Observer-Based Fault Estimation Applied to a Point Absorber Wave Energy Converter

Abstract

The economic viability of wave energy conversion systems is one of the open points among the research community. To lower the energy cost, the devices in charge of extracting the wave power, called wave energy converters (WEC), are often controlled by means of optimal control (OC) strategies. Such OC systems, which have proven their effectiveness in wave energy applications, often rely on a mathematical model of the device (and, in some cases, on the wave excitation force) to optimize the control action provided to the system. Nevertheless, the marine environment results hostile for general device safe operations, potentially triggering a variety of faults in the WEC system. Such condition (for example, a sensor failure or additional friction inside a gearing) directly affects the system dynamics. If this deviation is not considered by the control algorithm, the energy production performance can degrade considerably, or the control action itself can cause a more serious fault. A possible solution is that of designing an algorithm capable of compensating for eventual faults in the system, while still respecting the initial design performance or, when not possible, preserving the main device functionalities. Such a control strategies belong to the family of Fault Tolerant Control (FTC) techniques, which can be divided into two macro-categories: Passive (PFTC) and active (AFTC) algorithms. While PFTC systems are designed offline and can account only for a predefined set of system faults, AFTC algorithms are more suitable to tackle significant system deviations from the nominal model. For this purpose, such algorithms may require some routine to detect, isolate and eventually estimate the specific fault. This task is accomplished by Fault Detection and Identification (FDI) routines. According to the AFTC algorithm, the FDI module must accomplish different tasks. Furthermore, the FDI module accuracy plays a crucial role in some AFTC strategies, since the poor estimation of a faulty signal can induce the controller to behave incorrectly.
This paper presents an FDI algorithm applied to a point-absorber wave energy converter (WEC). The proposed structure consists of an observer-based strategy in charge of detecting, isolating, and tracking effectively faulty signals occurring in numerical simulations. The results demonstrate the proposed observer effectiveness for a predefined set of actuator and sensor faults, both in the case of independent, and simultaneous fault occurrence.