CN114362180B - Compensation method, system, storage medium and computing device for LCL type grid-connected inverter - Google Patents

Abstract

The invention discloses a compensation method, a compensation system, a storage medium and a computing device for an LCL type grid-connected inverter, wherein the method is based on a topological structure of an LCL type grid-connected inverter system under a weak power grid environment, and a small signal model of the LCL type grid-connected inverter considering PLL influence is established; secondly, a phase-locked loop structure containing a low-pass filter is proposed from the improvement of the phase-locked loop structure; and finally, considering the influence of power grid background harmonic waves, providing a self-adaptive small signal compensation method based on deep reinforcement learning, and compensating disturbance caused by a phase-locked loop. The invention can widen the power grid impedance adaptation range of the system and effectively improve the robustness of the grid-connected inverter under the condition of weak power grid.

Description

A LCL type grid-connected inverter compensation method, system, storage medium and computing device

Technical Field

The invention relates to a compensation method, system, storage medium and computing device for an LCL type grid-connected inverter, belonging to the technical field of grid-connected control.

Background Art

Grid-connected inverters usually serve as the interface between distributed power generation systems and the public power grid, injecting the electricity generated by renewable energy into the grid in the form of alternating current. As the locations of distributed power sources become more and more widespread, the power grid gradually exhibits weak grid characteristics. The short circuit ratio (SCR) is usually used to describe the strength of the power grid. The larger the grid impedance, the smaller the SCR value, and the stronger the weak grid characteristics of the system. At this time, the line voltage drop of the grid-connected current flowing through the grid impedance cannot be ignored, and the voltage at the point of common coupling (PCC) is no longer approximately the grid voltage.

In the case of weak grid, the impact of the phase-locked loop (PLL) on system stability will gradually increase with the continuous increase of grid impedance. For the PLL output phase, the grid impedance and grid-connected current are interference quantities. When the grid impedance increases and the grid-connected current increases, it will affect the dynamic and static performance of its output phase and reduce the robustness of the PLL. The deterioration of the phase-locked loop performance leads to the distortion of the reference current, which further causes the distortion of the grid current and the oscillation of the grid power. In severe cases, the entire inverter cannot operate stably and is cut off from the grid by the protection device. The existing grid-connected inverter control method considering the influence of PLL mainly improves the adaptability of the grid-connected system to weak grids from the perspective of reshaping the phase-locked loop structure or the inverter output impedance. However, in weak grids, due to the volatility of distributed new energy power generation units such as wind power and photovoltaic power generation units and the load, there are background harmonics in the grid, and the grid inductance varies in a wide range. It is difficult to ensure the quality of the grid-connected current and the stability of the grid-connected system only from the perspective of improving the phase-locked loop structure or reshaping the inverter output impedance.

Summary of the invention

The purpose of the present invention is to propose a LCL type grid-connected inverter compensation method, system, storage medium and computing device, introduce a low-pass filter in the traditional phase-locked loop structure, and adopt the Deep Deterministic Policy Gradient (DDPG) algorithm in deep reinforcement learning to realize adaptive small signal compensation, so as to suppress the influence of grid background harmonics on grid-connected current to the greatest extent.

To achieve the above object, the technical solution adopted by the present invention is as follows:

The present invention provides a compensation method for an LCL type grid-connected inverter, comprising: adding a low-pass filter before a phase-locked loop q-axis PI controller of the LCL type grid-connected inverter system to filter out a high-frequency disturbance component of a common coupling point PCC voltage;

And, the low-frequency harmonic disturbance component of the power grid is compensated in the following ways:

和

The phase-locked loop frequency, phase-locked loop frequency error and integral of frequency error of the LCL grid-connected inverter system are used as the input of the DDPG agent to obtain the optimal compensation coefficient.

and

和

经锁相环后得到并网电流扰动补偿量和调制电压扰动补偿量；Optimal compensation coefficient

and

After the phase-locked loop, the grid-connected current disturbance compensation amount and the modulation voltage disturbance compensation amount are obtained;

The grid-connected current disturbance compensation amount is compensated into the q-axis grid-connected current, and the modulated voltage disturbance compensation amount is compensated into the grid-connected voltage.

Furthermore, the low-pass filter transfer function is:

Where G _LPF is the low-pass filter transfer function, ω _c is the low-pass filter cutoff frequency, and s is the Laplace operator.

Furthermore, the phase-locked loop transfer function is expressed as:

为锁相环传递函数，H_PLL(s)＝k_ppll+k_ipll/s是锁相环PI控制器，k_ppll、k_ipll分别为锁相环PI控制器的比例项系数和积分项系数，

是d轴公共耦合点电压稳态值。in,

is the phase-locked loop transfer function, _HPLL (s) = _kppll + _kipll /s is the phase-locked loop PI controller, _kppll and _kipll are the proportional term coefficient and the integral term coefficient of the phase-locked loop PI controller respectively,

is the steady-state value of the voltage at the common coupling point on the d-axis.

The present invention also provides a LCL type grid-connected inverter compensation system, comprising a DDPG intelligent body, a compensation module and a low-pass filter;

The low-pass filter is located before the phase-locked loop q-axis PI controller and is used to filter out the high-frequency disturbance component of the common coupling point PCC voltage;

和

The DDPG agent is used to generate the optimal compensation coefficient using the phase-locked loop frequency, phase-locked loop frequency error and the integral of the frequency error of the LCL type grid-connected inverter system as input signals

and

和

经锁相环后得到并网电流扰动补偿量和调制电压扰动补偿量；以及将并网电流扰动补偿量补偿入q轴并网电流处，将调制电压扰动补偿量补偿入并网电压处。The compensation module is used to set the optimal compensation coefficient

and

After the phase-locked loop, the grid-connected current disturbance compensation amount and the modulation voltage disturbance compensation amount are obtained; and the grid-connected current disturbance compensation amount is compensated into the q-axis grid-connected current, and the modulation voltage disturbance compensation amount is compensated into the grid-connected voltage.

Furthermore,

The low-pass filter transfer function is:

Where G _LPF is the low-pass filter transfer function, ω _c is the low-pass filter cutoff frequency, and s is the Laplace operator.

Furthermore, the cut-off frequency _ωc of the low-pass filter is 2200 Hz.

Furthermore, in the compensation module, the phase-locked loop transfer function is expressed as:

为锁相环传递函数，H_PLL(s)＝k_ppll+k_ipll/s是锁相环PI控制器，k_ppll、k_ipll分别为锁相环PI控制器的比例项系数和积分项系数，

是d轴公共耦合点电压稳态值。in,

is the phase-locked loop transfer function, _HPLL (s) = _kppll + _kipll /s is the phase-locked loop PI controller, _kppll and _kipll are the proportional term coefficient and the integral term coefficient of the phase-locked loop PI controller respectively,

is the steady-state value of the voltage at the common coupling point on the d-axis.

Furthermore, the DDPG agent uses the following formula as the reward function:

Where r is the reward function, f is the phase-locked loop frequency, and _f0 is the power frequency.

A third aspect of the present invention provides a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any of the methods described above.

A fourth aspect of the present invention provides a computing device, comprising:

One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the aforementioned methods.

The beneficial effects achieved by the present invention are:

The present invention provides a LCL type grid-connected inverter compensation method and system, which adds a low-pass filter before the traditional phase-locked loop q-axis PI controller, adopts the deep determination policy gradient algorithm in deep reinforcement learning, and adaptively compensates the modulation voltage and grid-connected current, thereby suppressing the influence of the background harmonics of the power grid on the grid-connected current to the greatest extent. The present invention can effectively solve the problems of grid-connected current distortion and grid-connected system instability caused by the phase-locked loop under weak power grid conditions, enhances the adaptability of the grid-connected inverter to wide range impedance changes, and even in the presence of background harmonics of the power grid, the phase-locked loop frequency can still maintain a very small steady-state error, effectively improving the reliability and robustness of the grid-connected inverter system.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is a topological diagram of an LCL-type grid-connected inverter system;

Figure 2 is a block diagram of the q-axis control of the LCL type grid-connected inverter;

FIG3 is a simplified block diagram of the q-axis control of the LCL type grid-connected inverter;

FIG4 is a topological structure of the improved PLL of the present invention;

FIG5 is a small signal compensation control block diagram of an LCL type grid-connected inverter provided by the present invention;

FIG6 is a block diagram of an adaptive small signal compensation control of an LCL type grid-connected inverter provided by the present invention;

Figure 7 is a diagram of the internal network architecture of DDPG;

FIG8 is a frequency simulation diagram of a conventional phase-locked loop before compensation is added in an embodiment of the present invention;

9 is a simulation diagram of the grid-connected three-phase voltage and current using a conventional phase-locked loop before adding compensation in an embodiment of the present invention;

FIG10 is a frequency simulation diagram of an improved phase-locked loop with adaptive small signal compensation added in an embodiment of the present invention;

FIG. 11 is a simulation diagram of the grid-connected three-phase voltage and current after an improved phase-locked loop is used to add adaptive small signal compensation in an embodiment of the present invention.

