Improved Sliding Mode Active Disturbance Rejection Control Method for Two-Stage Energy Storage Converters

Abstract

This invention discloses an improved sliding mode active disturbance rejection control method for a two-stage energy storage converter, specifically including the following steps: Step 1, establishing a mathematical model of the two-stage energy storage converter in the d-q rotating coordinate system; Step 2, converting the mathematical model in Step 1 into a second-order active disturbance rejection paradigm, and then designing a controller to achieve control of the two-stage energy storage converter. This invention solves the problems of unstable converter output signal and system chattering in traditional control methods when the energy storage converter experiences power surges.

Description

Technical Field

[0001] This invention belongs to the field of electronic power technology and relates to an improved sliding mode active disturbance rejection control method for a two-stage energy storage converter. Background Technology

[0002] In the context of new energy sources, energy storage converters, as key devices connecting renewable energy to the power grid, are of paramount importance in terms of performance stability and efficiency. With the large-scale integration of new energy sources such as wind power and photovoltaics into the grid, energy storage converters need to cope with the complex and ever-changing power environment, including challenges such as intermittency, volatility, and uncertainty.

[0003] Active disturbance rejection control (ADRC), as an advanced control method, has shown great potential in the control of energy storage converters due to its independence from precise mathematical models, strong robustness, and adaptive capabilities. This control method estimates and compensates for uncertain disturbances inside and outside the system, achieving precise control of the output voltage and current of the energy storage converter, thereby improving system stability and power quality. Energy storage converters not only need to efficiently convert energy but also possess rapid response and intelligent adjustment capabilities to cope with complex operating conditions such as grid frequency fluctuations and sudden load changes. Sliding mode ADRC, through its unique control logic and algorithm design, can significantly improve the control performance of energy storage converters in complex environments, providing strong support for the stable grid connection of new energy sources and the safe operation of the power grid. However, due to the high-frequency switching of control signals in sliding mode control, the converter output control signal may oscillate when the system power changes abruptly. Effectively suppressing system chattering during the control process remains a key challenge. Summary of the Invention

[0004] The purpose of this invention is to provide an improved sliding mode active disturbance rejection control method for a two-stage energy storage converter, which solves the problems of unstable converter output signal and system chattering when the power of the energy storage converter changes suddenly in the traditional control method.

[0005] The technical solution adopted in this invention is an improved sliding mode active disturbance rejection control method for a two-stage energy storage converter, which specifically includes the following steps:

[0006] Step 1: Establish a mathematical model of the two-stage energy storage converter in the dq rotating coordinate system; Step 2: Convert the mathematical model in Step 1 into a second-order active disturbance rejection paradigm, and then design the controller to achieve control of the two-stage energy storage converter.

[0007] The invention is further characterized by: The specific process of step 1 is as follows: During the grid-connected operation of the converter, the operating states of the two switches on each bridge arm are complementary. Therefore, the switching function is defined as follows: (1) In equation (1), j = a , b , c Represents AC power grid a , b , c Three phases; According to Kirchhoff's laws, the voltage and current relationship equations of the converter in the abc three-phase stationary coordinate system can be obtained as follows: (2) In equation (2), u a , u b , u c This refers to the three-phase voltage on the AC side of the converter. i sa , i sb , i sc This refers to the three-phase current on the AC side. e a , e b , e c AC mains voltage; L For filtering inductors; C dc These are the parameters of the DC-side capacitor; R These are the equivalent resistance parameters of the line; i r This is the output current of the DC-side converter; u dc , i C DC bus voltage and current; s ka , s kb , s kc They are respectively a , b , c Three-phase bridge arm switch control signals; Performing the Park transformation on equation (2), we obtain the mathematical model of the converter in the dq coordinate system as follows: (3) In equation (3), ω The angular frequency of the grid voltage; e d 、i d 、s d and eq 、i q 、s q d-axis 、 The q-axis represents the grid voltage, current, and switching functions; u dc This is the DC bus voltage.

[0008] The specific process of step 2 is as follows: Step 2.1, design of the power outer loop controller, which provides a current reference value for the current inner loop; Step 2.2, design of the inner current loop controller, the observer adopts a finite-time extended state observer, and the controller adopts a quasi-continuous integral terminal sliding mode controller.

[0009] The specific process of step 2.1 is as follows: According to the instantaneous power theory, in order for the energy storage converter to be connected to the grid with near unity power factor, the reactive current must be always equal to 0. Under the condition of three-phase grid voltage balance, the active power on the grid side of the two-stage energy storage converter is... P and reactive power Q The instantaneous value is expressed as: (4) From equation (4), we know that through i d and i q Control separately P and Q This allows for independent regulation of active and reactive power. By transforming equation (4), when... P ref , Q ref Given a constant, the reference values ​​of the d-axis and q-axis currents are directly obtained through calculation; a PI regulator is introduced to eliminate steady-state errors. P , Q and P , Q The deviation of the command value is adjusted by a PI controller, and then, based on instantaneous power theory, the system... P and Q The dynamic response is then used to obtain the active current reference value. i dref and reactive current reference value i qref As shown in the following formula: (5) In equation (5), P ref , Q ref Given reference values ​​for active and reactive power, idref , i qref These are the reference values ​​for the inner loop of the d-axis and q-axis currents. k dp , k di and k qp , k qi These are the PI adjustment parameters for the d and q axes, respectively.

