Investigation of demand pattern and identification of markets niches for products family: Wind turbines for offshore wind farms

A Pattern Classification Based Market Demand Analysis (PC-MDA) technique is developed to investigate the demand pattern and identify the market niches for a family of products. The global market in the case of offshore wind turbines is analyzed. In a family of products, individual product types exhibit certain unique features; these features have significant influence on the market demand of each product. K-means clustering method is used, in conjunction with the Elbow Criterion, to identify the clusters among the different products available in the market. The Batch Perceptron algorithm and the Minimum Squared Error algorithm are used to determine the linear discriminant boundaries (hyperplanes) among clusters, followed by a sensitivity analysis of product features. PC-MDA technique is performed on the offshore wind turbines by using six critical features: (i) rated power, (ii) rotor diameter, (iii) rated wind speed, (iv) cut-in wind speed, (v) cut-out wind speed, and (vi) hub height. The subsequent sensitivity analysis allowed us to identify the cut-in wind speed as a non-classifying product feature. An investigation into the evolution of wind turbines over years is also performed to explore and predict the market requirements for the upcoming line of products (turbines). © 2018, Asian Association for Agricultural Engineering. All rights reserved.