Wave power generation power control method based on time delay compensation

Abstract

The invention relates to a wave power generation power control method based on time delay compensation, which is suitable for a vibration water column type wave power generation device control system. The method comprises three parts of a feed-forward control layer, a rolling optimization layer and a feedback correction layer, the feed-forward control layer predicts and compensates future time lag influence by using an LSTM network, the rolling optimization layer optimizes a system control rate in real time by using a model prediction control algorithm MPC, and the feedback correction layer provides system robustness by using sliding mode control. According to the method, the problems of power fluctuation and efficiency loss caused by time-varying delay in a wave power generation system are solved, and the robustness of the system is ensured while global optimization power tracking is realized by adopting a mode of combining a model predictive control algorithm MPC and sliding mode control.

Description

Technical Field

[0001] This invention relates to the field of wave energy power generation technology, and specifically to a wave energy power generation control method based on time delay compensation. Background Technology

[0002] Wave energy, as a clean and renewable energy source, has enormous development potential. Vibrating water column wave energy generation devices are one of the more mature technologies currently available. They generate electricity by driving an air turbine through the vibration of a water column within an air chamber, offering advantages such as simple structure and ease of implementation. However, this device still faces many technical challenges in practical operation, especially in terms of power generation control.

[0003] The output power of a vibrating water column wave energy generator is directly affected by wave characteristics. The randomness and time-varying nature of waves cause instability in the input energy of the device, leading to fluctuations in power generation. Furthermore, wave energy generation systems exhibit various time delay phenomena, such as fluid inertial time delay within the gas chamber, turbomachinery response time delay, and wide-area signal transmission time delay. These time delays cause system response delays, further exacerbating power fluctuations and reducing power generation efficiency. Existing wave energy generation control systems mainly employ traditional PID control or simple feedback control strategies. While these methods can stabilize the system output to some extent, their control performance is often insufficient when facing complex marine environments and time-varying delays. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a wave energy power generation control method based on time delay compensation to solve the aforementioned problems.

[0005] This invention provides the following technical solution:

[0006] A wave energy power generation control method based on time-delay compensation comprises three parts: a feedforward control layer, a rolling optimization layer, and a feedback correction layer. The feedforward control layer uses an LSTM network to predict and compensate for future time-delay effects; the rolling optimization layer employs a Model Predictive Control (MPC) algorithm to optimize the system control law in real time; and the feedback correction layer utilizes sliding mode control to provide system robustness. Specifically, it includes the following steps:

[0007] Step S1: Real-time acquisition of the pressure P inside the wave energy generator's air chamber using an air chamber pressure transmitter, speed encoder, and wave height radar. c Generator speed ω g and wave height H in the air chamber w .

[0008] Step S2: Perform time-delay dynamic modeling on the time-delay source of the wave energy power generation device, and use an LSTM network to predict the future valve opening and turbine speed to generate the feedforward compensation quantity u. ff(t), the specific steps are as follows:

[0009] Step S21: The time delay sources include the inertial time delay of the fluid in the gas chamber, the time delay of the turbomachinery response, and the time delay of the wide-area signal transmission. The unified state equation for multiple time delays is established as follows:

[0010]

[0011] Where x(t) = [P] c (t),ω(t),H ω (t)] T ;τ i For time-varying time delays, τ1(t), τ2(t), and τ3(t) represent the fluid inertia time delay, turbine response time delay, and signal transmission time delay, respectively, all of which are time-varying time delays; A is the system matrix, describing the dynamic relationship between system state variables; Bi is the control input matrix, describing the influence of control input u(t) on system state variables; D is the disturbance term coefficient matrix, describing the influence of disturbance term ω(t) on system state variables; ω(t) represents wave-excited disturbance, modeled as colored noise as follows:

[0012]

[0013] α: Attenuation coefficient, used to describe the attenuation characteristics of the disturbance, W(s) is Brownian motion.

