A major oil and gas transmission plant, generating an annual revenue of $10 billion, faced frequent reliability issues with its compressor pumps, adversely affecting the plant’s output to downstream refineries.

  • Although equipment anomalies were flagged with their in-house legacy software (without self-learning AI), it lacked the ability to take domain knowledge into account. This resulted in a heavy reliance on subject matter expert (SME) consultants, and their corresponding RCA (Root Cause Analysis) delays led to increased downtime.
  • The use of multiple models for tracking the same equipment made it difficult to monitor anomalies accurately, leading to numerous false positives.
  • Analyzing correlations between different operating conditions in the compressors revealed that the sensor data deviated from the expected behaviour.
  • The Predictive Maintenance (PdM) solution previously implemented at the plant triggered many false alarms and necessitated manual training of AI models, reducing plant operations’ efficiency.

According to Deloitte, Fortune Global 500 manufacturing firms lose 3.3 million production hours annually due to machine failures—a staggering $864B loss despite investments in Predictive Maintenance [PdM] solutions. These challenges directly impact Asset Utilization, making it difficult for senior executives to achieve annual Operational Excellence [OEx] goals.

 

  

Get a copy of the Customer Story: