Opportunistic maintenance for wind farms considering multi-level imperfect maintenance thresholds

Abstract

Few methods are available for optimizing corrective maintenance and time-based maintenance for wind farms, although these strategies are currently widely used in practice. Economic dependencies exist among wind turbine systems and their components in the wind farm. That is, it may be more economical to maintain multiple turbines or turbine components when a corrective or preventive maintenance opportunity presents. In this paper, opportunistic maintenance approaches are developed for wind farms to take advantage of the maintenance opportunities. Imperfect maintenance actions are considered, which addresses the practical issue that preventive maintenance does not always return components to as-good-as-new status. The proposed opportunistic maintenance policies are defined by the component's age threshold values, and different imperfect maintenance thresholds are introduced for failure turbines and working turbines. Three types of preventive maintenance actions are considered, including perfect, imperfect and two-level action. Simulation methods are developed to evaluate the costs of proposed opportunistic maintenance policies. Numerical examples are provided to illustrate the proposed approaches. Comparative study with the widely used corrective maintenance policy demonstrates the advantage of the proposed opportunistic maintenance methods in significantly reducing the maintenance cost. The developed methods are expected to bring immediate benefits to wind power industry.

Introduction

Global warming and high oil and gas price present urgent needs to explore competitive clean and renewable energy. Wind energy is one of most important renewable energy sources in the world, and its installed capacity worldwide has grown significantly in recent years. Operation & maintenance have drawn increasing interests for reducing the significant investment in the wind power projects, and appropriate and practical maintenance strategies need to be heavily studied for successful future developments. Maintenance management aims at improving the availability of the systems and reducing the overall maintenance cost. The existing maintenance methods for wind power systems can be classified into failure-based (corrective), time-based, and condition-based maintenance (CBM). Failure-based maintenance is carried out only after a failure occurs. In time-based maintenance, preventive maintenance is performed at predetermined time intervals. CBM applies the health condition prediction techniques to continuously monitor the components so that components are used the most effectively. However, the availability of condition monitoring data is a big challenge for CBM applications in wind turbine systems today. Currently corrective maintenance and time-based preventive maintenance are widely used in wind power industry, which take advantage of ease of management, particularly in the case of extreme conditions and high load associated with offshore farms. However, they have not been studied adequately and few models and methods are developed to optimize the time-based maintenance strategies.

Europe Wind Energy Report (2001) proposed four maintenance strategies for European offshore wind farms, and one of them is opportunistic maintenance. In opportunistic maintenance, whenever a failure occurs in the wind farm, the maintenance team is sent onsite to perform corrective maintenance, and take this opportunity to simultaneously perform preventive maintenance on the other components in the failed turbines and the running turbines and their components which show relatively high risks. There are typically multiple wind turbines in a wind farm and a wind turbine has multiple components. Economic dependencies exist among various components and systems in the farm. When a down time opportunity is created by the failed components, maintenance team may perform preventive maintenance for other components satisfying pre-specified decision conditions, such as certain age thresholds. As a result, substantial cost can be saved comparing to separate maintenance for the components.

In the general maintenance engineering field, various opportunistic maintenance policies and applications have been reported. Laggoune [2] considered hydrogen compressors with different component failure distributions, and made maintenance decisions based on if performing replacements can lower the expected costs. An age-based policy was used by Crocker [3] to optimize the maintenance of a military aero-engine, and they concluded that opportunistic maintenance should be performed on relatively cheap components in their application. Mohamed-Salah et al. [4] proposed an opportunistic maintenance policy for ball bearings based on the time difference between expected preventive maintenance time and failure instant. Kabir et al. [5] assumed that the components are identical following the same Weibull distribution, and presented a maintenance method for multi-unit systems. However, very few studies are reported on opportunistic maintenance for wind power systems. Besnard [7] proposed an opportunistic maintenance method for offshore wind turbine systems based on both failure chance and real wind data. They presented an optimization model with a series of constraints aiming at minimizing the cost, and an optimal maintenance schedule for a 5 turbines wind farm was presented. Tian et al. [6] developed a CBM method for wind farms by considering the economic dependencies among components, and determined the maintenance actions based on the optimized failure probability threshold values and the condition monitoring data.

In most existing studies on preventive maintenance of wind turbines, one disadvantage is that preventive maintenance actions are generally considered to be replacement, which is the perfect action to return a component to the as-good-as-new state. In practice, however, preventive maintenance does not always return components to the as-good-as-new status. According to Spinato et al. [1], repair actions for wind turbine components may include addition of a new part, exchange of parts, removal of a damaged part, changes or adjustment to the settings, software update, lubrication or cleaning, etc. Ding and Tian [10] developed opportunistic maintenance methods for wind farms considering imperfect maintenance actions. However, they did not distinguish between the failed turbines and working turbines regarding if preventive maintenance should be performed, and used the same maintenance thresholds for all the wind turbines.

To address the issues above, in this paper, opportunistic maintenance approaches are developed for wind farms to take advantage of the maintenance opportunities and consider imperfect maintenance actions. The proposed opportunistic maintenance policies are defined by the component's age threshold values, and different imperfect maintenance thresholds are introduced for failure turbines and working turbines. Three types of preventive maintenance actions are considered, including perfect, imperfect and two-level action. Simulation methods are developed to evaluate the costs of proposed opportunistic maintenance policies. Numerical examples will be provided to illustrate the proposed approaches.