[0010] The specific process of step 2.2 is as follows: Step 2.2.1: Based on the feedforward error system, design a finite-time extended state observer; Step 2.2.2, Design of a second-order quasi-continuous adaptive integral terminal sliding mode controller.

[0011] The specific process of step 2.2.1 is as follows: First, by transforming equation (3), we obtain the state variable dq-axis current. i d , i q The first-order differential form is as follows: (6) The unmodeled portion of the dq-axis current inner loop coupling term, as well as the internal and external disturbances of the inner loop, are considered as the total inner loop disturbance, and are respectively used as... f d1 , f q1 This means that by differentiating equation (6), we can convert it into ADRC normal form: (7) In equation (7), , , b To control the gain of the quantity; definition b o To estimate the control gain, equation (7) is transformed into the following form: (8) In equation (8), , , b o This is an estimate of the control quantity gain; For the above second-order system, the feedforward error signal is taken. , i ref Given the current reference value, calculate the second derivative and substitute it into equation (8) to obtain: (9) Selecting state variables , , Then the system shown in equation (9) is transformed into: (10) In equation (10), , representing the lumped disturbance along the dq axis; For the feedforward error system shown in equation (10), the lumped disturbance of the dq axis is... f d , f q Expand into new state variables z d3 , z q3 The observers for the dq axes were designed separately.

[0012] In step 2.2.1, the d-axis finite-time observer is designed as follows: (11) The q-axis observer is designed as follows: (12) In equations (11)-(12), e d1 , e q1 It represents the difference between the observed value and the actual value. z d1 , z q1 for x d1 , x q1 The estimated value; z d2 , z q2 for x d1 , x q1 An estimate of the derivative; z d3 , z q3 The total disturbance including coupling terms, internal disturbances, and unmodeled parts. f d , f q Estimated value; , , ， , , This represents the observer gain, and all values ​​are greater than 0. , , for the observer adjustment parameters, F( e i ) is a hyperbolic function.

[0013] In step 2.2.1, due to the sign function sign( x The discontinuity exists, so the discontinuity switching term sign function in the traditional disturbance observer is replaced by a hyperbolic function F( e i ) instead, hyperbola F( e i The function is represented as: (13) In the formula, n The size determines F( e i The slope of ).

[0014] The specific process of step 2.2.2 is as follows: Let e ​​be the system state error d, the deviation between the actual q-axis current and the reference current, as shown below: (14) According to equation (14), the current deviation e Design of the sliding surface of the integral terminal S as follows: (15) In equation (15), p 1, p 2, q 1, q Both 2 are positive odd numbers, and , ; , , , The adaptive control parameters for the sliding surface are expressed as follows: (16) In equation (16), , These are adaptive gain parameters, all of which are real numbers greater than 0; The sliding surface of equation (15) S Differentiating, we get: (17) (18) For the feedforward error system shown in equation (10), substituting into the above differential result, we get: (19) (20) Since the system converges in a finite time, according to equivalent control theory, let Substituting the finite-time ESO variable observations of the d and q axes in equations (11) and (12) into equations (18) and (20), the equivalent control law of the system is as follows: (twenty one) As shown in equations (18) and (20), the sliding surface S The control quantity first appears in the second-order differential. s Therefore, the relative order of the system is 2. A second-order quasi-continuous algorithm is introduced as the sliding mode switching control law, which takes the following form: (twenty two) In equation (22), l The proportional gain for switching controllers, sgn( x ) is a symbolic function; Using the saturation function sat( S ) replaces the symbolic function sgn( S The dq axis switching control law is designed in the following form: (twenty three) In equation (23), l d , l d These are the proportional coefficients of the dq axis switching controller, and sat( S ) is a saturation function; Combining equations (21) and (23), we obtain the overall control law of the system. s d , s q The expression is as follows: (twenty four) Substituting equations (21) and (23) into equation (24), we obtain the second-order quasi-continuous sliding mode controller of the energy storage converter system as shown in equation (25): (25).

[0015] The beneficial effects of this invention are as follows: 1. In order to reduce system chattering and improve the observer's ability to observe total inner-loop disturbances, the discontinuous switching term sign function in the traditional finite-time extended state observer is replaced by a hyperbolic function, thereby improving the observer's performance and its ability to observe disturbances.

[0016] 2. Compared to traditional sliding mode control or other disturbance observation methods, the finite-time ESO observer has the characteristic of "finite-time convergence". It can accurately estimate the disturbance in the system within a finite time, achieve high-precision estimation of the disturbance in the system, and compensate for it, so that the system can reach a steady state more quickly, improving the response speed and dynamic performance of the converter system when faced with power surges.

[0017] 3. This invention introduces a feedforward error system. Since feedforward control is performed before the system output variable deviates, it is not affected by system lag, can respond to system changes more quickly, and can better handle nonlinear systems and model uncertainties.

[0018] 4. This invention introduces a quasi-continuous algorithm as a sliding mode switching controller. As a high-order sliding mode control algorithm, it can generate continuous control signals and effectively suppress chattering, thus regulating the system state more efficiently and stably. Simultaneously, it employs a saturation function `sat(` x ) replace the sign function sign( x This further suppressed high-frequency jitter in the system and improved the system's operational stability.

