MPC-ADMM-based hierarchical coordinated voltage control strategy for coupled systems

Abstract

The application discloses a coupling system voltage layered collaborative control strategy based on MPC-ADMM, the coupling system comprises an upper MPC control system and a lower MPC control system, the control of the upper MPC control system and the lower MPC control system satisfies a safety constraint condition, and the application is specifically implemented according to the following steps: analyzing the 'unit-station-system' hierarchical characteristics of the coupling system, determining the layered voltage control structure of the coupling system; considering the operation network loss optimization and the voltage control requirements of various grid-connected points of the coupling system, considering various reactive power source resources in the coupling system, constructing a layered collaborative control mathematical model of a dual-mode adaptive switching of network loss optimization control and voltage correction control based on MPC-ADMM; the parameters in the layered collaborative control mathematical model are optimized by using an ADMM algorithm, and the grid-connected point voltage control result is calculated; the control strategy can make the coupling system satisfy the grid-connected voltage safety requirements and optimize the operation network loss of the system.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of voltage stability analysis and control of high-proportion renewable energy power systems, and particularly relates to a coupled system voltage hierarchical collaborative control strategy based on MPC-ADMM. BACKGROUND

[0002] The coupling mode of wind and light new energy stations and traditional thermal power units accessing the power grid through the grid connection point has been widely used in northern regions. The coupled system not only solves the energy crisis, but also brings severe challenges to the operation of the system: the new energy units, stations and system grid connection points of the coupled system all need to meet their own grid connection voltage requirements, which increases the difficulty of voltage control of the coupled system, and how to coordinate various types of reactive power control resources within the coupled system and fully utilize the voltage regulation capability of various types of reactive power equipment is also an important problem to be solved in the voltage control process.

[0003] According to the structure of the control system, the voltage control method can be roughly divided into centralized control and decentralized control, as well as a distributed control structure that combines the advantages of centralized control and decentralized control. Many achievements have been made in the research on the voltage control problem of the above control methods, which provides guarantee for the safe and stable operation of the system. However, there are still some deficiencies, mainly reflected in the following aspects: 1) the research on the voltage of the new energy station grid connection point or the active distribution network is lack of research on the voltage control problem of the "unit-station-system" grid connection point of the coupled system; 2) a single control objective function is set, which cannot meet the system loss optimization and voltage control requirements; 3) the centralized control and decentralized control cannot meet the technical and economic requirements in the voltage control process.

[0004] Therefore, it is urgent to propose a voltage control method suitable for the coupled system, so that the various grid connection points in the coupled system can meet the grid connection voltage safety requirements and optimize the operating loss of the coupled system. SUMMARY

[0005] The purpose of the present application is to provide a coupled system voltage hierarchical collaborative control strategy based on MPC-ADMM, which can make the coupled system meet the grid connection voltage safety requirements and optimize the operating loss of the system.

[0006] The technical solution adopted by the present application is a coupled system voltage hierarchical collaborative control strategy based on MPC-ADMM. The coupled system includes an upper MPC control system and a lower MPC control system. The control of the upper MPC control system and the lower MPC control system meets the safety constraint conditions, and is specifically implemented according to the following steps:

[0007] Step 1, analyze the "unit-station-system" hierarchical characteristics of the coupled system, and determine the hierarchical voltage control structure of the coupled system;

[0008] Step 2, considering the coupling system operation loss optimization and various grid-connected point voltage control requirements, considering various reactive power source resources in the coupling system, a hierarchical collaborative control mathematical model of dual-mode adaptive switching of network loss optimization control and voltage correction control based on MPC-ADMM is constructed;

[0009] Step 3, the ADMM algorithm is used to optimize the parameters in the hierarchical collaborative control mathematical model, and the grid-connected point voltage control result is calculated.

[0010] The application also has the characteristics of:

[0011] The specific process of step 1 is: the upper MPC control system includes traditional thermal power units, reactive power compensation devices, and upper control centers; the upper control center connects the HMPC controller through the coupling system grid-connected point and inputs the coupling system grid-connected point reference voltage instruction to the HMPC controller; the traditional thermal power units and the reactive power compensation devices are connected to the HMPC controller and receive the coupling system grid-connected point reference voltage instruction through the HMPC controller; the lower MPC control system is divided into multiple sub-station systems controlled by LMPC controllers according to the types of new energy stations; the HMPC controller connects the LMPC controller through the grid-connected point of each station and issues the grid-connected point voltage reference instruction of each station to the LMPC controller; the LMPC controller directly issues the grid-connected point voltage reference instruction of each station to each new energy station; and the HMPC controller receives the remaining adjustable reactive power range of each new energy station.

[0012] The types of new energy stations include wind farms and photovoltaic farms.

[0013] The safety constraint condition is expressed as:

[0014]

[0015] In formula (1), is the reference voltage instruction of the wind farm grid-connected point issued by the HMPC controller, is the reference voltage instruction of the photovoltaic power station grid-connected point issued by the HMPC controller, is the reference voltage instruction of the wind farm grid-connected point issued by the LMPC controller, is the reference voltage instruction of the photovoltaic power station grid-connected point issued by the LMPC controller.

[0016] The specific process of step 2 is:

[0017] The network loss optimization and various grid-connected point voltage control requirements are operated on the control target voltage setting, two control modes, voltage correction control and normal network loss control, are designed, the two control modes realize adaptive switching of control state by judging the operation state of the coupling system grid-connected point voltage, and the target function and prediction model of the upper MPC control system and the lower MPC control system are established in the two modes, that is, the hierarchical collaborative control mathematical model.

[0018] The specific process of adaptive switching of control state by the two control modes through judging the operation state of the coupling system grid-connected point voltage is as follows:

[0019] The coupling system grid-connected point is taken as a node i, and a voltage safety range is set, and the controller control switching logic rule is as follows:

[0020] ①The node i voltage is in the preset voltage safety range, that is: i -U i,ref ||≤U i,th , the HMPC controller or the LMPC controller is in the normal network loss control mode, the system network loss is optimal under the condition of guaranteeing the voltage safety constraint;

[0021] ②The node i voltage is out of the set voltage safety range, that is: i -U i,ref ||>U i,th , the HMPC controller or the LMPC controller turns to the voltage correction control mode, and the node i voltage deviation is minimum;

[0022] Wherein, U i represents the node i voltage, U i,ref represents the corresponding node voltage reference value; and U i,th represents the node i voltage threshold.

