Inverter model prediction current amplitude limiting control method under unbalanced power grid

Abstract

The invention belongs to the technical field of power electronics, and particularly relates to a model prediction control method and a T-type three-level three-phase inverter device thereof. The invention provides a model prediction current amplitude limiting control method for a T-type three-level three-phase inverter under an unbalanced power grid based on four-vector synthesis space vector pulse width modulation. A reference voltage is obtained through model prediction, after the reference voltage is obtained, based on a space vector pulse width modulation strategy, a vector synthesis mode is adopted, a voltage vector is divided into 36 sectors according to different switching states, a switching state corresponding to the generated reference voltage is obtained through a four-vector synthesis mode, and a fixed switching frequency is obtained; and a multi-target cost function is established according to a reference voltage and a direct-current side capacitor voltage difference, a resultant vector is symmetrically distributed through a time inverse proportion function, finally, the output current of the inverter is controlled to be kept at a low harmonic distortion rate under balanced and unbalanced working conditions and is effectively limited within a specified range, and meanwhile, the direct-current side neutral-point potential is kept balanced.

Description

Technical Field

[0001] The invention belongs to the technical field of power electronics, and in particular relates to an inverter model prediction current limiting control method under an unbalanced power grid. Background Art

[0002] Since Japanese scholars proposed the three-level inverter structure in the 1880s, multi-level inverters have developed rapidly. Multi-level inverters have many output levels, small harmonics, low voltage change rate caused by switching state changes, and small filter inductance to reduce system losses and costs, making them widely used in high-voltage and high-power applications.

[0003] The three-phase T-type three-level inverter is a three-phase three-level inverter. By adding a clamping branch to the traditional half-bridge circuit, the improvement from two levels to three levels is realized. A pair of switch tubes are used to form a bidirectional switch to realize the clamping function of the two switch tubes connected to the positive and negative poles of the DC power supply. Compared with the I-type three-level inverter, the T-type three-level inverter uses two fewer diodes, which reduces the space of the inverter, and there is no problem of the switch tube shutdown sequence, which is easier to control; at the same time, the I-type circuit requires 4 independent power supplies to power the drive, while the T-type circuit only needs three. Moreover, the commutation paths in the T-type inverter circuit are consistent. When discrete devices are used in the circuit, there is no problem of stray inductance and voltage spikes. Due to the above advantages, the T-type three-level inverter is very suitable for scenarios such as photovoltaic grid-connected power generation.

[0004] When the power system is running normally and stably, the traditional PI control can achieve good control effects and realize efficient operation of the grid-connected inverter. However, the power grid often suffers from imbalance, such as single-phase grounding, high-power load access, etc. At this time, the active and reactive power output by the grid-connected inverter will oscillate at double frequency, and the current will also be distorted, resulting in reduced power quality of the grid and affecting the safe operation of the grid-connected inverter. At the same time, the grid-connected rules require the grid-connected inverter to have low voltage ride-through capability, but during the low voltage ride-through process, due to the presence of negative sequence voltage, the inverter grid-connected current will increase significantly, and in severe cases, the switching devices will be damaged, resulting in fault ride-through failure.

[0005] Model predictive control builds a mathematical model of the optimization problem and controls the state of the switch tube by comparing the cost function. It is simple to achieve and has fast dynamic response. It does not require multiple PI controllers, which reduces the difficulty of adjusting and coordinating the various parameters. Since the traditional finite control set model predictive control strategy has a large amount of calculation and the switching frequency is not fixed, the harmonic components of the output current increase. Therefore, model predictive control based on space vector modulation strategy is adopted to achieve fast response, small amount of calculation, and low harmonic distortion rate control requirements. Considering the problem of a significant increase in the output current of the grid-connected inverter under an unbalanced power grid fault, in order to avoid the trouble of using the calculation method to obtain the reference power and reference current, the power outer loop is used and the limiting strategy is added to obtain the reference current as the input of the model predictive control. The structure is simple and easy to debug. Summary of the invention

