Multi-step model predictive control method based on hysteresis control and DSP controller

Abstract

The present application relates to a multi-step model predictive control method and a DSP controller based on hysteresis control, and belongs to the field of high-performance model predictive control technology. The present application can use a linear extrapolation method to predict the prediction parameters corresponding to the voltage vector in multiple steps, if any parameter of the prediction parameters corresponding to the voltage vector at the previous moment does not meet the preset hysteresis range condition, and there are prediction parameters corresponding to the voltage vector in the voltage vector set corresponding to the target moment that meet the hysteresis range condition, until any parameter of the prediction parameters corresponding to the voltage vector does not meet the hysteresis range condition, and obtain the prediction step corresponding to the voltage vector; construct a cost function according to the prediction step corresponding to the voltage vector, and determine the target voltage vector that matches the switching range of the switch sequence; use the target voltage vector to output pulses to drive the inverter switch tube, thereby improving the model predictive control effect and making it more practical.

Description

Technical Field

[0001] The present application belongs to the technical field of high-performance model predictive control, and specifically relates to a multi-step model predictive control method based on hysteresis control and a DSP (Digital Signal Processing) controller. Background Art

[0002] In the field of electric drive, the finite control set-model predictive control (FCS-MPC) method has been widely used due to its simple concept, easy consideration of system constraints and excellent multivariable control capability.

[0003] The existing finite state set model predictive control method selects the optimal voltage vector within a single step. In practice, it is found that this method has problems such as high switching frequency and poor control performance. In order to solve this problem, existing scholars have proposed some solutions, but most of these methods are relatively complex and not very practical.

[0004] In response to the above problems, there is an urgent need for a model predictive control method with better control effect and higher practicality. Summary of the invention

[0005] To this end, the present application provides a multi-step model predictive control method based on hysteresis control and a DSP controller, which helps to improve the model predictive control effect and has higher practicality.

[0006] To achieve the above objectives, this application adopts the following technical solutions:

[0007] In a first aspect, the present application provides a multi-step model predictive control method based on hysteresis control, the method comprising:

[0008] Determine a switch sequence switching range, and control the switch sequence according to the switch sequence switching range; wherein the switch sequence switching range includes a plurality of switch sequence states;

[0009] Calculate the prediction parameters corresponding to the voltage vector at the previous moment according to the grid-side voltage, grid-side current and DC bus voltage difference at the target moment; wherein the prediction parameters include active power prediction value, reactive power prediction value and DC bus voltage difference prediction value;

[0010] If any parameter of the prediction parameters corresponding to the voltage vector at the previous moment does not satisfy the preset hysteresis range condition, then for each voltage vector in the voltage vector set corresponding to the target moment, determine the prediction parameter corresponding to the voltage vector;

[0011] If all prediction parameters corresponding to the voltage vector satisfy the hysteresis range condition, then the prediction parameters corresponding to the voltage vector are predicted in multiple steps by using a linear extrapolation method until any parameter among the prediction parameters corresponding to the voltage vector does not satisfy the hysteresis range condition, thereby obtaining a prediction step length corresponding to the voltage vector;

[0012] Constructing a first cost function corresponding to the voltage vector according to the prediction step length corresponding to the voltage vector;

[0013] Determining a target voltage vector matching the switching sequence switching range based on a first cost function corresponding to each voltage vector in the voltage vector set;

[0014] The target voltage vector is used to output pulses to drive the inverter switch tube.

[0015] Furthermore, the method further comprises:

[0016] If the prediction parameters corresponding to the voltage vector at the previous moment all satisfy the hysteresis range condition, the voltage vector at the previous moment is determined as the target voltage vector.

[0017] Furthermore, the method further comprises:

[0018] If there is no voltage vector whose prediction parameters all satisfy the hysteresis range condition in the voltage vector set corresponding to the target time, for each voltage vector in the voltage vector set, determining a second cost function corresponding to the voltage vector based on a single-step model predictive control category;

[0019] Based on a second cost function corresponding to each voltage vector in the voltage vector set, a target voltage vector matching the switching sequence switching range is determined.

