Power distribution network station area inverter switching control method and system based on double-layer MPC

Abstract

The invention discloses a double-layer MPC-based power distribution network area inverter switching control method and system, and the method comprises the steps: obtaining the operation parameters and equipment parameters of an area power distribution network, and building a voltage source type inverter system model considering the frequency fluctuation; constructing an outer-layer model, and generating an output voltage reference value of an inverter grid-connected mode and an output voltage reference value of a grid construction mode according to the operation parameters and the equipment parameters; calculating a synchronous decision signal, and switching the working mode of the inverter; calculating a parallel system bus voltage reference value and an output current reference value of the current sampling period; and constructing an inner-layer model, calculating a predicted value and a reference value of voltage and current in a next sampling period, constructing a value function cost optimization model according to a deviation degree between the predicted value and the reference value, generating an optimal switching state, and controlling the power distribution network area inverter. Dynamic response and stability are both considered, seamless switching of the voltage source type inverter is achieved, meanwhile, voltage out-of-limit is avoided, and the stability of a power system is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power quality control of power systems, and more specifically, relates to a distribution network substation inverter switching control method and system based on a double-layer MPC. Background Art

[0002] As the penetration of power electronics continues to increase in the power system, the existing grid has undergone substantial changes, seriously affecting voltage quality. Low-voltage distribution network loads exhibit significant regional, seasonal, and time-varying characteristics. The addition of large amounts of intermittent loads can lead to reduced voltage levels and, in severe cases, even voltage overshoots.

[0003] Grid-connected and off-grid operating modes are key to demonstrating the technical and economic advantages of microgrids. Therefore, seamless converter switching technology is crucial for ensuring the stable operation of a coordinated PV-storage distribution network. A grid-connected converter can be externally equivalent to a low-impedance, controllable voltage source connected in series. Due to its voltage source characteristics, instantaneous voltage differences can cause a sudden increase in the converter's output current, leading to overcurrent in the grid-connected converter and, in severe cases, even damage to the grid-connected device.

[0004] When the grid-side voltage drops under the traditional grid-type control strategy, the grid-type converter will have output overcurrent, power oscillation, and be unable to provide reactive power support to the grid in a timely manner, affecting the stable operation of the power system.

[0005] Authorized prior art document 1 (CN113991748B) discloses a seamless switching control method for a voltage source inverter based on droop control. However, prior art document 1 has shortcomings: limited by the time constant of the low-pass filter, it has a slow dynamic response speed; it is sensitive to grid impedance, has low system robustness, and is susceptible to interference in weak grid / island microgrid scenarios with high penetration of new energy.

[0006] The paper "Model Predictive Control Strategy for Voltage Source Three-Phase Inverters" discloses an MPC (Model Predictive Control) controller for a three-phase inverter. However, its shortcomings are as follows: the MPC controller in the paper is a single-layer MPC, centered on current tracking and optimizing a single voltage vector; the dynamic response time is long, making it only applicable to steady-state scenarios and unable to meet the requirements of new power systems for rapid frequency and voltage regulation; and the lack of a mode switching mechanism makes it difficult to adapt to the frequent switching requirements of photovoltaic-energy storage hybrid stations. Summary of the Invention

[0007] In order to solve the shortcomings of the existing technology, the present invention provides a seamless switching control method for voltage source inverters in distribution network substations based on double-layer model predictive control to adapt to the requirements of new power system construction, improve the voltage over-limit problem of traditional control methods, and realize seamless and smooth switching of voltage source inverters in distribution network substations coordinated by photovoltaic and energy storage.

[0008] The present invention adopts the following technical solutions.

[0009] A first aspect of the present invention provides a method for controlling inverter switching in a distribution network area based on a double-layer MPC, comprising the following steps:

[0010] Step 1: Obtain the operating parameters and equipment parameters of the distribution network in the substation area and establish a voltage source inverter system model that takes frequency fluctuations into account;

[0011] Step 2: construct an outer model; based on the operating parameters and device parameters, the outer model generates an output voltage reference value of the inverter grid-connected mode and an output voltage reference value of the grid-forming mode;

[0012] Step 3: Calculate a synchronization decision signal based on the deviation of the output voltage reference values of the grid-connected mode and the grid-forming mode, and switch the inverter's operating mode according to the synchronization decision signal; and calculate the system bus voltage reference value and the inverter output current reference value for the current sampling period by the outer model based on the synchronization decision signal and the output voltage reference value of the grid-connected mode.

[0013] Step 4, construct an inner model; based on the reference value of the bus voltage of the system and the reference value of the output current of the inverter in the current sampling period, the inner model calculates the predicted value and reference value of the bus voltage of the system and the output current of the inverter in the next sampling period; based on the degree of deviation between the predicted value and the reference value, a value function cost optimization model is constructed to generate the optimal switching state, and a pulse signal is generated according to the optimal switching state to control the inverter in the distribution network substation.

