Energy storage converter controller optimization method and device, storage medium and product

Abstract

The invention discloses an energy storage converter controller optimization method and device, a storage medium and a product, and the method comprises the steps: designing an energy storage converter controller which comprises an outer ring and an inner ring, and the inner ring is a deadbeat controller; establishing a mathematical model of the dead-beat controller; determining a recursive expression of the identification parameters according to the constructed optimization objective function; when parameter identification calculation is judged to be started according to the instruction current and the actual output current at the current moment, an identification parameter at the current moment is calculated according to current data and voltage data of a main circuit of the energy storage converter at the current moment and a recursive expression of the identification parameter; when convergence is judged according to the identification parameters of the current moment and the previous moment, the identification parameters of the current moment are substituted into a mathematical model of the dead-beat controller, and online optimization processing of the dead-beat controller is achieved. According to the method, the mathematical model of the dead-beat controller is optimized on line, and the system control precision is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy storage control, and in particular relates to an energy storage converter controller optimization method, device, storage medium and product based on inductance parameter identification. Background Art

[0002] The basic function of the energy storage system is to realize the two-way flow of energy between the energy storage battery and the power grid. The key equipment to realize the two-way conversion of energy in the energy storage system is the energy storage inverter. Therefore, the energy storage inverter is the energy conversion core of the entire energy storage system, and its importance is self-evident. The energy storage inverter is a power electronic device whose main function is to convert electrical energy between the energy storage battery and the power grid. It can realize the charging and discharging process of electrical energy and ensure the efficient operation of the energy storage system. When the energy storage system is charging, the energy storage inverter converts the three-phase AC power of the power grid into DC power to charge the energy storage battery; when the energy storage system is discharging, the energy storage inverter inverts the DC power of the battery into three-phase AC power and sends it to the power grid, thereby realizing the two-way flow of energy between the power grid and the energy storage battery.

[0003] At present, the main circuit topologies of energy storage converters used in the industry mainly include two-level topology, three-level topology, cascaded H-bridge topology, modular multi-level topology and full-bridge converter, and the hardware circuit structure is basically mature. The key lies in the research of energy storage converter control algorithm. The wide application of high-speed digital signal processing chip DSP in the field of power electronics provides a hardware foundation for the implementation of more intelligent algorithms. Therefore, the controller of energy storage converter mostly adopts DSP control chip with high-speed digital signal processing capability, which can achieve high-precision tracking of target signals, and has the advantages of fast dynamic response, simple parameter setting and easy algorithm implementation.

[0004] The controller design of energy storage converters mostly adopts dual closed-loop controllers. Depending on the operating conditions of the equipment, the outer loop can be a DC bus voltage outer loop or a power outer loop, and the inner loop is a current inner loop. The function of the DC bus voltage outer loop is to maintain the voltage across the bus capacitor of the energy storage converter at the set value, and the function of the power outer loop is to track the reference power change of the actual output / input power of the system. Whether it is the voltage outer loop or the power outer loop, the control quantity is the DC quantity. According to the classical control theory, it can be known that the use of PID control can achieve error-free tracking control of the DC quantity. In the dual closed-loop controller, the output of the outer loop is the reference current value of the energy storage converter. The current inner loop receives the reference current value given by the outer loop, and also receives the real-time output current value of the AC side of the energy storage converter collected by the current transformer. The reference current value and the real-time output current value are simultaneously sent to the current inner loop controller for control algorithm calculation, and the PWM control instruction of the energy storage converter is output. The power module IGBT of the main circuit receives the PWM control instruction, and the energy storage converter outputs the three-phase grid-connected current. At the next moment, the collected DC bus voltage value or power value of the main circuit and the three-phase grid-connected current value are sent to the control system, thus forming a complete feedback control system.

[0005] There are many design methods for the current inner loop controller, such as PI control, PR control, deadbeat control, and more advanced intelligent control algorithms. The deadbeat control algorithm has the advantage of being easy to implement digitally. With the development of digital technology, the deadbeat control algorithm can be easily implemented through a digital controller (DSP), reducing the complexity and cost of system design, and its engineering applications are becoming more and more extensive. The energy storage converter adopts the power or voltage outer loop PI control and the current inner loop deadbeat control strategy, which ensures a stable control system and convenient parameter adjustment. The outer loop PI control ensures the stability of the DC bus voltage or output power through proportional and integral regulation, eliminates steady-state errors, and improves system stability; while the current inner loop deadbeat control uses an accurate mathematical model to predict future currents, achieves fast response and precise control, effectively solves the system delay problem, and improves dynamic performance. The combination of the two enables the energy storage converter to achieve a high level of voltage, power, and current control, meeting the needs of basic application scenarios.

