a b s t r a c tAn optimisation model and a solution method for maintenance routing and scheduling at offshore wind farms are proposed. The model finds the optimal schedule for maintaining the turbines and the optimal routes for the crew transfer vessels to service the turbines along with the number of technicians required for each vessel. The model takes into account multiple vessels, multiple periods (days), multiple Operation & Maintenance (O&M) bases, and multiple wind farms. We develop an algorithm based on the Dantzig-Wolfe decomposition method, where a mixed integer linear program is solved for each subset of turbines to generate all feasible routes and maintenance schedules for the vessels for each period. The routes have to consider several constraints such as weather conditions, the availability of vessels, and the number of technicians available at the O&M base. An integer linear program model is then proposed to find the optimal route configuration along with the maintenance schedules that minimise maintenance costs, including travel, technician and penalty costs. The computational experiments show that the proposed optimisation model and solution method find optimal solutions to the problem in reasonable computing times.

Due to lack of operating experience in the field of offshore wind energy and large costs associated with maintaining offshore wind farms, there is a need to develop accurate operation and maintenance models for strategic planning purposes. This paper provides an approach for verifying such simulation models and demonstrates it by describing the verification process for four models. A reference offshore wind farm is defined and simulated using these models to provide test cases and benchmark results for verification for wind farm availability and O&M costs. This paper also identifies key modelling assumptions that impact the results. The calculated availabilities for the four models show good agreement apart from cases where maintenance resources are heavily constrained. There are also larger discrepancies between the cost results. All the differences in the results can be explained by different modelling assumptions. Therefore, the models can be regarded as verified based on the presented approach. INTRODUCTION MotivationOffshore wind energy is a new area for operation and maintenance (O&M) research, and the operational legacy of the industry is only just over a decade. Operation and maintenance cost modelling software tools are being developed to support activities in this field. Because of the novelty of offshore wind energy generation and lack of real data, there are limited options for validation and verification of these models. Verification and validation of a simulation model is essential if the model is to say something useful about the system it is meant to represent. We define verification as ensuring that the simulation model is implemented according to the specifications of the conceptual model of the system; validation is defined as ensuring that this conceptual model is in fact a faithful representation of the real system for the purposes of the model [1]. It may prove difficult for researchers to acquire suitable data to perform model validation. For full operational validation [1], necessary historical data would include repair and logistical costs, statistical information on component reliability and performance indicators such as total operations costs or availability. This type of information is possessed by the farm owner/operator, turbine manufacturer or non-existent for new generation wind turbines. BackgroundSeveral O&M simulation models for offshore wind farms have been developed, of which Hofmann [2] provides a thorough overview. Often, the intended applications of the models differ slightly. For example, one model will focus on assessing heavy-lift vessels, whereas another will be used for maintenance strategy optimisation. [3]. One position is that models are never entirely validated because it is not practicable to assess correspondence between the system and the model for its entire domain of applicability [1]. Even if the system is observable and a comparison of model output and system output is possible, one is often interested in predicting system behaviour under circumst...

No abstract

Flexible distributed energy resources, such as energy storage systems (ESSs), are increasingly considered as means for mitigating challenges introduced by the integration of stochastic, variable distributed generation (DG). The optimal operation of a distribution system with ESS can be formulated as a multi-period optimal power flow (MPOPF) problem which involves scheduling of the charging/discharging of the ESS over an extended planning horizon, e.g., for day-ahead operational planning. Although such problems have been the subject of many works in recent years, these works very rarely consider uncertainties in DG, and almost never explicitly consider uncertainties beyond the current operational planning horizon. This article presents a framework of methods and models for accounting for uncertainties due to distributed wind and solar photovoltaic power generation beyond the planning horizon in an AC MPOPF model for distribution systems with ESS. The expected future value of energy stored at the end of the planning horizon is determined as a function of the stochastic DG resource variables and is explicitly included in the objective function. Results for a case study based on a real distribution system in Norway demonstrate the effectiveness of an operational strategy for ESS scheduling accounting for DG uncertainties. The case study compares the application of the framework to wind and solar power generation. Thus, this work also gives insight into how different approaches are appropriate for modeling DG uncertainty for these two forms of variable DG, due to their inherent differences in terms of variability and stochasticity.

No abstract