Øivind Kåre Kjerstad

Disturbance rejection by acceleration feedforward : Application to dynamic positioning

Abstract This paper addresses environmental disturbance rejection in dynamic systems by means of acceleration feedforward. A feedforward structure calculates the environmental force and magnitude by comparing the filtered acceleration measurement with the acceleration due to actuation and known forces. Direct compensation of the environment by feedforward will alleviate the position and velocity feedback terms. The presented scheme compensates environmental disturbances affecting the system in addition to model uncertainties. First, the design is used for a mechanical system and illustrated for an inverted pendulum. Then, the main application, being a dynamic positioning system affected by ice forces in an Arctic operation, is controlled by acceleration feedforward to give enhanced performance.

Weather Optimal Positioning Control for Marine Surface Vessels

Abstract This paper addresses the problem of energy-efficient dynamic positioning (DP) for marine surface vessels by means of a weather optimal control scheme. Fundamental geometric relations associated with a pendulum in a force field are utilized to obtain fuel-efficient DP functionality. First, weather optimal heading control (WOHC) is achieved through the use of specific surge and yaw controllers which make a vessel behave like a force-field pendulum, employing a virtual suspension point and an associated pendulum length. This design is particularly suitable for sway-unactuated vessels. Second, an intuitive and geometrically-motivated update law is assigned to the virtual suspension point to facilitate weather optimal positioning control (WOPC), which enables a vessel to be positioned at a particular location with a weather optimal heading. Computer simulations incorporating realistic environmental disturbances are used to illustrate the performance of the suggested WOHC and WOPC systems.

Recursive nullspace-based control allocation with strict prioritization for marine craft

Abstract In control allocation it is often important to prioritize among the directions to produce control effort, especially in cases of limited capacity. Motivated by a requirement for strict prioritization among the control directions, a recursive nullspace-based allocation design based on direct use of the generalized pseudoinverses and nullspace matrices is proposed. The allocation design divides the overall problem into r subproblems according to a prioritization sequence, and a recursive method is proposed to solve the allocation subproblem in each step. By hiding the allocated controls from subsequent steps in the nullspace corresponding to the present step, the influence of lower priority control actions onto higher priority directions are nullified to achieve the specified prioritization. The method is verified by analyzing the thruster capacity of an Arctic intervention vessel based on experimental ice towing data.

Disturbance Rejection by Acceleration Feedforward for Marine Surface Vessels

Acceleration signals have a powerful disturbance rejection potential in rigid body motion control, as they carry a measure proportional to the resulting force. Yet, they are seldom used, since measuring, decoupling, and utilizing the dynamic acceleration in the control design is not trivial. This paper discusses these topics and presents a solution for marine vessels building on conventional methods together with a novel control law design, where the dynamic acceleration signals are used to form a dynamic referenceless disturbance feedforward compensation. This replaces conventional integral action and enables unmeasured external loads and unmodel dynamics to be counteracted with low time lag. A case study shows the feasibility of the proposed design using experimental data and closed-loop high fidelity simulations of dynamic positioning in a harsh cold climate environment with sea-ice.

Observer design with disturbance rejection by acceleration feedforward

Abstract This paper addresses observer design of mechanical systems subject to unknown external disturbances and unmodeled dynamics. For fast-acting disturbance forces, realistic and accurate physical models do not always exist, such as for dynamic positioning of marine vessels in ice-infested waters. To accommodate such forces and provide disturbance rejection, a delayed acceleration measurement is used to form an acceleration feedforward which captures the perturbations and enhances the observer phase margins. Through the acceleration measurement, an inherent sensor bias is introduced, creating a pseudo-force in the system. This is tracked and compensates by a bias estimate. The theoretical foundation is exemplified by the derivation of a DP observer and a simulation study featuring an inverted pendulum.

Feedforward linearization and disturbance rejection of mechanical systems using acceleration measurements

Abstract In this paper a time-delayed acceleration measurement is used to linearize a second order nonlinear system. A theoretical foundation is developed, and a control strategy is presented which handles the challenges related to classical linearization. Utilizing the acceleration measurement to directly cancel the nonlinearities constitute a feedforward approach which increases robustness with respect to unmodeled dynamics and model uncertainties. Due to the time-delay, one or more perturbation terms dependent on a previous state of the system appears. This work outlines a stability criterion which these terms must satisfy in order for the linearization to be valid, where it is shown that the performance of the scheme is dependent on the measurement time-delay. The proposed method proves feasible when compared to two other nonlinear control strategies, including a conventional linearization control law and a sliding mode control law. In comparison, the feedforward linearization features robust performance without aggressive control input.