[0006] The controllers of voltage / power outer loop PI control and current inner loop deadbeat control used in energy storage converters are studied. It is found that the controllers designed based on this method have great advantages in DSP application, but also have obvious disadvantages, mainly the following two points:

[0007] (1) The core of deadbeat control lies in an accurate mathematical model. The lack of an accurate model will make it difficult for the control algorithm to accurately predict the system behavior, which will lead to poor control effect and decreased system stability. Therefore, when applying deadbeat control, it is necessary to ensure that the mathematical model relied on is as accurate as possible. In the process of establishing the mathematical model of deadbeat control of energy storage converters, the measured value of the inductance of the main circuit of the energy storage converter is required. The traditional method directly substitutes the inductance parameters provided by the inductor supplier as known quantities into the mathematical model. The inductance parameters provided by the supplier are the inductance when the inductance is unloaded, which is very different from the inductance value after the inductance is loaded. At the same time, when the inductance is running under actual working conditions, its inductance changes in real time. At this time, the deadbeat mathematical model established based on the inductance parameters provided by the supplier has a large deviation, which directly causes the output current of the energy storage converter to be unable to track the target current changes. There is a large tracking error, which leads to distortion of the output current waveform and increased harmonic content, affecting the output power quality of the energy storage converter and the overall performance of the system. Therefore, ensuring the accuracy of the inductance parameters is the key to achieving efficient deadbeat control. There are two main sources of inductance parameter errors:

[0008] (a) The inductance parameters provided by the supplier deviate from the actual value and are not realistic enough;

[0009] (b) The inductance parameters provided by the supplier are no-load parameters. When the energy storage inverter is running, the inductance will face overload and magnetic saturation, and the inductance parameters will continue to change. Specifically, in the actual operation of the energy storage inverter, the inductance value is not static, but changes dynamically due to the combined influence of various external and internal factors such as temperature rise, equipment aging, and current fluctuations. However, in zero-beat control, if the inductance parameters fail to reflect these real-time changes in a timely manner, but maintain a static, fixed value, then there will inevitably be a deviation between the mathematical model of zero-beat control and the actual operating conditions. This mathematical model deviation will directly lead to a deviation in the zero-beat control instruction, resulting in the system output current being unable to track the target current changes, which will further aggravate the current waveform distortion, increase the harmonic content of the current, and adversely affect the overall performance and stability of the system.

[0010] (2) When applying parameter identification technology to the parameter optimization of the deadbeat controller, the actual code implementation needs to be considered. If the optimal parameters cannot converge as quickly as possible, the processor operation pressure will be high and the data storage volume will be large. Therefore, it is necessary to design the convergence conditions of the parameter identification operation. In addition, during the operation of the equipment, according to the actual working conditions, there must be judgment conditions for the input and output of the parameter identification operation. Otherwise, when the equipment is started with pre-charge resistance or when a transient fault or even a permanent fault occurs, the mathematical model of the system has changed because the system is in a non-steady-state working state. At this time, the inductance parameter identification is inaccurate, and the wrong optimal inductance parameters will cause the entire control system to crash. Therefore, under non-steady-state conditions, the execution of the inductance parameter identification needs to be stopped. Summary of the invention

[0011] The purpose of the present invention is to provide an energy storage converter controller optimization method, device, storage medium and product to solve the problem that when the current inner loop is in deadbeat control, the mathematical model of the deadbeat control is difficult to establish accurately due to the change of inductance parameters, which leads to poor control effect and reduced performance and stability of the energy storage system.

[0012] The present invention solves the above technical problems through the following technical solutions: a method for optimizing an energy storage converter controller, comprising:

[0013] According to the main circuit topology of the energy storage converter and Kirchhoff's voltage law, a mathematical model of the energy storage converter is established;

[0014] Designing an energy storage converter controller, wherein the energy storage converter controller is a double closed-loop controller, the double closed-loop controller comprises an outer loop and an inner loop, and the inner loop is a deadbeat controller;

[0015] According to the mathematical model of the energy storage converter and the discretization principle, a mathematical model of the deadbeat controller is established;

[0016] Constructing an optimization objective function according to the mathematical model of the energy storage converter, and determining a recursive expression of an identification parameter according to the optimization objective function;

[0017] Obtain the current command current and the actual output current of the energy storage converter at the current moment, and determine whether to start parameter identification calculation according to the current command current and the actual output current;

[0018] When the parameter identification calculation is turned on, the current data and voltage data of the main circuit of the energy storage converter at the current moment are obtained;

[0019] Calculating the identification parameters at the current moment according to the current data and voltage data of the main circuit of the energy storage converter at the current moment and the recursive expression of the identification parameters;

[0020] Whether convergence is achieved is determined based on the identification parameters at the current moment and the previous moment; if so, the identification parameters at the current moment are substituted into the mathematical model of the deadbeat controller to achieve online optimization of the deadbeat controller; if not, the parameter identification calculation process is repeated.

