Client Background
The client is a company operating within the energy manufacturing industry. Due to a Non-Disclosure Agreement (NDA), further details regarding the specific company cannot be disclosed.
Challenges
Despite regular maintenance, equipment used in energy production remains prone to unexpected failures and breakdowns. When equipment fails unexpectedly, monthly production targets cannot be met, leading to operational disruptions and financial losses.
Typically, failures stem from a single faulty component rather than the entire machine. However, the non-functional part renders the entire piece of equipment inoperable until repairs are completed, and the system is restored to its regular operating mode.
Our Solution
To tackle these challenges while ensuring security, we created an on-premise solution utilizing SQL, Python, Machine Learning, and IoT. Telemeters collect real-time equipment data, transmitted via mobile networks to the SCADA system.
The solution comprises two core components:
Anomaly Detection
Each device operates within a predefined regular operating mode, with expected telemetry readings such as voltage, resistance, and power.
The system detects anomalies as early warnings of potential issues, distinguishing between defects and external factors (e.g., power outages) to minimize false alarms.
Failure Prediction
Anomalies are analyzed to predict potential failures within up to 4 weeks without preventive action.
An ML model estimates the likelihood of specific equipment failures, processing data via SCADA in near-real-time.
The solution enables preemptive recovery planning, minimizing downtime and avoiding extra costs.
Outcomes
Cost Savings
By avoiding failures, missed production targets, and reducing procurement costs, we saved an estimated €300,000 per incident. A retrospective analysis further uncovered an additional €250,000 in savings by identifying previously undetected anomalies.
Improved Operational Insights
A better understanding of regular operating parameters led to adjustments that significantly enhanced equipment performance and reliability. This improved insight allowed for proactive measures, ensuring smoother operations.
Optimized Production Planning
Failure predictions enabled more efficient production planning, minimizing disruptions and allowing for effective management of reduced production. This strategic foresight ensured a smoother workflow and optimized resource allocation.
At a Glance
Predictive Maintenance: On-premise solution using IoT, Machine Learning, and SCADA predicts failures 120 days in advance, minimizing downtime and disruptions.
Cost Efficiency: Saved €300,000 per incident and an additional €250,000 through anomaly detection and analysis.
Operational Improvements: Enhanced equipment performance, reliability, and production planning with better insights and preemptive recovery strategies.