[0019] 5. In traditional sliding mode control, the system may experience chattering due to the presence of high-frequency switching variables. However, integral terminal sliding mode uses integration to obtain the actual control variable, which does not contain high-frequency switching variables. Therefore, there is no chattering in the system, and it can guarantee finite-time convergence and has a faster convergence speed.

[0020] 6. By designing appropriate adaptive parameters for the sliding surface, the sliding variables can be better adjusted, thereby improving the system control accuracy and convergence speed.

[0021] 7. The improved finite-time ESO used in this invention can accurately estimate disturbances in the system within a finite time, providing real-time disturbance information to the quasi-continuous integral terminal sliding mode controller. The sliding mode controller then uses these estimates to precisely control the system, achieving real-time compensation for disturbances. The combined effect of these two mechanisms enables the system to respond more accurately to external disturbances, improving control accuracy and stability, shortening convergence time, and reducing steady-state error. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the overall control principle of the two-stage PCS in the improved sliding mode active disturbance rejection control method for the two-stage energy storage converter of the present invention. Figure 2 This is the main circuit topology of the energy storage converter in the improved sliding mode active disturbance rejection control method for the two-stage energy storage converter of the present invention; Figure 3 The hyperbola F( ) in the improved sliding mode active disturbance rejection control method for the two-stage energy storage converter of this invention.e i ) function and sign( x )contrast; Figure 4 This is a block diagram of the energy storage converter control in the improved sliding mode active disturbance rejection control method for the two-stage energy storage converter of the present invention; Figure 5(a) shows the simulation comparison curves of current response under the PI control method; Figure 5(b) shows the simulation comparison curves of the current response under the control of the improved sliding mode active disturbance rejection control method of the two-stage energy storage converter of the present invention. Figure 6(a) shows the transient current response of P when it jumps from 30kW to 80kW under the PI control method; Figure 6(b) shows the transient current response of P jumping from 30kW to 80kW under the control of the improved sliding mode active disturbance rejection control method of the two-stage energy storage converter of the present invention. Figure 7(a) shows the transient response when the active power P changes from 30kW to -30kW under PI control. Figure 7(b) shows the transient response of P when the active power direction switches from 30kW to -30kW under the control of the improved sliding mode active disturbance rejection control method of the two-stage energy storage converter of the present invention. Figure 8(a) shows the current harmonic analysis under PI control; Figure 8(b) shows the harmonic analysis during control under the improved sliding mode active disturbance rejection control method of the two-stage energy storage converter of the present invention. Detailed Implementation

[0023] The following detailed description is provided in conjunction with specific implementation methods.

[0024] Example 1 This invention relates to an improved sliding mode active disturbance rejection control method for a two-stage energy storage converter, such as... Figure 1As shown, this paper presents an improved sliding mode active disturbance rejection control method for a two-stage energy storage converter based on a feedforward finite-time ESO (Extended State Observer). First, a mathematical model of the energy storage converter in the dq rotating coordinate system is established. To simplify the controller design, the current inner loop coupling term and unmodeled portion along with the internal and external disturbances of the inner loop are considered as the total inner loop disturbance and expanded into a new state variable. Based on this, it is converted into a second-order active disturbance rejection paradigm. For the aforementioned second-order system, a feedforward error signal is selected, and a feedforward error system is established. Based on this, a finite-time ESO is designed to observe and compensate for the lumped disturbance. The discontinuous switching term sign function in the traditional observer is replaced with a hyperbolic function to improve the disturbance observation capability. Then, a second-order quasi-continuous integral terminal sliding mode is designed as a feedback controller. As a high-order sliding mode controller, the quasi-continuous controller can generate continuous control signals and effectively suppress chattering. Integral-terminal sliding mode ensures finite-time convergence and offers faster convergence speed. Furthermore, by using integration to obtain the actual control input, it eliminates high-frequency switching, thus preventing chattering. In summary, to further improve system stability and reliability, this invention combines a quasi-continuous algorithm with the integral-terminal sliding mode concept to design a second-order quasi-continuous integral-terminal sliding mode controller. This invention maintains stable converter output when power surges occur in the energy storage converter, effectively suppressing excessive DC bus voltage overshoot and coupling between the d and q axes, while reducing transient time, allowing the system to reach a stable state more quickly.

[0025] Example 2 This invention relates to an improved sliding mode active disturbance rejection control method for a two-stage energy storage converter. The overall frame is as follows: Figure 1 As shown, the specific steps are as follows: Step 1: Establish a mathematical model of the two-stage energy storage converter in the dq rotating coordinate system; Step 2: Convert the mathematical model from Step 1 into a second-order active disturbance rejection paradigm, and design the controller accordingly.

[0026] Example 3 The specific process of step 1 is as follows: A two-stage energy storage converter is a widely used device in the fields of new energy and power systems. It enables bidirectional conversion of electrical energy and plays a crucial role in both energy storage and power supply. It consists of two cascaded converters, including a grid-side converter and a DC-DC converter, with the circuit topology as follows: Figure 2As shown. When the battery pack operates in discharge mode, the energy stored inside is first boosted to the preset DC bus voltage by the front-end DC / DC converter. Then, the boosted DC power is inverted into AC power by the subsequent converter to feed back to the grid or supply the load to compensate for insufficient grid power supply. In charging mode, the grid AC power is first rectified by the subsequent converter to become DC power. This DC power enters the front-end DC / DC module, where its voltage is reduced to a level suitable for charging the battery pack, thus charging the energy storage device. The converter section adopts... PQ Control (constant power control). PQ Active power control is an important method in power systems, primarily used to maintain the stability and reliability of the system. Its core principle is to control the current to ensure that the output power reaches a preset value, thereby achieving precise control of the active and reactive power in the power system.