[0021] Furthermore, when the main circuit topology of the energy storage converter is a two-level three-phase three-bridge-arm split capacitor topology, the mathematical model of the energy storage converter is:

[0022]

[0023]

[0024] Among them, u jo represents the voltage of the midpoint of the jth bridge arm relative to the midpoint O of the DC bus capacitor, j = a, b, c, and the bridge arms a, b, c correspond to the phases A, B, and C of the distribution network one by one; e A 、e B 、e C Respectively represent the three-phase grid voltage of the distribution network; R LA , R LB , R LC Respectively represent the internal resistance of the three-phase filter inductor on the AC side; i aA 、i bB 、i cC They represent the current of the corresponding phase grid input into the midpoint of each bridge arm; L A , L B , L c Respectively represent the inductance of the three-phase filter inductance on the AC side; C 1 , C 2 Respectively represent the capacity of the upper and lower bus capacitors on the DC side; U dc1 , U dc2 Respectively represent the terminal voltage of the upper and lower bus capacitors on the DC side; S j Indicates the switch state of the jth bridge arm; U dc Indicates the DC bus voltage.

[0025] Furthermore, the outer loop is a power outer loop or a voltage outer loop, and the outer loop adopts a PI controller.

[0026] Furthermore, when the main circuit topology of the energy storage converter is a two-level three-phase three-bridge-arm split capacitor topology, the mathematical model of the deadbeat controller is:

[0027]

[0028] Among them, S j(k) represents the switch state of the jth bridge arm at the kth moment, j = a, b, c, and the bridge arms a, b, c correspond to the A, B, C phases of the distribution network one by one; U dc Indicates the DC bus voltage; L A , L B , L c Respectively represent the inductance of the three-phase filter inductance on the AC side; T S Indicates the sampling period; They represent the three-phase command current at the kth moment respectively; i aA (k), i bB (k), i cC (k) represents the current of the corresponding phase grid input into the midpoint of each bridge arm at the kth moment; R LA , R LB , R LC They represent the internal resistance of the three-phase filter inductance on the AC side; e A (k), e B (k), e C (k) represent the three-phase grid voltage of the distribution network at the kth moment respectively.

[0029] Furthermore, a recursive least square method with a forgetting factor is used to determine a recursive expression of the identification parameter; wherein the recursive expression of the identification parameter is:

[0030]

[0031]

[0032] in, Represents θ km The recursive estimate of θ km represents the identification parameter matrix of the mth phase at the kth moment, Represents θ k-1,m The recursive estimate of θ k-1,m represents the identification parameter matrix of the mth phase at the k-1th moment, m = A, B, C; β km , H k-1,m , Both represent intermediate quantities; p km represents the input quantity of the mth phase at the kth moment; q km represents the output of the mth phase at the kth moment; λ represents the forgetting factor; p 1m Represents the input quantity of the mth phase at the first moment; the superscript T represents transposition.

[0033] Furthermore, the judgment condition for starting the parameter identification calculation is:

[0034] When 0≤I mse <δ, start parameter identification calculation; when I mseWhen ≥δ, parameter identification calculation is prohibited;

[0035] Among them, δ represents the error threshold, I mse represents the real-time current tracking error of the deadbeat controller; the real-time current tracking error I mse for:

[0036]

[0037] in, represents the m-th phase command current of the energy storage converter at the k-th moment, n represents the number of sampling points in one power frequency cycle, n=f / 50, f represents the sampling frequency; i jm (k) represents the current input into the m-phase grid at the midpoint of the j-th bridge arm at the k-th moment, and the bridge arms a, b, and c correspond to the A, B, and C phases of the distribution network one by one.

[0038] Furthermore, the convergence is determined based on the identification parameters at the current moment and the previous moment, and the convergence condition is:

[0039]

[0040] in, Represents θ km The recursive estimate of (i), θ km (i) represents the i-th parameter in the identification parameter matrix of the m-th phase at the k-th moment, Represents θ k-1,m The recursive estimate of (i), θ k-1,m (i) represents the i-th parameter in the identification parameter matrix of the m-th phase at the k-1-th time; T S represents the sampling period; ζ represents the convergence threshold.

[0041] Based on the same concept, the present invention provides an electronic device, including a memory, a processor, and a computer program / instruction stored in the memory, wherein the processor executes the computer program / instruction to implement the energy storage converter controller optimization method as described above.

[0042] Based on the same concept, the present invention provides a computer-readable storage medium having a computer program / instruction stored thereon, which implements the energy storage converter controller optimization method as described above when the computer program / instruction is executed by a processor.

[0043] Based on the same concept, the present invention provides a computer program product, including a computer program / instruction, which implements the energy storage converter controller optimization method as described above when executed by a processor.

[0044] Beneficial Effects

[0045] Compared with the prior art, the advantages of the present invention are:

[0046] The present invention constructs a mathematical model of a deadbeat controller based on a mathematical model of an energy storage converter and a discretization principle, and then uses a parameter identification algorithm to perform online optimization on the mathematical model of the deadbeat controller, thereby improving the control accuracy of the system and solving the problem that the mathematical model of the deadbeat control is difficult to accurately establish due to changes in inductance parameters. The present invention determines a start-up judgment condition of the parameter identification algorithm based on a command current and an actual output current at the current moment, and also determines a convergence condition based on identification parameters at the current moment and the previous moment, that is, by judging the change in the identification parameter within a unit control cycle, a balance between optimal parameter acquisition and operation speed is achieved.

