Presentations

Abstract

Wind plant owners and operators lack control of their assets. Despite potential solutions to industry problems, e.g., * Collective control wake steering to address wind plant wake losses (anywhere between 5 and 15% for most onshore wind plants and significantly higher offshore) * Retrofit solutions to meet evolving compliance requirements * Sensor input integration (lightning and blade icing detection) for turbine protection owners and operators must frequently rely on their wind turbine OEM or a network of 3rd party vendors for a solution. However, OEM solutions are often a black box, and leveraging a network of 3rd party vendors can be difficult to manage without a unified platform and central datasource. However, through IoT that enables actuation (e.g. for collective control wake steering) and a centralized platform for sensor input integration, wind plant owners and operators can gain more control of their assets to achieve site objectives for performance, reliability, and compliance. A major benefit to this IoT platform is actuation, which enables collective control wake steering of the wind plant. Collective control through wake steering is needed to promote wind plant optimization and to achieve a paradigm shift in plant output (+1% in AEP). These improvements in energy production are imperative for an industry that is increasingly subject to margin pressure. Results are presented of wake steering at a 165 turbine wind plant, wherein a novel model-in-the-loop wake steering control system was developed and deployed. The system employs a novel model architecture using a data-driven input estimation and calibration of an engineering wake model along with a neural network-based output correction to provide improved performance over a static lookup table. A multi-month data collection period is used for validation and to demonstrate successful cases of wake steering and opportunities for improvement. While collective control wake steering can be used to facilitate production improvements by transitioning away from OEM defined turbine-centric control, and will be a component of this talk, there is not a one-size-fits-all solution for wind plant optimization. Objectives, whether those be performance, reliability, or compliance, depend on the wind plant, which is why IoT platforms enabling actuation and sensor integration need to be flexible (i.e., beyond actuation for collective control wake steering). Use cases will be presented for actuation and sensor input integration that can be used to meet these objectives. For example, compliance in some geographic territories requires that turbines operate in a wider range of temperature conditions (outside OEM bounds) given climate change, and thus, how can sensor input and actuation be used to achieve these compliance requirements in a cost effective manner.