Power station hybrid direct current system control method based on generalized over-torque algorithm and high-order extended state observation algorithm

Abstract

The invention discloses a power station hybrid DC system control method based on a generalized over-torque algorithm and a high-order extended state observation algorithm. The method comprises the following steps: S1, constructing a bidirectional DC-DC power conversion model; s2, performing generalized over-torque calculation on the model; s3, constructing a current inner loop control link by adopting a generalized over-torque algorithm; and S4, constructing a voltage outer loop control link by adopting a high-order state observation algorithm. Compared with other control strategies, the method for controlling the hybrid direct current system of the power station based on the generalized over-torque algorithm and the high-order extended state observation algorithm, provided by the invention, has the advantages that accurate adjustment of the output voltage of the converter is realized under the condition that unknown load impact and disturbance exist; and a smaller overshoot and a higher dynamic response speed are obtained.

Description

Technical Field

[0001] The invention relates to the technical field of power station control, and in particular to a power station hybrid DC system control method based on a generalized super torque algorithm and a high-order extended state observation algorithm. Background Art

[0002] The battery-based DC system is very important for the safe and stable operation of the power station. During the system operation, it is necessary to ensure the stable operation and reliable power supply of the power station DC system. In order to prevent the battery voltage from being greatly impacted and rapidly dropping when the DC system of the substation generates an impact load, resulting in battery damage, a hybrid DC system operation mode is often used with supercapacitors to quickly respond when the system generates an impact load to effectively buffer the battery discharge. The typical hybrid DC system structure of a power station is as follows: Figure 1 shown.

[0003] In order to improve energy efficiency, a bidirectional DC-DC power converter is often connected between the battery, supercapacitor and DC bus to achieve deep discharge. In order to effectively buffer the battery charge and discharge and improve the system's ability to cope with impact voltage and current, the bidirectional DC-DC converter is required to have stable output and rapid response to impact loads. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a control method for a hybrid DC system of a power station based on a generalized supertorque algorithm and a high-order extended state observation algorithm, and propose a voltage and current control model for a bidirectional DC-DC power converter of a hybrid DC system, which can effectively stabilize the impact voltage and current load, improve the rapid response capability, and ensure the stability and reliability of the DC system.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: a control method of a hybrid DC system of a power station based on a generalized super torque algorithm and a high-order extended state observation algorithm, comprising the following steps: S1, construct a bidirectional DC-DC power conversion model; S2, perform generalized supertorque calculation on the model; S3, using the generalized super torque algorithm to build the current inner loop control link; S4. Use high-order state observation algorithm to build the voltage outer loop control link.

[0006] Preferably, in step S1, the bidirectional DC-DC power conversion model includes an input module, a first interleaved parallel module, a second interleaved parallel module and a load module, and the first interleaved parallel module and the second interleaved parallel module are directly connected between the input module and the load module; The input module includes an input power supply v inIn the first staggered parallel module, the current forms a loop through the first inductor L1 and the first synchronous rectifier switch S1, and the first rectifier diode D1 is connected in series with the first filter capacitor C1 and then connected in parallel with the first synchronous rectifier switch S1; in the second staggered parallel module, the current forms a loop through the second inductor L2 and the second synchronous rectifier switch S2, and the second rectifier diode D2 is connected in series with the second filter capacitor C2 and then connected in parallel with S2; the load module is a load resistor R.

[0007] Preferably, the step S2 is specifically as follows: Consider the following first-order single-input single-output nonlinear system (1); in θ represents a sliding variable, f ( θ , t ) represents the unknown disturbance, which is given by limited, is an unknown positive constant, β To control the gain, u is the control input; the control function based on the generalized super torque algorithm of formula (1) is given by: (2); in and is the normal number to be designed, and Given by: (3); in , , symbolic function Given by the following formula (4); Based on formula (3) and The definition of can be derived as follows: (5); in express Relative to The partial derivative of; According to formula (1) and formula (5), the control function of the closed-loop system can be expressed as: (6).