[0014] Preferably, in step 1, the system model of the voltage source inverter is as follows:

[0015]

[0016] Where i od 、i oq are the d-axis and q-axis components of the inverter output current, I od , I oq i od 、i oq The steady-state value, v Ld 、v Lq are the d-axis and q-axis components of the inverter-integrated system bus voltage, V Ld 、V Lq vLd 、v Lq The steady-state value, v od 、v oq are the d-axis and q-axis components of the inverter output voltage, i Sd 、i Sq are the d-axis and q-axis components of the AC grid current, i Ld 、i Lq are the d-axis and q-axis components of the inverter-integrated system bus current, ω is the grid angular frequency, ω N is the steady-state value of ω, L V and C V are the filter inductance and filter capacitance of the LC filter, R V is the equivalent resistance.

[0017] Preferably, in step 2, the output voltage reference value includes: an amplitude reference value, a frequency reference value, and a phase reference value;

[0018] Among them, the reference value of the inverter output voltage amplitude in grid-connected mode is and frequency reference value f k Keep consistent with the rated voltage and rated frequency of the AC power grid respectively;

[0019] The output voltage phase θ on both sides of the parallel switch of the voltage source inverter is collected through the phase-locked loop o , as the phase reference value of the inverter output voltage in grid-connected mode.

[0020] Preferably, in step 2, generating an output voltage reference value of the networking mode includes:

[0021] According to the inverter droop control equation, the initial value ω2 of the output voltage angular frequency reference value and the initial value V of the amplitude reference value in the grid-forming mode are calculated. o ,as follows:

[0022]

[0023] Where, ω n =2πf n , f n is the frequency reference value, V n is the reference value of the voltage, m and n are the Pf and QV droop coefficients respectively, p and q are the instantaneous active power and instantaneous reactive power output by the inverter respectively, P ref and Q ref are the average active power and average reactive power respectively;

[0024] A recursive formula is constructed to use the reference value of the previous sampling period to calculate the reference value of the current sampling period, and the prediction model of the output voltage amplitude reference value and angular frequency reference value in the networking mode is obtained as follows:

[0025]

[0026] Where, are the output voltage angular frequency reference values in the kth and k-1th sampling period networking mode, respectively. are the d-axis components of the output voltage amplitude reference values in the k-th and k-1-th sampling period networking modes, respectively. are the q-axis components of the output voltage amplitude reference values in the kth and k-1th sampling period networking mode, T s is the sampling period, p k-1 is the instantaneous active power output by the inverter in the k-1th sampling period, q k-1 is the instantaneous reactive power output by the inverter in the k-1th sampling period;

[0027] right and Perform dq / abc transformation to obtain the reference value of the inverter output voltage amplitude in the grid-forming mode according to Get the frequency reference value of the inverter output voltage in the grid-connected mode right Integrate to obtain the phase reference value θ of the inverter output voltage in the grid-forming mode L .

[0028] Preferably, in step 3, calculating the synchronization decision signal, and switching the working mode of the inverter according to the synchronization decision signal includes:

[0029] Calculate the synchronization decision signal l, which is expressed as follows:

[0030] u=[sign(|Δv|-0.8)+sign(|Δf|-0.02)+sign(|Δθ|-3)]

[0031]

[0032] Where |Δv| is the deviation of the output voltage amplitude reference value between the grid-connected mode and the grid-forming mode, |Δf| is the deviation of the frequency reference value of the output voltage between the grid-connected mode and the grid-forming mode |Δθ| is the deviation of the phase reference value of the output voltage between the grid-connected mode and the grid-forming mode, |Δθ|=|θ o -θ L |, θ L are the amplitude reference value, frequency reference value, and phase reference value of the inverter output voltage in the grid-forming mode, respectively. f k ,θo They are respectively the amplitude reference value, frequency reference value and phase reference value of the inverter output voltage in grid-connected mode;

[0033] When l=1, the inverter is switched to the grid-connected working mode; when l=0, the inverter is switched to the grid-building working mode.

[0034] Preferably, in step 3, the reference value of the bus voltage of the system is calculated and expressed as follows:

[0035]

[0036] In the formula, |v Lr |、ω Lr ,θ Lr are the amplitude reference value, angular frequency reference value, and phase reference value of the bus voltage incorporated into the system; the amplitude reference value|v Lr |, frequency reference value ω Lr , phase reference value θ Lr , which is incorporated into the system bus voltage reference value v Lr (k)=|v Lr |sin(2πω Lr +θ Lr ).

[0037] Preferably, in step 3, the output current reference value of the inverter is calculated and expressed as the following formula:

[0038]

[0039] Where i odr (k), i oqr (k) are the d-axis and q-axis components of the output current reference value of the inverter in the kth sampling period, are the d-axis component and q-axis component of the output voltage amplitude reference value of the k-th sampling period, respectively. ref is the average active power, Q ref is the average reactive power.

[0040] Preferably, in step 4, calculating the predicted values of the system bus voltage and the inverter output current to be incorporated in the next sampling period includes:

[0041] The system model of the voltage source inverter in the dq rotating coordinate system is discretized and substituted into the voltage basic vector to obtain the predicted values of the output current and the bus voltage of the system in the k+1th cycle, which are expressed as the following formula:

[0042]

[0043] Where, The inverter output current and the predicted value of the k+1th cycle are respectively incorporated into the system bus voltage predicted value, i ox (k), v Lx (k),u ox (k), i Sx (k), i Lx (k)(x∈{d,q}) are the inverter output current sampling value, the bus voltage sampling value, the voltage basic vector, the AC grid current, and the bus current sampling value in the kth cycle, respectively. sh is the state matrix, B sho is the input matrix, T s is the sampling period, and ω(k) is the angular frequency of the kth period.

