/ Case Studies

Operational Excellence

Reduce operating costs for predictive maintenance using Kobai’s semantic graph integrated with Databricks.

Helping energy companies with better resource planning, fewermandatory onsite visits, increasing overallproductivity ofpartners and technical services teams.

About the Client

A leading Energy Company and one ofthe major generators of renewableenergy the US. The company has been
recognized for its commitment tosustainability, corporate responsibility,and operational excellence. Thecompany's focus on meetingsustainability goals can have a hugeimpact over the energy industry.

Figure shows siloed data flowing to a cloud of knowledge.

The Challenge

Frequent Outages: Hundreds of geographically dispersed sites experience unplanned equipment failures, material shortages, and planned maintenance, leading to downtime.
‍
‍Availability Loss: Outages (planned and unplanned) disrupt production, impacting overall output and equipment lifespan.

Performance Loss: Substandard materials, machine wear, and minor stoppages reduce efficiency and power generation.
‍
Resource Constraints: Difficulty in ensuring proper resource coverage for outages and maintenance activities, leading to inefficiencies and increased costs.

Data Challenges

Data Disparity: Diverse data sources exist, including real-time IoT data, historical equipment data, and resource/service information.
‍
Data Integration Needed: Techniques for integrating data were already being used, but as the scope and diversity of data increased, they simply took too long. Expertise from different teams was needed as part of the solution but translating that into a spec for a developer required many iterations.
‍
Predictive Maintenance: Leveraging historical and real-time data to predict and prevent equipment failures is essential for reducing downtime. While some progress was being made integrating machine data with asset details, widening the scope to resource management, supply chain planning and historical RCA added complexity that strained the current tools.
‍
Analytics at Scale: Combining these data sets is crucial for analyzing root causes, predicting outages, and optimizing resource allocation. Databricks' Data Lakehouse platform played a key role in enabling the integration and analysis of these diverse data sources.

Kobai's Approach

Kobai enabled the quick integration of diverse data sources (IoT, historical, asset, resource) into a single, flexible semantic model, in a collaborative and self-service way. By analyzing the data complexity within the context of asset relationships, we identified patterns and failure modes, leading to improved resource planning and better equipment maintenance strategies. These newly implemented data capabilities enabled the creation of simulations and "what-if" scenarios, empowering stakeholders to explore potential outcomes and make informed decisions.
‍
Moreover, optimization opportunities for asset upgrades and replacements improved overall performance and efficiency.

With Kobai’s intervention, the system now predicts maintenance needs based on advanced scenario planning and data analysis, allowing for proactive interventions and avoiding potential failures. Not just that, continuous monitoring of asset performance enables proactive maintenance decisions. Contextual linking of asset data with resource information for predicted outages meant - the right people being at the right place at the right time.

By linking asset data with procedures, regulations, and best practices, Kobai ensures informed maintenance planning that adheres to industry standards.

The Impact

The implemented semantic graph delivered significant benefits:

Proactive Maintenance & Reduced Downtime: Predictive maintenance strategies, enabled by Kobai, ensured proactive interventions, minimizing unplanned outages and downtime.
‍
Optimized Resource Allocation: Predicted outages were identified, and proper resource coverage was provided, eliminating inefficiencies and associated costs. Kobai provided supervisors with valuable insights enabling them to ensure proper resource coverage for sites where a predicted outage is likely to occur.
‍
Improved Efficiency & Reduced Losses: Reduced downtime and optimized resource allocation minimized the availability and performance losses, boosting overall efficiency.
‍
Ready for the Next Challenge: Now that the domain expertise needed to improve Predictive Maintenance has been captured in a visual semantic graph, it is primed for reuse and extension as needs evolve or new challenges pop up.

Build your AI Focused Semantic Data Fabric

Request Demo