[0020] Furthermore, the second cost function is as follows:

[0021]

[0022] in, is the second cost function, p ref is the preset active power reference value, p j (k+1) is the predicted value of active power, q ref is the preset reactive power reference value, q j (k+1) is the reactive power prediction value, K r is the preset weight coefficient used to adjust the midpoint potential balance. is the predicted value of DC bus voltage difference.

[0023] Furthermore, the method further comprises:

[0024] According to the voltage vector at the previous moment, a voltage vector set corresponding to the target moment is determined.

[0025] Further, the calculation of the prediction parameter corresponding to the voltage vector at the previous moment according to the grid-side voltage, grid-side current and DC bus voltage difference at the target moment includes:

[0026] According to a preset complex power prediction formula, the grid-side voltage at the target moment, the grid-side current and the voltage vector at the previous moment are substituted to obtain a complex power prediction value; wherein the complex power prediction formula is as follows:

[0027]

[0028] Wherein, S(k+1) is the complex power prediction value, S(k) is the complex power at the target time, and t sc is the preset control period, e(k) is the grid-side voltage, i(k) is the grid-side current, u(k) is the voltage vector at the previous moment, L is the inductance, R is the resistance, and ω is the fundamental angular frequency;

[0029] Determine the real part of the complex power prediction value as the active power prediction value corresponding to the voltage vector at the previous moment; and determine the imaginary part of the complex power prediction value as the reactive power prediction value corresponding to the voltage vector at the previous moment;

[0030] According to the preset DC bus voltage difference calculation formula, the DC bus voltage difference prediction value corresponding to the voltage vector at the previous moment is calculated; wherein the preset DC bus voltage difference calculation formula is as follows:

[0031]

[0032] Among them, V orr (k+1) refers to the DC bus voltage difference prediction value corresponding to the voltage vector at the previous moment, t sc is the preset control period, S a (k), S b (k) and S c (k) is the switch state function of each channel at the target time, i a (k), i b (k) and i c (k) is the current of each channel at the target time, C is the DC bus capacitance, V orr (k) is the DC bus voltage difference at the target moment.

[0033] Furthermore, the switch sequence state includes a switch sequence switching state, a switch sequence holding state and a switch sequence initial holding state.

[0034] Furthermore, the first cost function is as follows:

[0035]

[0036] Among them, c j is the first cost function, is the prediction step length, ||u j (l)-u j (l-1)||1 represents u j (l) and u j (l-1), k is the target time, u j (l) refers to the voltage vector at time l, u j (l-1) refers to the voltage vector at time l-1.

[0037] Further, the target voltage vector is used to output pulses to drive inverter switches, including:

[0038] The target voltage vector is used to output a pulse to a driving circuit, so that the driving circuit generates a driving signal for the inverter switch tube based on the pulse, so as to drive the inverter switch tube based on the driving signal.

[0039] In a second aspect, the present application provides a DSP controller, which is configured to execute the above-mentioned multi-step model predictive control method based on hysteresis control.

[0040] This application adopts the above technical solution, which has at least the following beneficial effects:

[0041] This application uses the linear extrapolation method to predict the prediction parameters corresponding to the voltage vector in multiple steps, thereby determining the optimal target voltage vector. Compared with the single-step model predictive control method, the linear extrapolation method is used for multi-step prediction, which can expand the prediction range, thereby helping to improve the model predictive control effect. In addition, this solution only needs to calculate the prediction parameters and construct the cost function to determine the optimal target voltage vector, and the calculation complexity is relatively reasonable and more practical.