[0006] In view of the problems existing in the background technology, the purpose of the present invention is to provide a model prediction current limiting control method for grid-connected inverters under unbalanced power grids, mainly based on 4-vector synthesis space vector pulse width modulation. The method is applied to the scene of unbalanced fault of grid voltage and grid-side load switching. The reference current is obtained by adding a limiting strategy to the PQ power outer loop, and the reference voltage is obtained by model prediction. After obtaining the reference voltage, based on the space vector pulse width modulation strategy, the voltage vector is divided into 36 sectors according to different switch states by using a vector synthesis method, and the switch state corresponding to the generated reference voltage is obtained by using a 4-vector synthesis method, and a multi-objective cost function is established with the reference voltage and the DC side capacitor voltage difference, and the synthetic vector is symmetrically distributed by the time inverse function, and finally the inverter is controlled to maintain a low harmonic distortion rate under balanced and unbalanced working conditions and effectively limit it within the specified range. At the same time, the DC side midpoint potential is kept balanced. This control method avoids the complex parameter setting process, and improves the dynamic response speed while reducing harmonic distortion.

[0007] To achieve the above object, the technical solution of the present invention is as follows:

[0008] A model predictive current limiting control method for a T-type three-level three-phase inverter under an unbalanced power grid based on four-vector synthesis space vector pulse width modulation, which is characterized by comprising the following steps:

[0009] Step 1: Obtain the physical parameters of the T-type three-level three-phase inverter and collect the three-phase grid voltage u through sensors g , inverter output three-phase voltage u abc , output current i abc , and use the decoupled dual synchronous reference coordinate system phase-locked loop to detect the phase θ of the grid voltage and perform abc-dq coordinate system transformation, and calculate the active and reactive power output by the inverter.

[0010] Step 2: Use the PQ power outer loop to calculate the grid-connected current reference value, and use the current limiting instruction to ensure that the current reference value does not exceed the set value.

[0011] Step 3: Construct the circuit differential equation in the two-phase stationary coordinate system, and use the forward difference method to discretize the circuit mathematical model to obtain the output current prediction value i at time k+1 p (k+1). Then use the model predictive control principle to inversely calculate the reference voltage vector u output by the inverter ref (k).

[0012] Step 4: The DC side capacitor current is represented by the DC side output current and the switch state. The dynamic model equations of the upper and lower capacitors on the DC side are discretized using the forward difference method and the difference is calculated, which is the predicted value of the DC side capacitor voltage difference.

[0013] Step 5: Based on the space vector pulse width modulation strategy, the voltage vector is divided into 36 regions. Each region contains 4 voltage vectors, that is, 4 different switching states. According to the average value equivalent principle, u s =d1u1+d2u2+d3u3+d4u4, where u s is the resultant vector, u i is the basic vector in the corresponding area, and d represents the duty cycle of the corresponding basic vector.

[0014] Step 6: According to the model predictive control principle, it is necessary to construct a cost function g between each basic vector and the reference voltage vector i , which contains additional control items of the absolute value of the switch voltage error and the absolute value of the DC link capacitor voltage difference.

[0015] Step 7: Calculate the duty cycle of each vector in the region and synthesize the cost function G for each region j The minimum cost function is obtained by comparison and the corresponding switching pulses are output to control the inverter operation and realize model predictive control.

[0016] Furthermore, the current limiting instruction expression in step 2 is as follows:

[0017]

[0018] Among them, i' kref Represents the reference current after current limiting, i max Indicates the maximum value set for the grid current.

[0019] Furthermore, the circuit differential equation in step 3 is expressed as follows:

[0020]

[0021] Among them, i α 、u α 、e α are the inverter output three-phase current, inverter output phase voltage and grid-connected point voltage in the α-β coordinate system, and L and R are the three-phase load inductance and resistance respectively.