[0023] The specific process of establishing the target function and prediction model of the upper MPC control system and the lower MPC control system in the two modes is as follows:

[0024] (1) For the control model of the upper MPC control system:

[0025] ①In the normal network loss control, if the key grid-connected point voltage of the upper system is in the preset range, that is: The HMPC controller switches to the normal network loss control mode; in this control mode, the system active network loss and control cost are taken as optimization objectives; based on the ADMM optimization algorithm, the target function is represented as:

[0026]

[0027] In formula (2), ω Ploss >0, ωcost >0 represent the weight coefficients of system network loss and upper system control cost, λ WF ,λ PV are all Lagrange multipliers, ρ WF , ρ PV are penalty factors, Nc represents the control step size in the model predictive control algorithm, where the prediction step size N p =N c =3;

[0028] System active network loss affected by generator terminal voltage U g , output power P g Wind and solar energy stations provide reactive power to the upper MPC control system The linear prediction model of the system active network loss is:

[0029]

[0030] In formula (3), is the measured value of the system active network loss at the current moment, is a column vector of control variables for the next k moments; is the sensitivity matrix of the system active network loss to the upper control variables;

[0031] The reactive power that a new energy station can provide can be expressed as:

[0032]

[0033] In formula (4), ΔQ WF,SVC , ΔQ PV,SVC Expressed as:

[0034]

[0035] In formula (4), S N,WF 、S N,PV is the rated capacity of wind turbine and photovoltaic unit, P m,WF 、P n,pv N is the output of the mth and nth units in the wind farm and photovoltaic power station at the current moment; WF , N PV The number of units corresponding to wind farms and photovoltaic power stations;

[0036] In formula (5), Q N,SVC is the rated capacity of SVC; Q WF,SVC , Q PV,SVC They represent the reactive power consumed by the reactive compensation equipment of the lower-level wind farm and photovoltaic power station respectively;

[0037] ② When the voltage correction control is in place, if the voltage of the key grid-connected point of the upper system exceeds the preset range, i.e., The HMPC controller switches to the voltage correction control mode. In this mode, the controller's objective function is:

[0038]

[0039] In equation (6), is the predicted value of the voltage of the key node i of the upper system; similarly affected by the reference voltage, output of the generator and the reactive power provided by the new energy station, the prediction model for the voltage of the key node i of the upper system is:

[0040]

[0041] wherein, is the measured value of the voltage of the node i of the upper system at the current time, is the sensitivity matrix of the voltage of the node i of the upper system to the control variable;

[0042] The objective functions (2) and (6) of the upper MPC control system satisfy the following constraint conditions:

[0043]

[0044] In addition to satisfying the inequality constraint (8) about the voltage and the range of the control variable, the objective function of the upper MPC control system also needs to satisfy the equality constraint in equations (3) and (6) about the system loss and the voltage of the key node of the upper system.

[0045] (2) Control model of the lower MPC control system

[0046] The control objective function of the new energy station is also set in two control modes: voltage correction control and normal loss control, in the same way as the setting of the objective function of the upper optimization control model.

[0047] ① When the normal loss control is in place, if the voltage of each key grid-connected point of the new energy station of the lower system is within the preset range, i.e., The LMPC controller operates in the normal control mode to ensure the economic operation of the new energy station, considering the minimum active loss and control cost of the station; wherein, the reference value of the node voltage in the lower MPC control system is the output result of the HMPC controller;

[0048] The objective function of the wind farm is:

[0049]

[0050] Photovoltaic power station:

[0051]

[0052] The network loss of the lower station is affected by the reactive power Q of the new energy unit and the static VAR compensation device. WFGU (Q PVGU ), Q SVC The linearized prediction model for active network loss of wind and solar power stations is:

[0053]

[0054]

[0055] In formulas (11) and (12), Corresponding to the measured value of the active network loss of the wind and solar power stations at the current moment; Δu WF (k) = [ΔQ WFGU (k),ΔQ SVC (k)] T is the column vector consisting of the control variables of the wind farm station at time k in the future; Δu PV (k) = [ΔQ PVGU (k),ΔQ SVC (k)] T is a column vector consisting of control variables corresponding to the optical field station at time k in the future; Sensitivity matrix of active network loss of wind and solar power stations to each control variable;

[0056] ② During voltage correction control, when the voltage at the key grid connection point of the lower-level new energy station exceeds the preset range: The LMPC controller switches to the correction control mode; with voltage safety as the goal, considering the offset between the grid connection point i voltage and the corresponding reference voltage and the minimum control cost, its controller objective function is:

[0057] The objective function of the wind farm is:

[0058]

[0059] The objective function of the photovoltaic power station is:

[0060]

[0061] In formulas (13) and (14), The voltage prediction value of the key grid connection point i in the wind and solar power stations at time k is also affected by the reactive power of the new energy unit and the static VAR compensation device. The prediction model of the voltage of the key node i in the lower-level MPC control system is expressed as:

[0062]

[0063]

[0064] wherein, is the wind, light field station node voltage at the current time measurement value; is the lower layer wind, light field station node voltage sensitivity matrix of control variable;

[0065] The new energy field station target functions (9), (10) and (13), (14) in the lower layer MPC control system satisfy the following conditions:

[0066] The lower layer MPC control system new energy field station target function (9), (10) and (13), (14) satisfy the following conditions:

[0067]

[0068]

[0069] In addition to satisfying the inequality constraints, the new energy field station target function in the lower layer MPC control system also needs to satisfy the equality constraints in equations (11), (12) and (15), (16) about the lower layer field station network loss, node voltage.