[0042] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 is a flow chart of a multi-step model predictive control method based on hysteresis control according to an exemplary embodiment;

[0045] Figure 2 is a schematic diagram of the hardware structure of a converter control system including a DSP controller according to an exemplary embodiment;

[0046] Figure 3 is a structural block diagram of a converter control system including a DSP controller executing a multi-step model predictive control method based on hysteresis control according to an exemplary embodiment;

[0047] Figure 4 This is the experimental result of the traditional single-step model predictive control at a sampling rate of 20kHz and a load of 1kW;

[0048] Figure 5 This is the experimental result of using multi-step model predictive control based on hysteresis control at a sampling rate of 20kHz with a load of 1kW;

[0049] Figure 6 This is the experimental result of the traditional single-step model predictive control with a power step from 600W to 1000W at a 20kHz sampling rate;

[0050] Figure 7 It is the experimental result of using multi-step model predictive control based on hysteresis control to step the power from 600W to 1000W at a sampling rate of 20kHz. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other implementation methods obtained by ordinary technicians in the field without creative work belong to the scope of protection of the present application.

[0052] See also Figure 1 , Figure 1 : is a flowchart of a multi-step model predictive control method based on hysteresis control according to an exemplary embodiment. The multi-step model predictive control method based on hysteresis control includes the following steps:

[0053] Step S11, determining a switch sequence switching range, and controlling the switch sequence according to the switch sequence switching range; wherein the switch sequence switching range includes a plurality of switch sequence states.

[0054] In this embodiment, the execution subject may be a DSP controller. In addition, this solution may be applied to the hardware circuit of the converter control system for circuit control. For a detailed description of the hardware circuit of the converter control system, please refer to Figure 2 The detailed description is omitted here. Specifically, the optimal voltage vector can be determined based on hysteresis control and multi-step model prediction. The driving signal for driving the inverter switch tube is obtained using the optimal voltage vector.

[0055] In this embodiment, the switching range of the switching sequence refers to the switching range of the switching sequence in the hardware circuit of the converter control system, specifically, a range consisting of multiple switching sequence states. For example, the switching sequence state may include a switching sequence switching state, denoted as "S", a switching sequence holding state, denoted as "E", and a switching sequence initial holding state, denoted as "e". Wherein e refers to the switching sequence holding state set in the initialization of the program. Further, the switching sequence switching range composed of these switching sequence states may include but is not limited to "eSE", "eSESE", "eSESESE", etc., which is not limited in this embodiment. For example, for the switching sequence switching range "eSE", it means that the switching sequence is from the initial holding state to the switching sequence switching state, and then from the switching sequence switching state to the switching sequence holding state. For the "eSE" whose switching range of the switching sequence includes one or more "SE", the optimal target voltage vector determined is a vector, wherein the more "SE", the longer the prediction range considered.

[0056] Optionally, the execution subject may determine a first requirement for control accuracy and a second requirement for control efficiency of the current control requirement, and determine a corresponding switch sequence switching range based on the first requirement and the second requirement. The state of the switch sequence is controlled according to the switch sequence switching range, and a subsequent optimal voltage vector determination operation is performed under the corresponding switch sequence state, so that the determination of the voltage vector is more flexible and adaptable to different control accuracy and control efficiency requirements.

[0057] Step S12, calculating the prediction parameters corresponding to the voltage vector at the previous moment according to the grid-side voltage, grid-side current and DC bus voltage difference at the target moment; wherein the prediction parameters include active power prediction value, reactive power prediction value and DC bus voltage difference prediction value.

[0058] In this embodiment, the execution subject can collect the grid-side voltage, grid-side current and DC bus voltage difference at the target time, and substitute the grid-side voltage, grid-side current, DC bus voltage difference and voltage vector at the previous time into the prediction formula to calculate the prediction parameters corresponding to the voltage vector at the previous time. The prediction parameters here can include the predicted active power prediction value, reactive power prediction value and DC bus voltage difference prediction value at the next time.

[0059] It should be noted that the target moment here can be the current sampling moment, and the sampling control period can be pre-set in the execution subject. The next moment relative to the target moment can be the moment after the sampling control period. The previous moment relative to the target moment can be the moment before the sampling control period. The setting of the sampling control period can be flexibly set based on different requirements, and this embodiment does not limit this.

[0060] Step S13: if any parameter of the prediction parameters corresponding to the voltage vector at the previous moment does not satisfy the preset hysteresis range condition, then for each voltage vector in the voltage vector set corresponding to the target moment, determine the prediction parameter corresponding to the voltage vector.

