An adaptive control system for an inverter

Abstract

The application provides an adaptive control system of an inverter, comprising a model-free adaptive controller, an inverter and a dynamic data correction filter; the model-free adaptive controller obtains a control signal for controlling on-off of a switch tube in the inverter and a prediction signal required by the dynamic data correction filter at the next moment according to a sine wave signal of an external signal source at the current moment and a filtering signal of the dynamic data correction filter at the current moment; when receiving a measurement signal formed by an output signal of the inverter at the current moment and external disturbance and measurement noise interference, the dynamic data correction filter filters the measurement signal in combination with the prediction signal at the current moment, forms a filtering signal feedback without measurement noise and external disturbance, and reduces deviation between actual and theoretical values of the inverter. By implementing the application, the influence of measurement noise on performance can be effectively reduced, the signal fed back to the inverter is more accurate, and the accuracy of the inverter output is improved.

Description

Technical Field

[0001] This invention relates to the field of inverter control technology, and more particularly to an adaptive control system for inverters. Background Technology

[0002] An inverter is an energy conversion device that converts direct current (DC) to alternating current (AC), playing a crucial role in energy conversion. Inverters allow DC sources to power AC loads and transfer energy in different forms between various devices to meet the needs of different locations in factories and homes. Initially, inverters only needed to stabilize frequency and voltage. However, with the development of power systems and the increasing diversity of electrical loads, improved inverter performance is necessary to meet these demands. To enhance the inverter's response speed, control accuracy, tracking performance, anti-interference performance, and stability, under given circuit topology conditions, control technology plays a decisive role in the inverter's power output, leading to the emergence of various control technologies. The development of control theory has also greatly promoted the development of inverter control technology, making the research of more effective control methods an important aspect of inverter technology.

[0003] Currently, both domestic and international research on inverter control technology has yielded substantial theoretical achievements and successful engineering applications. Inverter control encompasses various methods, such as state feedback control, dual-loop control, deadbeat control, hysteresis control, and repetitive control. Each method has its own characteristics, collectively forming a system that provides precise control schemes for power electronic converters. Among these, state feedback control and dual-loop control belong to linear feedback control strategies and are the most fundamental control methods in inverter control schemes. They can achieve rapid control of variables and effectively improve the steady-state dynamic response of the system. However, they require very accurate modeling of the inverter's electrical model, and even small deviations can lead to inaccurate output. Compared to state feedback control and dual-loop control, deadbeat control introduces the concept of periodic control, making it a discrete control scheme for specific control systems. Compared to basic inverter control strategies such as dual-loop control, it has the advantages of fast dynamic response and no overshoot. However, due to the highly nonlinear characteristics of the inverter, it cannot achieve zero-error regulation, thus failing to meet the requirement of high-precision control performance under rectified loads. Hysteresis control is an inverter control technology aimed at direct current control. It can give the inverter system good robustness and dynamic response. However, the variable switching frequency of hysteresis control leads to poor controllability. At the same time, increasing the switching frequency will significantly shorten the service life of the switching devices. In addition, under heavy load and narrow hysteresis conditions, problems such as excessively high switching device frequency and excessive electrical stress will bring greater difficulty to control. Repetitive control constructs an internal model of a periodic signal and embeds it into the control loop. It can achieve steady-state errorless tracking or disturbance elimination of the periodic signal (i.e., the superposition signal of DC component, sinusoidal fundamental component and its harmonic components). It is very suitable for eliminating harmonics in power systems. However, the existence of a delay element leads to a slow dynamic response, which is difficult to meet the requirements of inverters.

[0004] Furthermore, the various control methods for inverters mentioned above all neglect the impact of measurement noise. In fact, measurement noise is difficult to avoid in inverter control systems; it can be generated due to sensor damage, environmental changes, and system malfunctions, adversely affecting the inverter's output waveform. Therefore, the more reasonable the control method, the lower the measurement noise in the system, and the more ideal the inverter's performance. Conversely, inadequate control methods will lead to the inverter's output failing to meet expectations.

[0005] Therefore, there is an urgent need for a new inverter control system that can effectively reduce the impact of measurement noise on performance and make the signals fed back to the inverter more accurate, thereby improving the accuracy of the inverter output. Summary of the Invention

[0006] The technical problem to be solved by the embodiments of the present invention is to provide an adaptive control system for an inverter, which can effectively reduce the impact of measurement noise on performance, make the signal fed back to the inverter more accurate, and thus improve the accuracy of the inverter output.