[0070] The specific process of step 3 is as follows:

[0071] Step 3.1: Set the power daily fluctuation scene, the time scale is 15 min, and the scene cycle iteration flag m = m + 1;

[0072] Step 3.2: Initialize ADMM algorithm parameters, algorithm multipliers, and system internal control variables, lower layer field station grid-connected point voltage Set the iteration flag k = 0;

[0073] Step 3.3: Let k = k + 1, after the upper layer MPC control system accepts the grid-connected point reference voltage instruction from the upper control center, judge the voltage state of the key grid-connected point of the upper layer MPC control system, the HMPC controller switches to the corresponding control mode, and solves the grid-connected point voltage of the upper layer field station controlled by the upper layer MPC control system as the corresponding new energy field station grid-connected point voltage control instruction, which is passed to the lower layer corresponding field station LMPC controller;

[0074] Step 3.4: After the lower layer MPC control system accepts the grid-connected point voltage reference instruction from the upper layer MPC control system, according to the voltage state of the lower layer wind, light field station grid-connected point, select the corresponding LMPC controller control mode, respectively solve the grid-connected point voltage of the lower layer field station, and upload the remaining controllable reactive power range of the new energy field station to the HMPC controller system;

[0075] Step 3.5: Judge the grid-connected point voltage of the wind, light field station of the HMPC control system and the grid-connected point voltage of the corresponding LMPC control system field station whether the residual error between the two satisfies a set tolerance limit, if yes, output the system node voltage control result, update the algorithm multiplier and return to step 3.1, otherwise, update the algorithm multiplier and return to step 3.2.

[0076] The update formula for updating the algorithm multiplier is:

[0077]

[0078] wherein, μ, σ are adjustable parameters, wherein, μ>0, τ inc >1.0, τ dec >1.0.

[0079] The present application has the following advantages:

[0080] (1) By analyzing the physical level characteristics of the coupling system and the distribution characteristics of the reactive power source resources, the layered voltage control structure of the coupling system is determined, the global optimization control problem of the coupling system is divided into multiple sub-optimization control problems, and the calculation scale and complexity of the optimization control problem are reduced;

[0081] (2) Considering the influence of the control process on the future state and various types of reactive power sources in the coupling system, a layered collaborative control strategy based on MPC-ADMM is constructed, and the control delay problem is avoided;

[0082] (3) Considering the operation loss optimization of the coupling system and the voltage control demand of various types of grid-connected points, the normal loss control and voltage correction control modes are set in the optimization control objective function, the voltage out-of-limit problem of various types of grid-connected points in the coupling system is solved, and the operation loss of the system is optimized. BRIEF DESCRIPTION OF DRAWINGS

[0083] Figure 1 is a schematic diagram of the voltage layered collaborative control architecture;

[0084] Figure 2 is a schematic diagram of the layered voltage control structure of the coupling system;

[0085] Figure 3 is a schematic diagram of the modified IEEE-14 node system;

[0086] Figure 4 is a schematic diagram of the power fluctuation of the modified IEEE-14 node system;

[0087] Figure 5 is a schematic diagram of the voltage fluctuation of the key nodes of the modified IEEE-14 node system before control;

[0088] Figure 6 is a schematic diagram of the voltage fluctuation of the key nodes of the modified IEEE-14 node system after control;

[0089] Figure 7 For comparison of network loss optimization results before and after control;

[0090] Figure 8 For a local power grid in Liaoning;

[0091] Figure 9 For a local power grid in Liaoning;

[0092] Figure 10 For a local power grid in Liaoning;

[0093] Figure 11 For a local power grid in Liaoning;

[0094] Figure 12 For comparison of network loss optimization results before and after control. DETAILED DESCRIPTION

[0095] The application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0096] As Figure 1 shown is the voltage hierarchical collaborative control architecture suitable for coupling systems proposed by the application, a multi-power system in which wind, light and other new energy stations and traditional thermal power generation are coupled through a grid connection point in the same region is defined as a coupling system, and the coupling system has a "unit-station-system" hierarchical feature in physical structure. Various types of reactive power source resources of the system are considered, wherein Y represents the admittance matrix of the coupling system, U and θ represent the node voltage and phase angle of the coupling system, respectively; represent the reference active power and reference voltage of the thermal power unit, respectively; represent the reference voltage of the wind and light station grid connection point, respectively; and correspond to the reactive power reference values of the reactive power compensation device, the wind farm and the photovoltaic power station, respectively; and respectively represent the upper and lower limits of reactive power that can be provided by wind and light new energy stations. In the voltage hierarchical coordination architecture, the overall coupled system control system is composed of an upper MPC control system and a lower MPC control system. The LMPC control system is a sub-station control system according to different new energy stations. The sensitivity calculation module is used to calculate the required sensitivity coefficient of each MPC controller and is sent to the HMPC and LMPC controllers. After receiving the reference voltage instruction of the coupled system grid connection point issued by the upper control center, the HMPC controller solves the control scheme that meets the upper layer safety constraint condition by coordinating the upper layer system control resources such as thermal power units and reactive power compensation devices and the lower layer controllable reactive power resources of new energy stations, and sends the grid connection point voltage of each station in the upper layer as the reference instruction of the grid connection point voltage of each station to the corresponding LMPC controller of the new energy station. After receiving the corresponding station grid connection point voltage reference instruction, the sub-station LMPC controller adjusts the control resources such as unit groups and reactive power compensation devices in the station, solves the control scheme that meets the lower layer safety constraint condition, and uploads the remaining adjustable reactive power range of each new energy station to the HMPC controller, and so on. The iterative convergence is repeated to the global optimal solution. Specifically:

[0097] Step 1, analyze the "unit-station-system" hierarchical characteristics of the coupled system, and determine the hierarchical voltage control structure of the coupled system; the specific process is as follows:

[0098] The upper MPC control system includes traditional thermal power units, reactive power compensation devices, and upper control centers. The upper control center connects the HMPC controller through the coupled system grid connection point and inputs the coupled system grid connection point reference voltage instruction to the HMPC controller. The traditional thermal power units and the reactive power compensation devices are connected to the HMPC controller and receive the coupled system grid connection point reference voltage instruction through the HMPC controller. The lower MPC control system is divided into multiple sub-station systems controlled by LMPC controllers according to the types of new energy stations, including wind power stations and photovoltaic power stations. The HMPC controller connects the LMPC controller through the station grid connection point and sends the station grid connection point voltage reference instruction to the LMPC controller. The LMPC controller directly sends the station grid connection point voltage reference instruction to each new energy station, and the HMPC controller receives the remaining adjustable reactive power range of each new energy station.

