A method for power system load modeling based on big data

Abstract

The present invention discloses a method for power system load modeling based on big data, which relates to the field of power systems and includes the following steps: First, historical data samples of the load nodes to be identified are obtained, and the data includes bus voltage U, active power P, reactive power Q, and bus frequency f; then, load models of static ZIP and induction motors are established respectively; next, a comprehensive load model of the power grid is established to determine the proportion of each load component in the power grid; finally, each load parameter in the power grid is determined. The present invention is simple to implement and accurate in results, and can be applied to power system load modeling and parameter identification, providing a basis for subsequent power system security and stability analysis and control.

Description

Technical Field

[0001] The present invention relates to the field of power systems, and particularly to a method for power system load modeling based on big data. Background Art

[0002] As one of the important components in the power system, the load has a great impact on the analysis and simulation calculations of the static, dynamic, and transient characteristics and stability of the power system. However, the widely used load models are still relatively simplified and rough. The excessive roughness of the load models has become a key factor restricting the accuracy of power system analysis and simulation calculations. Establishing a dynamic load model that conforms to the actual situation and can accurately reflect the actual important characteristics has very important practical significance.

[0003] With the continuous development of the power system, the grid structure has become increasingly complex, and the types of power loads have gradually become diverse. The traditional load models and research methods can no longer meet the needs of operation and dispatch personnel. In recent years, phasor measurement units (PMUs) and wide area measurement systems (WAMS) have gradually become popular and developed in the power system. A large amount of measurement data has been summarized into the grid dispatching system, which provides big data support for further load modeling and parameter identification research under complex grid conditions. The combination of big data and machine learning algorithms can make relatively comprehensive use of these massive measurement data, and through continuous learning and correction, achieve the identification of load model parameters. With the support of big data, machine learning can well make up for the deficiencies of traditional parameter identification methods, and also provide a more accurate load model for simulation calculations to adapt to the continuously developing grid conditions in the future. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method for power system load modeling based on big data, which provides a basis for the safety and stability analysis and control of the power grid.

[0005] To achieve the above object, the present invention adopts the following technical solutions: A method for power system load modeling based on big data, the method comprising the following steps:

[0006] (1) Obtain historical data samples of the load nodes to be identified, and the data includes bus voltage U, active power P, reactive power Q, and bus frequency f;

[0007] (2) Establish a load model

[0008] 1) Static ZIP load model

[0009] The static ZIP load model adopts a polynomial model and belongs to an input-output model. Therefore, a static ZIP load model as shown in Equation (1) is established;

[0010]

[0011] Where: P is the active power of the static load, Q is the reactive power of the static load, U is the voltage, U0 is the initial voltage value, P0 is the initial active power value, Q0 is the initial reactive power value, p1, p2, p3 are the proportions of constant impedance, constant current, and constant power in the active load, and q1, q2, q3 are the proportions of constant impedance, constant current, and constant power in the reactive load;

[0012] In Equation (1), p1, p2, p3, q1, q2, q3 are parameters to be identified, and the parameters to be identified need to satisfy the constraint conditions shown in Equation (2);

[0013]

[0014] 2) Induction motor load model

[0015] The induction motor model should consider both the mechanical transient and electromagnetic transient processes of the motor. Therefore, the model of the induction motor is described by a third-order model, and the dynamic behavior of the induction motor is often represented by a differential equation, as shown in Equation (3);

[0016]

[0017] Where: e′ d and e q ′ are the d-axis and q-axis components of the transient electromotive force of the wind turbine, T d0 ′ is the rotor circuit time constant when the stator circuit is open, x is the sum of the leakage reactance of the stator winding and the excitation reactance, x′ is the transient reactance of the stator winding when the rotor circuit is short-circuited, i d and i q are the d-axis and q-axis components of the stator current respectively, s is the slip ratio, ω is the rotor speed, T J is the equivalent mechanical inertia time constant of the generator, M m and M e are the mechanical load torque and electromagnetic torque of the doubly-fed induction motor respectively.

[0018] The output equation of the induction motor is:

[0019]

[0020] The steady-state constraint is:

[0021]

[0022] Then the parameters to be identified in the electromagnetic transient part of the induction motor are r1, x′, T d0 ′, and the parameters to be identified in the electromechanical transient part are T J , α, β, M e , s.

[0023] (3) Determine the proportion of each load component in the power grid

[0024] 1) Set a threshold for the bus voltage of the load node in the power grid, compare the measured bus voltage of the load node in the power grid with the threshold. If the measured value of the bus voltage of the load node in the power grid is less than the threshold, it is determined that the power grid is disturbed, and 4 groups of bus voltages U i , active power P i , reactive power Q i and the initial phase θ of the bus voltage i are repeatedly recorded, where i = 1, 2, 3, 4; if the measured value of the bus voltage of the load node in the power grid is greater than or equal to the threshold, it is determined that the power grid is in steady-state operation, and the bus voltage U0, active power P0, reactive power Q0 and the initial phase θ0 of the bus voltage during steady-state operation are recorded.