DETAILED DESCRIPTION

The present invention is further described below. The following examples are only used to more clearly illustrate the technical solution of the present invention, and are not intended to limit the protection scope of the present invention.

The present invention provides a LCL type grid-connected inverter compensation method, comprising:

By adding a low-pass filter before the phase-locked loop q-axis PI controller of the LCL type grid-connected inverter system, the high-frequency disturbance component of the PCC voltage at the common coupling point is filtered out;

And, the low-frequency harmonic disturbance component of the power grid is compensated in the following ways:

和

The phase-locked loop frequency, phase-locked loop frequency error and integral of frequency error of the LCL grid-connected inverter system are used as the input of the DDPG agent to obtain the optimal compensation coefficient.

and

和

经锁相环后得到并网电流扰动补偿量和调制电压扰动补偿量；Optimal compensation coefficient

and

After the phase-locked loop, the grid-connected current disturbance compensation amount and the modulation voltage disturbance compensation amount are obtained;

The grid-connected current disturbance compensation amount is compensated into the q-axis grid-connected current, and the modulated voltage disturbance compensation amount is compensated into the grid-connected voltage.

As a preferred implementation, this embodiment provides an adaptive small signal compensation method for an LCL type grid-connected inverter, which is specifically as follows:

作差后输入PI控制器，对PI控制器输出信号经解耦后进行派克反变换，为抑制由LCL引入的高频谐振，本发明采用电容电流反馈有源阻尼法将派克反变换后的信号与电容电流反馈信号相减得到SVPWM的输入信号U_rα和U_rβ，通过SVPWM模块输出PWM信号来控制IGBT的开通和关断，从而实现并网。由于弱电网条件下，公共耦合点(PCC)电压易受干扰从而影响锁相环的动态特性，因此电网电压相角θ与锁相环输出相角θ_PLL之间存在偏差Δθ。Δθ的存在将dq坐标轴分为系统dq坐标轴和控制器dq坐标轴，现将θ_PLL和θ的参考坐标轴分别定义为c域和s域。变量x在这两个域间的关系可表示为：Firstly, the topological structure of the LCL grid-connected inverter system under weak power grid environment is described, and the small signal model of the LCL grid-connected inverter considering the influence of PLL is established. As shown in Figure 1, the LCL grid-connected inverter system is based on the hybrid coordinate system structure, where L ₁ , L ₂ , C _f are the inverter side inductance, grid side inductance and filter capacitor respectively, U _dc is the DC side voltage, u _a , u _b , u _c are the bridge arm voltages of the three phases a, b, c respectively, U _ga , U _gb , U _gc are the three phase grid voltages a, b, c respectively, ω ₀ is the grid angular frequency, the voltage U _{pcc_abc} of the three phases a, b, c common coupling point is θ after PLL output, K _c is the capacitor current feedback coefficient, the current loop adopts PI controller, and its expression is G _i = (k _p + k _i / s), k _p , k _i are the proportional term coefficient and integral term coefficient of the PI controller respectively, I ₁ , I ₂ , I _c are the bridge arm current, grid current and capacitor current respectively. The control principle is: the three-phase grid-connected current I _{2_abc} obtained after Parker transformation is I _2d and I _2q, which are the actual values of the d-axis and q-axis grid-connected currents respectively. I _2d and I _2q are the reference values of the d-axis and q-axis grid-connected currents.

