| The wave compensation device is a mechanical device that provides a relatively stable construction operation environment for offshore personnel and equipment in the turbulent offshore construction environment.According to the driving mode,it can be divided into three types:active,passive and semi-active.Compared with passive wave compensation device and semi-active wave compensation device,active wave compensation device has the advantages of high control precision and good multi-degree-of-freedom compensation coupling effect.The active wave compensation device is very sensitive to the influence of time delay in the working process,and it is a typical time delay control system.However,compared with the conventional time-delay control system,there are two core problems in the control of the active wave compensation system:First,due to the lack of effective reference objects at sea,the motion measurement of the active wave compensation system can usually only be performed by the inertial navigation system(Inertial navigation system).Navigation System,INS),but the measurement of multi-degree-of-freedom motion system by INS has problems such as signal drift and slow response,which seriously interferes with the accuracy of ship motion measurement;second,during the working process of the active wave compensation system,the system delay will be With the load,operation frequency,continuous operation time and other reasons changing at any time,it brings many uncertain influences to the control of the active wave compensation system.Based on the above reasons,the conventional predictive control scheme is often difficult to achieve the ideal control effect in the control process of the active wave compensation system.Aiming at this problem,this paper analyzes the limitations and deficiencies of predictive control schemes in practical engineering applications based on the working characteristics of active wave compensation systems,and proposes a feedforward time domain prediction-feedback frequency domain monitoring measurement control Program.In this paper,a 3-UPU_UP active wave compensation device test machine is used as the research object,the motion modeling and mechanical modeling of the 3-UPU_UP active wave compensation platform are carried out,and the simulation is carried out on the ship simulation platform of Shanghai Maritime University.Simulation.Through multi-group model analysis and experimental verification,it is proved that the control technology proposed in this paper has a significant improvement in measurement accuracy and delay adaptability compared with conventional control technology,and has high feasibility and application in practical applications.value.According to the working principle of the active wave compensation platform,the kinematics model of the 3-UPU UP compensation platform is established.It is proved that the 3-UPU UP active wave compensation platform can compensate the motion in the heave,roll and pitch directions of the ship.This paper discusses the 3-UPU_UP active wave compensation platform and compensation interval,and gives the driving target stroke of the electric cylinder when the attitude position is known.In this paper,based on the absolute sensing system and the relative sensing system,a scheme for monitoring and calculating the motion and attitude of the simulated offshore motion platform is proposed.According to the working principle of different sensors,the calculation formula of multi-sensor sensor data fusion correction is given.Through experiments,it is confirmed that this method can significantly improve and enhance the measurement accuracy.In this paper,an Absolute Sensors Data Fusion(ASDF)network that can fuse the sensing signals of multiple sensors is designed as a real-time predictor of the feedforward loop.Multiple sets of experiments show that the correction accuracy of ASDF network for multi-degree-of-freedom motion measurement is 11.4%~22.1%higher than that of conventional LSTM network.The larger the pitch and roll angle,the more obvious the correction effect of ASDF network.This paper analyzes the dynamic response of the compensation system in the presence of time-delay prediction residuals for ideal compensation control operations.And a method for calculating and correcting the delay of the predicted system in reverse in real time by compensating the actual response of the system.Then,based on the model,a Delay Model Prediction(DMP)recurrent neural network model is designed,which can monitor the residual prediction delay of the system in real time.Experiments show that the DMP network can predict and compensate the system time delay of the platform in real time.Using the DMP network as the time-delay closed-loop feedback of the ASDF neural network can effectively realize the time-delay adaptive compensation control of the absolute wave compensation system.In this paper,through simulation and experimental verification,it is verified that the active wave compensation technology based on the DMP-ASDF neural network predictive control method has the advantages of higher adaptability and high stability compensation accuracy.Through the experimental verification of multiple groups of different time delay variables,the error rate of the active compensation technology proposed in this paper under the influence of gradual time delay and sudden time delay decreased by 4.9%~8.5%and 3.7%~9.7%,respectively.The greater the time delay change The more the error drops.It provides a new technology with stronger adaptability,better stability and higher compensation precision for practical engineering applications. |