[0099] The present application adopts the hierarchical voltage control structure of the coupled system as shown in Figure 2 The lower system is divided into sub-station systems according to different new energy stations. The upper system and the lower sub-station system need to interact through the transmission of station grid connection point voltage information, and the consistency constraint condition is as follows:

[0100]

[0101] In formula (1), is the reference voltage instruction of the wind farm point of common coupling (PCC) issued by the HMPC controller, is the reference voltage instruction of the photovoltaic (PV) plant PCC issued by the HMPC controller, is the reference voltage instruction of the wind farm PCC issued by the LMPC controller, is the reference voltage instruction of the PV plant PCC issued by the LMPC controller.

[0102] Step 2, considering the operation loss optimization and voltage control requirements of various PCCs, considering various reactive power sources in the coupling system, a hierarchical collaborative control mathematical model based on MPC-ADMM is constructed for the dual-mode adaptive switching of the loss optimization control and voltage correction control; the specific process is as follows:

[0103] In the control target voltage setting, the operation loss optimization and various PCC voltage control requirements are considered, and two control modes are designed: voltage correction control and normal loss control. The two control modes realize adaptive switching of the control state by judging the operating state of the coupling system PCC voltage. The target function and prediction model of the upper MPC control system and the lower MPC control system are established in the two modes, that is, the hierarchical collaborative control mathematical model.

[0104] The specific process of adaptive switching of the control state by the two control modes through judging the operating state of the coupling system PCC voltage is as follows:

[0105] The coupling system PCC is regarded as a node i, and a voltage safety range is set. The controller control switching logic rule is as follows:

[0106] ① When the voltage of node i is within the preset voltage safety range, that is, ||U i -U i,ref ||≤U i,th , the HMPC controller or the LMPC controller is in the normal loss control mode, and the system loss is optimized under the condition of ensuring voltage safety constraints;

[0107] ② When the voltage of node i is outside the set voltage safety range, that is, ||U i -U i,ref ||>U i,th , the HMPC controller or the LMPC controller switches to the voltage correction control mode to minimize the voltage deviation of node i;

[0108] Wherein, U i represents the voltage of node i, U i,ref represents the corresponding node voltage reference value (usually set to 1.0 p.u.); U i,th represents the voltage threshold of node i.

[0109] The specific process of establishing the objective function and prediction model of the upper MPC control system and the lower MPC control system in two modes is as follows:

[0110] (1) For the control model of the upper MPC control system:

[0111] ① When in normal loss control, if the key point voltage of the upper system is within the preset range, that is: The HMPC controller switches to the normal loss control mode; in this control mode, the active power loss and the control cost of the system are taken as the optimization objectives; based on the ADMM optimization algorithm, the objective function is represented as:

[0112]

[0113] In formula (2), ω Ploss >0, ω cost >0 represent the weight coefficients of the system loss and the control cost of the upper system, λ WF , λ PV are Lagrange multipliers, ρ WF , ρ PV are penalty factors, and Nc represents the control step length in the model predictive control algorithm, wherein the prediction step length N p = N c = 3.

[0114] The active power loss of the system is affected by the generator terminal voltage U g , the output power P g , and the reactive power provided by the wind and light new energy stations to the upper MPC control system, so the linear prediction model of the active power loss of the system is:

[0115]

[0116] In formula (3), P N,WF is the measured value of the active power loss of the system at the current time, and is a column vector composed of control variables at future k time points; is the sensitivity matrix of the active power loss of the system to the upper control variables;

[0117] The reactive power provided by the new energy station can be represented as:

[0118]

[0119] In formula (4), ΔQ WF,SVC , ΔQ PV,SVC are represented as:

[0120] ​

[0121] In formula (4), S N,WF , S N,PV is the rated capacity of the fan, the photovoltaic unit, P m,WF , P n,pv is the output size of the mth and nth unit groups in the wind farm and photovoltaic power station at the current moment; N WF , N PV corresponds to the number of unit groups in the wind farm and photovoltaic power station;

[0122] In formula (5), Q N,SVC is the rated capacity of the SVC; Q WF,SVC , Q PV,SVC respectively represent the reactive power consumed by the lower layer wind farm and photovoltaic power station;

[0123] ② When the upper layer system key grid-connected point voltage exceeds the voltage preset range, i.e.: the HMPC controller turns into the voltage correction control mode; in this operation mode, the controller target function is:

[0124]

[0125] In formula (6), V is the predicted value of the voltage of the key node i of the upper layer system; similarly affected by the generator reference voltage, output and the reactive power that can be provided by the new energy field station, the prediction model for the voltage of the key node i of the upper layer is:

[0126]

[0127] wherein, is the measured value of the voltage of the node i of the upper layer system at the current moment, is the sensitivity matrix of the voltage of the node i of the upper layer to the control variable;

[0128] The target functions (2) and (6) of the upper layer MPC control system satisfy the following constraint conditions:

[0129]

[0130] In addition to satisfying the above inequality constraints (8) about the voltage and the range of the control variable, the target function of the upper layer MPC control system also needs to satisfy the equality constraints in formula (3) and (6) about the system loss and the voltage of the key node of the upper layer;

[0131] (2) For the control model of the lower layer MPC control system

[0132] Following the setting of the objective function of the upper-level optimization control model, the control objective function of the new energy station also sets two control modes: corrective voltage control and normal network loss control;

[0133] ① In normal grid loss control, if the voltages of key grid connection points of each new energy station in the lower layer system are within the preset range, that is: The LMPC controller operates in normal control mode, with the goal of ensuring the economic operation of the new energy station, taking into account the station's active network loss and the minimum control cost; among them, the node voltage reference value in the lower MPC control system Output results for the HMPC controller;

[0134] The objective function of the wind farm is:

[0135]

[0136] Photovoltaic power station:

[0137]

[0138] The network loss of the lower station is affected by the reactive power Q of the new energy unit and the static VAR compensation device. WFGU (Q PVGU ), Q SVC The linearized prediction model for active network loss of wind and solar power stations is:

[0139]

[0140]