[0014] Step S22: Use an LSTM network to predict future valve opening and turbine speed, where the input-output relationship of the LSTM network is as follows:

[0015] Input layer: Historical time series data {u(t-kΔt), x(t-kΔt)}, k=0,1,…,N (Δt is the time interval, N is the total prediction time); Hidden layer: LSTM unit state update formula is as follows:

[0016]

[0017] Among them, f t i t o t These represent the forget gate, input gate, and output gate, respectively; W is the weight matrix; and b is the bias term. Candidate values; ⊙ represents element-wise multiplication;

[0018] Output layer: Predict future time delay compensation amount:

[0019] T p For the predicted duration;

[0020] Step S23: Based on the multi-delay unified state equation and the predicted time delay output by the LSTM network, obtain the feedforward compensation command:

[0021]

[0022] Step S3: Design the Model Predictive Control (MPC) algorithm for the wave energy power generation device to obtain the initial control quantity u of the system. mpc (t), the specific steps are as follows:

[0023] Step S31, the discretized state equation is:

[0024] x[k+1]=A d x[k]+B d u[k]+D d ω[k];

[0025] Among them, A d =e AΔt ,

[0026] Step S32, define the objective function and constraints as follows:

[0027]

[0028]

[0029] Where, N p With N c These represent the prediction and control steps, respectively, P. max ω gmax ω gmin u max These are the upper limit of the pressure inside the air chamber, the upper limit of the generator speed, the lower limit of the generator speed, and the maximum change rate of the valve opening.

[0030] Step S33, use a quadratic programming (QP) solver.

[0031]

[0032] Among them, U=[u[0],u[1],…,u[Nc-1]] T The matrices H, f, G, and h are generated from the state equations and constraints.

[0033] Step S34: Obtain the initial control law u using the solver. mpc (t).

[0034] Step S4: Design a sliding mode compensation algorithm for the wave energy power generation device model to obtain the feedback compensation control quantity u. comp (t), the specific steps are as follows.

[0035] Step S41, define the tracking error e(t) = P ref (t)-Pc (t), the integral sliding surface is designed as follows:

[0036]

[0037] Among them, P ref (t) represents the desired power, P c (t) represents the actual power, and β represents the control adjustment parameter.

[0038] Step S42: Design the sliding mode control law to obtain the compensation control quantity u. comp (t) is:

[0039]

[0040] Among them, K p and K i To control the gain.

[0041] Step S5, feedforward compensation amount u ff (t), initial control quantity of the system u mpc (t) and feedback compensation control quantity u comp The final control quantity u(t) is obtained by superimposing the values ​​of (t) together:

[0042] u(t)=u comp (t)+u ff (t)+u mpc (t);

[0043] Power control is achieved by adjusting the air intake volume of the air chamber by driving the proportional-integral valve based on the final control quantity u(t).

[0044] The present invention has the following beneficial technical effects:

[0045] This invention utilizes an LSTM network in the feedforward control layer to predict and compensate for future time delays, proactively offsetting these delays and significantly reducing power fluctuations caused by time-varying delays in wave energy generation systems, thereby improving power generation stability. The MPC (Model Predictive Control) algorithm in the rolling optimization layer optimizes the system control law in real time, achieving global power point tracking. This enables the system to efficiently capture wave energy under various sea states, significantly improving power generation efficiency. The feedback correction layer employs a sliding mode control algorithm, providing strong robustness to the system. Sliding mode control can quickly suppress disturbances, ensuring stable operation of the system in complex marine environments (such as extreme sea states and random wave fluctuations), improving the system's reliability and adaptability.

[0046] This invention employs a multi-layered control architecture (feedforward control layer, rolling optimization layer, and feedback correction layer) to achieve refined control of wave energy generation power. Feedforward compensation corrects time lag effects in advance, MPC optimizes global power tracking, and sliding mode control quickly suppresses disturbances; the combination of these three significantly improves the performance of the control system. Attached Figure Description

[0047] Figure 1 This is a control principle diagram of a wave energy power generation control method based on time delay compensation according to the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Example

[0050] A wave energy power generation control method based on time delay compensation is applicable to the control system of a vibrating water column wave energy generation device. The control system of the vibrating water column wave energy generation device is existing technology, which achieves power generation control by adjusting the opening of the air chamber inlet valve in the vibrating water column wave energy generation device, and will not be described in detail here.

[0051] The hardware of the control system for the vibrating water column wave energy generation device includes data acquisition equipment, actuators, and a controller. The data acquisition equipment includes a chamber pressure transmitter, a speed encoder, and a wave height radar; the actuators include a proportional-integral valve (for controlling the opening of the chamber valve) and a permanent magnet synchronous generator; the controller includes a PLC controller and an edge AI chip (for carrying LSTM inference).

