https://ailabsinc.com/success-story/4998978b-7d50-11ee-94b2-7cd30...

https://ailabsinc.com/success-story/4998978b-7d50-11ee-94b2-7cd30ab5/predictive-maintenance-for-substation-using-minskytm-and-brainchips-akidatm-edge-ai

Predictive Maintenance For Substation Using
Minsky^TM And BrainChip’s Akida^TM(Edge Ai)

BACKGROUND/PROBLEM STATEMENT:

In general, Utility companies face significant challenges in effectively maintaining their critical electrical substations. They lack a proactive maintenance strategy, resulting in costly downtime and unplanned outages. The initial problem included limited historical data for analysis, making it challenging to predict maintenance needs accurately. Additionally, the absence of real-time monitoring capabilities hindered their ability to respond promptly to maintenance issues. The aim of this project is to implement predictive maintenance in a power substation using our Ai proprietary engine Minsky and Edge Ai using Akida platform

SOLUTION OVERVIEW:

To address the above challenges, the solution was implemented in two phases

Phase-1: On Minsky

After comprehensively analysing the data, we designed the solution using our Ai proprietary Engine Minsky for predictive maintenance on substations. We streamlined the real time data collection process from various sensors attached to substation equipment which includes voltage levels, temperature and other environmental conditions and were submitted to our Minsky predictive models. These predictive models utilized machine learning anomaly detection techniques to analyse historical data records which comprised of date, High Useful Load, High UseLess Load, Middle Useful Load, Middle Useless Load, Low Useful load, Low Useless load and real time data. These models can predict when specific equipment is likely to fail or required maintenance based on oil temperature

TYPICAL CHALLENGES:

• Increased downtime and unplanned outages
• Higher maintenance costs
• Accelerated asset degradation
• Safety risks to workers and the public
• Customer dissatisfaction and potential churn
• Inefficient resource allocation
• Lack of data-driven decision-making
• Missed cost savings opportunities
• Potential regulatory non-compliance
• Competitive disadvantage in the market

KEY BENEFITS (MINSKY):

• User-Friendly cloud-based AI platform
• Scalable across various domains/data.
• Provides you a list of % dependency features that can be used to optimize the outcomes.
• Ability to fine tune or optimize the models by trying different algorithms / prediction attributes
• Easy integration with other third-party solutions such as TABLEAU for data visualization

RESULTS:

• Reduced downtime and increase in cost savings
• Improved safety to workers
• Improved Asset life Span
• Better utilization of resources which increased productivity

Minsky^TM Results/Screenshots:

PHASE -2: DEPLOYMENT ON AKD-1000

These models are deployed on Linux servers in AKD-1000 environment for testing using client-provided Test Data with known results.

AKD^TM-1000 Results/Screenshots

Collectively, the above two phases provide a comprehensive solution that shifts the client from a reactive maintenance approach to a proactive one. The advanced hardware and AI-driven models empower real-time decision-making, reduce downtime, enhance equipment longevity, and ultimately optimize the client's substation maintenance operations.