[0141] In formulas (11) and (12), Corresponding to the measured value of the active network loss of the wind and solar power stations at the current moment; Δu WF (k) = [ΔQ WFGU (k),ΔQ SVC (k)] T is the column vector consisting of the control variables of the wind farm station at time k in the future; Δu PV (k) = [ΔQ PVGU (k),ΔQ SVC (k)] T is a column vector consisting of control variables corresponding to the optical field station at time k in the future; Sensitivity matrix of active network loss of wind and solar power stations to each control variable;

[0142] ② During voltage correction control, when the voltage at the key grid connection point of the lower-level new energy station exceeds the preset range: LMPC controller turns to correction control mode; considering the voltage safety as the target, the grid-connected point i voltage and the corresponding reference voltage offset and the minimum control cost, the controller objective function is:

[0143] The wind farm objective function is:

[0144]

[0145] The photovoltaic power station objective function is:

[0146]

[0147] In formula (13)-(14), The corresponding k moment wind, light field station in the key grid-connected point i voltage prediction value, also affected by the new energy unit and static var compensation device reactive power; about the lower MPC control system in the key node i voltage prediction model is expressed as:

[0148]

[0149]

[0150] Wherein, The wind, light field station node i voltage in the current moment measured value; The lower wind, light field station node voltage sensitivity matrix to control variables;

[0151] The lower MPC control system in the new energy field station objective function (9), (10) and (13), (14) meet the

[0152] The lower voltage, control variable inequality constraint:

[0153]

[0154]

[0155] In addition to satisfying the inequality constraint, the lower MPC control system in the new energy field station objective function also needs to satisfy the equation constraint of the lower field station network loss, node voltage in formula (11), (12) and (15), (16).

[0156] Step 3, the ADMM algorithm is used to optimize the parameters in the hierarchical collaborative control mathematical model, and the grid-connected point voltage control result is calculated. The specific process is as follows:

[0157] Step 3.1: set the power daily fluctuation scene, the time scale is 15 min, and the scene cycle iteration flag m=m+1;

[0158] Step 3.2: initialize ADMM algorithm parameters, algorithm multipliers and system internal control variables, lower-level field station and grid-connected point voltage Set iteration flag k = 0;

[0159] ADMM algorithm parameters include Lagrange multiplier lambda, penalty factor rho.

[0160] System internal control variables include: upper level: synchronous machine terminal voltage Vref, output power Pg, static reactive power compensation device provides reactive power, new energy field station provides reactive power lower level: static reactive power compensation device provides reactive power, new energy unit machine group provides reactive power.

[0161] Step 3.3: let k = k + 1, after the upper-level MPC control system accepts the reference voltage instruction of the grid-connected point from the upper control center, judge the key grid-connected point voltage state of the upper-level MPC control system, the HMPC controller switches to the corresponding control mode, and solves the upper-level field station grid-connected point voltage controlled by the upper-level MPC control system as the corresponding new energy field station grid-connected point voltage control instruction, which is passed to the lower-level corresponding field station LMPC controller;

[0162] Step 3.4: after the lower-level MPC control system accepts the field station grid-connected point voltage reference instruction from the upper-level MPC control system, according to the state of the lower-level wind and light field station grid-connected point voltage, select the corresponding LMPC controller control mode, respectively solve the lower-level field station grid-connected point voltage, and upload the remaining controllable reactive power range of the new energy field station to the HMPC controller system;

[0163] Step 3.5: judge whether the residual error between the wind and light field station grid-connected point voltage of the HMPC control system and the corresponding LMPC control system field station grid-connected point voltage satisfies the set tolerance limit, if yes, output the system node voltage control result, update the algorithm multiplier and return to step 3.1, otherwise, update the algorithm multiplier and return to step 3.2.

[0164] The update formula of the algorithm multiplier is:

[0165]

[0166] In the formula, μ, σ are adjustable parameters, wherein μ > 0, τ inc > 1.0, τ dec > 1.0.

[0167] Example 1:

[0168] Modified IEEE-14 node system, structure as Figure 3The modified IEEE-14 node system is composed of thermal power units and wind power and photovoltaic power stations, and is connected to form a coupling system through different grid connection points. The original nodes 9 and 14 of the IEEE-14 node system are set as wind power and photovoltaic grid connection points, and the internal structure of the station is as shown in Figure 3 (b) and (c), and the station output is designed based on the original load data. The wind power and photovoltaic stations are each composed of 8 wind turbines and photovoltaic units with a rated capacity of 6.25 MW and 5 MW, connected to the 13.8 kV power grid through a transformer, and an SVC is installed at each new energy station grid connection point as a reactive power compensation device. Nodes 35 and 36 are set as external grid nodes based on nodes 12 and 13, and nodes 12 and 13 are improved as IEEE-14 node system grid connection points.

[0169] The internal load nodes of the modified IEEE-14 node system are set according to the load fluctuation of a certain area in Nanjing, and the new energy power fluctuation is set according to the fluctuation of a certain area. The power fluctuation is as shown in Figure 4 According to the voltage requirements of the new energy grid connection and the voltage requirements of the IEEE-14 node system grid connection, the voltage range of the system grid connection point and the new energy grid connection safety voltage is set to [0.99, 1.01] (p.u.) and [0.97, 1.07] (p.u.) respectively. It can be seen from Figure 5 that under the action of new energy power and load power, the voltage at the IEEE-14 node coupling system grid connection points 1 and 2 (nodes 12 and 13) exceeds the upper limit within 24 hours. For the station grid connection points (nodes 9 and 14), the photovoltaic grid connection point has a small new energy output and a large load demand in the time period of 10:00-11:00 and 16:00-21:00, and the voltage exceeds the lower limit; the voltage of the wind power grid connection point meets the grid connection requirements within 24 hours. For the grid connection points of the units inside the station, the voltage of the wind power unit grid connection point exceeds the upper limit in the time period of 00:00-01:00, and exceeds the lower limit in the time period of 17:00-19:00; the photovoltaic unit has a large new energy output and a light load in the time period of 11:00-13:00, and the voltage exceeds the upper limit. Therefore, under the influence of random fluctuations of new energy and load power, the voltage of the key nodes of the IEEE-14 node system exceeds the preset grid connection voltage safety range, which seriously affects the voltage safety of the modified IEEE-14 node system.

