Optimaestro: 6 phases

• Process and business understanding

  1. Process and business understanding

  Which Key Performance Indicators (KPIs) measure my plant performances such as yield, energy efficiency, throughput, product quality, equipment availability?

  The first phase of any analytics project allows our team to understand your process and business challenges.

• Flash analysis

  2. Flash analysis

  What are the potential savings after implementing a pilot project?

  The Flash Analysis is a preliminary analysis on historical production data to quickly quantify the potential improvements and main sources of variability that might explain low performances.

  The goal is to refine the expected savings, identify possible roadblocks and prepare the brainstorming sessions with plant staff.

• Workshop with operators

  3. Workshop with operators

  How can we involve operators to ensure a successful project?

  At this stage, brainstorming workshops are organized with production and maintenance teams (operators, engineers, managers). These sessions are designed to pinpoint root causes underlying a particular performance problem and to understand operating constraints.

  Experience has shown that this phase plays a key role in the successful implementation and engagement of plant staff to trust AI based decision support tools.

• Advanced Analytics and Modeling

  4. Advanced Analytics and Modeling

  Which parameters are key to maintain optimal performance?

  In close collaboration with the production team and thanks to advanced analytics tools based on artificial intelligence, PEPITe’s team:

  - explores past operations variability and identify key parameters

  - understands conditions leading to best performances

  - predicts best performance depending on plant conditions

  - prescribes a range of manipulable parameters to maintain optimal performance

• Implementation

  5. Implementation

  What tools, dashboards or reports are required to facilitate and support operators in maintaining optimal performance?

  Models built with machine learning feed in real-time dashboards and reports to ensure the right information gets to the right person at the right time. The decision support tools are adapted specifically for each client (dashboards, daily/weekly/monthly reports, etc.) and typically use existing technology solutions already operational in your plant.

• Training and review

  6. Training and review

  How can I ensure long-term follow-up and sustainable results?

  Following implementation, plant staff is trained to take the appropriate actions based on live dashboards and reports. A follow-up review period ensures that models are performing correctly and that the reports and dashboards are understood and well used.