[0007] To address the aforementioned technical problems, embodiments of the present invention provide an adaptive control system for an inverter, comprising a model-free adaptive controller, an inverter, and a dynamic data correction filter; wherein...

[0008] The first input terminal of the model-free adaptive controller is connected to an external signal source, the second input terminal is connected to the output terminal of the dynamic data correction filter, the first output terminal is connected to the input terminal of the inverter, and the second output terminal is connected to the first input terminal of the dynamic data correction filter. The model-free adaptive controller is used to obtain a control signal for controlling the switching transistors in the inverter at the current moment, and to obtain a prediction signal required for the dynamic data correction filter at the next moment, based on the sinusoidal signal provided by the external signal source at the current moment and the filtered signal output by the dynamic data correction filter at the current moment.

[0009] The second input terminal of the dynamic data correction filter is connected to the output terminal of the inverter. The dynamic data correction filter is used to filter the received measurement signal when it receives the output signal of the inverter at the current moment, which is formed by external disturbances and measurement noise interference, in combination with the prediction signal provided by the model-free adaptive controller at the current moment, so as to form a filtered signal without measurement noise and external disturbances and feed it back to the model-free adaptive controller, so as to reduce the deviation between the actual output and the theoretical output of the inverter.

[0010] The filtered signal of the dynamic data correction filter is calculated based on the probability density distribution function of the received measurement signal and prediction signal.

[0011] Wherein, the probability density function of the filtered signal of the dynamic data correction filter is: and Among them, y r (k) represents the filtered signal output by the dynamic data correction filter at time k; y m (k) is the measurement signal received by the dynamic data correction filter at time k; y(k) is the predicted signal received by the dynamic data correction filter at time k; y(k) is the output signal of the inverter at time k.

[0012] The measurement signal y received by the dynamic data correction filter at time k mThe external disturbance ζ(k) in (k) follows a Gaussian distribution with the following probability density function:

[0013] The measurement signal y received by the dynamic data correction filter at time k m The probability density fractional function of the measurement noise ε(k) in (k) is:

[0014] Wherein, if the measurement noise ε(k) follows a Gaussian distribution, the filtered signal y of the dynamic data correction filter at time k is... r (k) is The value of y(k) when it reaches its maximum is... in,

[0015] The variance function of the external disturbance ζ(k) is ζ(k)~(0,ρ) 2 The variance function of the measured noise ε(k) is ε(k)~(0,δ). 2 ).

[0016] Wherein, if the measurement noise ε(k) does not follow a Gaussian distribution, the filtered signal y of the dynamic data correction filter at time k... r (k) is The mathematical expectation, i.e. in,

[0017] The measurement noise ε(k) is ε(t)=ωε1(t)+(1-ω)ε2(t); ε1(k)~(0,δ1 2 ); ε2(k)~(b,δ2) 2 ); η = ws1 + (1 - w)s2;

[0018] Wherein, the model-free adaptive controller provides the dynamic data correction filter with the prediction signal y(k+1) required for the next time step as y(k+1) = y(k) + φ(k)Δu(k); where,

[0019] The control rate of the model-free adaptive controller is J(u(k))=|r(k+1)-y(k+1)| 2 +λ|u(k)-u(k-1)| 2 You can get ρ is the step size factor and 0 < ρ ≤ 1, λ is the weight factor and λ > 0;

[0020] The feature parameter estimation criterion of the model-free adaptive controller is as follows: It can be obtained μ is the weighting factor, and η is the step size sequence.

[0021] Implementing the embodiments of the present invention has the following beneficial effects:

[0022] The adaptive control system of the inverter of this invention includes external disturbances and measurement noise that are more consistent with the actual system, which has higher research value. Furthermore, the use of a model-free adaptive controller enables the inverter output to track a standard sine wave within a finite time, ensuring output performance. At the same time, it avoids the parameter adjustment problems of traditional controllers and effectively reduces the impact of measurement noise through dynamic data correction filtering technology, making the signal fed back to the model-free adaptive controller more accurate and improving the accuracy of the inverter output. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, obtaining other drawings based on these drawings without creative effort still falls within the scope of the present invention.