[0052] A wave energy power generation control method based on time-delay compensation comprises three parts: a feedforward control layer, a rolling optimization layer, and a feedback correction layer. The feedforward control layer utilizes a Long Short-Term Memory (LSTM) network to predict and compensate for future time-delay effects. The rolling optimization layer employs a Model Predictive Control (MPC) algorithm to optimize the system control law in real time. The feedback correction layer utilizes sliding mode control to provide system robustness. Specifically, it includes the following steps:

[0053] Step S1: Real-time acquisition of the pressure P inside the wave energy generator's air chamber using an air chamber pressure transmitter, speed encoder, and wave height radar. c Generator speed ω g and wave height H in the air chamber w .

[0054] Step S2: Perform time-delay dynamic modeling on the time-delay source of the wave energy power generation device, and use an LSTM network to predict the future valve opening and turbine speed to generate the feedforward compensation quantity u. ff (t), the specific steps are as follows:

[0055] Step S21: The time delay sources include the inertial time delay of the fluid in the gas chamber, the time delay of the turbomachinery response, and the time delay of the wide-area signal transmission. The unified state equation for multiple time delays is established as follows:

[0056]

[0057] Where, x(t)=[P c (t),ω g (t),H ω (t)] T τ1(t), τ2(t), and τ3(t) represent the fluid inertial time delay, turbine response time delay, and signal transmission time delay, respectively, all of which are time-varying time delays; ω(t) represents the wave excitation disturbance, modeled as colored noise as follows:

[0058]

[0059] W(s) represents Brownian motion.

[0060] Step S22: A Long Short-Term Memory (LSTM) network is used to predict future valve opening and turbine speed. The input-output relationship of the LSTM network is as follows:

[0061] Input layer: historical time series data {u(t-kΔt),x(t-kΔt)}, k=0,1,…,N (Δt is the time interval, N is the total prediction time length);

[0062] Hidden layer: The LSTM cell state update formula is as follows:

[0063]

[0064] Among them, f t i t o t These represent the forget gate, input gate, and output gate, respectively; W is the weight matrix; and b is the bias term. y is a candidate value, and ⊙ represents element-wise multiplication.

[0065] Output layer: Predict future time delay compensation amount:

[0066]

[0067] in, T is the predicted value. p For the predicted duration;

[0068] Step S23: Based on the multi-delay unified state equation and the predicted time delay output by the LSTM network, obtain the feedforward compensation command:

[0069]

[0070] Step S3: Design the Model Predictive Control (MPC) algorithm for the wave energy power generation device to obtain the initial control quantity u of the system. mpc (t), the specific steps are as follows:

[0071] Step S31, Discretize the state equation as follows:

[0072] x[k+1]=A d x[k]+B d u[k]+D d ω[k];

[0073]

[0074] Step S32, define the objective function and constraints as follows:

[0075]

[0076]

[0077] Where, N p With N c These represent the number of prediction steps and the number of control steps, respectively. max ω gmax ω gmin u max These are the upper limit of the pressure inside the air chamber, the upper limit of the generator speed, the lower limit of the generator speed, and the maximum change rate of the valve opening.

[0078] Step S33, using a quadratic programming (QP) solver:

[0079]

[0080] Among them, U=[u[0],u[1],…,u[Nc-1]] T The matrices H, f, G, and h are generated from the state equations and constraints.

[0081] Step S34: Obtain the initial control law u using the solver. mpc (t).

[0082] Step S4: Design a sliding mode compensation algorithm for the wave energy power generation device model to obtain the feedback compensation control quantity u. comp (t), the specific steps are as follows.

[0083] Step S41, define the tracking error e(t) = Pref (t)-P c (t), the integral sliding surface is designed as

[0084]

[0085] Among them, P ref (t) represents the desired power, P c (t) represents the actual power, and β represents the control adjustment parameter.

[0086] Step S42: Design the sliding mode control law to obtain the compensation control quantity u. comp (t) is:

[0087]

[0088] Among them, K p and K i To control the gain.

[0089] Step S5, the feedforward compensation amount u ff (t), initial control quantity of the system u mpc (t) and feedback compensation control quantity u comp The final control quantity u(t) is obtained by superimposing the control quantities (t).

[0090] u(t)=u comp (t)+u ff (t)+u mpc (t);

[0091] Power control is achieved by adjusting the air intake volume of the air chamber by driving the proportional-integral valve based on the final control quantity u(t).