[0061] In this embodiment, for each parameter in the active power prediction value, the reactive power prediction value, and the DC bus voltage difference prediction value, the parameter can be compared with the hysteresis range corresponding to the parameter to determine whether the parameter is in the hysteresis range corresponding to the parameter. If any parameter is not in the hysteresis range corresponding to the parameter, it is determined that the parameter does not meet the preset hysteresis range condition, and the voltage vector set corresponding to the target time is further obtained. The execution subject can pre-store the hysteresis range corresponding to each parameter.

[0062] Afterwards, the execution subject can traverse each voltage vector in the voltage vector set. Specifically, the execution subject can initialize the optimal objective function, let c opt = 0. Then select the first voltage vector u from the voltage vector set F(k) j (k), remove u from the voltage vector set F(k) j (k). And, j (k), determine its corresponding prediction parameters.

[0063] Step S14: If the prediction parameters corresponding to the voltage vector all satisfy the hysteresis range condition, the prediction parameters corresponding to the voltage vector are predicted in multiple steps using a linear extrapolation method until any of the prediction parameters corresponding to the voltage vector does not satisfy the hysteresis range condition, thereby obtaining a prediction step length corresponding to the voltage vector.

[0064] In this embodiment, if the above u j (k) The corresponding prediction parameters are all within the corresponding hysteresis range, then u is determined j(k) If the corresponding prediction parameters all meet the hysteresis range conditions, the linear extrapolation method is used to predict the prediction parameters of the voltage vector in multiple steps until any parameter in the prediction parameters corresponding to the voltage vector is not within the corresponding hysteresis range, and the prediction step length is obtained. Among them, the linear extrapolation method refers to a method used to study things that change at a constant growth rate over time. Specifically, in a coordinate diagram with time as the horizontal axis, the changes of things are close to a straight line. Based on this straight line, the future changes of things can be inferred. The multi-step here refers to calculating the above prediction parameters again from the current moment k to the above sampling period. For example, the moment at which the current moment k is separated by one sampling period can be k+1, and the moment at which the current moment is separated by two sampling periods can be k+2. The execution subject can predict the prediction parameters of the voltage vector at the moment k+1 and the prediction parameters at the moment k+2. If there are parameters in the prediction parameters at the moment k+3 that are not within the corresponding hysteresis range, the extrapolation is terminated, and the step length at this time is 2.

[0065] Step S15: construct a first cost function corresponding to the voltage vector according to the prediction step size corresponding to the voltage vector.

[0066] In this embodiment, the execution entity may construct a first cost function according to the prediction step length, wherein the first cost function may also be preferably constructed based on the number of switch jumps and the prediction step length to measure the cost of the voltage vector.

[0067] Step S16: determining a target voltage vector matching the switching range of the switching sequence based on a first cost function corresponding to each voltage vector in the voltage vector set.

[0068] In this embodiment, the execution subject may determine the first cost value c corresponding to the voltage vector based on the first cost function corresponding to the voltage vector. j , if c j >c opt , then let c opt =c j And u(k)=u j (k). Here u(k) is the optimal target voltage vector finally determined. Then, the next voltage vector u is selected from the voltage vector set F(k). j (k), until all voltage vectors in the voltage vector set F(k) are traversed. By looping through the voltage vectors in the voltage vector set, calculating the first cost function, and updating the target voltage vector, the voltage vector with the largest cost value can be selected. At this time, the maximum cost value indicates that the cost of the voltage vector is the smallest. Specifically, the cost value may be proportional to the prediction step size and inversely proportional to the number of switch jumps.

[0069] Step S17: Utilize the target voltage vector to output pulses to drive inverter switches.

[0070] In this embodiment, after determining the optimal target voltage vector, the execution subject may use the target voltage vector to output pulses to drive the inverter switch tube.

[0071] Furthermore, the method further comprises:

[0072] If the prediction parameters corresponding to the voltage vector at the previous moment all satisfy the hysteresis range condition, the voltage vector at the previous moment is determined as the target voltage vector.