[0008] Preferably, the step S3 collects the output voltage and the inductor current as state quantities, and generates the duty cycle of each switch of the power converter module through a pulse width modulation strategy (PWM); Since the bus voltage is dynamically affected by the inductor current, the sliding variable is defined as (7); in I Lref is the inner loop reference current value; the generalized ST algorithm is used to construct the current control function to ensure that the inductor current follows the reference in a finite time regardless of the change of the input voltage vin; (8); in is the controller gain, is the auxiliary control variable, function and Given by: (9); in and is a positive constant; in the sliding model, we have: (10); Based on equations (7)-(10), the equivalent control input can be expressed as: (11).

[0009] Preferably, the step S4 is specifically as follows: The voltage during the average cycle satisfies: (12); Substituting the current inner loop equivalent control input into the above formula, we can get (13); Put this equation at the equilibrium point ( ) can be linearized to obtain (14); Therefore, the formula can be rewritten as follows: (15); in is the virtual control input, coefficient , and Given by the following formula (16); It can be seen from equation (16) that the output voltage of the two-phase floating interleaved parallel converter is not only affected by the input voltage change, but also by the load change. In addition, the control gain is unknown as it depends on the varying input voltage, so an active interference rejection control framework should be adopted to compensate for the adverse effects of the above interferences and achieve accurate voltage regulation performance.

[0010] In order to solve the unknown control gain, equation (15) is rewritten as (17) in is the control gain The known positive constant part determined by the steady-state system parameters, is the total disturbance term, Characterize the effects of input voltage and load changes; Define the state vector , then formula (17) can be transformed into (18); in and is the observer parameter to be designed, and the control function is defined as (19); Can be designed as: (20); in is the parameter that needs to be designed; from equation (18), the transfer function of the observer can be deduced as: (twenty one); In order to better estimate the output voltage and total disturbance f, a specific parameter tuning method is adopted and the characteristic polynomial of the observer is taken as (twenty two); in, , is defined as the bandwidth of the observer. The three controller parameters that need to be tuned satisfy , , ; Mainly affects the ability to suppress interference, The larger it is, the faster the total interference can be estimated and suppressed, and its upper limit is limited by the sensor noise and sampling rate; Mainly affects the system response speed, The larger it is, the faster the system responds. Usually selected as the control gain The calculated value of the control strategy proposed by the present invention is as follows Figure 3 As shown, the equivalent transfer function based on the traditional first-order ESO voltage outer loop control is obtained as follows: (twenty three); The closed-loop transfer function of the system is (twenty four); In order to achieve error-free estimation of step disturbances, the above control algorithm is extended to high-order state observation to obtain stronger anti-interference ability. The generalized N-order state observation algorithm can be expressed as (25); in is the observed state vector, and the tracking form is , , ,in represents the (n-2)th order derivative of the variable f. The design of the final control function is the same as the control design of the state observation algorithm given by equations (16) and (17). From equation (25), it can be deduced that the transfer function of the Nth order state observation algorithm from total disturbance to disturbance estimation error is: (26); Observer parameters You can choose (27); Therefore, the steady-state error of the generalized N-order state observation algorithm for step disturbance is (28).

[0011] The present invention provides a control method for a hybrid DC system of a power station based on a generalized supertorque algorithm and a high-order extended state observation algorithm. Compared with other control strategies, the method can achieve precise regulation of the output voltage of the converter in the presence of unknown load shocks and disturbances, thereby obtaining a smaller overshoot and a faster dynamic response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 This is the structural diagram of a typical power station hybrid DC system; Figure 2 This is a power conversion model diagram of the present invention; Figure 3 It is the control strategy structure diagram of the present invention; Figure 4 This is the simulation result diagram when the input voltage is step disturbed; Figure 5 This is the simulation result diagram when the load current is step disturbed. DETAILED DESCRIPTION

[0013] like Figure 1-2 As shown, a control method for a hybrid DC system of a power station based on a generalized super torque algorithm and a high-order extended state observation algorithm comprises the following steps: S1, construct a bidirectional DC-DC power conversion model; S2, perform generalized supertorque calculation on the model; S3, using the generalized super torque algorithm to build the current inner loop control link; S4. Use high-order state observation algorithm to build the voltage outer loop control link.