[0047] The present invention adopts the recursive least square method with forgetting factor to perform parameter identification calculation, and significantly reduces the amount of data storage and calculation by introducing the recursive mechanism, thereby effectively improving the efficiency of parameter identification; on this basis, the forgetting factor is introduced to dynamically adjust the weight of historical data and give new data a higher importance, thereby giving the parameter identification algorithm the ability of self-correction and optimization, ensuring the dual improvement of the accuracy and reliability of the identification parameters, optimizing the deadbeat control of the inner loop of the current, and solving the problem of reduced accuracy of the mathematical model caused by the change of the inductance parameter in the deadbeat control of the inner loop. The recursive least square method with forgetting factor is easy to program and implement, which not only simplifies the data processing process, but also significantly enhances the adaptability of the system to dynamic changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only one embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 is a flow chart of an energy storage converter controller optimization method according to an embodiment of the present invention;

[0050] Figure 2 is a structural principle diagram of a two-level three-phase three-bridge-arm split capacitor topology structure in an embodiment of the present invention;

[0051] Figure 3 It is a principle block diagram of a dual closed-loop controller in an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The following is a clear and complete description of the technical solutions in the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] The technical solution of the present application is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0054] like Figure 1 As shown, an energy storage converter controller optimization method provided by an embodiment of the present invention comprises the following steps:

[0055] Step 1: According to the main circuit topology of the energy storage converter and Kirchhoff's voltage law, a mathematical model of the energy storage converter is established.

[0056] Taking the main circuit topology of the energy storage converter as a two-level three-phase three-bridge arm split capacitor topology as an example, its structural principle diagram is as follows Figure 2 shown. Figure 2 in, u jo It represents the voltage of the midpoint of the jth bridge arm relative to the midpoint O of the DC bus capacitor, j = a, b, c, that is, u ao 、u bo 、u co Respectively represent the voltage of the midpoint of bridge arms a, b, c relative to the midpoint O of the DC bus capacitor; e m Represents the voltage of the mth phase of the distribution network, m = A, B, C, that is, e A 、e B 、e C Respectively represent the A, B, and C phase grid voltages of the distribution network; R Lm Represents the internal resistance of the mth phase filter inductor on the AC side, m = A, B, C, that is, R LA , R LB , R LC Respectively represent the internal resistance of the filter inductor of phase A, B, and C on the AC side; L m Indicates the inductance of the mth phase filter inductor on the AC side, m = A, B, C, that is, L A , L B , L c Respectively represent the inductance of the A, B, and C phase filter inductors on the AC side; i jm It represents the current of the mth phase grid inputted into the midpoint of the jth bridge arm. The bridge arms a, b, c correspond to the A, B, and C phases of the distribution network one by one, that is, when j = a, m = A; when j = b, m = B; when j = c, m = C; i aAThe midpoint of bridge arm a inputs the current of phase A grid, i bB represents the current of the midpoint of bridge arm b inputting into the B-phase grid, i cC represents the current of the C-phase grid input at the midpoint of bridge arm c; i ON Indicates the current of the DC bus capacitor midpoint O input to the neutral line N; C 1 , C 2 Respectively represent the capacity of the upper and lower bus capacitors on the DC side; U dc1 , U dc2 Respectively represent the terminal voltage of the upper and lower bus capacitors on the DC side; U dc Indicates the DC bus voltage; VT 1 ~VT 6 Represent the power modules of each bridge arm respectively.

[0057] Assume that the direction of the current flowing into the grid is the positive direction of the current. According to the main circuit topology of the energy storage converter and the reference direction of voltage and current, the mathematical model of the three-phase four-wire energy storage converter is established using Kirchhoff's voltage law. First, the switch state of each bridge arm is defined:

[0058]

[0059] Among them, S j Indicates the switch state of the jth bridge arm. Taking bridge arm a as an example, when the power module VT of the upper bridge arm 1 The power module VT of the lower bridge arm is turned on. 4 When turned off, the switch state of bridge arm a is 1; when the power module VT of the upper bridge arm 1 Shut down, the power module VT of the lower bridge arm 4 When turned on, the switching state of bridge arm a is -1.

[0060] The voltage of the midpoint of each bridge arm on the AC side of the energy storage converter relative to the midpoint O of the DC side bus capacitor is:

[0061]

[0062] The switching states of the three bridge arms are defined as S a , S b , S c , each switch state can take two different values: 1 and -1. For a three-phase four-wire system, the line impedance is ignored. By writing Kirchhoff's voltage law for the main circuit of the energy storage converter, its mathematical model can be obtained as follows:

[0063]

[0064] It can be seen from the mathematical model of the energy storage converter that for a three-phase four-wire system, there is no coupling relationship between the phases, and independent control of the current of each phase can be achieved.

[0065] Step 2: Design the energy storage converter controller. The energy storage converter controller is a dual closed-loop controller. The dual closed-loop controller includes an outer loop and an inner loop. The inner loop is a deadbeat controller.

[0066] According to different equipment working conditions, the outer loop can adopt voltage outer loop or power outer loop, and the outer loop controller adopts PI controller; the inner loop is designed as current inner loop, and the inner loop adopts deadbeat controller.