[0073] In this embodiment, if the prediction parameters obtained by using the voltage vector u(k-1) at the previous moment are all within the hysteresis range, the voltage vector u(k-1) at the previous moment is determined as the target voltage vector u(k).

[0074] Furthermore, the method further comprises:

[0075] If there is no voltage vector whose prediction parameters all satisfy the hysteresis range condition in the voltage vector set corresponding to the target time, for each voltage vector in the voltage vector set, determining a second cost function corresponding to the voltage vector based on a single-step model predictive control category;

[0076] Based on a second cost function corresponding to each voltage vector in the voltage vector set, a target voltage vector matching the switching sequence switching range is determined.

[0077] In this embodiment, if there is no voltage vector whose prediction parameters are all within the hysteresis range in the voltage vector set F(k) corresponding to the target time, the second cost function of each voltage vector can be determined based on the principle of the single-step model predictive control category.

[0078] Furthermore, the second cost function is as follows:

[0079]

[0080] in, is the second cost function, p ref is the preset active power reference value, p j (k+1) is the predicted value of active power, q ref is the preset reactive power reference value, q j (k+1) is the reactive power prediction value, K r is the preset weight coefficient used to adjust the midpoint potential balance. is the predicted value of DC bus voltage difference.

[0081] Afterwards, the execution subject may select the voltage vector with the smallest second cost value as the optimal voltage vector based on the second cost function, wherein the smallest second cost value means the smallest cost.

[0082] Specifically, select the first element u from the voltage vector set F(k) j (k) and remove it from F(k) to perform element traversal, initialize the optimal objective function, and let c opt =0 and To u j (k), if its output variable p at time k+1 j (k+1),q j (k+1) and If they are both within their respective hysteresis ranges, the linear extrapolation method is used to record the prediction step length. And calculate the first cost function c j , in this branch you can update c opt and u(k). Among them, p j (k+1) is the predicted value of active power, q j (k+1) is the reactive power prediction value, is the predicted value of the DC bus voltage difference.

[0083] If its output variable p at time k+1 j (k+1),q j (k+1) and If they are not within their respective hysteresis ranges, the traditional single-step model predictive control principle is used to calculate the second cost function In this branch, you can make And u dl (k) = u j (k), that is, update and u dl (k).

[0084] By implementing this embodiment, at time k, if at least one voltage vector in u(k-1) or the voltage vector set F(k) can make the output variables all within their respective hysteresis ranges, then the voltage vector u(k) is applied to the system; otherwise, the voltage vector u dl (k) acts on the system, thereby improving the accuracy of determining the optimal voltage vector.

[0085] Furthermore, the method further comprises:

[0086] According to the voltage vector at the previous moment, a voltage vector set corresponding to the target moment is determined.

[0087] In this embodiment, the execution subject determines the voltage vector set F(k) allowed at time k according to the voltage vector u(k-1) at time k-1. Assuming that the voltage vector acting at time k-1 is v1, the voltage vector allowed at time k is F(k)={v2,v 12 ,v 13 ,v 14 The rule for determining the voltage vector set here is to select a voltage vector that will not undergo a large jump as the voltage vector allowed at the next moment. The execution subject can pre-store multiple voltage vectors. Some voltage vectors will cause a large jump when implemented adjacently. For this situation, this solution determines the voltage vector F(k) allowed at moment k based on whether a large jump will occur. In this way, the reliability of determining the voltage vector set can be improved.

[0088] Further, the calculation of the prediction parameter corresponding to the voltage vector at the previous moment according to the grid-side voltage, grid-side current and DC bus voltage difference at the target moment includes:

[0089] According to a preset complex power prediction formula, the grid-side voltage at the target moment, the grid-side current and the voltage vector at the previous moment are substituted to obtain a complex power prediction value; wherein the complex power prediction formula is as follows:

[0090]

[0091] Wherein, S(k+1) is the complex power prediction value, S(k) is the complex power at the target time, and t ec is the preset control period, s(k) is the grid-side voltage, i(k) is the grid-side current, u(k) is the voltage vector at the previous moment, L is the inductance, R is the resistance, and ω is the fundamental angular frequency;