[0027] Figure 2 middle, u in This refers to the DC-side power supply voltage of the converter. u a , u b , u c This refers to the three-phase voltage on the AC side of the converter. i sa , i sb , i sc This refers to the three-phase current on the AC side. e a , e b , e c AC mains voltage; L , C These are the parameters of the filter inductor and capacitor; R These are the equivalent resistance parameters of the line; s 1~ s 6 represents the six switching transistors of the upper and lower bridge arms of the converter; i r This is the output current of the DC-side converter; i o Input current to the converter; u dc , i C DC bus voltage and current; L dc , C dc These are the DC-side inductor and capacitor, respectively, and the currents flowing through them are respectively... i Land i C Assume that the three-phase voltage waveforms of the power grid are symmetrical and without distortion.

[0028] During grid-connected operation, the operating states of the two switches on each bridge arm are complementary. Therefore, the switching function is defined as follows: (1) In equation (1), j = a , b , c Represents AC power grid a , b , c Three phases.

[0029] Depend on Figure 2 According to Kirchhoff's laws, the voltage and current relationship equations of the converter in the abc three-phase stationary coordinate system can be obtained as follows: (2) In equation (2), u a , u b , u c This refers to the three-phase voltage on the AC side of the converter. i sa , i sb , i sc This refers to the three-phase current on the AC side. e a , e b , e c AC mains voltage; L For filtering inductors; C dc These are the parameters of the DC-side capacitor; R These are the equivalent resistance parameters of the line; i r This is the output current of the DC-side converter; u dc , i C DC bus voltage and current; s ka , s kb , s kc They are respectively a , b , c Three-phase bridge arm switch control signal.

[0030] By performing the Park transformation on equation (2), we can obtain the mathematical model of the converter in the dq coordinate system as follows: (3) In equation (3), ω The angular frequency of the grid voltage; e d 、i d 、s d and e q 、i q 、s q d-axis 、 The q-axis represents the grid voltage, current, and switching functions; u dc This is the DC bus voltage.

[0031] Example 4 Step 2.1, design of the power outer loop controller, which provides a current reference value for the current inner loop; Step 2.2, design of the inner current loop controller, the observer adopts a finite-time extended state observer, and the controller adopts a quasi-continuous integral terminal sliding mode controller.

[0032] Example 5 The specific process of step 2.1 is as follows: This invention employs converter control PQ Control methods. PQ Control typically involves two key control loops: the power outer loop and the current inner loop. The outer loop determines the active and reactive power that the converter should provide based on the system's operating requirements, according to the set parameters. P and Q The corresponding dq-axis current reference value is calculated from the value. i dref , i qref The inner current loop monitors the current component at the converter output in real time. i d and i q It is compared with the current reference value generated by the outer loop to guide the converter to output specific active and reactive power. PQThe core of control lies in decoupling active and reactive power, enabling them to be adjusted independently. The main function of the outer power loop is to ensure power conservation during grid connection of the converter, providing a reference value for the subsequent inner current loop. According to instantaneous power theory, in order for the energy storage converter to achieve near-unity power factor grid connection, the reactive current must be constant at zero. Under the condition of three-phase grid voltage balance, the active power on the grid side of the two-stage energy storage converter... P and reactive power Q The instantaneous value is expressed as: (4) As can be seen from equation (4), it can be achieved through... i d and i q Control separately P and Q This allows for independent regulation of active and reactive power. Transforming equation (4), when... P ref , Q ref Given a constant current, the reference values ​​for the d-axis and q-axis currents can be directly obtained through calculation; however, to eliminate steady-state errors, a PI controller is introduced. P , Q and P , Q The deviation of the command value is adjusted by a PI controller, and then, based on instantaneous power theory, the system... P and Q The dynamic response is then used to obtain the active current reference value. i dref and reactive current reference value i qref As shown in the following formula: (5) In equation (5), P ref , Q ref Given reference values ​​for active and reactive power, i dref , i qref These are the reference values ​​for the inner loop of the d-axis and q-axis currents. k dp , k di and k qp , k qi These are the PI adjustment parameters for the d and q axes, respectively.

[0033] Example 6 The specific process of step 2.2 is as follows: Step 2.2.1: Based on the feedforward error system, design a finite-time extended state observer; Step 2.2.2, Design of a second-order quasi-continuous adaptive integral terminal sliding mode controller.

[0034] Example 7 The specific process of step 2.2.1 is as follows: To reduce system chattering and improve the disturbance observer's ability to observe the total inner-loop disturbance, the discontinuous switching term `sign` in the traditional finite-time extended state observer is replaced with a hyperbolic function. This improves the observer's performance and enhances its ability to observe disturbances. The following section describes an improved design of the finite-time extended state observer based on feedforward theory: First, by transforming equation (3), we can obtain the state variable dq-axis current. i d , i q The first-order differential form is as follows: (6) To simplify controller design, the inner loop coupling term of the dq axis current, the unmodeled portion, and the internal and external disturbances of the inner loop are considered as the total inner loop disturbance, and are respectively represented by... f d1 , f q1 This means that differentiating equation (6) transforms it into the ADRC (Active Disturbance Rejection Control) paradigm: (7) In equation (7), , , b To control the gain of the quantity.