[0044] Preferably, the state matrix and input matrix are as follows:

[0045]

[0046] Preferably, in step 4, calculating the reference values of the system bus voltage and the inverter output current to be incorporated into the next sampling period includes:

[0047] Based on the reference values of the inverter output current and the bus voltage connected to the system in the k-2th, k-1th, and kth cycles, the reference values of the bus voltage connected to the system and the inverter output current in the k+1th cycle are calculated and expressed as follows:

[0048]

[0049] Where i oxr (k+1), v Lxr (k+1)(x∈{d,q}) is the inverter output current reference value and the bus voltage reference value of the system in the k+1th cycle.

[0050] Preferably, in the value function cost optimization model constructed in step 4, the designed value evaluation function is as follows:

[0051]

[0052] Where g a is the evaluation function of the model prediction current error, g b is the evaluation function of the model prediction voltage error, |*| represents the Manhattan norm, is the predicted value of the inverter output current and the bus voltage of the connected system in the k+1th cycle, i oxr (k+1), v Lxr (k+1)(x∈{d,q}) is the inverter output current reference value and the bus voltage reference value of the connected system in the k+1th cycle;

[0053] The objective function of the designed inner model is expressed as follows:

[0054]

[0055] Where u o is the voltage vector, J is the weighted error, N is the prediction time domain defined by the prediction system, λ x and λ y are the first weight coefficient and the second weight coefficient respectively.

[0056] A second aspect of the present invention provides a distribution network substation inverter switching control system based on a double-layer MPC, comprising:

[0057] The parameter acquisition module is used to obtain the operating parameters of the distribution network and equipment parameters in the substation area and establish a voltage source inverter system model that takes frequency fluctuations into account;

[0058] The reference value calculation module is used to construct an outer model; based on the operating parameters and equipment parameters, the outer model generates the output voltage reference value of the inverter grid-connected mode and the output voltage reference value of the grid-forming mode;

[0059] A synchronization decision module is used to calculate a synchronization decision signal based on the deviation of the output voltage reference values of the grid-connected mode and the grid-forming mode, and switch the operating mode of the inverter according to the synchronization decision signal; based on the synchronization decision signal and the output voltage reference value of the grid-connected mode, the outer model calculates the bus voltage reference value of the current sampling period and the output current reference value of the inverter;

[0060] The cost optimization module is used to construct the inner model; based on the reference value of the bus voltage of the system and the reference value of the output current of the inverter in the current sampling period, the inner model calculates the predicted value and reference value of the bus voltage of the system and the output current of the inverter in the next sampling period; based on the degree of deviation between the predicted value and the reference value, a value function cost optimization model is constructed to generate the optimal switching state, and a pulse signal is generated according to the optimal switching state to control the inverter in the distribution network substation.

[0061] Compared with the prior art, the beneficial effects of the present invention include at least:

[0062] 1. The present invention adopts internal and external double-layer model predictive control, taking into account both dynamic response and stability, achieving seamless switching of voltage source inverters while avoiding voltage over-limit;

[0063] 2. The inner model of the present invention selects the optimal switching state of the inverter by solving the cost optimization model of the value function, with a short prediction step length and a dynamic response speed of 5 to 10 ms, thus achieving rapid tracking of the reference values of the load voltage and grid current;

[0064] 3. The outer model of the present invention predicts and corrects the reference value of the inverter output voltage in the grid-forming mode through the droop control strategy, and can control the voltage and frequency output of the inverter within the desired range even under interference;

[0065] 4. The present invention smoothly switches between the grid-connected mode and the grid-building mode based on synchronous decision signals, which can meet the frequent switching needs of photovoltaic-energy storage hybrid areas and improve the stability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic flow chart of a seamless switching control strategy for voltage source inverters in distribution network areas based on a two-layer model predictive control according to an embodiment of the present invention;

[0067] Figure 2 The overall control block diagram of the distribution area based on two-layer model predictive control for PV-storage coordination;

[0068] Figure 3 The flowchart for obtaining the synchronous decision signal. DETAILED DESCRIPTION

[0069] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0070] The present invention proposes a seamless switching control method and system for voltage source inverters in distribution network substations based on a two-layer model predictive control (MPC). Based on the coordinated control strategy for power outage recovery in the distribution network substation with photovoltaic and energy storage collaboration, the inner and outer layer MPC models of the inverter in the parallel grid mode are established with frequency deviation, phase deviation and voltage amplitude deviation as optimization targets.

[0071] like Figure 1 As shown, embodiment 1 of the present invention provides a method for seamless switching control of voltage source inverters in distribution network areas based on photovoltaic and energy storage coordination based on a two-layer model predictive control, comprising the following steps:

[0072] Step 1: Obtain the operating parameters and equipment parameters of the distribution network in the substation coordinated by photovoltaic and energy storage, and establish a voltage source inverter system model that takes frequency fluctuations into account.

