Power hardware-in-loop interface compensation method based on power amplifier parameter identification
By estimating the power amplifier filter parameters using the gradient descent method and designing a compensator, the amplitude and phase offset problems caused by the filter in the power amplifier are solved, thereby improving the power transmission accuracy and stability of the system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CENT SOUTH UNIV
- Filing Date
- 2026-01-26
- Publication Date
- 2026-05-05
AI Technical Summary
The LC low-pass filter of existing power amplifiers causes amplitude and phase differences between the actual power transfer characteristics and the ideal power transfer characteristics, and is affected by device aging and parameter offset, which affects the stability and reliability of the system.
The filter parameters in the power amplifier are estimated by gradient descent, and a corresponding compensator is designed to compensate in the interface method, thereby reducing the system's sensitivity to changes in filter parameters and improving system reliability.
It achieves effective compensation for changes in filter parameters, improves the power transmission accuracy and stability of the system, reduces the sensitivity to parameter changes, and enhances the reliability of the system.
Smart Images

Figure CN121984449A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of inverter control, specifically a power hardware-in-the-loop interface compensation method based on power amplifier parameter identification. Background Technology
[0002] With the rapid development of computer technology, especially parallel computing and various real-time simulation tools, hardware-in-the-loop simulation is increasingly widely used in power electronics and power systems. Power hardware-in-the-loop, as a special type of hardware-in-the-loop testing, provides power to the hardware under test, allowing for repeatable, safe, and economical studies of new equipment or control methods under realistic conditions within a simulated power grid system. Due to the order-of-magnitude difference between the simulation system and the hardware system, appropriate power amplification and conversion devices are needed as the interface between the simulation and the hardware. Ideal transformer-type power amplifiers are widely used due to their simplicity and good accuracy. However, the LC low-pass filter in a real power amplifier introduces amplitude and phase differences compared to an ideal amplifier and is affected by device aging and parameter shifts, thus deteriorating the system's power transfer characteristics. Therefore, designing an interface method capable of handling variations in filter parameters within the power amplifier is crucial for improving system stability and reliability. Summary of the Invention
[0003] To address the aforementioned technical problems, this invention proposes a power hardware-in-the-loop interface compensation method based on power amplifier parameter identification. This method estimates the filter parameters in the power amplifier using the gradient descent method and designs a corresponding compensator in the interface method to compensate for the amplitude and phase shifts caused by the filter, thereby reducing the system's sensitivity to changes in filter parameters and improving system reliability.
[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:
[0005] The power hardware-in-the-loop interface compensation method based on power amplifier parameter identification has the following specific steps:
[0006] S1, through data filtering, obtains the excitation for the parameter identification method;
[0007] S2, Based on the gradient descent method, identify the parameter information of passive devices in the power amplifier;
[0008] S3. Based on the identified parameters, design a power amplifier compensation interface method on the software side.
[0009] In one embodiment of the present invention, the data filtering in step S1 is specifically performed as follows:
[0010] According to Kirchhoff's laws, the average model of a power amplifier can be expressed as:
[0011] (1)
[0012] in and These represent the amplifier's filter inductance and filter capacitor, respectively. Indicates filter inductance The equivalent resistance. , as well as These represent the output voltage, inductor current, and output current, respectively. Indicates the DC side voltage. This indicates a control input.
[0013] definition For output variables, For state variables, For the coefficient matrix, (1) can be written in the following form:
[0014] (2)
[0015] in , , .
[0016] Since differentiation cannot be achieved in practical control, the signal is processed in the frequency domain as follows:
[0017] (3)
[0018] Where s is the Laplace operator and λ is the decay coefficient;
[0019] The equivalent form of (2) can be obtained as:
[0020] (4)
[0021] in , , .
[0022] In one embodiment of the present invention, step S2 involves identifying the parameter information of passive devices in the power amplifier based on the gradient descent method. The specific process is as follows:
[0023] Error is estimated by defining parameters:
[0024] (5)
[0025] in This represents the reference value of the parameter to be identified. This represents the estimated value of the parameter to be identified.
[0026] Therefore, the parameter estimation dynamic equation can be designed as follows:
[0027] (6)
[0028] in This represents the derivative of the estimated value of the parameter to be identified. This indicates the convergence constant for identification.
[0029] According to (6), combined with the Lyapunov function We can obtain:
[0030] (7)
[0031] By selecting appropriate parameters Ensure the convergence speed of the tracking error system.
[0032] As one embodiment of the present invention, the power amplifier compensation interface method in step S3 is specifically as follows:
[0033] The compensator designed based on the identified parameters has the following form.
[0034] (8)
[0035] in Let be the impedance of the device under test. To ensure the physical feasibility of the compensator, a second-order low-pass filter needs to be connected in series.
[0036] (9)
[0037] in The resonant frequency, is the damping ratio.
[0038] Therefore, the total amplitude compensator is obtained as follows:
[0039] (10)
[0040] At this point, an additional phase delay is introduced.
[0041] (11)
[0042] Among them, w r ζ represents the resonant frequency of the filter. op Given the open-loop damping ratio, k is the adjustable gain, and w is the input signal frequency, phase compensation is also required. In the interface method, phase compensation can be achieved using a harmonic successive compensation scheme based on a frequency sampling filter.