[0035] Due to practical engineering calculations b The value of is generally unknown and cannot be accurately determined, therefore it is defined as . b o To estimate the gain of the control quantity, equation (7) can be transformed into the following form: (8) In equation (8), , , b o This is an estimate of the control gain.

[0036] For the above second-order system, the feedforward error signal is taken. , i refGiven the current reference value, we take its second derivative and substitute it into equation (8) to obtain: (9) Selecting state variables , , Then the system shown in equation (9) can be transformed into: (10) In equation (10), , representing the lumped disturbance along the dq axis.

[0037] For the feedforward error system shown in equation (10), the lumped disturbances fd and fq along the dq axis are expanded into new state variables. z d3 , z q3 Therefore, the observers for the d and q axes can be designed separately. The finite-time observer for the d axis is designed as follows: (11) The q-axis observer is designed as follows: (12) In equations (11)-(12), ed1 and eq1 are the differences between the observed and actual values; zd1 and zq1 are the estimated values ​​of xd1 and xq1; zd2 and zq2 are the estimated values ​​of the derivatives of xd1 and xq1; and zd3 and zq3 are the estimated values ​​of the total disturbance fd and fq, including coupling terms, internal disturbances, and unmodeled parts. , , , , , This represents the observer gain, and all its values ​​are greater than 0. , Let F(ei) be the observer adjustment parameter, and F(ei) be a hyperbolic function. Because the sign function sign(x) is discontinuous, causing the system to be prone to chattering, this invention replaces the discontinuous switching term sign function in the traditional disturbance observer with the hyperbolic function F(ei). e i This is replaced by a hyperbola F( ) to improve the ability to observe disturbances. e i) The function is represented as: (13) hyperbolic function F( e i The image is as follows Figure 3 As shown, by Figure 3 It can be seen that, n The size determines F(e i The slope of ) n The larger F( e i The closer it gets to the sign function sign( x ) image, and switching function F( e i ) It is continuous and smooth, with no discontinuities, which theoretically can reduce the system chattering problem and improve the observation capability of the disturbance observer.

[0038] Example 7 Step 2.2.2 specifically describes the process as follows: To meet the high-precision control design requirements of the energy storage converter, this invention introduces a quasi-continuous integral terminal sliding mode controller. As a high-order sliding mode control algorithm, the quasi-continuous algorithm can generate continuous control signals and effectively suppress chattering, thus regulating the system state more efficiently and stably. In traditional sliding mode control, chattering may occur due to the presence of high-frequency switching quantities. However, integral terminal sliding mode uses integration to obtain the actual control quantity, which does not contain high-frequency switching quantities. Therefore, there is no chattering in the system, and it can guarantee finite-time convergence with a faster convergence speed.

[0039] Therefore, in order to further improve the stability and reliability of the system, a second-order quasi-continuous integral terminal sliding mode controller was designed by combining the quasi-continuous algorithm with the sliding mode concept of the integral terminal. The design process is as follows: First, let e be the system state error d, the deviation between the actual q-axis current and the reference current, as shown below: (14) According to equation (14), the current deviation e Design of the sliding surface of the integral terminal S as follows: (15) In equation (15), p 1, p 2, q 1, q Both 2 are positive odd numbers, and , ; , , , The adaptive control parameters for the sliding surface are expressed as follows: (16) In equation (16), , These are adaptive gain parameters, all real numbers greater than 0. To better adjust the sliding variable, an appropriate adaptive gain needs to be selected. and . and The larger the value, the faster the system convergence rate and the smaller the system steady-state error; but and If the value is too large, it will cause the system to be unstable and affect the system control performance.

[0040] The sliding surface of equation (15) S By differentiation, we get: (17) (18) For the feedforward error system shown in equation (10), substituting it into the above differential result yields: (19) (20) Since the system converges in a finite time, according to equivalent control theory, let Substituting the finite-time ESO variable observations of the d and q axes in equations (11) and (12) into equations (18) and (20), the equivalent control law of the system is as follows: (twenty one) As shown in equations (18) and (20), the sliding surface S The control quantity first appears in the second-order differential. s Therefore, the relative order of the system is 2. To generate a continuous control signal and effectively suppress system chattering, this invention introduces a second-order quasi-continuous algorithm as the sliding mode switching control law, with the following form: (twenty two) In equation (22), l The proportional gain for switching controllers, sgn( x ) is a symbolic function.

[0041] To further suppress high-frequency jitter and improve system stability, this invention employs a saturation function sat( S ) replaces the symbolic function sgn( S Therefore, the dq axis switching control law can be designed in the following form: (twenty three) In equation (23), l d , l d These are the proportional coefficients of the dq axis switching controller, and sat( S ) is a saturation function.