[0073] The specific control process of the distribution area coordinated by photovoltaic and storage based on the two-layer model predictive control is as follows: Figure 2In a preferred but non-limiting embodiment of the present invention, step 1 specifically includes:

[0074] Step 1.1: Collect the distribution network operation parameters of the photovoltaic and energy storage coordinated substation, including the voltage source inverter switch output voltage v o And the output current i o , collect the voltage v of the voltage source inverter connected to the system bus L and current i L , measure the current i of the AC grid S ; Obtain the parameters of the distribution network equipment in the area coordinated by solar and energy storage, including the filter inductance value L of the LC filter V , capacitance value C V And the equivalent resistance value R V .

[0075] Step 1.2: Based on the basic data collected in step 1.1, Kirchhoff's voltage law is used to construct an average model of the voltage source inverter, which is expressed as the following formula:

[0076]

[0077] Where i o is the output current of the inverter, v o is the output voltage of the inverter; L V and C V are the filter inductance and filter capacitance of the LC filter, R V is the equivalent resistance; v L and i L Represent the voltage and current values of the busbar connected to the system, i S is the grid current.

[0078] In step 1.3, taking frequency variation into account, the Clark-Park transformation is applied to rewrite the average model into a system model of a voltage source inverter in the dq rotating coordinate system, expressed as follows:

[0079]

[0080] Where, the subscripts d and q represent the d-axis and q-axis components of the corresponding variables, respectively; I od , I oq Represents the current i od 、i oq The steady-state value, V Ld 、V Lq Represents voltage v Ld 、v Lq The steady-state value of ω is the grid angular frequency, ω N is the steady-state value of ω.

[0081] Step 2: Based on the voltage source inverter system model, an outer MPC model is constructed. The outer MPC model generates an output voltage reference value for the grid-connected mode according to the rated voltage and rated frequency of the AC power grid, and generates an output voltage reference value for the grid-forming mode according to the droop control. The output voltage reference value includes reference values for the amplitude, frequency, and phase of the output voltage.

[0082] In a preferred but non-limiting embodiment of the present invention, step 2 specifically comprises:

[0083] Step 2.1: Design the outer MPC model of the inverter in grid-connected mode to generate the reference value of the inverter output voltage in grid-connected mode, including the reference values of the output voltage amplitude, frequency, and phase. The specific principles are as follows:

[0084] In grid-connected mode, the amplitude and frequency of the output voltage of the voltage source inverter should be consistent with those of the AC grid.

[0085] In grid-connected mode, the amplitude reference value of the inverter's output voltage should track the grid's rated voltage as closely as possible.

[0086] In grid-connected mode, the frequency reference value of the inverter's output voltage should track the grid frequency as closely as possible.

[0087] In a further preferred embodiment of the present invention, according to the above principles, step 2.1 specifically includes:

[0088] Step 2.1.1, reference value of the inverter output voltage amplitude in grid-connected mode and frequency reference value f k They should be consistent with the rated voltage and rated frequency of the AC power grid respectively.

[0089] Step 2.1.2: Use the phase-locked loop to collect the output voltage phase θ of the parallel switches on both sides of the voltage source inverter. o , as the phase reference value of the inverter output voltage in grid-connected mode.

[0090] Step 2.2: Design the outer MPC model of the inverter in the grid-forming mode and generate the reference value of the inverter output voltage in the grid-forming mode, including the reference values of the output voltage amplitude, frequency, and phase.

[0091] In a further preferred embodiment of the present invention, droop control is used as the control strategy in the inverter grid mode. Droop control can achieve reasonable power distribution between inverters simply by detecting its own voltage and frequency parameters. Droop control can suppress power backflow and frequency mutations at the moment of mode switching of the voltage source inverter, achieving a smooth transition of mode switching. Step 2.2 specifically includes:

[0092] Step 2.2.1: In grid-building mode, calculate the instantaneous active power p and instantaneous reactive power q output by the inverter through the LC filter, expressed as follows:

[0093]

[0094] Step 2.2.2: The instantaneous active power p and instantaneous reactive power q in step 2.2.1 are converted to c The low-pass filter obtains the average active power P ref and average reactive power Q ref , expressed as follows:

[0095]

[0096] Step 2.2.3, according to the inverter droop control equation, calculate the initial value ω2 of the output voltage angular frequency reference value and the initial value V of the amplitude reference value in the grid-forming mode. o , expressed as follows:

[0097]

[0098] Where, ω n =2πf n , where f n is the frequency reference value. In an exemplary but non-limiting embodiment of the present invention, the frequency reference value is taken as the rated frequency of the AC power grid, that is, f n =50Hz, V n is the voltage reference value. In an exemplary but non-limiting embodiment of the present invention, the voltage reference value is taken as the rated voltage of the AC grid, V n =380V, m and n are Pf and QV droop coefficients respectively.

[0099] Step 2.2.4: Construct a recursive formula and use the reference value of the previous sampling period to calculate the reference value of the current sampling period. This will yield a prediction model for the output voltage amplitude reference value and angular frequency reference value in the networking mode, which can be expressed as follows:

[0100]

[0101] Where, are the output voltage angular frequency reference values in the kth and k-1th sampling period networking mode, respectively. are the d-axis components of the output voltage amplitude reference values in the k-th and k-1-th sampling period networking modes, respectively. are the q-axis components of the output voltage amplitude reference values in the kth and k-1th sampling period networking mode, T s is the sampling period, p k-1is the instantaneous active power output by the inverter in the k-1th sampling period, q k-1 is the instantaneous reactive power output by the inverter in the k-1th sampling period.