[0025] 2) Establish a set of equations for the power grid comprehensive load model based on the disturbance measurement values and the steady-state measurement values;

[0026]

[0027] The expressions of K1 to K8 in Equation (6) are:

[0028]

[0029] In Equation (7): K M is the proportion of the active power of the induction motor in the total power of the power grid comprehensive load, K Z is the proportion of the active power of the constant impedance load in the total power of the power grid comprehensive load, K I is the proportion of the active power of the constant current load in the total power of the power grid comprehensive load, K P is the proportion of the active power of the constant power load in the total power of the power grid comprehensive load, and α is the internal potential angle of the motor in the power grid comprehensive load.

[0030] 3) Combining Equation (6) and Equation (7), the proportions K M , K Z , K I , K P of the components of the required power grid comprehensive load model can be obtained:

[0031]

[0032] (4) Determine the power grid load parameters according to the load model in step (2) and the proportion of each load component in the power grid in step (3)

[0033] 1) Determination of static ZIP load model parameters

[0034] Γ zip = [p1K Z , p2K I , p3K P , q1K Z , q2K I , q3K P (9)

[0036] 2) Determination of parameters of the induction motor load model

[0037] Determination of parameters of the electromagnetic transient part:

[0038] Γ e = [r1 x′ T d0 ′]·K M (10)

[0040] Determination of parameters of the electromechanical transient part:

[0041] Γ m = [T J α β M e s]·K M (11) Description of the drawings

[0043] Figure 1 This is the flow chart of the present invention.

[0044] Figure 2 This is the structure diagram of the load model including the static ZIP load and the induction motor. Detailed implementation manners

[0045] The present invention includes the following steps:

[0046] (1) Obtain the historical data samples of the load nodes to be identified, and the data includes the bus voltage U, active power P, reactive power Q, and bus frequency f;

[0047] (2) Establish a load model

[0048] 1) Static ZIP load model

[0049] The static ZIP load model adopts a polynomial model, which belongs to an input-output model. Therefore, a static ZIP load model as shown in Equation (1) is established;

[0050]

[0051] ​Where: P is the active power of the static load, Q is the reactive power of the static load, U is the voltage, U0 is the initial voltage value, P0 is the initial active power value, Q0 is the initial reactive power value, p1, p2, p3 are the proportions of constant impedance, constant current, and constant power in the active load, and q1, q2, q3 are the proportions of constant impedance, constant current, and constant power in the reactive load;

[0052] In the above formula, p1, p2, p3, q1, q2, q3 are parameters to be identified, and the parameters to be identified need to satisfy the following constraints;

[0053]

[0054] 2) Induction motor load model

[0055] The induction motor model should consider both the mechanical transient and electromagnetic transient processes of the motor. Therefore, the model of the induction motor is described by a third-order model, and the dynamic behavior of the induction motor is often represented by a differential equation, as shown in the following formula;

[0056]

[0057] Where: e′ d and e′ q are the d-axis and q-axis components of the transient electromotive force of the wind turbine, respectively, T d0 ′ is the rotor circuit time constant when the stator circuit is open, x is the sum of the leakage reactance of the stator winding and the excitation reactance, x′ is the transient reactance of the stator winding when the rotor circuit is short-circuited, i d and i q are the d-axis and q-axis components of the stator current, respectively, s is the slip, ω is the rotor speed, T J is the equivalent mechanical inertia time constant of the generator, M m and M e are the mechanical load torque and electromagnetic torque of the doubly-fed induction motor, respectively.

[0058] The output equation of the induction motor is:

[0059]

[0060] The steady-state constraint is:

[0061]

[0062] Converting the above two formulas into state equations gives:

[0063]

[0064] According to Equation (14), e d0 ′, e q0 ′

[0065]

[0066] Then, calculate x and s using Equation (15).

[0067]

[0068] Mechanical torque of induction motor:

[0069]

[0070] Electromagnetic torque of induction motor:

[0071]

[0072] From Equation (16) and Equation (17), the electromechanical transient part of Equation (3) can be gradually transformed into the following equation:

[0073]

[0074] Then, the parameters to be identified in the electromagnetic transient part of the induction motor are r1, x′, T d0 ′, and the parameters to be identified in the electromechanical transient part are T J , α, β, M e , s.

[0075] (3) Determine the proportion of each load component in the power grid

[0076] 1) Set a threshold for the bus voltage of the power grid load node, compare the measured bus voltage of the power grid load node with the threshold. If the measured value of the bus voltage of the power grid load node is less than the threshold, it is determined that the power grid has a disturbance, and repeat to record 4 sets of bus voltages U i , active power P i , reactive power Q i and the initial phase θ of the bus voltage i , where i = 1, 2, 3, 4; if the measured value of the bus voltage of the power grid load node is greater than or equal to the threshold, it is determined that the power grid is in steady-state operation, and record the bus voltage U0, active power P0, reactive power Q0 and the initial phase θ0 of the bus voltage during steady-state operation.