[0043] Compared with the prior art, the present invention has the following beneficial effects:
[0044] The power hardware-in-the-loop interface compensation method based on power amplifier parameter identification described in this invention estimates the filter parameters in the power amplifier using the gradient descent method. It eliminates the need for prior knowledge of the filter parameters, requiring only voltage and current sampling to compensate the filter, thus reducing the system's sensitivity to parameter changes and improving system reliability. Furthermore, this invention compensates for the phase lag introduced by the compensator, improving the system's stability margin. By compensating for amplitude and phase shifts introduced by the power amplifier in the interface method, power transmission accuracy is improved. Attached Figure Description
[0045] Figure 1 Schematic diagram of power amplifier interface compensation according to an embodiment of the present invention;
[0046] Figure 2 The control structure diagram of the hardware-in-the-loop method for accurate power transfer based on interface compensation in this invention is shown in the embodiment.
[0047] Figure 3 The parameter estimation results of the accurate power transmission hardware-in-the-loop method based on interface compensation in this invention under unknown / offset filter parameters;
[0048] Figure 4 The simulation results of the hardware-in-the-loop method for accurate power transfer based on interface compensation in this invention are shown in the case of unknown / offset filter parameters. Detailed Implementation
[0049] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:
[0050] This invention proposes a power hardware-in-the-loop interface compensation method based on power amplifier parameter identification. By identifying the parameters of the power amplifier system, the parameter information of passive components in the power amplifier is estimated, improving the system's robustness to parameter changes. Since the presence of filters in the power amplifier typically causes a deviation between the actual power delivery and the ideal power delivery characteristics, this invention compensates for the amplitude and phase shifts caused by the filters using parameters identified in the interface method, thereby improving power delivery accuracy and the reliability of the power hardware-in-the-loop. Furthermore, the compensated system is closer to the ideal system and is more suitable for power hardware-in-the-loop stability analysis and verification, especially showing significant advantages in cases of power amplifier parameter aging and shift.
[0051] like Figure 1 The diagram shows a power amplifier interface compensation schematic. It consists of four parts: a digital real-time simulation section, a three-phase full-bridge inverter section, an LC filter section, and the hardware under test section.
[0052] like Figure 2The diagram shows the control structure of the hardware-in-the-loop method for accurate power transfer based on interface compensation. The gradient descent method estimates the filter parameters, and the interface method uses these estimated parameters to design a compensation method. The compensated signal is then phase-calibrated, and finally, the calibrated signal is used for modulation.
[0053] like Figure 3 The image shows the parameter estimation results of the hardware-in-the-loop method for accurate power transfer based on interface compensation under conditions of unknown / offset filter parameters. The results demonstrate that when filter parameters are unknown / offset, accurate estimation of the filter parameters can be achieved simply by sampling the voltage and current, and the convergence coefficient can be adjusted. To accelerate dynamic response.
[0054] like Figure 4 The simulation results of the hardware-in-the-loop method for accurate power transfer based on interface compensation under unknown / offset filter parameters are shown. The simulation results demonstrate that, when the filter parameters are unknown / offset, by designing a compensator using the estimated parameters, the amplitude and phase offset of the output reference voltage can be compensated, approximating the voltage waveform under ideal conditions.
Claims
1. A power hardware-in-the-loop interface compensation method based on power amplifier parameter identification, characterized in that, The specific steps are as follows: S1, through data filtering, obtains the excitation for the parameter identification method; S2, Based on the gradient descent method, identify the parameter information of passive devices in the power amplifier; S3. Based on the identified parameters, design a power amplifier compensation interface method on the software side.
2. The power hardware-in-the-loop interface compensation method based on power amplifier parameter identification according to claim 1, characterized in that: The data filtering process in step S1 is as follows: According to Kirchhoff's laws, the average model of a power amplifier can be expressed as: (1); in and These represent the amplifier's filter inductance and filter capacitor, respectively. Indicates filter inductance The equivalent resistance; , as well as These represent the output voltage, inductor current, and output current, respectively. Indicates the DC side voltage. Indicates control input; definition For output variables, For state variables, For the coefficient matrix, (1) can be written in the following form: (2); in , , ; Since differentiation cannot be achieved in practical control, the signal is processed in the frequency domain as follows: (3); Where s is the Laplace operator and λ is the decay coefficient; The equivalent form of (2) can be obtained as: (4); in , , .
3. The power hardware-in-the-loop interface compensation method based on power amplifier parameter identification according to claim 1, characterized in that: In step S2, the parameter information of passive devices in the power amplifier is identified based on the gradient descent method. The specific process is as follows: Error is estimated by defining parameters: (5); in This represents the reference value of the parameter to be identified. This represents the estimated value of the parameter to be identified; Therefore, the parameter estimation dynamic equation can be designed as follows: (6); in This represents the derivative of the estimated value of the parameter to be identified. Indicates the convergence constant; According to (6), combined with the Lyapunov function ,get: (7); By selecting appropriate parameters Ensure the convergence speed of the tracking error system.
4. The power hardware-in-the-loop interface compensation method based on power amplifier parameter identification according to claim 1, characterized in that: The power amplifier compensation interface method in step S3 is as follows: The compensator designed based on the identified parameters has the following form. (8); in The impedance of the device under test is given; to ensure the physical realizability of the compensator, a second-order low-pass filter needs to be connected in series. (9); in The resonant frequency, The damping ratio; Therefore, the total amplitude compensator is obtained as follows: (10); At this point, an additional phase delay is introduced. (11); Among them, w r ζ represents the resonant frequency of the filter. op Given the open-loop damping ratio, k is the adjustable gain, and w is the input signal frequency, phase compensation is also required. In the interface method, phase compensation can be achieved using a harmonic successive compensation scheme based on a frequency sampling filter.