Client Background
Our client is a leader in creating energy technologies and creates a third of the world’s electricity. Its technology equips 90% of power transmission utilities worldwide.
Business Objectives
- To create a Recommendation Engine to assist Outage Planners and Field Service Teams in their efforts to more effectively guide and inform maintenance interactions with their clients.
- To source data from a variety of internal (example; outage reports, equipment data, sensor data, technical lists, resource lists, scope lists, operational checklists, tribal knowledge) and possibly external sources to create the Outage Risk Assessment Platform. This data set will be leveraged to create a recommendation engine that flags and prioritizes risks at the client level and equipment level.
Solution
Solution development journey was divided into following phases:
PHASE 1: Data Collection, Integration and Transformation
During this phase our consultants worked with key stakeholders and subject matter experts in the client’s business to understand and document the current baseline process Outage Planners take to plan for scheduled maintenance sessions with clients. This included understanding data sources used, link information to decisions made and document how the Outage Planners use the information to prioritize risk issues for their client maintenance interactions today. Our team then collected and assembled information from various data sources provided by client.
PHASE 2: Data Analysis and Model Development
During Phase 2, our team conducted basic data exploration, defined key events relevant to the planning process and communicated preliminary findings. Following the delivery of the initial set of insights, we started the modelling process. The purpose of the recommendation engine was to help Outage Planners identify and prioritize key risks to inform the agenda of their scheduled maintenance reviews. The range of recommendations included direct step actions the Outage Planner should take to address obvious issues to more general risks that might require further drill down to identify root cause issues. Our team worked with the client to identify several Outage Planners for a beta test of the recommendation engine. In early testing, the recommendation engine was able to identify 50%-60% of unplanned outages and impacting factors across 4 categories. The results of the beta enabled us to modify the recommendation engine to ensure greater relevance of the product.
PHASE 3: System Implementation
Once the model was finalized, we implemented the algorithm and the report that features the model output/recommendations on the client’s internal system. Outage Planners had access to the information/report and could lift the prioritized risk related issues and use the information in their mitigation plans as they prepared for maintenance meetings with their clients.
Outcome
From engineer notes that were stored as PDF, our recommendation engine discovered reasons for maintenance delays and machine breakdown. These insights were used to identify best practices for outage planning, scheduling maintenance, procuring repair parts.