[0014] Preferably, in step S1, the bidirectional DC-DC power conversion model is specifically as follows: In order to obtain a higher step-up ratio, lower switch voltage stress, lower current ripple and higher conversion efficiency, the present invention adopts a floating interleaved parallel converter as a choice for connecting batteries, supercapacitors and DC busbars and undertaking power conversion. The proposed power converter structure is as follows Figure 2 shown.

[0015] The bidirectional DC-DC power conversion model includes an input module, a first interleaved parallel module, a second interleaved parallel module and a load module, wherein the first interleaved parallel module and the second interleaved parallel module are directly connected to the input module and the load module; The input module includes an input power supply v in In the first staggered parallel module, the current forms a loop through the first inductor L1 and the first synchronous rectifier switch S1, and the first rectifier diode D1 is connected in series with the first filter capacitor C1 and then connected in parallel with the first synchronous rectifier switch S1; in the second staggered parallel module, the current forms a loop through the second inductor L2 and the second synchronous rectifier switch S2, and the second rectifier diode D2 is connected in series with the second filter capacitor C2 and then connected in parallel with S2; the load module is a load resistor R.

[0016] Preferably, the step S2 is specifically as follows: The generalized supertorque algorithm is an improvement on the traditional supertorque algorithm. By adding some simple correction terms, when the system trajectory moves away from the origin, the added linear terms are stronger than the nonlinear terms, so the robustness and convergence speed of the closed-loop system are improved accordingly. Consider the following first-order single-input single-output nonlinear system (1); in θ represents a sliding variable, f ( θ , t ) represents the unknown disturbance, which is given by limited, is an unknown positive constant, β To control the gain, u is the control input; the control function based on the generalized super torque algorithm of formula (1) is given by: (2); in and is the normal number to be designed, and Given by: (3); in , , symbolic function Given by the following formula (4); Based on formula (3) and The definition of can be derived as follows: (5); in express Relative to The partial derivative of; According to formula (1) and formula (5), the control function of the closed-loop system can be expressed as: (6).

[0017] Preferably, the step S3 collects the output voltage and the inductor current as state quantities, and generates the duty cycle of each switch of the power converter module through a pulse width modulation strategy (PWM); the goal of the inner loop controller is to adjust the inductor current to a reference value.

[0018] Since the bus voltage is dynamically affected by the inductor current, the sliding variable is defined as (7); in I Lref is the inner loop reference current value; the generalized ST algorithm is used to construct the current control function to ensure that the inductor current follows the reference in a finite time regardless of the change of the input voltage vin; (8); in is the controller gain, is the auxiliary control variable, function and Given by: (9); in and is a positive constant; in the sliding model, we have: (10); Based on equations (7)-(10), the equivalent control input can be expressed as: (11).

[0019] Preferably, the step S4 is specifically as follows: The voltage during the average cycle satisfies: (12); Substituting the current inner loop equivalent control input into the above formula, we can get (13); Put this equation at the equilibrium point ( ) can be linearized to obtain (14); Therefore, the formula can be rewritten as follows: (15); in is the virtual control input, coefficient , and Given by the following formula (16); It can be seen from equation (16) that the output voltage of the two-phase floating interleaved parallel converter is not only affected by the input voltage change, but also by the load change. In addition, the control gain is unknown as it depends on the varying input voltage, so an active interference rejection control framework should be adopted to compensate for the adverse effects of the above interferences and achieve accurate voltage regulation performance.