[0022] The output current prediction value obtained by discretization using the forward difference method in step 3 is as follows:

[0023]

[0024] Among them, T s Represents the sampling time. Using the model predictive control principle, the reference voltage vector u output by the inverter is inversely calculated. ref (k) are as follows:

[0025]

[0026] Furthermore, the space vector modulation sector allocation in step 5 is to first divide the space into 6 large sectors, and in each small sector, the areas surrounded by small vectors, small vectors and medium vectors, and medium vectors and large vectors are divided into 4 areas, and the areas surrounded by small vectors, small vectors and medium vectors are divided into two equal parts at 30° in the middle, thus dividing 36 areas.

[0027] Furthermore, the vector action duty cycle and the vector cost function g described in step 5 are i Inversely proportional, the specific expression is as follows

[0028]

[0029] Furthermore, the cost function between each basic vector and the voltage reference vector in step 6 is expressed as follows:

[0030] g i =||v ref_α -v i_α || 2 +||v ref_β -v i_β || 2 +λ(Δvc(k+1)) 2

[0031] Among them, the subscripts α and β represent the two phases in the two-phase stationary coordinate system, and λ represents the weight factor of the additional control term of the midpoint potential.

[0032] Furthermore, the cost function G of each region in step 7 is j It is expressed as follows

[0033]

[0034] Furthermore, the voltage vector pulse output described in step 7 is in each sampling period T s In the inverter, the four basic vector sequences composed of the synthetic vectors are output to the inverter in a forward and backward symmetrical manner in a certain order.

[0035] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0036] The present invention takes into account the serious distortion of the output current of the unbalanced power grid inverter, the sharp increase of the inverter output current under unbalanced fault, and the unbalanced potential of the midpoint of the DC side capacitor, and changes the way of obtaining the reference current by the traditional calculation method, adopts the PQ power outer loop plus limiting strategy to output the reference current, and has simple debugging and good effect; changes the traditional T-type inverter model predictive control method, and applies the model predictive control based on the space vector pulse width modulation strategy to the unbalanced power grid control, while realizing the fixing of the switching frequency and the synchronous prediction of the current, and well controls the DC side midpoint potential to maintain balance, thereby ensuring the rapid response and stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A control structure block diagram of a T-type three-level three-phase inverter model prediction current limiting control method applied to an unbalanced power grid provided by the present invention.

[0038] Figure 2 This is the circuit topology diagram of a T-type three-level three-phase inverter.

[0039] Figure 3 This is a schematic diagram of the a-phase bridge arm model and switch state encoding of a T-type three-level three-phase inverter.

[0040] Figure 4 It is the complete vector diagram of space vector pulse width modulation of T-type three-level three-phase inverter. DETAILED DESCRIPTION

[0041] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the implementation modes and the accompanying drawings.

[0042] The block diagram of the model predictive control method of T-type three-level three-phase inverter under unbalanced power grid based on 4-vector synthesis space vector pulse width modulation of the present invention is as follows Figure 1 As shown, the block diagram mainly includes a main circuit module, a phase-locked loop, a power calculation module, a PQ power outer loop plus a limiting current instruction, a reference voltage prediction model based on zero-beat control theory and a DC side capacitor voltage difference prediction model, a model predictive control based on space vector pulse width modulation, and a drive output of a three-phase inverter.

[0043] The control process is: sampling grid voltage V g, inverter output voltage V o , output current I o , through the phase-locked loop, all three-phase AC signals are transformed into a synchronous rotating coordinate system through coordinate system transformation, and the power output of the inverter is calculated. The reference current is generated by adding a limiting current instruction to the power outer loop. The predicted output voltage model and the DC side capacitor voltage difference prediction model are obtained based on the model predictive control theory. The predicted output voltage model and the DC side capacitor voltage difference prediction model are substituted into the cost function to obtain the optimal vector combination and its action time. The obtained optimal pulse sequence is used to control the inverter output to realize the model predictive control of the inverter.