After the difference is made, it is input into the PI controller. The output signal of the PI controller is decoupled and then subjected to Parker's inverse transformation. In order to suppress the high-frequency resonance introduced by LCL, the present invention adopts the capacitor current feedback active damping method to subtract the signal after Parker's inverse transformation from the capacitor current feedback signal to obtain the SVPWM input signals _Urα and _Urβ . The SVPWM module outputs the PWM signal to control the opening and closing of the IGBT, thereby achieving grid connection. Due to the weak grid condition, the voltage at the common coupling point (PCC) is susceptible to interference, thereby affecting the dynamic characteristics of the phase-locked loop. Therefore, there is a deviation Δθ between the grid voltage phase angle θ and the phase-locked loop output phase angle θ _PLL. The existence of Δθ divides the dq coordinate axis into the system dq coordinate axis and the controller dq coordinate axis. The reference coordinate axes of θ _PLL and θ are now defined as the c domain and s domain, respectively. The relationship between the variable x in these two domains can be expressed as:

Where: the subscripts “d” and “q” represent the d-axis and q-axis respectively; the subscript “0” represents the steady-state value; Δx represents the small signal disturbance variable.

The q-axis control block diagram of the grid-connected inverter current loop is shown in Figure 2, where U _pccq , _Urq and I _cq are the q-axis PCC voltage, q-axis modulation voltage and q-axis capacitor current respectively, G _de (s) is the control delay, and the expression is as follows:

Where _Ts is the sampling period.

Through the equivalent transformation of the control block diagram, Figure 2 can be simplified to Figure 3, where:

The system loop gain _TA can be expressed as:

T _A (s)＝G _x1 (s)G _x2 (s)G _de (s)G _i (s) (5)

The small signal model of the phase-locked loop is:

H_PLL(s)＝k_ppll+k_ipll/s是锁相环PI控制器，k_ppll、k_ipll分别为锁相环PI控制器的比例项系数和积分项系数，

是d轴PCC电压稳态值，

为PCC电压小信号扰动量。将式(6)代入式(1)，得：Where G _PLL (s) is the phase-locked loop transfer function, which is expressed as

H _PLL (s) = k _ppll + k _ipll /s is a phase-locked loop PI controller, k _ppll and k _ipll are the proportional term coefficient and the integral term coefficient of the phase-locked loop PI controller respectively.

is the steady-state value of the d-axis PCC voltage,

is the PCC voltage small signal disturbance. Substituting equation (6) into equation (1), we get:

为0，且PLL的动态特性仅会影响电网电流I_2q和调制电压U_rq，即，Assuming the power factor is unity, the system q-axis steady-state value is

is 0, and the dynamic characteristics of PLL will only affect the grid current I _2q and the modulation voltage U _rq , that is,

分别为控制器q轴调制电压小信号扰动量和系统q轴调制电压小信号扰动量，

为系统d轴调制电压稳态值，

为控制器q轴并网电流小信号扰动量，

为系统d轴并网电流稳态值。in,

are the controller q-axis modulation voltage small signal disturbance and the system q-axis modulation voltage small signal disturbance, respectively.

is the steady-state value of the system d-axis modulation voltage,

is the controller q-axis grid-connected current small signal disturbance,

is the steady-state value of the system d-axis grid-connected current.

Under weak power grid, the q-axis output impedance of the grid-connected inverter can be expressed as:

in,

The q-axis grid-connected current small signal disturbance can be expressed as:

It can be seen from formula (13) that the grid current disturbance caused by the phase-locked loop can be effectively reduced by compensating the q-axis grid-connected current and the modulation voltage.

对锁相环输出相角的影响，根据前面建立的LCL型并网逆变器的小信号模型，从改造锁相环结构出发，提出含有低通滤波器的锁相环结构，如图4所示。在传统锁相环的基础上，在锁相环q轴PI控制器前加入低通滤波器，其传递函数可表示为，In order to suppress the high-frequency disturbance component of PCC voltage and reduce the

In order to determine the influence of the phase-locked loop output phase angle, based on the small signal model of the LCL grid-connected inverter established above, a phase-locked loop structure with a low-pass filter is proposed, as shown in Figure 4. On the basis of the traditional phase-locked loop, a low-pass filter is added before the phase-locked loop q-axis PI controller, and its transfer function can be expressed as:

Wherein ω _c is the cut-off frequency, the smaller ω _c is, the better the anti-interference ability is, but a smaller ω _c will lead to a slower response speed. After comprehensive consideration, the present invention selects ω _c =2200.