[0020] In order to solve the unknown control gain, equation (15) is rewritten as (17) in is the control gain The known positive constant part determined by the steady-state system parameters, is the total disturbance term, Characterize the effects of input voltage and load changes; Define the state vector , then formula (17) can be transformed into (18); in and is the observer parameter to be designed, and the control function is defined as (19); Can be designed as: (20); in is the parameter that needs to be designed; from equation (18), the transfer function of the observer can be deduced as: (twenty one); In order to better estimate the output voltage and total disturbance f, a specific parameter tuning method is adopted and the characteristic polynomial of the observer is taken as (twenty two); in, , is defined as the bandwidth of the observer. The three controller parameters that need to be tuned satisfy , , ; Mainly affects the ability to suppress interference, The larger it is, the faster the total interference can be estimated and suppressed, and its upper limit is limited by the sensor noise and sampling rate; Mainly affects the system response speed, The larger it is, the faster the system responds. Usually selected as the control gain The calculated value of the control strategy proposed by the present invention is as follows Figure 3 As shown, the equivalent transfer function based on the traditional first-order ESO voltage outer loop control is obtained as follows: (twenty three); The closed-loop transfer function of the system is (twenty four); In order to achieve error-free estimation of step disturbances, the above control algorithm is extended to high-order state observation to obtain stronger anti-interference ability. The generalized N-order state observation algorithm can be expressed as (25); in is the observed state vector, and the tracking form is , , ,in represents the (n-2)th order derivative of the variable f. The design of the final control function is the same as the control design of the state observation algorithm given by equations (16) and (17). From equation (25), it can be deduced that the transfer function of the Nth order state observation algorithm from total disturbance to disturbance estimation error is: (26); Observer parameters You can choose (27); Therefore, the steady-state error of the generalized N-order state observation algorithm for step disturbance is (28).

[0021] It can be seen from equation (28) that the proposed generalized N-order state observation algorithm can accurately estimate the step disturbance. As the order increases, the interference suppression capability is enhanced, but the anti-noise capability will be weakened accordingly. Therefore, the second-order state observation algorithm control is the best choice given by weighing the anti-interference capability, anti-noise capability and stability margin.

[0022] In order to test the robustness of the controller designed in the present invention under input voltage disturbance, a step disturbance is applied to the converter at 0.45s. The output voltages of the three controllers under input change are shown in Figure 4 As shown in the figure, it can be seen that after a step disturbance occurs in the input voltage, the PI control has a 2.5V deviation for the steady-state output voltage, and the adjustment time is the longest, which is 0.17s. The traditional state observation algorithm based on the super-torque algorithm has a 1.4V deviation for the steady-state output voltage, and the adjustment time is 0.12s; while the control strategy based on the generalized super-torque control algorithm and the high-order state observation algorithm proposed in the present invention has only a 0.9V deviation for the steady-state output voltage, and the adjustment time is only 0.09s, showing the best control performance.

[0023] In order to test the transient performance of the three control methods under load impact disturbance, a step disturbance is applied to the load current of the converter to observe the converter's anti-disturbance ability to the load current step disturbance.

[0024] Figure 5 The simulation results of the output voltage under the load current step disturbance are given. It can be seen that the output voltage is maintained at 100V before 0.5s. At 0.5s, a positive step disturbance of 0.5A is applied to the converter. It can be seen that after experiencing a voltage drop process, the converters under the three control methods can return to a stable state. Since PI control lacks a disturbance compensation mechanism in the external voltage loop, the stabilization time of PI control is the longest compared with the other two control methods. Compared with the traditional state observation control method based on the super torque algorithm, the control strategy based on the generalized super torque control algorithm and the high-order state observation algorithm proposed in the present invention has the smallest overshoot and stabilization time.

[0025] The above embodiments are only preferred technical solutions of the present invention and should not be regarded as limiting the present invention. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A control method for a hybrid DC system of a power station based on a generalized super torque algorithm and a high-order extended state observation algorithm, characterized in that: The following steps are involved: S1, construct a bidirectional DC-DC power conversion model; S2, perform generalized supertorque calculation on the model; S3, using the generalized super torque algorithm to build the current inner loop control link; S4. Use high-order state observation algorithm to build the voltage outer loop control link.

2. According to claim 1, a power station hybrid DC system control method based on generalized super torque algorithm and high-order extended state observation algorithm is characterized in that: In step S1, the bidirectional DC-DC power conversion model includes an input module, a first interleaved parallel module, a second interleaved parallel module and a load module, and the first interleaved parallel module and the second interleaved parallel module are directly connected between the input module and the load module; The input module includes an input power supply v in In the first staggered parallel module, the current forms a loop through the first inductor L1 and the first synchronous rectifier switch S1, and the first rectifier diode D1 is connected in series with the first filter capacitor C1 and then connected in parallel with the first synchronous rectifier switch S1; In the second staggered parallel module, the current forms a loop through the second inductor L2 and the second synchronous rectifier switch S2, the second rectifier diode D2 is connected in series with the second filter capacitor C2 and then connected in parallel with S2; the load module is a load resistor R.