[0067] The schematic diagram of the dual closed-loop controller is as follows: Figure 3 As shown, where P s Represents the output power of the energy storage converter, P ref / U dcref Indicates the set active power command value / voltage command value, The difference between the three-phase command current (i.e., reference current value) of the energy storage converter and the output power of the energy storage converter (or the difference between the voltage command value and the DC bus voltage) is obtained by the PI controller to obtain the active current command value i dref , active current command value i dref After dq-abc coordinate conversion, the reference current value of the mth phase is obtained The output of the outer loop is the reference current value of the energy storage converter. The current inner loop receives the reference current value given by the outer loop, and also receives the real-time output current value i of the AC side of the energy storage converter collected by the current transformer. jm , the reference current value and the real-time output current value are simultaneously sent to the deadbeat controller for control calculation, and the PWM control instruction of the energy storage converter is output.

[0068] Step 3: According to the mathematical model of the energy storage converter and the discretization principle, a mathematical model of the deadbeat controller is established.

[0069] Combining formula (2) and formula (3), with sampling period T S By performing forward differential discretization, the mathematical model of the deadbeat controller of the energy storage converter is obtained as follows:

[0070]

[0071] Among them, S j (k) represents the switch state of the jth bridge arm at the kth moment; i aA (k), i bB (k), i cC (k) represents the current of the corresponding phase grid input into the midpoint of each bridge arm at the kth moment; i aA (k+1), i bB (k+1), i cC (k+1) represents the current of the corresponding phase grid input into the midpoint of each bridge arm at the k+1th moment; eA (k), e B (k), e C (k) respectively represent the A, B, and C phase grid voltages of the distribution network at the kth moment. Sampling period T S Small enough.

[0072] The control goal of deadbeat control is to make the output current at the next moment track the command current value calculated at the current moment without error. Figure 3 It can be seen that sampling is performed at the kth moment, and the command current value at the kth moment is obtained through the outer loop PI controller and dq transformation. If the parameters of the deadbeat controller are designed reasonably, after one control cycle, that is, at the k+1th moment, the output current of the energy storage converter will track the command current value. Therefore, the sampling and parallel push method is used, and:

[0073]

[0074] Substituting formula (5) into formula (4), the final mathematical model of the deadbeat controller is:

[0075]

[0076] in, They respectively represent the A, B, and C phase command currents at the kth moment.

[0077] Step 4: Construct an optimization objective function based on the mathematical model of the energy storage converter, and determine the recursive expression of the identification parameters based on the optimization objective function.

[0078] According to the mathematical model of the deadbeat controller, the main circuit inductance parameter value has the greatest impact on the accuracy of the model. However, a common practice in engineering is to introduce a deadbeat control coefficient ψ and multiply it with the inductance parameter L to achieve the purpose of correcting the inductance parameter. However, this method can only be tried and true through multiple experiments, and the obtained inductance parameter will lead to poor control accuracy of the entire control system. Therefore, the present invention uses a recursive least square method with a forgetting factor to identify the inductance parameter of the deadbeat controller.

[0079] The input and output of the system are expressed in vector form:

[0080] Pθ=Q (7)

[0081] P = [p 1 ,p 2 ,…,p k ] T , Q=[q 1 ,q 2 ,…,q k ] T (8)

[0082] Where P represents the input matrix, p k represents the input at the kth moment, Q represents the output matrix, q k represents the output at the kth moment, θ represents the identification parameter matrix, and θ can be obtained by parameter estimation through P and Q.

[0083] According to the mathematical model of the energy storage converter (i.e., formula (3)), in order to obtain the most accurate mathematical model, the inductance parameters of the three phases need to be identified separately. For a three-phase four-wire system, there is no coupling relationship between the phases, and the current of each phase can be independently controlled. Therefore, the three-phase inductance parameter identification is calculated separately, that is, the identification process is to first identify the inductance parameters of phase A, then the inductance parameters of phase B, and finally the inductance parameters of phase C. Taking the inductance parameter identification of the mth phase as an example, the output quantity q can be defined as m =u jo -e A , the input is Identification parameter θ m =[L m R Lm ] T To facilitate code implementation, the differential term of the input quantity needs to be discretized with a sampling period T S Perform backward differential discretization, where T s =1 / f (f is the sampling frequency of the control system):

[0084]

[0085] Among them, i jm (k) represents the current of the jth bridge arm midpoint input into the mth phase grid at the kth moment, i jm (k-1) represents the current of the jth bridge arm midpoint input into the mth phase grid at the k-1th moment. Therefore, formula (7) becomes:

[0086]

[0087] Among them, u jo (k) represents the voltage of the midpoint of the jth bridge arm relative to the midpoint O of the DC bus capacitor at the kth moment; e m (k) represents the m-th phase grid voltage of the distribution network at the k-th moment.

