Using Machine Learning and Google street View for Condition Assessment of Wood Poles

点击翻译按钮获取中文摘要

In today power systems, extensive progress is being made in the energy transition from centralized fossilfuel generation to renewable distributed generation sources namely wind and solar. However, this new development is limited by numerous existing challenges in the distribution systems including the aging infrastructure and policy issues. As a result, the bulk share of capital spending is directed towards renewal project hence leaving little or no room to invest in smart grid initiatives. Aging infrastructure employed on power distribution systems requires major capital investments into renewal and maintenance of aged assets to ascertain that the distribution systems continue to operate with increased level of safety and reliability. These capital investments are primarily driven by increased public scrutiny from interveners and energy regulators. The wood poles represent major fraction of asset total replacement value in any distribution system hence the condition assessment of wood poles is of great significance. To date, condition assessment of wood poles is performed manually by inspecting each pole which is a labor-intensive inspection data gathering process. The key objective of this project is to apply machine learning technique and google street view to estimate the condition of wood poles used in power distribution systems. The proposed smart approach would increase safety for employees and end customers while ensuring reliable power delivery to the community.