[0092] This invention employs a multi-layered control architecture (feedforward control layer, rolling optimization layer, and feedback correction layer) to achieve refined control of wave energy generation power. Feedforward compensation corrects time lag effects in advance, MPC optimizes global power tracking, and sliding mode control quickly suppresses disturbances; the combination of these three significantly improves the performance of the control system.

[0093] The embodiments described above are merely illustrative of specific implementations of the present invention, and while the descriptions are detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A wave energy power generation control method based on time-delay compensation, characterized in that, Includes the following steps: Step S1: Real-time acquisition of the pressure P inside the wave energy generator's air chamber using an air chamber pressure transmitter, speed encoder, and wave height radar. c Generator speed ω g and wave height H in the air chamber w ; Step S2: Perform time-delay dynamic modeling on the time-delay source of the wave energy power generation device, and use a Long Short-Term Memory (LSTM) network to predict future valve opening and turbine speed, generating feedforward compensation quantity u. ff (t); Step S3: Design the Model Predictive Control (MPC) algorithm for the wave energy power generation device to obtain the initial control quantity u of the system. mpc (t); Step S4: Design a sliding mode compensation algorithm for the wave energy power generation device model to obtain the feedback compensation control quantity u. comp (t); Step S5, adjust the feedforward compensation amount u ff (t), initial control quantity of the system u mpc (t) and feedback compensation control quantity u comp The control quantity u(t) is obtained by superimposing the control quantities u(t) and then driving the proportional-integral valve to adjust the air intake volume of the air chamber according to the final control quantity u(t), thereby realizing power control.

2. The wave energy power generation control method based on time delay compensation according to claim 1, characterized in that, Feedforward compensation amount u ff The specific steps of (t) include: Step S21: The time delay sources include the inertial time delay of the fluid in the gas chamber, the time delay of the turbomachinery response, and the time delay of the wide-area signal transmission. The unified state equation for multiple time delays is established as follows: Where, x(t)=[P c (t),ω g (t),H ω (t)] T ;τ i For time-varying time delays, τ1(t), τ2(t), and τ3(t) represent the fluid inertial time delay, turbine response time delay, and signal transmission time delay, respectively; A is the system matrix, B... i The control input matrix is ​​D; the disturbance term coefficient matrix is ​​D; ω(t) is the wave excitation disturbance, modeled as colored noise as follows: Where α is the attenuation coefficient, used to describe the attenuation characteristics of the disturbance; W(s) is Brownian motion.

3. The wave energy power generation control method based on time delay compensation according to claim 2, characterized in that, Feedforward compensation amount u ff The specific steps of (t) also include: Step S22: Use an LSTM network to predict future valve opening and turbine speed. The input-output relationship of the LSTM network is as follows: Input layer: historical time series data {u(t-kΔt), x(t-kΔt)}, k=0,1,…,N, Δt is the time interval, and N is the total prediction time length; Hidden layer: LSTM cell state update formula is as follows: Among them, f t i t o t These are the forget gate, input gate, and output gate, respectively; ⊙ represents element-wise multiplication; Output layer: predicts future time delay compensation. in, T is the predicted value. p For the predicted duration.

4. The wave energy power generation control method based on time delay compensation according to claim 1, characterized in that, The initial control quantity u of the system mpc The specific steps for (t) are as follows: Step S31, the discretized state equation is: x[k+1]=A d x[k]+B d u[k]+D d ω[k] Among them, a d =e AΔt , Step S32, define the objective function and constraints as follows: Where, N p With N c These represent the prediction and control steps, respectively, P. max ω max ω min u max These are the upper limit of the pressure inside the air chamber, the upper limit of the generator speed, the lower limit of the generator speed, and the maximum change rate of the valve opening; Step S33, using a quadratic programming (QP) solver: Among them, U=[u[0],u[1],…,u[Nc-1]] T The matrices H, f, G, and h are generated from the state equations and constraints.

5. The wave energy power generation control method based on time delay compensation according to claim 1, characterized in that, Feedback compensation control quantity u comp The specific steps for (t) are as follows: Step S41, define the tracking error e(t) = P ref (t)-P c (t), the integral sliding surface is designed as follows: Where β is the control adjustment parameter; Step S42: Design the sliding mode control law to obtain the compensation control quantity u. comp (t) is: Among them, K p and K i To control the gain.

Images