[0102] Step 2.2.5, for and Perform dq / abc transformation to obtain the reference value of the inverter output voltage amplitude in the grid-forming mode according to Get the frequency reference value of the inverter output voltage in the grid-connected mode right Integrate to obtain the phase reference value θ of the inverter output voltage in the grid-forming mode L .

[0103] Step 3: Calculate the synchronization decision signal based on the deviation between the output voltage reference value of the grid-forming mode and the output voltage reference value of the grid-connected mode, and switch the working mode of the inverter according to the synchronization decision signal, including the grid-forming mode and the grid-connected mode; and calculate the system bus voltage reference value and the inverter output current reference value of the current sampling period by the outer model based on the synchronization decision signal and the output voltage reference value of the grid-connected mode.

[0104] In a preferred but non-limiting embodiment of the present invention, the specific process of obtaining the networking mode synchronization decision signal 1 is as follows: Figure 3 As shown, step 3 specifically includes:

[0105] Step 3.1: Based on the reference values of the grid-connected mode in the kth sampling period obtained in step 2, the deviation of the reference values of the amplitude, frequency, and phase of the inverter output voltage is calculated as follows:

[0106]

[0107] Step 3.2: Switch the output mode using the synchronous decision signal 1 to control the power quality issues caused by voltage, frequency, and phase within the standard range according to the IEEE / IEC standard. The synchronous decision signal 1 is calculated according to the following formula:

[0108] u=[sign(|Δv|-0.8)+sign(|Δf|-0.02)+sign(|Δθ|-3)]

[0109]

[0110] When l=1, the inverter is switched to the grid-connected working mode; when l=0, the inverter is switched to the grid-building working mode.

[0111] Step 3.3, calculate the bus voltage reference value v of the current k-th cycle in the synchronization decision signal l determination mode Lr (k) and the output current reference value i or (k).

[0112] In a further preferred but non-limiting embodiment of the present invention, step 3.3 specifically comprises:

[0113] Step 3.3.1, obtain the voltage v of the inverter connected to the system bus through the synchronization decision signal l L Amplitude reference value |v Lr |, frequency reference value ω Lr , phase reference value θ Lr , which is incorporated into the system bus voltage reference value v Lr (k)=|v Lr |sin(2πω Lr +θ Lr ), which is then input into the inner MPC model after Clark-Park transformation.

[0114] More preferably, the inverter is incorporated into the system bus voltage v L The amplitude, frequency, and phase reference values are calculated using the following formulas:

[0115]

[0116] Step 3.3.2, calculate the inverter output current reference value i according to the inverter output voltage reference value in the grid mode in step 2.2 or (k), expressed as follows:

[0117]

[0118] Step 4, construct an inner MPC model; the inner MPC model receives the voltage and current reference values of the outer MPC model, and calculates the predicted value and reference value of the bus voltage and inverter output current of the next sampling period based on the reference value of the bus voltage of the system and the reference value of the inverter output in the current sampling period; constructs a value function cost optimization model based on the degree of deviation between the predicted value and the reference value, generates the optimal switching state, and generates a pulse signal based on the optimal switching state to control the inverter in the distribution network substation.

[0119] In a preferred but non-limiting embodiment of the present invention, step 4 specifically comprises:

[0120] In step 4.1, the system model of the voltage source inverter in the dq rotating coordinate system in step 1 is discretized using the Euler approximation method to obtain the prediction model of the voltage source inverter, which is expressed as the following formula:

[0121]

[0122] Where, T s is the sampling period, ω(k) is the angular frequency at the kth moment, A sh is the state matrix, B sho is the input matrix. Further preferably, the state matrix and the input matrix are expressed as follows:

[0123]

[0124] Step 4.2, consider the relationship between the three-phase 6-way pulse signal of inverter a, b, c, the bridge arm switch tube has 8 switching states, combined with the parallel converter output voltage v od 、v oq , corresponding to the generation of 8 basic voltage vectors u ox , expressed as follows:

[0125] u ox =v ox *S{1,i′},x∈{d,q}

[0126] In the formula, S{1,i′}, (i′=1,2,…,8) is the 8 switching states of the bridge arm switch tube, among which,

[0127]

[0128] It can be understood that each voltage vector u o Each corresponds to a switching state, according to the voltage vector u o A unique set of pulse signals for controlling the inverter bridge arm switch tubes can be generated.

[0129] Step 4.3: The basic vector u of the switch voltage in the kth cycle is od and u oq Substitute the inverter output voltage v into the prediction model in step 4.1 od and v oq , the predicted values of the output current and the bus voltage of the connected system in the k+1th cycle are obtained, which are expressed as follows:

[0130]

[0131] Where, It is the predicted value of the inverter output current and the bus voltage of the connected system in the k+1th cycle.

[0132] In step 4.4, based on the reference values of the system output current and the bus voltage connected to the system in the k-2th cycle, the k-1th cycle, and the kth cycle obtained in step 3.3, the second-order Lagrange extrapolation method is used to perform delay compensation on the reference voltage and current given by the actual model to obtain the reference values of the output current and the bus voltage connected to the system in the k+1th cycle of the photovoltaic-storage coordinated distribution network system, which can be expressed as the following formula:

[0133]

[0134] Where i oxr (k+1), v Lxr (k+1)(x∈{d,q}) is the inverter output current reference value and the bus voltage reference value of the system in the k+1th cycle.