[0042] By combining equations (21) and (23), the overall control law of the system can be obtained. s d , s q Its expression is as follows: (twenty four) In summary, substituting equations (21) and (23) into equation (24), we can obtain the second-order quasi-continuous sliding mode controller of the power conversion system (PCS) as shown in equation (25): (25) This invention employs SPWM (Sine Pulse Width Modulation) to modulate the converter signal. Its basic working principle is as follows: Based on the controller designed in the control section, the obtained control signal is used as the modulation wave, and a high-frequency triangular wave is selected as the carrier wave. The on / off state of the switching transistors in the converter is determined by comparing the magnitudes of the modulation wave and the carrier wave. When the modulation wave is higher than the carrier wave, the switching transistor is turned on; when the modulation wave is lower than the carrier wave, the switching transistor is turned off. Thus, at the output of the converter, a pulse sequence with a width varying according to a sine law, i.e., an SPWM wave, is obtained. Finally, the control block diagram of the energy storage converter can be drawn as follows: Figure 4 As shown.

[0043] Example 9 To verify the effectiveness of the control method designed in this invention, a corresponding simulation circuit was constructed in MATLAB / Simulink simulation software, and it was compared and analyzed with the traditional PI control method. The simulation parameters are set as shown in Table 1.

[0044] Table 1 Circuit Parameters

[0045] The DC bus voltage stability of a two-stage energy storage converter reflects the efficiency of energy conversion, distribution, and management in a DC microgrid system, as well as its ability to cope with the volatility and instability of distributed energy sources. A stable DC bus voltage is crucial for ensuring efficient and reliable system operation. Since the PCS is directly connected to the DC output of the DC / DC converter, the PCS system can be analogized to the load portion of the DC / DC converter. When the power output of the PCS changes, the corresponding load on the DC / DC converter changes. During power fluctuations, the DC bus voltage also experiences voltage fluctuations at the corresponding moment.

[0046] To verify the advantages of the control method of the present invention compared with the traditional PI control method, corresponding active power mutations were set in the simulation to visually compare the performance differences of the two control methods when facing dynamic changes. The simulation results are shown in Figure 5(a) and Figure 5(b).

[0047] Figures 5(a) and 5(b) show the d-axis current under different control conditions. i d q-axis current caused by sudden changes in active power i q The transient waveform diagram shows two active power transitions at 0.2s (P jumps from 30kW to 80kW) and 0.28s (P jumps from 80kW to 30kW). When the converter outputs 30kW of active power and 0kW of reactive power, i d When the current is 64.3A, the output active power is 80kW, and the reactive power is 0, i d It is 171.5A.

[0048] Analysis of the data in Figures 5(a) and 5(b) shows that when the active power output of the PCS jumps from 30kW to 80kW and from 80kW to 30kW, respectively, both control methods can effectively track the system state changes and quickly restore steady-state operation. However, under the PI control method, sudden changes in active power have a greater impact on... i q The coupling effect is quite significant, with current fluctuations of 23A and 21A caused by the two transient processes, and transition times of 22ms and 34ms, respectively, which will affect the stable operation of the system. In contrast, the control method of this invention effectively suppresses the d-axis coupling phenomenon when the system's active power jumps, and the coupling effect caused by the two transient processes is significantly reduced. i q The fluctuation is only about 1A, and the transition time is very short, which is more conducive to the stable operation of the system.

[0049] To verify the effectiveness of the proposed control method, this invention employs a hardware-in-the-loop (HIL) experimental platform with a sampling frequency set to 20kHz to ensure data acquisition accuracy. First, a simulation circuit is built using Simulink to simulate the circuit control logic of the actual system. Then, the simulated circuit is loaded into an MT6020 and RCP1050 microcontroller via a host computer to simulate the behavior and response of the actual hardware. Finally, an adapter board is used to physically connect the real-time signals generated by the experimental platform to an oscilloscope, which monitors and records key data during the experiment in real time for subsequent analysis and verification.

[0050] In the image below, i d ,i q These represent the current values ​​along the d and q axes, respectively. u dc This is the DC bus voltage. Since direct measurement of the converter output power is difficult during the experiment, the output power can be indirectly reflected by observing the voltage and current along the d-axis and q-axis after coordinate transformation, according to formula (4). Since the d-axis voltage is set to a constant value, i d , i q The change in the output power of the converter can directly reflect the change in the output power of the converter. By observation, the real-time status of the output power of the converter can be indirectly understood, thereby verifying the effectiveness of the control method. According to the calculation of equation (5), when the active power output of the converter is 30kW and the reactive power is 0, i d It is 64.5A. i q The value is 0A. This calculation result can be used as a reference value for the experiment. The experiment is set to use the steady-state DC bus voltage. u dc It is 800V. i d The accuracy is 100A / division. i q The accuracy is 200A / division. u dc The voltage is 100V / division. The comparison results of the power jump experiment under the two control methods are shown in Figures 6(a), 6(b) and 7(a), 7(b).

[0051] According to the data in Figure 6(a), under the PI control method, when the output active power of the energy storage converter jumps from 30kW to 80kW, the DC bus voltage... u dc The overshoot reached 51V, and after a transient process of 37ms, it gradually recovered to a stable state. i d The increase from 65.3A to 172.4A resulted in an overshoot of 23A due to current fluctuations, with a transition time of 22ms. Simultaneously, due to coupling effects, q-axis current fluctuations occurred. i q The current fluctuation reached as high as 78A, and after a transient process of 18ms, it returned to steady-state operation.