[0170] In view of the problem that the node voltage exceeds the grid connection voltage requirement under the power fluctuation scenario, the control strategy proposed in the application is applied to the IEEE-14 node coupling system. The internal control resource limit of the system is set as: U g = [0.95, 1.10] (p.u.), P g = [0, 2.0] (p.u.), and Q N,svc= [ -0.50, 0.50] (p.u.). The parameters in the ADMM algorithm are set as λ WF = 0, ρ PV = 0.0001; the original residual margin ε w = 0.05; the penalty factor updating parameters are set as σ = 10, μ = 2. s pri dual

[0171] It can be seen from Figure 6 that, under the power daily fluctuation scenario, the voltages of the system interconnection points 1 and 2 and the new energy station interconnection point are effectively controlled within the safe operation range after the layered cooperative control of the applied voltage is performed on the IEEE-14 node coupling system; for the new energy unit interconnection point, the voltage still exceeds the limit due to the limitation of the internal control resources of the system at a certain period in the control process. It can be seen from the comparison of the active power loss optimization results in Figure 7 that the active power loss of the IEEE-14 node coupling system is reduced after the control, meeting the economy of system operation; but the internal active power loss of the new energy station is higher than that before the control at some periods after the control strategy is applied. This is because the voltage level of the unit interconnection point in the station is high before the control strategy is applied, and the voltage level is low after the control strategy is applied. In the voltage over-limit scenario, the system voltage safety requirement needs to be ensured first, and the system is controlled to correct the voltage, so that the voltage of the corresponding node is reduced to the safe range, and thus the active power loss of the new energy station is higher at some periods.

[0172] Embodiment 2:

[0173] The practicability of the layered voltage cooperative control strategy of the application is further verified by taking a regional new energy and traditional power coupling system in Dalian as an example. The regional power grid has 1200 MW of 20 kV thermal power units, 300 MW of 0.7 kV wind power and 300 MW of 0.4 kV photovoltaic power, including voltage levels of 35 kV, 220 kV and 500 kV, and is connected to form a coupling system at the same interconnection point, and contains the voltage levels of the transmission and distribution networks. The SVC is installed at the new energy station interconnection point as a reactive power compensation device. The power grid structure is shown in Figure 8 , wherein node 1 is the coupling system interconnection point.

[0174] The new energy output fluctuation of a local coupling system in Liaoning is set as the new energy fluctuation, and the voltage fluctuation of the key nodes of the system before the control is shown in Figure 9 , 10 . It can be seen from Figure 10 that, under the power daily fluctuation scenario, the voltages of the system interconnection points 1 and 2 and the new energy station interconnection point are effectively controlled within the safe operation range after the layered cooperative control of the applied voltage is performed on the IEEE-14 node coupling system; for the new energy unit interconnection point, the voltage still exceeds the limit due to the limitation of the internal control resources of the system at a certain period in the control process. It can be seen from the comparison of the active power loss optimization results in Figure 7 that the active power loss of the IEEE-14 node coupling system is reduced after the control, meeting the economy of system operation; but the internal active power loss of the new energy station is higher than that before the control at some periods after the control strategy is applied. This is because the voltage level of the unit interconnection point in the station is high before the control strategy is applied, and the voltage level is low after the control strategy is applied. In the voltage over-limit scenario, the system voltage safety requirement needs to be ensured first, and the system is controlled to correct the voltage, so that the voltage of the corresponding node is reduced to the safe range, and thus the active power loss of the new energy station is higher at some periods.

[0172] Embodiment 2:

[0173] The practicability of the layered voltage cooperative control strategy of the application is further verified by taking a regional new energy and traditional power coupling system in Dalian as an example. The regional power grid has 1200 MW of 20 kV thermal power units, 300 MW of 0.7 kV wind power and 300 MW of 0.4 kV photovoltaic power, including voltage levels of 35 kV, 220 kV and 500 kV, and is connected to form a coupling system at the same interconnection point, and contains the voltage levels of the transmission and distribution networks. The SVC is installed at the new energy station interconnection point as a reactive power compensation device. The power grid structure is shown in Figure 8 , wherein node 1 is the coupling system interconnection point.

[0174] The new energy output fluctuation of a local coupling system in Liaoning is set as the new energy fluctuation, and the voltage fluctuation of the key nodes of the system before the control is shown in Figure 9 , 10 . It can be seen from Figure 10It can be seen that under the scenario of daily power fluctuations of renewable energy, the voltage at the grid-connected point of the coupled system exceeded the upper limit within 24 hours; although the voltage at the grid-connected point of wind power and photovoltaic stations met the voltage requirements for renewable energy grid connection, the voltage at the grid-connected point of the internal wind turbines and photovoltaic units exceeded the upper limit during the periods of 21:00-22:00 and 7:00-16:00, respectively, affecting the voltage safety of the system operation.

[0175] Aiming at the voltage over-limit problem under the scenario of new energy power fluctuation in the coupled system in Liaoning Province, the present invention proposes a reactive voltage hierarchical coordinated control strategy, and the voltage safety range of key nodes is set according to the node voltage safety range of the IEEE-14 node coupled system. The internal control resource limit of the coupled system in Liaoning Province is set as: U g =[0.95,1.05](pu),P g =[0.0,2.0](pu),Q N,svc =[-1.50,1.50](pu). The parameters in the ADMM algorithm are set to λ WF ,λ PV =0,ρ w , ρ s =0.0001; original residual margin ε pri , dual residual margin ε dual are 0.05 respectively; the penalty factor update parameters are set to σ=10 and μ=2.

[0176] Depend on Figure 11 It can be seen that in the power fluctuation scenario, after control, the voltage of the system grid connection point, the station grid connection point and the unit grid connection point are effectively controlled within the preset safety range, while the active network loss of the coupled system is reduced. Figure 12 It can be seen that the reason why the active power loss of the station increases in certain periods after control is the same as the reason for the increase in the power loss of the new energy station within the IEEE-14 node coupling system.