[0024] Figure 1 This is a schematic diagram of the structure of an adaptive control system for an inverter provided in an embodiment of the present invention;

[0025] Figure 2 A waveform comparison diagram of the inverter output signal based on measurement signal feedback and the inverter output signal based on dynamic data correction signal feedback in an adaptive control system for an inverter provided in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0027] like Figure 1 As shown in the figure, an adaptive control system for an inverter is proposed in an embodiment of the present invention, including a model-free adaptive controller 1, an inverter 2, and a dynamic data correction filter 3; wherein,

[0028] The first input terminal of the model-free adaptive controller 1 is connected to an external signal source, the second input terminal is connected to the output terminal of the dynamic data correction filter 3, the first output terminal is connected to the input terminal of the inverter 2, and the second output terminal is connected to the first input terminal of the dynamic data correction filter 3. The model-free adaptive controller 1 is used to adjust the input based on the sinusoidal signal r(k) provided by the external signal source at the current moment and the filtered signal y output by the dynamic data correction filter 3 at the current moment. r(k), to obtain the control signal u(k) used to control the switching on and off of the switching transistors in inverter 2 at the current moment, and to obtain the prediction signal required for the dynamic data correction filter 3 at the next moment.

[0029] The second input terminal of the dynamic data correction filter 3 is connected to the output terminal of the inverter 2; the dynamic data correction filter 3 is used to receive the measurement signal y(k) formed by the external disturbance ζ(k) and measurement noise ε(k) interference from the output signal y(k) of the inverter 2 at the current moment. m At time (k), the prediction signal provided by the model-free adaptive controller 1 at the current time is combined. For the received measurement signal y m (k) Filtering is performed to form a filtered signal y free from measurement noise and external disturbances. r (k) is fed back to the model-free adaptive controller 1 to reduce the deviation between the actual output and the theoretical output of the inverter 2. It can be understood that this dynamic data correction filter 3 corrects the original inverter control system for external disturbances ζ(k) and measurement noise ε(k), making the signal fed back to the model-free adaptive controller 1 more accurate than the directly measured signal, thereby reducing the impact of external disturbances ζ(k) and measurement noise ε(k) and improving the performance of the inverter control system.

[0030] In this embodiment of the invention, the filtered signal y of the dynamic data correction filter 3 r (k) is based on the measurement signal y it receives. m (k) and prediction signal The probability density distribution function is calculated from it.

[0031] Assume the dynamic data correction filter receives the measurement signal y at time 3k. m The external disturbance ζ(k) in (k) follows a Gaussian distribution with the following probability density function: And assuming the dynamic data correction filter receives the measurement signal y at time 3k. m The probability density fractional function of the measurement noise ε(k) in (k) is:

[0032] According to Bayes' theorem, the probability density function of the filtered signal of the dynamic data correction filter is: and Among them, y r (k) represents the filtered signal output at time k; y m (k) represents the measurement signal at time k; y(k) is the predicted signal at time k; y(k) is the output signal at time k.

[0033] At this point, if the measured noise ε(k) follows a Gaussian distribution, the filtered signal y of the dynamic data correction filter at time 3k... r (k) is The value of y(k) when it reaches its maximum is... The variance function of the external disturbance ζ(k) is ζ(k)~(0,ρ) 2 The variance function of the measured noise ε(k) is ε(k)~(0,δ). 2 ).

[0034] Alternatively, if the measured noise ε(k) does not follow a Gaussian distribution, the filtered signal y at time 3k of the dynamic data correction filter... r (k) is The mathematical expectation, i.e. in,

[0035] The measurement noise ε(k) is ε(t)=ωε1(t)+(1-ω)ε2(t); ε1(k)~(0,δ1 2 ); ε2(k)~(b,δ2) 2 ); η = ws1 + (1 - w)s2;

[0036] Through theoretical derivation, the filtered signal y in the above two cases... r The variance of the error between (k) and the system output must be less than the variance of the measurement noise ε(k) and the variance of the external disturbance ζ(k), thus the filtered signal y can be obtained. r (k) compared to the measured signal y m (k) is more accurate, and feeding it back to the model-free adaptive controller 1 can effectively suppress the influence of measurement noise ε(k) and improve the control performance of inverter 2.

[0037] In this embodiment of the invention, according to the definition of model-free adaptive control, the prediction signal y(k+1) required by the model-free adaptive controller 1 for the dynamic data correction filter 3 at the next moment can be obtained as y(k+1) = y(k) + φ(k)Δu(k); where,

[0038] At this point, the control rate of the model-free adaptive controller is J(u(k))=|r(k+1)-y(k+1)| 2 +λ|u(k)-u(k-1)| 2 You can get ρ is the step size factor and 0 < ρ ≤ 1, λ is the weight factor and λ > 0;

[0039] Then, the characteristic parameter estimation criterion for the model-free adaptive controller is designed as follows: It can be obtained μ is the weighting factor, and η is the step size sequence.