[0092] Determine the real part of the complex power prediction value as the active power prediction value corresponding to the voltage vector at the previous moment; and determine the imaginary part of the complex power prediction value as the reactive power prediction value corresponding to the voltage vector at the previous moment;

[0093] According to the preset DC bus voltage difference calculation formula, the DC bus voltage difference prediction value corresponding to the voltage vector at the previous moment is calculated; wherein the preset DC bus voltage difference calculation formula is as follows:

[0094]

[0095] Among them, V orr (k+1) refers to the DC bus voltage difference prediction value corresponding to the voltage vector at the previous moment, t sc is the preset control period, S a (k), Sb (k) and S c (k) is the switch state function of each channel at the target time, i a (k), i b (k) and i c (k) is the current of each channel at the target time, C is the DC bus capacitance, V orr (k) is the DC bus voltage difference at the target moment.

[0096] In this embodiment, when calculating the prediction parameters for the next moment, it is necessary to substitute the grid-side voltage, grid-side current and voltage vector into the above complex power prediction formula to calculate the complex power prediction value. Then, based on the complex power prediction value, the active power p(k+1)=Re(S(k+1)) and reactive power q(k+1)=Im(S(k+1)) at the moment k+1 are obtained. At the same time, when calculating the DC bus voltage difference V at the moment k+1, orr When (k+1), it is necessary to substitute the switch state functions of each channel at the target time and the currents of each channel at the target time into the preset DC bus voltage difference calculation formula to obtain V orr (k+1).

[0097] Each time the prediction parameter corresponding to each voltage vector in the voltage vector set is calculated, the above formula may be substituted.

[0098] Furthermore, the switch sequence state includes a switch sequence switching state, a switch sequence holding state and a switch sequence initial holding state.

[0099] Furthermore, the first cost function is as follows:

[0100]

[0101] Among them, c j is the first cost function, is the prediction step length, ||u j (l)-u j (l-1)||1 represents u j (l) and u j (l-1), k is the target time, u j (l) refers to the voltage vector at time l, u j (l-1) refers to the voltage vector at time l-1. Taking the number of switching transitions into account in the cost function can reduce the average switching frequency of the device.

[0102] Further, the target voltage vector is used to output pulses to drive inverter switches, including:

[0103] The target voltage vector is used to output a pulse to a driving circuit, so that the driving circuit generates a driving signal for the inverter switch tube based on the pulse, so as to drive the inverter switch tube based on the driving signal.

[0104] This application uses the linear extrapolation method to predict the prediction parameters corresponding to the voltage vector in multiple steps, thereby determining the optimal target voltage vector. Compared with the single-step model predictive control method, the linear extrapolation method is used for multi-step prediction, which can expand the prediction range, thereby helping to improve the model predictive control effect. In addition, this solution only needs to calculate the prediction parameters and construct the cost function to determine the optimal target voltage vector. The calculation complexity is reasonable and the practicality is higher. In addition, taking the number of switch jumps into account in the cost function can reduce the average switching frequency of the device. In addition, the multi-step model predictive control scheme based on hysteresis control can significantly improve the control effect while reducing the requirements for the sampling frequency.

[0105] See also Figure 2 , Figure 2 FIG. 1 is a schematic diagram of a hardware structure of a converter control system including a DSP controller according to an exemplary embodiment. Figure 2 As shown, the converter control system hardware circuit including the DSP controller may include a three-phase voltage source 201, a load 206, a three-level PWM converter 202, a DC side capacitor, a voltage and current sampling circuit 203, a DSP controller 204 and a drive circuit 205. The voltage and current sampling circuit 203 may use a voltage Hall sensor and a current Hall sensor to respectively collect the DC side voltage and the AC side a and b phase voltage and current, that is, e a 、e b 、i a 、i b The sampled signal passes through the signal conditioning circuit and enters the DSP controller 204 to be converted into a digital signal. The DSP controller 204 completes the operation of the method proposed by the present invention, outputs twelve switching pulses, and then passes through the driving circuit 205 to obtain the final driving signal of the twelve switching tubes of the inverter.