[0177] Through simulation analysis of the modified IEEE-14 node system and a local coupled system in Dalian, it can be seen that the hierarchical collaborative control strategy with dual-mode adaptive switching of voltage correction control and network loss optimization control based on MPC-ADMM proposed in this invention not only achieves a reasonable distribution of voltages among new energy units, stations, and system grid connection points within the coupled system, meeting the grid-connected voltage safety requirements of the coupled system, but also optimizes the active network losses during system operation and improves the economy of the coupled system.

Claims

1. A voltage hierarchical coordinated control strategy for coupled systems based on MPC-ADMM, characterized by: The coupling system includes an upper-layer MPC control system and a lower-layer MPC control system. The control of the upper-layer MPC control system and the lower-layer MPC control system satisfies safety constraints and is specifically implemented according to the following steps: Step 1: Analyze the "unit-station-system" hierarchical characteristics of the coupled system and determine the hierarchical voltage control structure of the coupled system; Step 2: Taking into account both the network loss optimization of the coupled system and the voltage control requirements of various grid connection points, and considering various reactive source resources within the coupled system, a hierarchical collaborative control mathematical model with dual-mode adaptive switching of network loss optimization control and voltage correction control based on MPC-ADMM is constructed; Step 3: Use the ADMM algorithm to optimize the parameters in the hierarchical cooperative control mathematical model and calculate the voltage control result of the grid connection point; The specific process of step 3 is as follows: Step 3.1: Set the power daily fluctuation scenario, the time scale is 15 minutes, and the scenario loop iteration flag m = m +1; Step 3.2: Initialize ADMM algorithm parameters, algorithm multipliers, system internal control variables, and lower-level station grid connection point voltage U L WF, U L PV, set the iteration flag k =0; Step 3.3: Make k = k+ 1. After receiving the grid connection point reference voltage instruction from the upper control center, the upper-level MPC control system determines the voltage status of the key grid connection points of the upper-level MPC control system. The HMPC controller switches to the corresponding control mode and solves the upper-level station grid connection point voltage controlled by the upper-level MPC control system as the corresponding new energy station grid connection point voltage control instruction, which is then passed to the lower-level corresponding station LMPC controller; Step 3.4: After receiving the voltage reference instruction of the station grid connection point from the upper-level MPC control system, the lower-level MPC control system selects the corresponding LMPC controller control mode according to the voltage status of the grid connection point of the lower-level wind and solar stations, solves the grid connection point voltage of the lower-level stations respectively, and uploads the remaining controllable reactive power range of the new energy station to the HMPC controller system; Step 3.5: Determine the voltage of the wind and solar power stations in the HMPC control system. U H WF, U Voltage at the grid connection point between HPV and the corresponding LMPC control system U L WF, U Whether the residual between L PV meets the set tolerance limit, if so, output the system node voltage control result, update the algorithm multiplier and return to step 3.1, otherwise, update the algorithm multiplier and return to step 3.

2.

2. The MPC-ADMM-based voltage hierarchical coordinated control strategy for coupled systems according to claim 1 is characterized in that: The specific process of step 1 is as follows: the upper-level MPC control system includes a traditional thermal power unit, a reactive power compensation device, and an upper-level control center. The upper-level control center controls the HMPC controller through the coupling system grid connection point and inputs the coupling system grid connection point reference voltage instruction to the HMPC controller. The traditional thermal power unit and the reactive power compensation device are both connected to the HMPC controller and receive the coupling system grid connection point reference voltage instruction through the HMPC controller; the lower-level MPC control system is decomposed into multiple sub-station systems controlled by LMPC controllers according to the type of new energy station. The HMPC controller is connected to the LMPC controller through the grid connection point of each station and issues the grid connection point voltage reference instruction of each station to the LMPC controller. The LMPC controller directly issues the grid connection point voltage reference instruction of each station to each new energy station, and the HMPC controller receives the remaining adjustable reactive power range of each new energy station.

3. The MPC-ADMM-based voltage hierarchical coordinated control strategy for coupled systems according to claim 2 is characterized in that: The types of new energy sites include wind farms and photovoltaic farms.

4. The MPC-ADMM-based voltage hierarchical coordinated control strategy for coupled systems according to claim 2 is characterized in that: The safety constraints are expressed as: （1） In formula (1), It is the reference voltage instruction sent by the HMPC controller to the wind farm grid connection point. It is the reference voltage instruction sent by the HMPC controller to the photovoltaic power station grid connection point. It is the reference voltage instruction sent by the LMPC controller to the wind farm grid connection point. It is the reference voltage instruction sent by the LMPC controller to the grid connection point of the photovoltaic power station.

5. The MPC-ADMM-based voltage hierarchical coordinated control strategy for coupled systems according to claim 2 is characterized in that: The specific process of step 2 is: Based on the control target voltage setting, network loss optimization and various grid connection point voltage control requirements are implemented, and two control modes are designed: voltage correction control and normal network loss control. The two control modes realize adaptive switching of control states by judging the operating state of the voltage at the grid connection point of the coupling system; and the objective functions and prediction models of the upper-level MPC control system and the lower-level MPC control system are established in the two modes, which is the hierarchical collaborative control mathematical model.

6. The MPC-ADMM-based voltage hierarchical coordinated control strategy for coupled systems according to claim 5 is characterized in that: The two control modes achieve adaptive switching of control states by judging the operating state of the voltage at the grid connection point of the coupling system. The specific process is as follows: The grid connection point of the coupled system is regarded as a node i , and set the voltage safety range. The controller control switching logic rules are as follows: ①Node i When the voltage is within the preset voltage safety range, that is: ||U i -U i,ref || ≤ U i,th , the HMPC controller or LMPC controller is in normal network loss control mode, which optimizes the system network loss under the condition of ensuring voltage safety constraints; ②Node i The voltage is outside the set voltage safety range, that is, ||U i -U i,ref || >U i,th , the HMPC controller or LMPC controller switches to the voltage correction control mode, so that the node i Minimum voltage deviation; in, U i Representation node i Voltage, U i,ref It represents the corresponding node voltage reference value; U i,th Representation node i Voltage threshold.