[0135] Step 4.5, by comparing the deviation between the prediction model and the actual model, a value assessment is performed to construct an inner layer MPC value function cost optimization model. In a further preferred embodiment, step 4.5 specifically includes:

[0136] In step 4.5.1, calculate the Manhattan norm of the difference between the reference and predicted values of voltage and current in the k+1th cycle.

[0137] In the MPC value optimization problem, the Manhattan norm acts as a regularization term to help the control algorithm avoid overfitting, reduce the complexity of the model and improve the interpretability of the control strategy, thereby affecting the structural characteristics of the solution to the optimization problem, reducing the fluctuation of the control input, and improving the robustness of the model.

[0138] In step 4.5.2, based on the Manhattan norm calculated in step 4.5.1, a value evaluation function is designed so that the inverter MPC system perfectly tracks the grid-connected bus voltage and current reference values.

[0139] The value evaluation function of the inverter is calculated according to the following formula:

[0140]

[0141] Where g a is the evaluation function of the model prediction current error, g b is the evaluation function of the model predicted voltage error, and |*| represents the Manhattan norm.

[0142] In step 4.5.3, the evaluation function of the model predicted current and voltage errors in step 4.5.2 is used as the optimization target, and weight factors are assigned respectively to obtain the objective function of the inner MPC algorithm, and the inner MPC value function cost optimization model is constructed based on the objective function.

[0143] In multi-objective optimization problems, weighting factors are used to balance the importance of different objectives, adjust the learning speed of the model, and reduce model complexity. By designing weighting factors, we can accelerate model convergence and reduce the model's sensitivity to noise in the training data, thereby improving the model's generalization ability.

[0144] More preferably, considering minimizing the error tracking amount, the objective function of the designed inner MPC model is expressed as follows:

[0145]

[0146] Where u o is the voltage vector, J is the weighted error, and N is the prediction time domain defined by the prediction system, that is, the predicted state of the control quantity applied to the next N cycles at the current moment; i and λ u are the first weight coefficient and the second weight coefficient respectively.

[0147] Preferably, the initial condition satisfies λ i / λ u =R / X, where R / X is the line impedance ratio. The first weight coefficient λ is then dynamically adjusted based on the model overshoot and the model adjustment reaction rate. i ∈[0.1,5] and the second weight coefficient λu i ∈[0.1~5]. In an exemplary but non-limiting embodiment, λ i =1,λ u =0.012.

[0148] Step 4.6: Solve the inner-layer MPC cost optimization model of step 4.5 to obtain the optimal voltage vector among the eight basic voltage vectors in step 4.2, and output the optimal switching state of the voltage source inverter bridge arm switch corresponding to the voltage vector. The optimal switching state is the optimal vector selected from the eight basic voltage vectors.

[0149] The present invention decouples the characteristics of the voltage source and the current source through hierarchical predictive optimization to form a dynamic balance mechanism of "voltage closed-loop forced tracking + current loop active damping injection". With multi-objective dynamic reconstruction as the core, the switching process is decomposed into an optimization problem of a continuous time series through a predictive model; CN113991748B is essentially still a rigid control based on the droop equation, and only compensates for the transient impact of mode switching through phase correction. It is a threshold-triggered mutation response and is essentially an improvement of the open-loop switching strategy. The above method makes the dynamic response speed of the present invention higher, the system robustness higher, and it is more suitable for high-penetration new energy access to weak power grids / island microgrids.

[0150] Embodiment 2 of the present invention provides a distribution network substation inverter seamless switching control system based on double-layer model predictive control, and runs the distribution network substation inverter seamless switching control method based on double-layer model predictive control described in embodiment 1, including: a parameter acquisition module, a reference value calculation module, a synchronous decision module and a cost optimization module.

[0151] The parameter acquisition module is used to obtain the operating parameters and equipment parameters of the distribution network in the substation area and establish a voltage source inverter system model taking into account frequency fluctuations;

[0152] The reference value calculation module is used to construct an outer model; based on the operating parameters and equipment parameters, the outer model generates the output voltage reference value of the inverter grid-connected mode and the output voltage reference value of the grid-forming mode;

[0153] The synchronization decision module is used to calculate the synchronization decision signal based on the deviation of the output voltage reference value of the grid-connected mode and the grid-forming mode, and switch the working mode of the inverter according to the synchronization decision signal; based on the synchronization decision signal and the output voltage reference value of the grid-connected mode, the outer model calculates the bus voltage reference value of the current sampling period and the output current reference value of the inverter;

[0154] The cost optimization module is used to construct the inner model; based on the reference value of the system bus voltage and the reference value of the inverter output current in the current sampling period, the inner model calculates the predicted value and reference value of the system bus voltage and the inverter output current in the next sampling period; based on the degree of deviation between the predicted value and the reference value, a value function cost optimization model is constructed to generate the optimal switching state, and a pulse signal is generated according to the optimal switching state to control the inverter in the distribution network substation.