[0088] The least squares method is an online learning method that minimizes the difference between the observed value Q (obtained by sampling) and the model estimate. (in ) is taken as the target, and the model parameters are continuously updated to seek the optimal solution. Specifically, the optimization objective function of the mth phase is:

[0089]

[0090] Where J represents the optimization target value; Indicates Q m The estimated value of Q m represents the output matrix of the mth phase, P m represents the input matrix of the mth phase, Represents the identification parameter matrix θ of the mth phase m The estimated value of .

[0091] Select the identification parameter matrix θ that minimizes the optimization objective function m , so that the estimated value output by the deadbeat controller can be closer to the system observation value (actual sampling value), that is, the optimal estimation is completed. Let the optimization objective function equal to 0, and solve the identification parameter matrix θ by calculation m The expression is:

[0092]

[0093] In order to implement the parameter identification algorithm in a computer, the matrix Q m and Q m The data in the matrix is ​​continuously expanded as new data is sampled. It is defined that after sampling and obtaining new data at the kth moment, the matrix P m P km , the matrix Q m Q km , essentially P km and Q km It is obtained by stacking the new data and the old data vertically, then:

[0094]

[0095] Among them, P km represents the input matrix of the mth phase at the kth moment, P k-1,m represents the input matrix of the mth phase at the k-1th moment, p km represents the input quantity of the mth phase at the kth moment, Q km represents the output matrix of the mth phase at the kth moment, Q k-1,m represents the output matrix of the mth phase at the k-1th moment, q km Represents the output of the mth phase at the kth moment.

[0096] In order to avoid the problem of recalculating the least squares method using formula (12) when using the least squares method for parameter identification, the recursive least squares method, which is more suitable for computer operation, is used. Recursive solution is used in each round to avoid large-scale matrix multiplication operations, minimize the use of chip computing power, and maintain real-time calculation.

[0097] Define the intermediate quantity H km for:

[0098]

[0099] According to formula (12), formula (13), formula (14) and formula (15), the recursive estimate of the identification parameter matrix of the mth phase at the kth moment is for:

[0100]

[0101] β km =H km p km (17)

[0102] in, Represents the recursive estimate of the identification parameter matrix of the mth phase at the k-1th moment.

[0103] In the parameter identification process of the recursive least squares method, in order to effectively prevent new data from being overwhelmed by a large amount of historical data (especially old data), so as to ensure that the system model can more accurately reflect the dynamic characteristics of the current system, the concept of forgetting factor is introduced. The recursive least squares method with forgetting factor adjusts the weight of historical data so that the newer data occupies a larger proportion in the parameter estimation, while the influence of the old data gradually weakens. The formula for parameter identification of the recursive least squares method with forgetting factor is:

[0104]

[0105] q km =u jo (k)-e m (k) (23)

[0106] in, Represents θ km The recursive estimate of θ km represents the identification parameter matrix of the mth phase at the kth moment, Represents θ k-1,m The recursive estimate of θ k-1,m represents the identification parameter matrix of the mth phase at the k-1th moment, m = A, B, C; β km , H k-1,m , All represent intermediate quantities; λ represents the forgetting factor, whose value is between 0 and 0.9; p 1m Represents the input quantity of the mth phase at the first moment.

[0107] A parameter identification algorithm based on the recursive least squares method is used. This algorithm seeks the optimal solution of parameters by minimizing the square of the error. Compared with search algorithms such as particle swarm, the algorithm is simpler and converges faster. In the process of continuous iteration, with the continuous input of sampled data, more and more new data are generated. Therefore, a forgetting factor is introduced on the basis of the recursive least squares method to reduce the weight of old data, enhance the role of new data, improve the adaptive ability of the parameter identification algorithm, and continuously improve the accuracy of parameter identification.

[0108] Step 5: Obtain the command current at the current moment and the actual output current of the energy storage converter at the current moment, and determine whether to start parameter identification calculation according to the command current and the actual output current at the current moment.

[0109] To identify the inductance parameters of the deadbeat controller, it is necessary to clarify the parameter identification startup conditions. During the startup of the energy storage converter or when the energy storage converter fails, the main circuit topology will change, resulting in a fundamental change in the mathematical model of the system. At this time, it is not very meaningful to optimize the deadbeat controller, and incorrect deadbeat parameter optimization may even accelerate the collapse of the entire control system and expand the scope of the fault. Therefore, the present invention uses the real-time current tracking error of the deadbeat controller as the judgment condition, that is, whether the actual current value obtained by the energy storage converter through the control instruction output by the deadbeat controller tracks the given command current value. When the judgment condition is met, the program executes the deadbeat inductance parameter identification, thereby optimizing the mathematical model of the deadbeat controller. When the judgment condition is not met, it is determined that the energy storage converter controller is in a non-steady-state operation mode, that is, a transient mode or a permanent fault condition under common startup states, overcurrent and other fault conditions. The error is defined as:

[0110]

[0111] Among them, I mse represents the real-time current tracking error of the deadbeat controller, represents the m-th phase command current of the energy storage converter at the k-th moment, n represents the number of sampling points in one power frequency cycle, n=f / 50; i jm (k) represents the current of the m-phase grid input into the midpoint of the j-th bridge arm at the k-th moment, that is, the actual output current sampling value of the energy storage converter at the k-th moment.