From the cut-off frequency, it can be seen that the low-pass filter can effectively filter out the 7th harmonics in the PCC voltage, but the power grid itself has abundant low-order background harmonics, which will also affect the grid-connected control system through the phase-locked loop. Only by improving the phase-locked loop structure, it is impossible to ensure that the grid-connected inverter outputs good grid-connected current quality. After adding the low-pass filter, the small signal disturbance at the modulation voltage and grid-connected current of the grid-connected inverter can be expressed as,

等于所设定的d轴电流参考值

调制电压稳态值

可近似等于电网电压

如图5所示，分别在并网电流反馈环和调制电压处加入相应的补偿即可消除锁相环对并网电流的影响。但在考虑存在电网背景谐波时，上述近似等效不再成立。From equation (13), it can be seen that by compensating the q-axis grid-connected current and the modulation voltage, the grid-connected current disturbance caused by the phase-locked loop can be effectively reduced. Assuming that the grid-connected inverter operates at unity power factor and the grid voltage does not fluctuate, the steady-state value of the grid current is

Equal to the set d-axis current reference value

Modulation voltage steady state value

Can be approximately equal to the grid voltage

As shown in Figure 5, adding corresponding compensation to the grid-connected current feedback loop and the modulation voltage can eliminate the influence of the phase-locked loop on the grid-connected current. However, when considering the existence of grid background harmonics, the above approximate equivalence is no longer valid.

In order to improve the adaptability of the grid-connected inverter to wide range impedance changes in the presence of grid background harmonics, adaptive compensation is added on the basis of improving the phase-locked loop structure. The Deep Deterministic Policy Gradient (DDPG) algorithm in deep reinforcement learning is used to adaptively compensate the modulation voltage and grid current, thereby improving the grid current distortion under weak grid conditions.

The deep deterministic policy gradient strategy based on deep reinforcement learning is divided into two modules: environment and agent. The environment is the physical model of the grid-connected inverter, and the agent includes the strategy (deep neural network) and reinforcement learning algorithm. During the training process, the reinforcement learning algorithm continuously updates the strategy based on the observation value (Observation) and reward (Reward). The action command (Action) generated by the strategy acts on the grid-connected inverter, thus forming a closed-loop control between the grid-connected inverter and the agent.

和

来补偿锁相环引入的扰动分量，将补偿系数

经锁相环后得到并网电流扰动补偿量(式(17))，补偿入q轴并网电流I_2q处，再与q轴并网电流参考值

作差后输入q轴PI控制器；q轴PI控制器的输出经解耦后得到并网电压；The DDPG adaptive compensation control is shown in Figure 6. The modulation voltage disturbance and the grid-connected current disturbance (Equations 15, 16, and 17) are adaptively compensated. That is, the phase-locked loop frequency error and the integral of the frequency error are used as the input of the DDPG to obtain an optimal compensation coefficient through the algorithm.

and

To compensate for the disturbance component introduced by the phase-locked loop, the compensation coefficient

After the phase-locked loop, the grid-connected current disturbance compensation value (Equation (17)) is obtained, which is compensated into the q-axis grid-connected current I _2q and then compared with the q-axis grid-connected current reference value

The difference is input into the q-axis PI controller; the output of the q-axis PI controller is decoupled to obtain the grid-connected voltage;

经锁相环后得到调制电压扰动补偿量(式(16))，补偿入并网电压处后再进行派克反变换；The compensation factor

After the phase-locked loop, the modulation voltage disturbance compensation amount is obtained (Equation (16)), which is compensated into the grid-connected voltage and then subjected to the Park inverse transformation;

Finally, the signal after Parker inverse transformation is subtracted from the capacitor current feedback signal to obtain the input signal of SVPWM, and the PWM signal is output through the SVPWM module to control the opening and closing of the IGBT, thereby achieving grid connection.

将锁相环频率f、误差信号f₀-f以及误差信号积分∫(f₀-f)作为DDPG的观测信号，记为s＝{f,f₀-f,∫(f₀-f)}，其中，f₀为工频；The phase-locked loop transfer function is:

The phase-locked loop frequency f, the error signal f ₀ -f and the error signal integral ∫(f ₀ -f) are used as the observation signals of DDPG, denoted as s = {f, f ₀ -f, ∫(f ₀ -f)}, where f ₀ is the power frequency;

The value obtained by the controlled quantity f through the reward function module is used as the reward of DDPG, denoted as r; the reward function is shown in formula (18).