3. According to claim 2, a power station hybrid DC system control method based on generalized super torque algorithm and high-order extended state observation algorithm is characterized in that: The step S2 is specifically as follows: Consider the following first-order single-input single-output nonlinear system (1)； in θ represents a sliding variable, f ( θ , t ) represents the unknown disturbance, which is given by limited, is an unknown positive constant, β To control the gain, u is the control input; the control function based on the generalized super torque algorithm of formula (1) is given by: (2)； in and is the normal number to be designed, and Given by: (3)； in , , symbolic function Given by the following formula (4)； Based on formula (3) and The definition of can be derived as follows: (5)； in express Relative to The partial derivative of; According to formula (1) and formula (5), the control function of the closed-loop system can be expressed as: (6)。 4. According to claim 3, a power station hybrid DC system control method based on generalized super torque algorithm and high-order extended state observation algorithm is characterized in that: The step S3 collects the output voltage and the inductor current as state quantities, and generates the duty cycle of each switch of the power converter module through a pulse width modulation strategy (PWM); Since the bus voltage is dynamically affected by the inductor current, the sliding variable is defined as (7)； in I Lref is the inner loop reference current value; the generalized ST algorithm is used to construct the current control function to ensure that the inductor current follows the reference in a finite time regardless of the change of the input voltage vin; (8)； in is the controller gain, is the auxiliary control variable, function and Given by: (9)； in and is a positive constant; in the sliding model, we have: (10)； Based on equations (7)-(10), the equivalent control input can be expressed as: (11)。 5. The control method of a hybrid DC system of a power station based on a generalized super torque algorithm and a high-order extended state observation algorithm according to claim 4, characterized in that: The step S4 is specifically as follows: The voltage during the average cycle satisfies: (12); Substituting the current inner loop equivalent control input into the above formula, we can get (13)； Put this equation at the equilibrium point ( ) can be linearized to obtain (14)； Therefore, the formula can be rewritten as follows: (15)； in is the virtual control input, coefficient , and Given by the following formula (16)； In order to solve the unknown control gain, equation (15) is rewritten as (17) in is the control gain The known positive constant part determined by the steady-state system parameters, is the total disturbance term, Characterize the effects of input voltage and load changes; Define the state vector , then formula (17) can be transformed into (18)； in and is the observer parameter to be designed, and the control function is defined as (19)； Can be designed as: (20)； in is the parameter that needs to be designed; from equation (18), the transfer function of the observer can be deduced as: (21)； In order to better estimate the output voltage and total disturbance f, a specific parameter tuning method is adopted and the characteristic polynomial of the observer is taken as (22)； in, , Defined as the bandwidth of the observer; the three controller parameters that need to be tuned satisfy , , ; Mainly affects the ability to suppress interference, The larger it is, the faster the total interference can be estimated and suppressed, and its upper limit is limited by the sensor noise and sampling rate; Mainly affects the system response speed, The larger it is, the faster the system responds; Usually selected as the control gain The calculated value of; the equivalent transfer function based on the traditional first-order ESO voltage outer loop control is obtained as follows: (23)； The closed-loop transfer function of the system is (24)； In order to achieve error-free estimation of step disturbances, the above control algorithm is extended to high-order state observation to obtain stronger anti-interference ability. The generalized N-order state observation algorithm can be expressed as (25)； in is the observed state vector, and the tracking form is , , ,in represents the (n-2)th order derivative of the variable f. The design of the final control function is the same as the control design of the state observation algorithm given by equations (16) and (17). From equation (25), it can be deduced that the transfer function of the Nth order state observation algorithm from total disturbance to disturbance estimation error is: (26)； Observer parameters You can choose (27)； Therefore, the steady-state error of the generalized N-order state observation algorithm for step disturbance is (28)。

Images