Wind ITO Fulfillment Center: Capturing Proprietary Processes for Wind Farm Siting Analysis

This presentation was made at CAASE18, The Conference on Advancing Analysis & Simulation in Engineering. CAASE18 brought together the leading visionaries, developers, and practitioners of CAE-related technologies in an open forum, to share experiences, discuss relevant trends, discover common themes, and explore future issues.

Resource Abstract

When manual engineering processes are automated, it is crucial to make sure that these proprietary processes are properly captured. The siting analysis of a wind farm involves multiple steps to determine the feasibility of installing one or multiple models of wind turbines in specific locations. Among many environmental and legal considerations, it is essential to determine the engineering feasibility of installing wind turbines, which is based on ambient wind conditions, turbulence and geographical data in the selected location and the characteristics of the selected turbine models. Previously these evaluation processes for site-specific mechanical loads analysis (MLA) were done by geographically distributed experts on their local desktop computers with isolated software applications with multiple manually produced input files.

A system called Wind ITO Fulfillment Center (WIFC) was created to provide a streamlined approach to wind farm siting, enabling a unified method to execute a disparate set of specialized analyses and programs from a common web interface. WIFC serves as a centralized online portal for siting information, storage, review and analysis of applications for wind turbine MLA and general siting suitability. Site specific information on ambient wind conditions, environmental and geographical data along with the specifications of selected turbine models are the inputs to this system. All the engineering programs and related libraries used in the analyses are stored in a centralized file system. As algorithms and programs are enhanced or replaced, they can be updated to the live system without disrupting existing processes. WIFC framework was designed with several goals in mind: 1) Support a diverse set of new or legacy siting analysis applications on multiple OS platforms (Windows, Linux, Unix, etc.); 2) Modular to allow quick adaptation or plug-in of engineering or business tools that interact with data already in the system; 3) Offer the ability of quickly prototyping to add new functionalities; 4) Allow rapid deployment of new code and applications to the system.

WIFC was developed using proprietary algorithms, custom software, and commercial off-the-shelf tools. For its web frontend and some backend processes, Enterprise Accessible Software Applications (EASA) system was used. EASA allowed for rapid web user interface development, automated job queuing, and seamless connectivity to computational servers. Custom algorithms were developed to generate inputs, post-process outputs of engineering programs, and execute workflow between processes. This system has resulted in a siting process that provides unified, consistent and reproducible results, a common knowledge base that allows data validation and verification of past analyses, enhanced productivity and a quicker turn-around time on analyses.

Recently added functionalities to this system include component level analysis, Automated Configuration Release, and Turbine Allocation. Component level analysis enables users to run suitability analysis on critical components of a wind turbine model, like towers or foundations. Automated Configuration Release is a process to enable users to test the validity of tools for new wind turbine models prior to their formal release and usage in wind farm siting. Turbine Allocation allows engineers to rapidly vary turbine models on a turbine by turbine basis to calculate the predicted energy for a proposed wind farm site. Turbine Allocation runs on the Predix platform by GE Digital to utilize unified web widget design and backend microservices. Our future work aims at enriching this system with better parallel computation to reduce latency, more scalable architecture and load-balancing, and the possibility of employing machine learning techniques for automated optimization and selection of wind turbine models.

Document Details

AuthorKornfein. M
Date 6th June 2018
OrganisationGE Global Research & Development


Back to Search Results