7. The MPC-ADMM-based voltage hierarchical coordinated control strategy for coupled systems according to claim 6 is characterized in that: The specific process of establishing the objective function and prediction model of the upper-level MPC control system and the lower-level MPC control system in two modes is as follows: (1) For the control model of the upper-level MPC control system: ① In normal grid loss control, if the voltage of key grid-connected points of the upper system is within the preset range, that is: || U H i - U H i ,ref|| ≤ U H i ,th, the HMPC controller switches to the normal network loss control mode; in this control mode, the system active network loss and control cost are taken as optimization targets; Based on the ADMM optimization algorithm, the objective function is expressed as: （2） In formula (2), ω Ploss > 0, ω cost > 0 represent the weight coefficients of system network loss and upper system control cost, λ WF ,λ PV are all Lagrange multipliers, ρ WF , ρ PV are penalty factors, Nc Represents the control step size in the model predictive control algorithm, where the prediction step size N p = N c =3; System active network loss affected by generator terminal voltage U g , output power P g Wind and solar energy stations provide reactive power to the upper MPC control system Q WF sta、 Q Considering the influence of PV sta, the linearized prediction model of system active network loss is: （3） In formula (3), P H loss (0) is the measured value of the system active network loss at the current moment, ∆ u H ( k )=[∆ U g( k ) ,∆ P g( k ) ,∆ Q WF sta( k ) ,∆ Q PV sta( k )] T For the future k A column vector of moment control variables; ∂ P H loss / ∂ u H is the sensitivity matrix of the system active network loss to the upper control variables; The reactive power that a new energy station can provide can be expressed as: （4） In formula (4), ∆ Q WF,SVC ,∆ Q PV,SVC Expressed as: （5） In formula (4), S N,WF 、 S N,PV is the rated capacity of wind turbine and photovoltaic unit, P m,WF 、 P n,pv The current moment in the wind farm and photovoltaic power station m and n The output of each unit; N WF , N PV The number of units corresponding to wind farms and photovoltaic power stations; In formula (5), Q N,SVC is the rated capacity of the SVC; Q WF,SVC 、 Q PV,SVC They represent the reactive power consumed by the reactive compensation equipment of the lower-level wind farm and photovoltaic power station respectively; ② In voltage correction control, when the voltage of the key grid connection point of the upper system exceeds the voltage preset range, that is: || U H i - U H i ,ref|| ≥ U H i ,th, the HMPC controller switches to voltage correction control mode; in this operating mode, considering the offset between the grid connection point voltage and the corresponding reference voltage and the control cost, the controller objective function is: （6） In formula (6), U H i ( k ) is the key node of the upper system i Voltage prediction value; same Affected by the generator reference voltage, output and reactive power that can be provided by the new energy station, about the upper key nodes i The voltage prediction model is: （7） in, U H i (0) is the upper system node i The voltage measured at the current moment, ∂ U H i / ∂ u H For the upper node i Sensitivity matrix of voltage to control variables; The objective functions (2) and (6) of the upper-level MPC control system satisfy the following constraints: （8） In addition to satisfying the above-mentioned inequality constraints (8) on voltage and control variable range, the objective function of the upper-level MPC control system must also satisfy the equality constraints on system network loss and upper-level key node voltage in equations (3) and (6); (2) Control model for the lower-level MPC control system Following the setting of the objective function of the upper-level optimization control model, the control objective function of the new energy station also sets two control modes: corrective voltage control and normal network loss control; ① In normal grid loss control, if the voltage of key grid connection points of each new energy station in the lower layer system is within the preset range, that is: || U L i - U L i ,ref|| ≤ U L i ,th, the LMPC controller operates in the normal control mode, with the goal of ensuring the economic operation of the new energy station, considering the station's active network loss and the minimum control cost; among them, the node voltage reference value in the lower MPC control system U L i ,ref is the output result of HMPC controller; The objective function of the wind farm is: （9） Photovoltaic power station: （10） The network loss of the lower station is affected by the reactive power of the new energy unit and the static VAR compensation device. Q WFGU ( Q PVGU )、 Q SVC The linearized prediction model for active network loss of wind and solar power stations is: （11） （12） In formulas (11) and (12), P WF loss(0), P PV loss(0) corresponds to the measured value of the active power loss of the wind and solar power stations at the current moment; ∆ u WF ( k )=[∆ Q WFGU (k),∆ Q SVC ( k )] T It corresponds to the future k The column vector of control variables of the wind farm station at time ∆ u PV ( k )=[∆ Q PVGU ( k ),∆ Q SVC ( k )] T It corresponds to the future k The optical field station at the moment is a column vector consisting of control variables; ∂ P WF loss / ∂ u WF , ∂ P PV loss / ∂ u PV Sensitivity matrix of active network loss of wind and solar power stations to each control variable; ② During voltage correction control, when the voltage at the key grid connection point of the lower-level new energy station exceeds the preset range, that is: U L i - U L i ,ref|| ≥ U L i ,th, LMPC controller switches to correction control mode; with voltage safety as the goal, considering the grid connection point i The voltage offset from the corresponding reference voltage and the control cost are minimized, and the controller objective function is: The objective function of the wind farm is: （13） The objective function of the photovoltaic power station is: （14） In formulas (13) and (14), U WF i ( k ), U PV i ( k )correspond k Key grid connection points within wind and solar power stations at all times i The voltage prediction value is also affected by the reactive power of the new energy unit and the static VAR compensation device; about the key nodes in the lower MPC control system i The voltage prediction model is expressed as: （15） （16） in, U WF i (0), U PV i (0) is the wind and solar station node i The voltage measured at the current moment; ∂ U WF i / ∂ u WF , ∂ U PV i / ∂ u PV is the sensitivity matrix of the node voltage of the lower-level wind and solar field stations to the control variables; The objective functions (9), (10), (13), and (14) of the new energy station in the lower-level MPC control system satisfy the following inequality constraints on voltage and control variables: （17） （18） In addition to satisfying the inequality constraints, the objective function of the new energy station in the lower-level MPC control system must also satisfy the equality constraints on the network loss and node voltage of the lower-level station in Equations (11), (12), (15), and (16).

8. The MPC-ADMM-based voltage hierarchical coordinated control strategy for coupled systems according to claim 7 is characterized in that: The update formula of the update algorithm multiplier is: （19） Where, μ , σ is an adjustable parameter, where μ >0, τ inc >1.0, τ dec >1.0.

Images