whether the concavity and convexity index is greater than zero includes: The concavity of the current power angle-angular velocity phase trajectory is calculated according to the state trajectory of the equivalent single-machine infinite system, and the concavity of the phase trajectory is used to determine whether the system has power angle instability; Take the angular velocity and power at the last two moments in the state trajectory and substitute them into formula (11) to calculate the first-order derivative of the power angle-angular velocity phase trajectory at the current moment and the previous moment: Among them, d k is the first-order derivative of the power angle-angular velocity phase trajectory at the kth moment; Based on the first-order derivative of the power angle-angular velocity phase trajectory and the last two values ​​of the power angle of the equivalent system, substitute into formula (12) to calculate the second-order derivative of the power angle-angular velocity phase trajectory of the equivalent system at the current moment: Based on the second-order derivative of the power angle-angular velocity phase trajectory and the angular velocity at the current moment, the trajectory concavity index is calculated according to formula (13): in, is the current t af The concavity of the power angle-angular velocity phase trajectory at the moment; if Then the system is judged to enter the power angle instability state, otherwise the calculation if The system is determined to be in a stable state and all cache data is cleared; if Only the current fault is judged stable, and data collection at the next moment starts the next round of judgment.

5. The method for online decision-making for emergency machine cutting based on stabilization effect evaluation according to claim 4 is characterized in that: If the system is unstable, the acceleration area during the generator fault is calculated, and then the deceleration area before stable control is calculated using the equivalent system power angle curve from the time of fault removal to the time of stabilization judgment, and the equivalent system power angle characteristic prediction curve is formed by curve fitting and extrapolation to calculate the remaining deceleration area. The process includes: The acceleration area during generator fault is: Among them, S A is the acceleration area of ​​the equivalent system during the fault The deceleration area that has occurred from the time when the fault is cleared to the time when stability is determined: S D1 It is the deceleration area of ​​the equivalent system from the time when the fault is removed to the time when stability is judged; The deceleration area of ​​the equivalent system from the time of fault removal to the time of judging stability ΔS0=S A -S D1 ;(16) Assume that the equivalent system power angle curve and mechanical power curve conform to the following form: P e =α+βsin(γδ+C) P m =a+bδ;(17) P e =P e af ,δ=δ af and δ=δ af For mapping sequence, nonlinear and least square parameter identification methods are used to calculate the first parameter to be identified α, the second parameter to be identified β, the third parameter to be identified γ and the fourth parameter to be identified C, as well as the fifth parameter to be identified a and the sixth parameter to be identified b; exist The interval is equally divided into a preset number of power angle points to form a power angle prediction sequence Substitute into formula (16) to calculate the corresponding electromagnetic power and mechanical power prediction sequence P e predict , Search from front to back to meet The sequence elements of the conditions form the electromagnetic and mechanical power prediction sequence corresponding to the deceleration area and its corresponding power angle prediction sequence The predicted mechanical power is screened by taking the predicted power angle sequence as the independent variable. and electromagnetic power As the dependent variable, substitute it into formula (18) and use the rectangular method to calculate the remaining deceleration area: Among them, S D2 It is the remaining decelerable area of ​​the equivalent system from the stabilization moment to the point of no return.

6. The method for online decision-making for emergency machine cutting based on stabilization effect evaluation according to claim 5 is characterized in that: The theoretical cutting amount is calculated based on the principle that the acceleration area is not less than the sum of the total deceleration area, and finally a potential cutting plan set is formed based on the principle that the sum of the real-time output of the cut-off units is greater than the theoretical cutting amount; The remaining deceleration area shortfall is determined based on the principle that the acceleration area is not less than the sum of the total deceleration area: S D3 =λ D (S A -S D1 -S D2 ); (19) In the formula, S D3 The deceleration area required to maintain system stability, λ D It is the reliability factor that can guarantee the stabilization effect when taking into account the calculation error; Calculate the equivalent cutting amount ΔP m : in, They are the equivalent system power angle and the farthest point power angle at the time of judging stability respectively; Calculate the theoretical cutting capacity ΔP M : The process of forming a potential cutting plan set based on the principle that the sum of the real-time output of the cut-off units is greater than the theoretical cutting amount is as follows: Read the leading cluster generator index set Ω S , set the maximum allowed number of cutting machines N cmax ; Set the current allowed number of cutting machines to N c is 1; Formation containing only N c The set of elements Ω S All subsets of ; Calculate the sum of the real-time output of the generators in each subset. If the sum of the real-time output is greater than the theoretical cutting capacity, add the subset to the potential cutting solution set; until the number of cutting units reaches N c Increase to the maximum allowable number of cutting machines N cmax , complete the cutting machine solution set.

7. The method for online decision-making for emergency machine cutting based on stabilization effect evaluation according to claim 6 is characterized in that: The specific process of calculating the unbalanced energy change of the selected cutting scheme includes: Determine the remaining leading cluster index set Ω′ s ; According to Ω′ s and Ω A Calculate the set of eliminated units Ω c State trajectory of the post-equivalent system during fault periodδ bf' ,ω bf' , Calculate the acceleration area of ​​the equivalent system after removing the disconnected unit during the fault period: Among them, S' A It is the acceleration area of ​​the equivalent system after removing the cut-off units during the fault period; Determine the set of eliminated units Ω c State trajectory of the post-equivalent system during fault periodδ af' ,ω af' , Calculate the deceleration area of ​​the equivalent system from the time of fault removal to the time of judging stability: Among them, S' D1 It is the deceleration area of ​​the equivalent system from the time of fault removal to the time of judging stability; Calculate the change in unbalanced energy caused by cutting machine control: ΔS1=(S A -S D1 )-(S' A -S' D1 ); (24) Where ΔS1 is the change in the accumulated unbalanced energy before and after the machine cutting control.