[0044] A model predictive current limiting control method for a T-type three-level three-phase inverter under an unbalanced power grid based on four-vector synthesis space vector pulse width modulation, which is characterized by comprising the following steps:

[0045] Step 1: Obtain the physical parameters of the T-type three-level three-phase inverter and collect the three-phase grid voltage u through sensors g , inverter output three-phase voltage u abc , output current i abc , and use the decoupled dual synchronous reference frame phase-locked loop to detect the phase θ of the grid voltage and perform abc-dq coordinate system transformation, and calculate the active and reactive power output by the grid-connected inverter, as follows

[0046]

[0047] Among them, P and Q are active power and reactive power respectively, and the subscript αβ represents the two-phase stationary coordinate system. is the conjugate component of i. Decomposing equation (1), we can get the expressions of instantaneous active power p and reactive power q:

[0048]

[0049] Among them, P0 and Q0 are the average components of P and Q respectively, P c2 and Q c2 are the amplitudes of the cosine double frequency components of P and Q, respectively. s2 and Q s2 are the amplitudes of the sinusoidal double frequency components of P and Q respectively. The calculation formula is as follows

[0050]

[0051] in, are respectively the positive and negative sequence components of the grid voltage in the dq reference coordinate system, They are respectively the positive and negative sequence components of the grid-connected current in the dq reference coordinate system.

[0052] Step 2: Use the PQ power outer loop to calculate the grid-connected current reference value i ref The current limiting instruction is used to ensure that the current reference value does not exceed the set value, as follows

[0053]

[0054] Among them, i' kref represents the reference current after current limiting in the dq coordinate system, i max Indicates the maximum value of the grid current setting, let i max =25A.

[0055] Step 3: Construct the circuit differential equation in the two-phase stationary coordinate system as follows

[0056]

[0057] Among them, i α 、u α 、e α are the inverter output three-phase current, inverter output phase voltage and grid-connected point voltage in the α-β coordinate system, and L and R are the three-phase load inductance and resistance respectively.

[0058] The forward difference method is used to discretize the circuit mathematical model to obtain the output current prediction value i at time k+1 p (k+1), as follows

[0059]

[0060] Where i(k) is the current value at the current moment, i(k+1) is the current value at the next moment, T s is the sampling period, substituting (6) into (5) to obtain

[0061]

[0062] Then, the model predictive control principle is used to inversely calculate the reference voltage vector u output by the inverter. ref (k) as follows:

[0063]

[0064] Step 4: Express the DC side capacitor current ic using the DC side output current and switch state, as follows

[0065]

[0066] Among them, i dc is the inverter DC side current, i a 、i b 、i cRespectively represent the current output by each phase of the inverter, s a1 、s b1 ,

[0067] ---s c1 and a2 、s b2 、s c2 Indicates the state of the switch tube, 1 for on and 0 for off. Figure 2 The DC side capacitor voltage dynamic equation is as follows:

[0068]

[0069] The subscript i indicates the number of capacitors, i = 1, 2. The dynamic model equations of the upper and lower capacitors on the DC side are discretized using the forward difference method, as follows:

[0070]

[0071] Then calculate the difference between the two capacitance prediction values, which is the predicted value of the DC side capacitance voltage difference, as follows

[0072]

[0073] Since the upper and lower capacitors on the DC side are the same, the predicted value of the DC side capacitor voltage difference can be expressed by the DC side midpoint voltage difference and midpoint current, as follows:

[0074]

[0075] Among them, Δvc represents the DC side midpoint voltage difference, and io represents the midpoint current.

[0076] Step 5: Based on the space vector pulse width modulation strategy, the space is first divided into 6 large sectors. In each small sector, the area surrounded by the small vector, the small vector and the medium vector, and the medium vector and the large vector are divided into 4 areas, and the area surrounded by the small vector, the small vector and the medium vector are equally divided into two parts at 30° in the middle, thus dividing 36 areas. Each area contains 4 voltage vectors, that is, it contains 4 different switch state combinations, such as Figure 3 , Figure 4 Each basic voltage vector is represented by a combination of p, n, and o, where p, o, and n represent the corresponding three-phase output as high level, 0 level, and low level, respectively.