[0106] See also Figure 3 , Figure 3 FIG. 1 is a block diagram showing a converter control system including a DSP controller executing a multi-step model predictive control method based on hysteresis control according to an exemplary embodiment. Figure 3 As shown, the voltage and current sampling circuit 203 obtains the DC side voltage and the AC side a and b phase voltage and current, that is, e a 、e b 、i a 、i b, the DSP controller 204 can perform calculation steps 301 and 302, perform 3 / 2 transformation and calculate active power and reactive power based on these voltages and currents. Afterwards, after passing, calculation step 302 can be performed to calculate p k+1 ,q k+1 Among them, p k+1 ,q k+1 are the predicted values ​​of active power and reactive power respectively. Afterwards, the execution subject can determine the optimal target voltage vector according to the predicted parameters and the switching range of the switching sequence. Figure 3 An adder 207 is also included. Figure 3 p ref ,q ref They are active power reference value and reactive power reference value, voltage U dc For U dc1 , U dc2 The sum of S1, S2, and S3 corresponds to the above S a (k), S b (k) and S c (k) is the DC voltage reference value.

[0107] See also Figure 4 , Figure 4 This is the experimental result of the traditional single-step model predictive control at a sampling rate of 20kHz and a load of 1kW. Please refer to Figure 5 , Figure 5 The experimental results are based on the multi-step model predictive control with hysteresis control at a sampling rate of 20kHz and a load of 1kW. Figure 4 and Figure 5 As shown in the figure, the waveforms from top to bottom are active power reference value, actual active power, actual reactive power, DC bus voltage difference and phase a current. Figure 5 It can be seen that the active power, reactive power and DC bus voltage difference are all controlled within the hysteresis range. By recording the number of actions of the switching device, it can be obtained that the average switching frequency of the traditional method is 2000Hz, and the average switching frequency of the method described in the present invention is 600Hz. Therefore, it can be obtained that the present invention can effectively reduce the average switching frequency.

[0108] See also Figure 6 , Figure 6 This is the experimental result of the traditional single-step model predictive control with a power step from 600W to 1000W at a 20kHz sampling rate. Please refer to Figure 7 , Figure 7 The experimental results are as follows: The power is stepped from 600W to 1000W at a sampling rate of 20kHz using a multi-step model predictive control based on hysteresis control. Figure 6 and Figure 7 As shown, in the dynamic process, the method described in the present invention has a fast dynamic performance similar to the traditional method, and both active and reactive power accurately track the reference value.

[0109] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0110] It should be noted that, in the description of this application, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, the meaning of "plurality" and "multiple" refers to at least two.

[0111] It should be understood that when an element is referred to as being "fixed to" or "disposed on" another element, it can be directly on the other element or there may be an intermediate element at the same time; when an element is referred to as being "connected to" another element, it can be directly connected to the other element or there may be an intermediate element at the same time. In addition, the "connection" used here may include wireless connection; the wording "and / or" used includes any unit and all combinations of one or more associated listed items.

[0112] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0113] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0114] A person skilled in the art may understand that all or part of the steps in the above-mentioned embodiment method may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0115] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0116] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0117] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0118] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A multi-step model predictive control method based on hysteresis control, characterized in that: The method comprises: Determine a switch sequence switching range, and control the switch sequence according to the switch sequence switching range; wherein the switch sequence switching range includes a plurality of switch sequence states; Calculate the prediction parameters corresponding to the voltage vector at the previous moment according to the grid-side voltage, grid-side current and DC bus voltage difference at the target moment; wherein the prediction parameters include active power prediction value, reactive power prediction value and DC bus voltage difference prediction value; If any parameter of the prediction parameters corresponding to the voltage vector at the previous moment does not satisfy the preset hysteresis range condition, then for each voltage vector in the voltage vector set corresponding to the target moment, determine the prediction parameter corresponding to the voltage vector; If all prediction parameters corresponding to the voltage vector satisfy the hysteresis range condition, then the prediction parameters corresponding to the voltage vector are predicted in multiple steps by using a linear extrapolation method until any parameter among the prediction parameters corresponding to the voltage vector does not satisfy the hysteresis range condition, thereby obtaining a prediction step length corresponding to the voltage vector; Constructing a first cost function corresponding to the voltage vector according to the prediction step length corresponding to the voltage vector; Determining a target voltage vector matching the switching sequence switching range based on a first cost function corresponding to each voltage vector in the voltage vector set; The target voltage vector is used to output pulses to drive the inverter switch tube.