8. The method for online decision-making for emergency machine cutting based on stabilization effect evaluation according to claim 7 is characterized in that: Then, the specific process of calculating the remaining retarded area of ​​the equivalent system of the remaining units after the machine cutting and the change of the remaining retarded area caused by the machine cutting includes: Based on the elimination of the unit set Ω c State trajectory of the post-equivalent system during fault periodδ af' ,ω af' , P e af' The trajectory is fitted with power angle characteristics and extrapolated to calculate the remaining deceleration area; Assume that the equivalent system power angle curve and mechanical power curve after removing the cut-off unit are as shown in formula (17), respectively: and For mapping sequence, nonlinear and least square parameter identification methods are used to calculate the first parameter to be identified α, the second parameter to be identified β, the third parameter to be identified γ and the fourth parameter to be identified C, as well as the fifth parameter to be identified a and the sixth parameter to be identified b; exist The interval is equally divided into a preset number of power angle points to form a power angle prediction sequence Substitute into formula (23) to calculate the corresponding electromagnetic power and mechanical power prediction series Search from front to back to meet The sequence elements of the conditions form the electromagnetic and mechanical power prediction sequence corresponding to the deceleration area and its corresponding power angle prediction sequence The predicted mechanical power is screened by taking the predicted power angle sequence as the independent variable. and electromagnetic power As the dependent variable, substitute it into formula (25) and use the rectangular method to calculate the remaining deceleration area: In the formula, S' D2 It is the remaining deceleration area of ​​the equivalent system after eliminating the cut-off units from the stability judgment moment to the no-return point.

9. The method for online decision-making for emergency machine cutting based on stabilization effect evaluation according to claim 8 is characterized in that: The process of evaluating the overall stabilization effect of the cutting machine to determine whether the cutting machine solution can stabilize the system, and finally comparing all the cutting machine solutions that can stabilize the system, and selecting the solution with the smallest cutting amount as the optimal solution for the cutting machine solution includes: The uncut-off generator is modeled by a voltage source model, and the cut-off generator is modeled by the injected current change, forming a circuit model with the injected power change as input, and forming a node-related admittance matrix according to the circuit interconnection relationship; Calculate the injection current change of the uncut-off units caused by cutting the generator using the node voltage method are the injected currents of the rth and sth nodes before the removed unit is removed; is the change in the injected current of the i-th generator and the j-th generator after the cut-off unit is cut off; y ir ,y is are the associated admittance elements of the rth and sth nodes of the i-th generator of the removed unit and the unremoved unit respectively; y jr ,y js are the associated admittance elements of the rth and sth nodes of the jth generator of the removed unit and the unremoved unit respectively; The power change of the non-removed units is obtained by multiplying the current change of the non-removed units by the corresponding node voltage: In the formula, They are the internal potential of the uncut-off unit, the variation of injected power and the variation of injected current respectively; Formula (28) is used to calculate the total change of electromagnetic power of the equivalent system after cutting the machine: In the formula, ΔP e af' is the change in electromagnetic power of the equivalent system caused by cutting, M C is the sum of the equivalent inertia of the removed units; Formula (29) is used to calculate the change in the decelerable area caused by the change in electromagnetic power of the equivalent system: Among them, S' D3 is the change in the decelerable area caused by the change in the electromagnetic power of the equivalent system; The combined deceleration area that can be provided by the cutting machine control is obtained by summing the change in unbalanced area, the remaining deceleration area, and the change in deceleration area: S dec =ΔS1+S' D2 +S' D3 ; (30) Comparison of comprehensive deceleration area S dec and the initial unbalanced energy ΔS0: If S dec >ΔS0, then the generator cutting can stabilize the system, and the generator set included in the generator cutting scheme is added to the stable generator cutting scheme set Ω stable And record the corresponding cutting amount; Comparison of stable cutting machine solutions stable The total power of all cutting plans in the cutting plan is calculated, and the plan with the smallest total power is selected as the optimal cutting plan.

10. The emergency machine cutting online decision system based on stabilization effect evaluation is characterized by: It includes a judgment module, a calculation module and a decision-making module; The judgment module is used to collect generator inertia parameters, record generator working status data after a fault occurs and generator working status data after the fault is removed; Based on the data matrix of the working state of the generator after the fault is removed, a clustering algorithm is used to cluster the power angle curve to divide the unit into a leading group and a lagging group; the leading group and the lagging group are equal to a single-machine infinite system, and the convexity index of the angular velocity-power angle phase trajectory of the equivalent system is calculated, and whether the system is unstable is judged according to whether the convexity index is greater than zero; The calculation module is used to calculate the acceleration area during the generator fault period if the system is unstable, and then use the equivalent system power angle curve from the time of fault removal to the time of stabilization to calculate the deceleration area before stable control, and form an equivalent system power angle characteristic prediction curve through curve fitting and extrapolation to calculate the remaining deceleration area; calculate the theoretical machine cutting amount according to the principle that the acceleration area is not less than the sum of the total deceleration area, and finally form a potential machine cutting plan set based on the principle that the sum of the real-time output of the cut-off units is greater than the theoretical machine cutting amount; The decision module is used to calculate the unbalanced energy change of the selected machine cutting scheme, and then calculate the remaining decelerable area of ​​the equivalent system of the remaining units after the machine cutting and the change in the remaining decelerable area caused by the machine cutting, further evaluate the overall stabilization effect of the machine cutting to determine whether the machine cutting scheme can stabilize the system, and finally compare all the machine cutting schemes that can stabilize the system, and select the scheme with the smallest machine cutting amount as the optimal solution for the machine cutting scheme.

Images