[0077] 2) Establish a set of simultaneous equations for the power grid comprehensive load model based on the disturbance measurement values and the steady-state measurement values;

[0078]

[0079] The expressions of K1 to K8 in the above equation are:

[0080]

[0081] Where: K M is the proportion of the active power of induction motors in the comprehensive grid load to the total load power, and K Z is the proportion of the active power of constant-impedance loads in the comprehensive grid load to the total load power, and K I is the proportion of the active power of constant-current loads in the comprehensive grid load to the total load power, and K P is the proportion of the active power of constant-power loads in the comprehensive grid load to the total load power, and α is the internal potential angle of the motors in the comprehensive grid load.

[0082] 3) By combining Equation (6) and Equation (7), the proportions K M , K Z , K I , K P of the various components of the required comprehensive grid load model can be obtained:

[0083]

[0084] (4) Based on the load model in step (2) and the proportions of the various load components in the grid in step (3), determine the grid load parameters

[0085] 1) Determination of the parameters of the static ZIP load model

[0086] Γ zip = [p1K Z , p2K I , p3K P , q1K Z , q2K I , q3K P

[0087] 2) Determination of the parameters of the induction motor load model

[0088] Determination of the parameters of the electromagnetic transient part:

[0089] Γ e = [r1 x′ T d0 ′]·K M

[0090] Determination of the parameters of the electromechanical transient part:

[0091] Γ m = [T J α β M e s]·K M

[0092] ​The above embodiments are only used to illustrate the present invention and are not intended to limit the present invention. Those of ordinary skill in the relevant technical field can also make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, all equivalent technical solutions also fall within the protection scope of the present invention.

Claims

1. A method for power system load modeling based on big data, characterized in that: (1) Obtain historical data samples of the load nodes to be identified, and the data includes bus voltage U, active power P, reactive power Q, and bus frequency f; (2) Establish load models of static ZIP and induction motors respectively; (3) Determine the proportion of each load component in the power grid; (4) Determine the load parameters of each part in the power grid; Determination of the load parameters of each part in the power grid in step (4); 1) Determination of static ZIP load model parameters Γ zip = [p1K Z , p2K I , p3K P , q1K Z , q2K I , q3K P ​ K Z is the proportion of the active power of the constant impedance load in the total power of the grid's comprehensive load, K I is the proportion of the active power of the constant current load in the total power of the grid's comprehensive load, K P is the proportion of the active power of the constant power load in the total power of the grid's comprehensive load; 2) Determination of induction motor load model parameters Determination of parameters in the electromagnetic transient part: Γ e = [r1 x′T d0 ′]·K M Determination of parameters in the electromechanical transient part: Γ m = [T J αβM e s]·K M ; r1, x′, T d0 ′ are the parameters to be identified in the electromagnetic transient part of the induction motor, T J , α, β, M e , s are the parameters to be identified in the electromechanical transient part; K M is the proportion of the active power of the induction motor in the total power of the grid integrated load.

2. The method for power system load modeling based on big data according to claim 1, characterized in that, In step (2), establish load models of static ZIP and induction motors respectively; 1) Static ZIP load model: P is the active power of the static load, Q is the reactive power of the static load, U is the voltage, U0 is the initial voltage value, P0 is the initial active power value, Q0 is the initial reactive power value, p1, p2, p3 are the proportions of constant impedance, constant current, and constant power in the active load, and q1, q2, q3 are the proportions of constant impedance, constant current, and constant power in the reactive load; 2) Induction motor load model: e′ d and e′ q are the d-axis and q-axis components of the transient electromotive force of the wind turbine, respectively. T d0 ′ is the rotor circuit time constant when the stator circuit is open. x is the sum of the leakage reactance of the stator winding and the excitation reactance. x′ is the transient reactance of the stator winding when the rotor circuit is short-circuited. i d and i q are the d-axis and q-axis components of the stator current, respectively. s is the slip ratio, ω is the rotor speed, and T J is the equivalent mechanical inertia time constant of the generator. M m and M e are the mechanical load torque and electromagnetic torque of the doubly-fed induction machine, respectively.

3. A method for power system load modeling based on big data according to claim 1, characterized in that, In step (3), determine the proportion of each load component in the power grid; 1) Set a threshold for the bus voltage of the grid load node, compare the measured bus voltage of the grid load node with the threshold. If the measured value of the bus voltage of the grid load node is less than the threshold, it is determined that the grid has a disturbance, and repeat to record 4 sets of bus voltages U i , active power P i , reactive power Q i and the initial phase θ of the bus voltage i , where i = 1, 2, 3, 4; if the measured value of the bus voltage of the grid load node is greater than or equal to the threshold, it is determined that the grid is in steady-state operation, and record the bus voltage U0, active power P0, reactive power Q0 and the initial phase θ0 of the bus voltage during steady-state operation; 2) Establish a simultaneous equation of the power grid comprehensive load model according to the disturbance measurement value and the steady-state measurement value; The expressions of K1 to K8 in the above formula are: 3) By combining the above two equations, the proportions K M , K Z , K I , K P of the components of the grid composite load model to be calculated can be obtained:

Images