[0077] In order to make the switch state and action time distribution corresponding to the basic vector simple and consistent, the negative short vector is selected as the starting vector of each sampling cycle, and the action order of the basic vectors in each switch state combination follows the following rules: any change in the basic voltage vector can only have one bridge arm switch action, that is, only one bit in its three-phase level can change. Therefore, the selection of the 0 vector can be selected according to the combination of switch states. The centrosymmetric seven-segment space vector pulse width modulation method is used to allocate the action time of each basic vector to the corresponding vector state. Taking the first sector as an example, the synthetic vector combination number and pulse sequence are shown in the following table

[0078]

[0079] According to the volt-second balance principle, u s =d1u1+d2u2+d3u3+d4u4, where u s is the resultant vector, u i (i=1,2,3,4) is the basic vector in the corresponding sector, d represents the duty cycle of the corresponding basic vector, which is inversely proportional to the vector cost function. The specific expression is as follows

[0080]

[0081] Among them, g i represents the cost function between the base vector and the reference voltage vector.

[0082] Step 6: According to the model predictive control principle, it is necessary to construct a cost function g between each basic vector and the reference voltage vector i , which includes additional control items of the absolute value of the switch voltage error and the absolute value of the DC side capacitor voltage difference, as follows

[0083] g i =||v ref_α -v i_α || 2 +||v ref_β -v i_β || 2 +λ(Δvc(k+1)) 2 (17)

[0084] The subscripts α and β respectively represent the two phases in the two-phase stationary coordinate system.

[0085] Step 7: Calculate the duty cycle of each vector in the sector and synthesize the cost function G for each sector j , as follows

[0086]

[0087] By comparing and obtaining the minimum cost function and outputting the corresponding switching pulses to control the inverter operation, model predictive control is realized. The corresponding switching pulse action time calculation formula is as follows

[0088] t i =T s ·d i (19)

[0089] The above description is only a specific implementation mode of the present invention. Any feature disclosed in this specification, unless otherwise stated, can be replaced by other alternative features that are equivalent or have similar purposes; all the disclosed features, or all the steps in the methods or processes, except for mutually exclusive features and / or steps, can be combined in any way.

Claims

1. A method for controlling the current limiting of an inverter under an unbalanced power grid by using a model-based prediction method, the method comprising the following steps: Step 1: Obtain the physical parameters of the T-type three-level three-phase inverter and collect the three-phase grid voltage u through sensors g , inverter output three-phase voltage u abc , output current i abc , and use the decoupled dual synchronous reference frame phase-locked loop to detect the phase θ of the grid voltage and perform abc-dq frame transformation, and calculate the instantaneous active power P and reactive power Q output by the inverter; Step 2: Use the PQ power outer loop to calculate the grid-connected current reference value, and use the current limiting instruction to ensure that the current reference value does not exceed the set value; Step 3: Construct the circuit differential equation in the two-phase stationary coordinate system, and use the forward difference method to discretize the circuit mathematical model to obtain the output current prediction value i at time k+1 p (k+1), and then use the model predictive control principle to inversely calculate the reference voltage vector u output by the inverter ref (k); Step 4: The DC side capacitor current is expressed by the DC side output current and the switch state, and the dynamic model equations of the upper and lower capacitors of the DC side are discretized using the forward difference method and the difference is calculated, which is the predicted value of the DC side capacitor voltage difference; Step 5: Based on the space vector pulse width modulation strategy, the voltage vector is divided into 36 regions. Each region contains 4 voltage vectors, that is, 4 different switching states. According to the average value equivalent principle, u s =d1u1+d2u2+d3u3+d4u4, where u s is the resultant vector, u i is the basic vector in the corresponding area, and d represents the duty cycle of the corresponding basic vector; Step 6: According to the model predictive control principle, it is necessary to construct a cost function g between each basic vector and the reference voltage vector i , which contains additional control items of the absolute value of the switch voltage error and the absolute value of the DC side capacitor voltage difference; Step 7: Calculate the duty cycle of each vector in the region and synthesize the cost function G for each region j , by comparing and obtaining the minimum cost function and outputting the corresponding switching pulses to control the inverter operation and realize model predictive control.

Images