[0112] The judgment conditions for starting parameter identification calculation are:

[0113] When 0≤I mse When <δ, the system is judged to be in steady-state operation mode and the parameter identification calculation program is started; where δ represents the error threshold.

[0114] When I mseWhen ≥δ, the system is judged to be in a transient process or fault condition. At this time, parameter identification operation is prohibited, and the inductance value remains unchanged or is updated to the system initial setting value.

[0115] Step 6: When the parameter identification calculation is started, the current data and voltage data of the main circuit of the energy storage converter at the current moment are obtained.

[0116] The input matrix of the mth phase, the recursive estimate of the identification parameter matrix, the forgetting factor λ and the intermediate quantity β km Initialize; collect the current data and voltage data of the main circuit of the energy storage converter at the kth moment, and obtain p km and q km , and then get the input matrix P km And the output matrix Q km .

[0117] Step 7: Calculate the identification parameters at the current moment according to the current data and voltage data of the main circuit of the energy storage converter at the current moment and the recursive expression of the identification parameters.

[0118] Substitute the data obtained in step 6 into formulas (18) to (23) to calculate the identification parameter matrix at the current moment:

[0119] Step 8: Determine whether convergence is achieved based on the identification parameters at the current moment and the previous moment; if so, substitute the identification parameters at the current moment into the mathematical model of the deadbeat controller to achieve online optimization of the deadbeat controller; if not, repeat the parameter identification calculation process until convergence.

[0120] In order to accelerate the convergence of identification parameters and reduce the computing power of the processor, the convergence conditions of the given identification parameters are:

[0121]

[0122] Among them, ζ represents the convergence threshold; Represents θ km The recursive estimate of (i), θ km (i) represents the i-th parameter in the identification parameter matrix of the m-th phase at the k-th moment, Represents θ k-1,m The recursive estimate of (i), θ k-1,m (i) represents the i-th parameter in the identification parameter matrix of the m-th phase at the k-1th moment. and Each parameter in is judged to be convergent. When the parameter with the maximum rate of change satisfies formula (25), the entire identification parameter matrix is ​​judged to be The identification parameters in converge. When formula (25) is satisfied, the corresponding Substitute it into the mathematical model of the deadbeat controller to achieve online optimization processing of the deadbeat controller.

[0123] In order to reduce the number of algorithm iterations and speed up the convergence of the optimal solution of the inductor parameters, as well as to reduce the fluctuation of the inductor parameters and the parameter identification data jitter caused by sampling data errors, the convergence conditions of parameter identification (such as formula (25)) are added. After obtaining the real-time values ​​of the inductor parameters, the parameters of the deadbeat controller are updated to increase the accuracy of its mathematical model.

[0124] Example 2

[0125] An embodiment of the present invention further provides an electronic device, which includes: a memory, a processor, and a computer program / instructions stored in the memory, and the processor executes the computer program / instructions to implement the energy storage converter controller optimization method in the embodiment of the present application.

[0126] Although not shown, the electronic device includes a processor, which can perform various appropriate operations and processes according to the program and / or data stored in the read-only memory (ROM) or the program and / or data loaded from the storage portion into the random access memory (RAM). The processor can be a multi-core processor, or it can include multiple processors. In some embodiments, the processor can include a general main processor and one or more special coprocessors, such as a central processing unit, a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), etc. In RAM, various programs and data required for device operation are also stored. The processor, ROM and RAM are connected to each other via a bus. The input / output (I / O) interface is also connected to the bus.

[0127] The processor and memory are used together to execute the program / instructions stored in the memory. When the program / instructions are executed by the computer, the methods, steps or functions described in the above embodiments can be implemented.

[0128] Although not shown, an embodiment of the present invention further provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the energy storage converter controller optimization method in the embodiment of the present application.

[0129] Storage media in embodiments of the present invention include permanent and non-permanent, removable and non-removable items that can be implemented by any method or technology to store information. Examples of storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0130] Readable storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0131] Although not shown, an embodiment of the present invention further provides a computer program product, including: a computer program / instruction, which, when executed by a processor, implements the energy storage converter controller optimization method in the embodiment of the present application.

[0132] What is disclosed above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or modifications within the technical scope disclosed in the present invention, which should be covered within the protection scope of the present invention.

Claims

1. A method for optimizing an energy storage converter controller, characterized in that: The optimization method comprises: According to the main circuit topology of the energy storage converter and Kirchhoff's voltage law, a mathematical model of the energy storage converter is established; Designing an energy storage converter controller, wherein the energy storage converter controller is a double closed-loop controller, the double closed-loop controller comprises an outer loop and an inner loop, and the inner loop is a deadbeat controller; According to the mathematical model of the energy storage converter and the discretization principle, a mathematical model of the deadbeat controller is established; Constructing an optimization objective function according to the mathematical model of the energy storage converter, and determining a recursive expression of an identification parameter according to the optimization objective function; Obtain the current command current and the actual output current of the energy storage converter at the current moment, and determine whether to start parameter identification calculation according to the current command current and the actual output current; When the parameter identification calculation is turned on, the current data and voltage data of the main circuit of the energy storage converter at the current moment are obtained; Calculating the identification parameters at the current moment according to the current data and voltage data of the main circuit of the energy storage converter at the current moment and the recursive expression of the identification parameters; Whether convergence is achieved is determined based on the identification parameters at the current moment and the previous moment; if so, the identification parameters at the current moment are substituted into the mathematical model of the deadbeat controller to achieve online optimization of the deadbeat controller; if not, the parameter identification calculation process is repeated.

