시간지연을 고려한 해양크레인의 HIV보상시스템의 능동적 제어. 해양과학기술융합연구과 해양과학기술대학원. 본 논문은 해상크레인이 예상치 못한 외란 및 외력을 받았을 때의 HIV 보상 시스템을 제시한다.
For better performance, sliding mode control and nonlinear generalized predictive control algorithm are used due to the time delay. As a result, the oscillating amplitude of the payload is reduced by using the control algorithm. Considering the time delay assumed to be one second involved in the system, nonlinear generalized predictive controller with robust controller is a suitable control algorithm for this heave compensation system because it makes the position of the payload reach the desired position with at least error.
논문의 마지막 부분에서는 강인한 제어를 이용한 제어 알고리즘과 그 시뮬레이션 결과를 보여줍니다. 키워드: 활성 상동 보상 상동 활성 제어; 비선형 시스템 비선형 시스템; 글라이딩 모드 제어 글라이딩 모드 제어; 시간 지연 시간 지연; 비선형 일반화 예측 제어.
Introduction
- Background and History
- Recent research
- Coordinate of AHC system
- Dynamic relations among the forces
- Hydraulic-driven winch system dynamics
To ensure the functionality of the crane system in difficult sea conditions, such as the situation described in the figure, it is crucial to consider the tension of the rope to which the load is attached, because it causes damping dynamics in the underwater condition. In addition, the unwanted vertical movement of the cargo must be noticeably reduced regardless of the environmental load.
A control system is often based on a programmable logic controller and it calculates how the active parts of the system should react to movement. A lift compensator package capable of transferring heavy loads was devised to compensate for a precise control of good load positioning. A pressure transducer mounted on a support structure produced a pressure signal indicating depth so that the sonar array could be used to map the ocean floor using the circuit.
Then, with the help of the Kalman filter, he was able to determine a more accurate value of the high frequency, which is important for the accurate representation of the seabed profile. In order to acquire the pitching motion of the offshore crane in real time caused by the disturbances at sea, the modern system uses an inertial measurement unit (IMU) as shown in the figure. Since the load is attached to the end of the rope, there are time delays in the input-output signal system between the sensors and control systems.
On the other hand, Kuchler et al., (2011) used HIV prediction algorithms to estimate and predict the vessel's HIV motion based on previous measurements and then applied control action based on these predicted motions. The AHC system consists of the offshore crane, the hydraulically driven winch and the payload clings to the elastic rope. Since the AHC system concentrates on estimating and compensating the heave motion of the payload, only the vertical motion coordinate is defined.
The control purpose of AHC system is to maintain the position of the payload when it is located at the deep sea. The position of the payload consists of the length of rope , the elongation of elastic rope ∆ and payload movement effected by the offshore crane disturbances acting on. Because only 1 degree is considered and above assumption, the deviation from the origin of the queue changes so that it does not break any features of dynamic model.
Here, is Young's modulus of the rope and denotes the cross-sectional area of the rope. From the rope extension, the forces acting on the AHC system are described in this section.
Control Strategy
PD controller
Sliding mode controller
To satisfy the sliding surface equation , the equivalent control input is calculated as in equation (3.9). In the motion time frame, iterative differentiation up to times of the output with respect to time is needed. As the whole system consists of dynamic system, disturbance, controller, winch encoder and sensor, Fig.4.3 shows the whole flow of AHC system.
Due to the disturbance, which is assumed to be a normal wave, while at sea the actual disturbance is an irregular wave, which is not considered in this study, the movement of the cargo obviously fluctuates. Its maximum displacement is 0.58 m, where the origin of the payload is 2000 m below the sea. The effect of PD control is clearly visible as the load displacement is reduced as expected.
In this thesis, the heave compensation control system which consists of marine crane assumed to be a rigid body, hydraulically driven winch, elastic rope and load is mathematically modeled and implemented in a simulation. The purpose of the control is to maintain the relative movement of the load from the reference point to rest and the important variable in this system is the dynamic change of the length of the elastic rope. As for load position control, the PD control algorithm is the first option to choose.
Since the purpose of the control is matched, and the characteristics of the error dynamics are shown in the AHC system. From the point of view of error dynamics, robust controller, i.e., sliding mode controller is proposed here and used for simulation. From equation (3.4), the whole AHC system is transformed into equation (A.1) with the appropriate coordinate change. A.1) where indicates the non-linear change of the coordinate.
Assume that the output of the nonlinear system in the prediction horizon is approximated by its Taylor series expansion up to order . After the nonlinear system (A.16) is approximated by its Taylor series expansion up to the order , the highest derivative is the required control. However, for a general nonlinear under consideration as in (A.1), the control does not appear in a linear way, and it is difficult to give the explicit solution for .
Nonlinear generalised predictive controller
Simulation and Results
PD controller
Since the PD control originally reduces the amplitude of the variables, it seems that the PD control algorithm is well applied to this system.
Sliding mode controller
While the chatter exists, the amplitude is only in about 10% compared to the original dynamic characteristic that already has. It does not matter what kind of controller is used on this system because the oscillatory motion is not removed. The low frequency disturbance makes the controlled output stable at the end while the chatter remained a bit.
Since the dynamic parameters and variables are sensed by the sensors, calculated by the winch encoder and must be used to derive the control algorithms, there is time delay between the sensors and other devices. Actually, time delay is the main problem in the signal system, but in this paper, only the existing time delay of 1 second is taken into account for the simulation.
Nonlinear generalised predictive controller
Conclusions
Predictive control of general nonlinear systems
A general nonlinear single-input-single-output (SISO) system can be described by equation (A.16). The results developed in this section can be extended to general multivariable nonlinear systems. By repeating the above procedure, higher order derivatives can be calculated up to the th order derivative, which is given by.
By calling (A.27) into (A.20), the receding horizon output is approximated by its Taylor expansion to the order as. Similarly, the command in the receding horizon can also be approximated by its Taylor series expansion to order as. At time , the MPC tries to find the optimal control profile in the receding horizon.
Hyeong Sik Choi, Department of Convergence Study on the Ocean Science and Technology Korea Maritime and Ocean University, Korea for his precious advice and continuous encouragement for my study. I am very grateful for their help with all the things we have been through both good and bad moments. Finally, I would like to thank my family who always have faith in me and have also loved, encouraged and supported me.
