Case Study: Improving Reliability of Fermenter in Pharma Plant

Plant Maintenance has progressed from Reactive, through Preventive and on to Predictive methods. This is a disruptive development driven by the confluence of several factors – miniaturized sensors; ubiquitous availability of wireless bandwidth and cloud computing capacity. Remote monitoring of equipment 24X7 is possible as a result. Combined with deep understanding of the mechanics, it enables machines to run forever within operating conditions!

AccuPredict’s Approach to Predictive Maintenance

AccuPredict has nearly four decades of experience in providing Condition Based Monitoring of equipment with a long-term stable client base in Asia, Europe, the Middle East & Africa. We have upgraded the offering using IoT based sensors and ML algorithms to improve the efficiency of our Engineers. We approach Predictive Maintenance from Machine Fundamentals rather than pure Pattern Matching based solutions.

Here’s how customers benefit from our differentiated approach:
1. Accelerate rollout of Predictive Maintenance: We start providing corrective actions needed and MTBF from Day 1. Our trials do not take more than a month as we do not need to build and train
machine models.

2. Lead time to failure in terms of months rather than days: We do this by tracking vibration trends at specific frequencies of each component rather than an overall pattern. Customers benefit by not having to disrupt manufacturing schedules. They are also able to reduce inventory of spare parts. Most importantly by keeping equipment and components within their operating parameters we enable customers to extend machine & component life. This is a critical ingredient of their ESG strategy.

3. We use our deep experience in vibration analysis to help customers improve their equipment designs: We use our understanding of equipment vibrations to identify changes needed to machine design to eliminate causes of resonance leading to component failure.

Our Pharmaceutical customer’s Fermenter was developing excessive vibration. We identified the issue was with bearing ‘C’ out of the twelve bearings we were monitoring. The root cause was a shaft misalignment. This would have led to a failure in 9 months. We advised the engineering team to complete the alignment at their next planned shutdown. This led to a reduction of the vibration to normal levels and avoided an unplanned shutdown.

Milind Yedkar
Milind is the co-Founder & CEO of Singapore based startup AccuPredict Services Pte Ltd ( that provides Predictive Maintenance as a service to manufacturers Worldwide