[0052] As shown in Figure 6(b), under the control method of the present invention, the DC bus voltage caused by power surge is... u dc The fluctuation amplitude has been significantly reduced, with voltage fluctuation amplitude only around 9V and transient time reduced to around 10ms; power jump caused by id The overshoot is almost zero during power transients. i d The transition time is 11ms; the q-axis current is almost unaffected by the coupling effect caused by sudden changes in active power. During the transient transition, the current fluctuation is 5A, the transition time is 7ms, there is almost no chattering, and it quickly maintains steady-state operation in the new state, which is significantly lower than PI control. Therefore, during the process of active power jumps, the system transient performance under the control method of this invention is better, the response speed is faster, and it can effectively suppress reactive power coupling caused by active power switching.

[0053] As shown in Figure 7(a), under the PI control method, when the active power direction of the energy storage converter switches from 30kW to -30kW, the DC bus voltage... u dc The overshoot reached as high as 73V, and after a transient process of 39ms, it gradually recovered to a stable state; the current direction switched. i d The current jumps from 65.3A to -65.3A and recovers to steady state after 22ms, with an overshoot of 21A caused by the current fluctuation; simultaneously, due to the coupling effect, the q-axis current is affected. i q The current fluctuation reached as high as 81A, and after a transient process of 17ms, it recovered to steady-state operation.

[0054] As shown in Figure 7(b), under the control method of the present invention, during the process of switching of active power direction, u dc The overshoot is 13V, and the transition time is 12ms. i d The time required to restore steady state is 13ms; i q The current fluctuation is 7A, and the transition time is 12ms. Compared with PI control, the transition time required for the system to return to steady-state operation under the control method of this invention is significantly reduced, and the system operation is more stable. Therefore, taking all factors into consideration, the control method proposed in this invention is superior to the traditional PI control method in terms of transient response performance and operational stability, and also has a better effect on suppressing power coupling.

[0055] To verify the impact of the control method of this invention on the power quality of the converter output current, a Fourier analysis of the grid-connected current of the converter with an output active power of 30kW was performed in the Simulink simulation environment for two steady-state cycles.

[0056] As shown in Figures 8(a) and 8(b), the harmonic distortion rate under the PI control method is 3.75%, while the harmonic distortion rate is reduced to 0.79% when using the control method of the present invention. This difference indicates that the control method of the present invention can effectively reduce the harmonic content. Therefore, compared with traditional PI control, the control method of the present invention can provide higher quality power to the power grid.

[0057] Therefore, this invention designs a quasi-continuous adaptive integral terminal sliding mode active disturbance rejection control method based on an improved finite-time ESO (Electronic Stability of Variables). This control method uses ADRC (Advanced Dynamic Control Rejection Control) as the main controller, introducing finite-time ESO and quasi-continuous integral terminal sliding mode control into ADRC to improve the observer's disturbance estimation capability and the error convergence speed of feedback control, thereby achieving higher control accuracy and enhancing the system's robustness. This invention can also solve the problems of excessive DC bus voltage overshoot and coupling between the d and q axes, enabling the system to reach a stable state more quickly.

Claims

1. An improved sliding mode active disturbance rejection control method for a two-stage energy storage converter, characterized in that: Specifically, the steps include the following: Step 1: Establish a mathematical model of the two-stage energy storage converter in the dq rotating coordinate system; Step 2: Convert the mathematical model in Step 1 into a second-order active disturbance rejection paradigm, and then design the controller to achieve control of the two-stage energy storage converter.

2. The improved sliding mode active disturbance rejection control method for a two-stage energy storage converter according to claim 1, characterized in that: The specific process of step 1 is as follows: During the grid-connected operation of the converter, the operating states of the two switches on each bridge arm are complementary. Therefore, the switching function is defined as follows: （1） In equation (1), j = a , b , c Represents AC power grid a , b , c Three phases; According to Kirchhoff's laws, the voltage and current relationship equations of the converter in the abc three-phase stationary coordinate system can be obtained as follows: （2） In equation (2), u a , u b , u c This refers to the three-phase voltage on the AC side of the converter. i sa , i sb , i sc This refers to the three-phase current on the AC side. e a , e b , e c AC mains voltage; L For filtering inductors; C dc These are the parameters of the DC-side capacitor; R These are the equivalent resistance parameters of the line; i r This refers to the output current of the DC-side converter. u dc , i C DC bus voltage and current; s ka , s kb , s kc They are respectively a , b , c Three-phase bridge arm switch control signals; Performing the Park transformation on equation (2), we obtain the mathematical model of the converter in the dq coordinate system as follows: （3） In equation (3), ω The angular frequency of the grid voltage; e d 、i d 、s d and e q 、i q 、s q d-axis 、 The q-axis represents the grid voltage, current, and switching functions; u dc This is the DC bus voltage.

3. The improved sliding mode active disturbance rejection control method for a two-stage energy storage converter according to claim 2, characterized in that: The specific process of step 2 is as follows: Step 2.1, design of the power outer loop controller, which provides a current reference value for the current inner loop; Step 2.2, design of the inner current loop controller, the observer adopts a finite-time extended state observer, and the controller adopts a quasi-continuous integral terminal sliding mode controller.