[0155] Compared with the prior art, the beneficial effects of the present invention include at least:

[0156] 1. The present invention adopts internal and external double-layer model predictive control, taking into account both dynamic response and stability, achieving seamless switching of voltage source inverters while avoiding voltage over-limit;

[0157] 2. The inner model of the present invention selects the optimal switching state of the inverter by solving the cost optimization model of the value function, with a short prediction step length and a dynamic response speed of 5 to 10 ms, thus achieving rapid tracking of the reference values of the load voltage and grid current;

[0158] 3. The outer model of the present invention predicts and corrects the reference value of the inverter output voltage in the grid-forming mode through the droop control strategy, and can control the voltage and frequency output of the inverter within the desired range even under interference;

[0159] 4. The present invention smoothly switches between the grid-connected mode and the grid-building mode based on synchronous decision signals, which can meet the frequent switching needs of photovoltaic-energy storage hybrid areas and improve the stability of the power system.

[0160] The present disclosure may be a system, method and / or computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0161] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A distribution network substation inverter switching control method based on double-layer MPC, characterized in that: The following steps are involved: Step 1: Obtain the operating parameters and equipment parameters of the distribution network in the substation area and establish a voltage source inverter system model that takes frequency fluctuations into account; Step 2: construct an outer model; based on the operating parameters and device parameters, the outer model generates an output voltage reference value of the inverter grid-connected mode and an output voltage reference value of the grid-forming mode; Step 3: Calculate a synchronization decision signal based on the deviation of the output voltage reference values of the grid-connected mode and the grid-forming mode, and switch the inverter's operating mode according to the synchronization decision signal; and calculate the system bus voltage reference value and the inverter output current reference value for the current sampling period by the outer model based on the synchronization decision signal and the output voltage reference value of the grid-connected mode. Step 4: construct an inner model; based on the reference value of the bus voltage of the system and the reference value of the output current of the inverter in the current sampling period, the inner model calculates the predicted value and reference value of the bus voltage of the system and the output current of the inverter in the next sampling period; According to the degree of deviation between the predicted value and the reference value, a cost optimization model of the value function is constructed to generate the optimal switching state. According to the optimal switching state, a pulse signal is generated to control the inverter in the distribution network station area.

2. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 1 is characterized in that: In step 1, the system model of the voltage source inverter is as follows: Where i od 、i oq are the d-axis and q-axis components of the inverter output current, I od , I oq i od 、i oq The steady-state value, v Ld 、v Lq are the d-axis and q-axis components of the inverter-integrated system bus voltage, V Ld 、V Lq v Ld 、v Lq The steady-state value, v od 、v oq are the d-axis and q-axis components of the inverter output voltage, i Sd 、i Sq are the d-axis and q-axis components of the AC grid current, i Ld 、i Lq are the d-axis and q-axis components of the inverter-integrated system bus current, ω is the grid angular frequency, ω N is the steady-state value of ω, L V and C V are the filter inductance and filter capacitance of the LC filter, R V is the equivalent resistance.

3. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 1, characterized in that: In step 2, the output voltage reference value includes: an amplitude reference value, a frequency reference value, and a phase reference value; Among them, the reference value of the inverter output voltage amplitude in grid-connected mode is and frequency reference value f k Keep consistent with the rated voltage and rated frequency of the AC power grid respectively; The output voltage phase θ on both sides of the parallel switch of the voltage source inverter is collected through the phase-locked loop o , as the phase reference value of the inverter output voltage in grid-connected mode.

4. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 3 is characterized in that: In step 2, generating the output voltage reference value of the networking mode includes: According to the inverter droop control equation, the initial value ω2 of the output voltage angular frequency reference value and the initial value V of the amplitude reference value in the grid-forming mode are calculated. o ,as follows: Where, ω n =2πf n , f n is the frequency reference value, V n is the reference value of the voltage, m and n are the Pf and QV droop coefficients respectively, p and q are the instantaneous active power and instantaneous reactive power output by the inverter respectively, P ref and Q ref are the average active power and average reactive power respectively; A recursive formula is constructed to use the reference value of the previous sampling period to calculate the reference value of the current sampling period, and the prediction model of the output voltage amplitude reference value and angular frequency reference value in the networking mode is obtained as follows: Where, are the output voltage angular frequency reference values in the kth and k-1th sampling period networking mode, respectively. are the d-axis components of the output voltage amplitude reference values in the k-th and k-1-th sampling period networking modes, respectively. are the q-axis components of the output voltage amplitude reference values in the kth and k-1th sampling period networking mode, T s is the sampling period, p k-1 is the instantaneous active power output by the inverter in the k-1th sampling period, q k-1 is the instantaneous reactive power output by the inverter in the k-1th sampling period; right and Perform dq / abc transformation to obtain the reference value of the inverter output voltage amplitude in the grid-forming mode according to Get the frequency reference value of the inverter output voltage in the grid-connected mode right Integrate to obtain the phase reference value θ of the inverter output voltage in the grid-forming mode L .

5. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 4 is characterized in that: In step 3, calculating the synchronization decision signal and switching the inverter operation mode according to the synchronization decision signal include: Calculate the synchronization decision signal l, which is expressed as follows: u=[sign(|Δv|-0.8)+sign(|Δf|-0.02)+sign(|Δθ|-3)] Where |Δv| is the deviation of the output voltage amplitude reference value between the grid-connected mode and the grid-forming mode, |Δf| is the deviation of the frequency reference value of the output voltage between the grid-connected mode and the grid-forming mode |Δθ| is the deviation of the phase reference value of the output voltage between the grid-connected mode and the grid-forming mode, |Δθ|=|θ o -θ L |, θ L are the amplitude reference value, frequency reference value, and phase reference value of the inverter output voltage in the grid-forming mode, respectively. f k ,θ o They are respectively the amplitude reference value, frequency reference value and phase reference value of the inverter output voltage in grid-connected mode; When l=1, the inverter is switched to the grid-connected working mode; when l=0, the inverter is switched to the grid-building working mode.

6. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 5, characterized in that: In step 3, the reference value of the bus voltage of the system is calculated and expressed as follows: In the formula, |v Lr |、ω Lr ,θ Lr are the amplitude reference value, angular frequency reference value, and phase reference value of the bus voltage incorporated into the system; the amplitude reference value|v Lr |, frequency reference value ω Lr , phase reference value θ Lr , which is incorporated into the system bus voltage reference value v Lr (k)=|v Lr |sin(2πω Lr +θ Lr ).

7. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 5, characterized in that: In step 3, the output current reference value of the inverter is calculated and expressed as follows: Where i odr (k), i oqr (k) are the d-axis and q-axis components of the output current reference value of the inverter in the kth sampling period, are the d-axis component and q-axis component of the output voltage amplitude reference value of the k-th sampling period, respectively. ref is the average active power, Q ref is the average reactive power.

8. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 2, characterized in that: In step 4, the predicted values of the bus voltage and inverter output current to be incorporated into the system in the next sampling period are calculated as follows: The system model of the voltage source inverter in the dq rotating coordinate system is discretized and substituted into the voltage basic vector to obtain the predicted values of the output current and the bus voltage of the system in the k+1th cycle, which are expressed as the following formula: Where, (x∈{d,q}) are the inverter output current and the predicted value of the k+1th cycle respectively, and the predicted value is incorporated into the system bus voltage predicted value, i ox (k), v Lx (k),u ox (k), i Sx (k), i Lx (k) (x∈{d,q}) are the inverter output current sampling value, the bus voltage sampling value, the voltage basic vector, the AC grid current, and the bus current sampling value in the kth cycle, respectively. sh is the state matrix, B sho is the input matrix, T s is the sampling period, and ω(k) is the angular frequency of the kth period.

9. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 8, characterized in that: The state matrix and input matrix are as follows:

10. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 1, characterized in that: In step 4, the reference values of the bus voltage and the inverter output current to be incorporated into the system in the next sampling period are calculated as follows: Based on the reference values of the inverter output current and the bus voltage connected to the system in the k-2th, k-1th, and kth cycles, the reference values of the bus voltage connected to the system and the inverter output current in the k+1th cycle are calculated and expressed as follows: Where i oxr (k+1), v Lxr (k+1)(x∈{d,q}) is the inverter output current reference value and the bus voltage reference value of the system in the k+1th cycle.

11. The method for controlling inverter switching in a distribution network area based on a double-layer MPC according to claim 1, characterized in that: In the value function cost optimization model constructed in step 4, the designed value evaluation function is as follows: Where g a is the evaluation function of the model prediction current error, g b is the evaluation function of the model prediction voltage error, |*| represents the Manhattan norm, is the predicted value of the inverter output current and the bus voltage of the connected system in the k+1th cycle, i oxr (k+1), v Lxr (k+1)(x∈{d,q}) is the inverter output current reference value and the bus voltage reference value of the connected system in the k+1th cycle; The objective function of the designed inner model is expressed as follows: Where u o is the voltage vector, J is the weighted error, N is the prediction time domain defined by the prediction system, λ x and λ y are the first weight coefficient and the second weight coefficient respectively.

12. A distribution network substation inverter switching control system based on a double-layer MPC, characterized in that: include: The parameter acquisition module is used to obtain the operating parameters of the distribution network and equipment parameters in the substation area and establish a voltage source inverter system model that takes frequency fluctuations into account; The reference value calculation module is used to construct an outer model; based on the operating parameters and equipment parameters, the outer model generates the output voltage reference value of the inverter grid-connected mode and the output voltage reference value of the grid-forming mode; A synchronization decision module is used to calculate a synchronization decision signal based on the deviation of the output voltage reference values of the grid-connected mode and the grid-forming mode, and switch the operating mode of the inverter according to the synchronization decision signal; based on the synchronization decision signal and the output voltage reference value of the grid-connected mode, the outer model calculates the bus voltage reference value of the current sampling period and the output current reference value of the inverter; The cost optimization module is used to construct the inner model; based on the reference value of the bus voltage of the system and the reference value of the output current of the inverter in the current sampling period, the inner model calculates the predicted value and reference value of the bus voltage of the system and the output current of the inverter in the next sampling period; According to the degree of deviation between the predicted value and the reference value, a cost optimization model of the value function is constructed to generate the optimal switching state. According to the optimal switching state, a pulse signal is generated to control the inverter in the distribution network station area.

Images