和

记为a，基于策略梯度的方法，DDPG的计算策略函数如式(19)所示，The output action of DDPG is the compensation coefficient

and

Denoted as a, based on the policy gradient method, the calculation strategy function of DDPG is shown in formula (19),

J(π _θ )＝∫∫ρ(s)π _θ (s,a)r(s,a)dads＝E _s-ρ [r(s,a)] (19)

Further improve the gradient of J(π _θ ), that is,

为梯度；Q(s,a)为观测是s、动作是a的条件下，策略π的价值函数。经过迭代，能够给出最佳的补偿量。Where π _θ is the strategy with parameter θ; J is the strategy function; ρ is the probability distribution of the strategy; E _s-ρ is the expectation that s follows the ρ distribution;

is the gradient; Q(s,a) is the value function of strategy π under the condition that the observation is s and the action is a. After iteration, the optimal compensation amount can be given.

In order to obtain the best compensation amount within the stable range, the present invention adopts the DDPG algorithm, which is an important branch of deep reinforcement learning. The internal structure of DDPG is shown in Figure 7. The algorithm adopts an actuator-evaluator structure. The experience replay area is used to store the interaction information between the agent and the environment. During offline training, in order to avoid coupling between experiences, random batch sampling is performed from the experience replay area each time to train the network parameters. The evaluator's goal is to obtain the best estimate and calculate the value function from the reward obtained to correctly evaluate the current action of the actuator.

Example

Matlab/Simulink is used for simulation. The specific training parameters of the DDPG algorithm are shown in Table 1. Table 2 shows the parameter settings of the LCL type grid-connected inverter.

Table 1 DDPG algorithm training parameters

Table 2 Parameters of LCL grid-connected inverter

Under the same weak power grid environment, the method of the present invention is compared with the traditional method.

When the grid impedance is 15mH, after adding 3% of the 7th harmonic and 2% of the 5th harmonic through the three-phase programmable voltage source, the simulation results of the conventional phase-locked loop without compensation are shown in Figures 8 and 9. The PLL error is about 15Hz, and the grid-connected current is significantly distorted by the grid voltage distortion. At this time, the current distortion rate THD is 8.07%. The simulation results of the method proposed by the present invention are shown in Figures 10 and 11. The PLL error is about 0.7Hz, the grid-connected current waveform is improved, and the current distortion rate THD is 2.07%.

It can be seen from the simulation results that the adaptive small signal compensation method of the LCL type grid-connected inverter based on deep reinforcement learning proposed in the present invention can effectively solve the problems of grid-connected current distortion and grid-connected system instability caused by the phase-locked loop under weak grid conditions, and enhance the adaptability of the grid-connected inverter to wide range of impedance changes. Even in the presence of grid background harmonics, the phase-locked loop frequency can still maintain a very small steady-state error. The proposed control method can effectively improve the reliability and robustness of the grid-connected inverter control system.

Another embodiment of the present invention provides an adaptive small signal compensation system for an LCL type grid-connected inverter, comprising a DDPG agent and a compensation module;

和

The DDPG agent is used to generate the optimal compensation coefficient using the phase-locked loop frequency, phase-locked loop frequency error and the integral of the frequency error of the LCL type grid-connected inverter system as input signals

and

和

经锁相环后得到并网电流扰动补偿量和调制电压扰动补偿量；以及将并网电流扰动补偿量补偿入q轴并网电流处，将调制电压扰动补偿量补偿入并网电压处。The compensation module is used to set the optimal compensation coefficient

and

After the phase-locked loop, the grid-connected current disturbance compensation amount and the modulation voltage disturbance compensation amount are obtained; and the grid-connected current disturbance compensation amount is compensated into the q-axis grid-connected current, and the modulation voltage disturbance compensation amount is compensated into the grid-connected voltage.

和

In this embodiment, the DDPG agent uses a deep deterministic policy gradient algorithm to calculate the optimal compensation coefficient

and

In this embodiment, the DDPG agent uses the following formula as the reward function:

Where r is the reward function, f is the phase-locked loop frequency, and _f0 is the power frequency.

Preferably, it also includes a low-pass filter, wherein the low-pass filter is located before the phase-locked loop q-axis PI controller.