4. The improved sliding mode active disturbance rejection control method for a two-stage energy storage converter according to claim 3, characterized in that: The specific process of step 2.1 is as follows: According to the instantaneous power theory, in order for the energy storage converter to be connected to the grid with near unity power factor, the reactive current is always equal to 0. Under the condition of three-phase grid voltage balance, the active power on the grid side of the two-stage energy storage converter is... P and reactive power Q The instantaneous value is expressed as: （4） From equation (4), we know that through i d and i q Control separately P and Q This allows for independent regulation of active and reactive power. By transforming equation (4), when... P ref , Q ref Given a constant, the reference values ​​of the d-axis and q-axis currents are directly obtained through calculation; a PI regulator is introduced to eliminate steady-state errors. P , Q and P , Q The deviation of the command value is adjusted by a PI controller, and then, based on instantaneous power theory, the system... P and Q The dynamic response is then used to obtain the active current reference value. i dref and reactive current reference value i qref As shown in the following formula: （5） In equation (5), P ref , Q ref Given reference values ​​for active and reactive power, i dref , i qref These are the reference values ​​for the inner loop of the d-axis and q-axis currents. k dp , k di and k qp , k qi These are the PI adjustment parameters for the d and q axes, respectively.

5. The improved sliding mode active disturbance rejection control method for a two-stage energy storage converter according to claim 4, characterized in that: The specific process of step 2.2 is as follows: Step 2.2.1: Based on the feedforward error system, design a finite-time extended state observer; Step 2.2.2, Design of a second-order quasi-continuous adaptive integral terminal sliding mode controller.

6. The improved sliding mode active disturbance rejection control method for a two-stage energy storage converter according to claim 5, characterized in that: The specific process of step 2.2.1 is as follows: First, by transforming equation (3), we obtain the state variable dq-axis current. i d , i q The first-order differential form is as follows: （6） The unmodeled portion of the dq-axis current inner loop coupling term, as well as the internal and external disturbances of the inner loop, are considered as the total inner loop disturbance, and are respectively used as... f d1 , f q1 This means that by differentiating equation (6), we can convert it into ADRC normal form: （7） In equation (7), , , b To control the gain of the quantity; definition b o To estimate the control gain, equation (7) is transformed into the following form: （8） In equation (8), , , b o This is an estimate of the control quantity gain; For the above second-order system, the feedforward error signal is taken. , i ref Given the current reference value, calculate the second derivative and substitute it into equation (8) to obtain: （9） Selecting state variables , , Then the system shown in equation (9) is transformed into: （10） In equation (10), , representing the lumped disturbance along the dq axis; For the feedforward error system shown in equation (10), the lumped disturbance of the dq axis is... f d , f q Expand into new state variables z d3 , z q3 The observers for the dq axes were designed separately.

7. The improved sliding mode active disturbance rejection control method for a two-stage energy storage converter according to claim 6, characterized in that: In step 2.2.1, the d-axis finite-time observer is designed as follows: （11） The q-axis observer is designed as follows: （12） In equations (11)-(12), e d1 , e q1 It represents the difference between the observed value and the actual value. z d1 , z q1 for x d1 , x q1 The estimated value; z d2 , z q2 for x d1 , x q1 An estimate of the derivative; z d3 , z q3 The total disturbance including coupling terms, internal disturbances, and unmodeled parts. f d , f q Estimated value; , , ， , , This represents the observer gain, and all values ​​are greater than 0. , , for the observer adjustment parameters, F( e i ) is a hyperbolic function.

8. The improved sliding mode active disturbance rejection control method for a two-stage energy storage converter according to claim 7, characterized in that: In step 2.2.1, due to the sign function sign( x The discontinuity exists, so the discontinuity switching term sign function in the traditional disturbance observer is replaced by a hyperbolic function F( e i ) instead, hyperbola F( e i The function is represented as: （13） In the formula, n The size determines F( e i The slope of ).

9. The improved sliding mode active disturbance rejection control method for a two-stage energy storage converter according to claim 8, characterized in that: The specific process of step 2.2.2 is as follows: Let e ​​be the system state error d, the deviation between the actual q-axis current and the reference current, as shown below: （14） According to equation (14), the current deviation e Design of the sliding surface of the integral terminal S as follows: （15） In equation (15), p 1, p 2, q 1, q Both 2 are positive odd numbers, and , ; , , , The adaptive control parameters for the sliding surface are expressed as follows: （16） In equation (16), , These are adaptive gain parameters, all of which are real numbers greater than 0; The sliding surface of equation (15) S Differentiating, we get: （17） （18） For the feedforward error system shown in equation (10), substituting into the above differential result, we get: （19） （20） Since the system converges in a finite time, according to equivalent control theory, let Substituting the finite-time ESO variable observations of the d and q axes in equations (11) and (12) into equations (18) and (20), the equivalent control law of the system is as follows: （21） As shown in equations (18) and (20), the sliding surface S The control quantity first appears in the second-order differential. s Therefore, the relative order of the system is 2. A second-order quasi-continuous algorithm is introduced as the sliding mode switching control law, which takes the following form: （22） In equation (22), l The proportional gain for switching controllers, sgn( x ) is a symbolic function; Using the saturation function sat( S ) replaces the symbolic function sgn( S The dq axis switching control law is designed in the following form: （23） In equation (23), l d , l d These are the proportional coefficients of the dq axis switching controller, and sat( S ) is a saturation function; Combining equations (21) and (23), we obtain the overall control law of the system. s d , s q The expression is as follows: （24） Substituting equations (21) and (23) into equation (24), we obtain the second-order quasi-continuous sliding mode controller of the energy storage converter system as shown in equation (25): （25）。

Images