2. The method according to claim 1, characterized in that The method further comprises: If the prediction parameters corresponding to the voltage vector at the previous moment all satisfy the hysteresis range condition, the voltage vector at the previous moment is determined as the target voltage vector.

3. The method according to claim 1, characterized in that The method further comprises: If there is no voltage vector whose prediction parameters all satisfy the hysteresis range condition in the voltage vector set corresponding to the target time, for each voltage vector in the voltage vector set, determining a second cost function corresponding to the voltage vector based on a single-step model predictive control category; Based on a second cost function corresponding to each voltage vector in the voltage vector set, a target voltage vector matching the switching sequence switching range is determined.

4. The method according to claim 3, characterized in that The second cost function is as follows: in, is the second cost function, p ref is the preset active power reference value, p j (k+1) is the predicted value of active power, q ref is the preset reactive power reference value, q j (k+1) is the reactive power prediction value, K r is the preset weight coefficient used to adjust the midpoint potential balance. is the predicted value of DC bus voltage difference.

5. The method according to claim 1, characterized in that The method further comprises: According to the voltage vector at the previous moment, a voltage vector set corresponding to the target moment is determined.

6. The method according to claim 1, characterized in that The step of calculating the prediction parameter corresponding to the voltage vector at the previous moment according to the grid-side voltage, grid-side current and DC bus voltage difference at the target moment includes: According to a preset complex power prediction formula, the grid-side voltage at the target moment, the grid-side current and the voltage vector at the previous moment are substituted to obtain a complex power prediction value; wherein the complex power prediction formula is as follows: Wherein, S(k+1) is the complex power prediction value, S(k) is the complex power at the target time, and t sc is the preset control period, e(k) is the grid-side voltage, i(k) is the grid-side current, u(k) is the voltage vector at the previous moment, L is the inductance, R is the resistance, and ω is the fundamental angular frequency; Determine the real part of the complex power prediction value as the active power prediction value corresponding to the voltage vector at the previous moment; and determine the imaginary part of the complex power prediction value as the reactive power prediction value corresponding to the voltage vector at the previous moment; According to the preset DC bus voltage difference calculation formula, the DC bus voltage difference prediction value corresponding to the voltage vector at the previous moment is calculated; wherein the preset DC bus voltage difference calculation formula is as follows: Among them, V orr (k+1) refers to the DC bus voltage difference prediction value corresponding to the voltage vector at the previous moment, t sc is the preset control period, S a (k), S b (k) and S c (k) is the switch state function of each channel at the target time, i a (k), i b (k) and i c (k) is the current of each channel at the target time, C is the DC bus capacitance, V orr (k) is the DC bus voltage difference at the target moment.

7. The method according to claim 1, characterized in that The switch sequence state includes a switch sequence switching state, a switch sequence holding state and a switch sequence initial holding state.

8. The method according to claim 1, characterized in that The first cost function is as follows: Among them, c j is the first cost function, is the prediction step length, Indicates u j (l) and u j (l-1), k is the target time, u j (l) refers to the voltage vector at time l, u j (l-1) refers to the voltage vector at time l-1.

9. The method according to claim 1, characterized in that: The step of using the target voltage vector to output pulses to drive inverter switches includes: The target voltage vector is used to output a pulse to a driving circuit, so that the driving circuit generates a driving signal for the inverter switch tube based on the pulse, so as to drive the inverter switch tube based on the driving signal.

10. A DSP controller, characterized in that: The DSP controller is configured to execute the multi-step model predictive control method based on hysteresis control as claimed in any one of claims 1 to 9.

Images