2. The energy storage converter controller optimization method according to claim 1, characterized in that: When the main circuit topology of the energy storage converter is a two-level three-phase three-bridge-arm split capacitor topology, the mathematical model of the energy storage converter is: Among them, u jo represents the voltage of the midpoint of the jth bridge arm relative to the midpoint O of the DC bus capacitor, j = a, b, c, and the bridge arms a, b, c correspond to the phases A, B, and C of the distribution network one by one; e A 、e B 、e C Respectively represent the three-phase grid voltage of the distribution network; R LA , R LB , R LC Respectively represent the internal resistance of the three-phase filter inductor on the AC side; i aA 、i bB 、i cC They represent the current of the corresponding phase grid input into the midpoint of each bridge arm; L A , L B , L c They represent the inductance of the three-phase filter inductance on the AC side; C1 and C2 represent the capacity of the upper and lower bus capacitors on the DC side; U dc1 , U dc2 Respectively represent the terminal voltage of the upper and lower bus capacitors on the DC side; S j Indicates the switch state of the jth bridge arm; U dc Indicates the DC bus voltage.

3. The energy storage converter controller optimization method according to claim 1, characterized in that: The outer loop is a power outer loop or a voltage outer loop, and the outer loop adopts a PI controller.

4. The energy storage converter controller optimization method according to claim 1, characterized in that: When the main circuit topology of the energy storage converter is a two-level three-phase three-bridge-arm split capacitor topology, the mathematical model of the deadbeat controller is: Among them, S j (k) represents the switch state of the jth bridge arm at the kth moment, j = a, b, c, and the bridge arms a, b, c correspond to the A, B, C phases of the distribution network one by one; U dc Indicates the DC bus voltage; L A , L B , L c Respectively represent the inductance of the three-phase filter inductance on the AC side; T S Indicates the sampling period; They represent the three-phase command current at the kth moment respectively; i aA (k), i bB (k), i cC (k) represents the current of the corresponding phase grid input into the midpoint of each bridge arm at the kth moment; R LA , R LB , R LC They represent the internal resistance of the three-phase filter inductance on the AC side; e A (k), e B (k), e C (k) represent the three-phase grid voltage of the distribution network at the kth moment respectively.

5. The energy storage converter controller optimization method according to claim 1, characterized in that: The recursive least square method with forgetting factor is used to determine the recursive expression of the identification parameter; wherein the recursive expression of the identification parameter is: in, Represents θ km The recursive estimate of θ km represents the identification parameter matrix of the mth phase at the kth moment, Represents θ k-1,m The recursive estimate of θ k-1,m represents the identification parameter matrix of the mth phase at the k-1th moment, m = A, B, C; β km , H k-1,m , Both represent intermediate quantities; p km represents the input quantity of the mth phase at the kth moment; q km represents the output of the mth phase at the kth moment; λ represents the forgetting factor; p 1m Represents the input quantity of the mth phase at the first moment; the superscript T represents transposition.

6. The energy storage converter controller optimization method according to claim 1, characterized in that: The judgment condition for starting parameter identification calculation is: When 0≤I mse <δ, start parameter identification calculation; when I mse When ≥δ, parameter identification calculation is prohibited; Among them, δ represents the error threshold, I mse represents the real-time current tracking error of the deadbeat controller; the real-time current tracking error I mse for: in, represents the m-th phase command current of the energy storage converter at the k-th moment, n represents the number of sampling points in one power frequency cycle, n=f / 50, f represents the sampling frequency; i jm (k) represents the current input into the m-phase grid at the midpoint of the j-th bridge arm at the k-th moment, and the bridge arms a, b, and c correspond to the A, B, and C phases of the distribution network one by one.

7. The energy storage converter controller optimization method according to any one of claims 1 to 6, characterized in that: The convergence is determined based on the identification parameters at the current moment and the previous moment, and the convergence condition is: in, Represents θ km The recursive estimate of (i), θ km (i) represents the i-th parameter in the identification parameter matrix of the m-th phase at the k-th moment, Represents θ k-1,m The recursive estimate of (i), θ k-1,m (i) represents the i-th parameter in the identification parameter matrix of the m-th phase at the k-1-th time; T S represents the sampling period; ζ represents the convergence threshold.

8. An electronic device comprising a memory, a processor, and a computer program / instruction stored in the memory, characterized in that: The processor executes the computer program / instructions to implement the energy storage converter controller optimization method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the energy storage converter controller optimization method according to any one of claims 1 to 7 is implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the energy storage converter controller optimization method according to any one of claims 1 to 7 is implemented.

Images