The low-pass filter transfer function is:

Where G _LBF is the low-pass filter transfer function, ω _c is the low-pass filter cutoff frequency, and s is the Laplace operator.

Preferably, the cut-off frequency ω _c of the low-pass filter is 2200 Hz.

Another embodiment of the present invention provides a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any of the aforementioned methods.

The present invention further provides a computing device, comprising:

One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the aforementioned methods.

Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

The present application is described with reference to the flowchart and/or block diagram of the method, device (system) and computer program product according to the embodiment of the present application. It should be understood that each process and/or box in the flowchart and/or block diagram, and the combination of the process and/or box in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one process or multiple processes in the flowchart and/or one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims (10)

1. A LCL type grid-connected inverter compensation method, characterized by comprising: By adding a low-pass filter before the phase-locked loop q-axis PI controller of the LCL type grid-connected inverter system, the high-frequency disturbance component of the PCC voltage at the common coupling point is filtered out; And, the low-frequency harmonic disturbance component of the power grid is compensated in the following ways: The phase-locked loop frequency, phase-locked loop frequency error and integral of frequency error of the LCL grid-connected inverter system are used as the input of the DDPG agent to obtain the optimal compensation coefficient.

and

Optimal compensation coefficient

and

After the phase-locked loop, the grid-connected current disturbance compensation amount and the modulation voltage disturbance compensation amount are obtained; The grid-connected current disturbance compensation amount is compensated into the q-axis grid-connected current, and the modulated voltage disturbance compensation amount is compensated into the grid-connected voltage. 2. A LCL type grid-connected inverter compensation method according to claim 1, characterized in that: The low-pass filter transfer function is:

Where G _LPF is the low-pass filter transfer function, ω _c is the low-pass filter cutoff frequency, and s is the Laplace operator. 3. The LCL type grid-connected inverter compensation method according to claim 2, characterized in that the phase-locked loop transfer function is expressed as:

in,

is the phase-locked loop transfer function, _HPLL (s) = _kppll + _kipll /s is the phase-locked loop PI controller, _kppll and _kipll are the proportional term coefficient and the integral term coefficient of the phase-locked loop PI controller respectively,

is the steady-state value of the voltage at the common coupling point on the d-axis. 4. An LCL type grid-connected inverter compensation system, characterized in that it includes a DDPG agent, a compensation module and a low-pass filter, The low-pass filter is located before the phase-locked loop q-axis PI controller and is used to filter out the high-frequency disturbance component of the common coupling point PCC voltage; The DDPG agent is used to generate the optimal compensation coefficient using the phase-locked loop frequency, phase-locked loop frequency error and the integral of the frequency error of the LCL type grid-connected inverter system as input signals

and

The compensation module is used to set the optimal compensation coefficient

and

After the phase-locked loop, the grid-connected current disturbance compensation amount and the modulation voltage disturbance compensation amount are obtained; and the grid-connected current disturbance compensation amount is compensated into the q-axis grid-connected current, and the modulation voltage disturbance compensation amount is compensated into the grid-connected voltage. 5. The LCL type grid-connected inverter compensation system according to claim 4, characterized in that: The low-pass filter transfer function is:

Where G _LPF is the low-pass filter transfer function, ω _c is the low-pass filter cutoff frequency, and s is the Laplace operator. 6. The LCL type grid-connected inverter compensation system according to claim 5, characterized in that in the compensation module, the phase-locked loop transfer function is expressed as:

in,

is the phase-locked loop transfer function, _HPLL (s) = _kppll + _kipll /s is the phase-locked loop PI controller, _kppll and _kipll are the proportional term coefficient and the integral term coefficient of the phase-locked loop PI controller respectively,

is the steady-state value of the voltage at the common coupling point on the d-axis. 7. The LCL grid-connected inverter compensation system according to claim 4, characterized in that the DDPG agent adopts the following formula as the reward function:

Where r is the reward function, f is the phase-locked loop frequency, and _f0 is the power frequency. 8. The LCL grid-connected inverter compensation system according to claim 5, characterized in that the cut-off frequency _ωc of the low-pass filter is 2200 Hz. 9. A computer-readable storage medium storing one or more programs, characterized in that: the one or more programs include instructions, and when the instructions are executed by a computing device, the computing device performs any one of the methods according to claims 1 to 3. 10. A computing device, comprising: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods according to claims 1 to 3.

Images