[0040] As can be seen from the above, by designing the control law and characteristic parameter estimation criteria, the system output can accurately track the setpoint. Furthermore, the model-free adaptive controller 1 effectively avoids the parameter adjustment problems of traditional controllers, and when internal parameters change, it can also make the system output track the setpoint again within a finite time.

[0041] like Figure 2 As shown, a waveform comparison between the inverter output signal based on measurement signal feedback and the inverter output signal based on dynamic data correction filter signal feedback reveals that the model-free adaptive controller 1 used in this embodiment enables the inverter 2 to effectively track the set standard sine wave. When external disturbances ζ(k) and measurement noise ε(k) exist in the system, the measurement signal y is directly used. m (k) The filtered waveform of the feedback signal is severely disturbed, but after adding the dynamic data correction filter 3, the output waveform of the inverter 2 is significantly improved, thus improving the control accuracy of the inverter 2.

[0042] Implementing the embodiments of the present invention has the following beneficial effects:

[0043] The adaptive control system of the inverter of this invention includes external disturbances and measurement noise that are more consistent with the actual system, which has higher research value. Furthermore, the use of a model-free adaptive controller enables the inverter output to track a standard sine wave within a finite time, ensuring output performance. At the same time, it avoids the parameter adjustment problems of traditional controllers and effectively reduces the impact of measurement noise through dynamic data correction filtering technology, making the signal fed back to the model-free adaptive controller more accurate and improving the accuracy of the inverter output.

[0044] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as ROM / RAM, disk, optical disk, etc.

[0045] The above description is merely a preferred embodiment of the present invention. It should be noted that this description should not be construed as limiting the scope of the present invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. An adaptive control system for an inverter, characterized in that, This includes a model-free adaptive controller, an inverter, and a dynamic data correction filter; among which, The first input terminal of the model-free adaptive controller is connected to an external signal source, the second input terminal is connected to the output terminal of the dynamic data correction filter, the first output terminal is connected to the input terminal of the inverter, and the second output terminal is connected to the first input terminal of the dynamic data correction filter. The model-free adaptive controller is used to obtain a control signal for controlling the switching transistors in the inverter at the current moment, and to obtain a prediction signal required for the dynamic data correction filter at the next moment, based on the sinusoidal signal provided by the external signal source at the current moment and the filtered signal output by the dynamic data correction filter at the current moment. The second input terminal of the dynamic data correction filter is connected to the output terminal of the inverter. The dynamic data correction filter is used to filter the received measurement signal when it receives the output signal of the inverter at the current moment, which is formed by external disturbances and measurement noise interference, in combination with the prediction signal provided by the model-free adaptive controller at the current moment, so as to form a filtered signal without measurement noise and external disturbances and feed it back to the model-free adaptive controller, so as to reduce the deviation between the actual output and the theoretical output of the inverter.

2. The adaptive control system for the inverter as described in claim 1, characterized in that, The filtered signal of the dynamic data correction filter is calculated based on the probability density distribution function of the received measurement signal and prediction signal.

3. The adaptive control system for the inverter as described in claim 2, characterized in that, The probability density function of the filtered signal of the dynamic data correction filter is: ,and ;in, For the dynamic data correction filter The filtered signal output at any given time; For the dynamic data correction filter The measurement signal received at any time; For the dynamic data correction filter The predicted signal received at any given time; For the inverter Output signal at time; ; The dynamic data correction filter Measurement signals received at any time External disturbances It follows a Gaussian distribution, and its probability density function is: ; The dynamic data correction filter Measurement signals received at any time Measurement noise The probability density fractional function is .

4. The adaptive control system for the inverter as described in claim 3, characterized in that, If the measurement noise When the dynamic data correction filter follows a Gaussian distribution, Filtered signal at time for When the maximum is reached ,Right now ;in, External disturbances The variance function is ; Measurement noise The variance function is .

5. The adaptive control system for the inverter as described in claim 4, characterized in that, If the measurement noise When the dynamic data correction filter does not follow a Gaussian distribution, Filtered signal at time for The mathematical expectation, i.e. ;in, Noise measurement for ; ; ; ; ; .

6. The adaptive control system for the inverter as described in claim 5, characterized in that, The model-free adaptive controller provides the dynamic data correction filter with the prediction signal required for the next time step. for ;in, The control rate of the model-free adaptive controller is You can get ; is the step size factor and , Adjusting the weighting factor for the control quantity and ; The feature parameter estimation criterion of the model-free adaptive controller is as follows: You can get ; Adjusting the weighting factor for pseudo-partial derivatives It is a step-size sequence.

Images