Our Approach

Challenges for today's industry

Process monitoring and analysis may provide partial answers, but...

“On average, between 60% and 73% of all data within an enterprise goes unused for analytics” - Forrester

The huge volume of data is too big, or moves too fast or exceeds current capacity

Our Solution

Traditional monitoring provides insufficient information and a lot of data remains unused. Performance for Assets offers a solution that includes the following steps:

  • 1. Detect: Identify the anomaly
  • 2. Diagnose: Why does the anomaly happen?
  • 3. Prognose: How does the anomaly evolve?
  • 4. Extract intelligence and turn it into actions (short to long term)

Our combination of process/asset knowledge and analytics through dedicated hybrid models (physical & data driven) is the key to success!


How do we make use of big data?

Raw data used

Many companies have an online monitoring system in place to track machine behavior and to stop when pre-set alarm values are reached. Some companies also use physical models basic on thermodynamics, mechanical dynamics or meteorology.


Live Situation | Online Monitoring & Alarm Values

Alarms based on norms or specifications triggered by simple online measurement filtered to avoid wrong detections + offline analysis by specialists to steer the search of potential causes

Examples : OEM generic power curves, OEM generic flow rates, linear analytics models based on measured parameters, vibration analysis norms, …

Ideal Situation | Physical Models

Fine-tuned physical model developed off-line for each machine in its environment and working conditions

Examples : Thermodynamic balanced models, mechanical dynamic models, meteorological models, …

Past Situation | Analysis of Historic Data

Pure data driven model based on advanced analytics and steered by process/asset specialist

Examples : simulation of optimal behaviour based on historical data and machine condition, detection of parameters combination which leads to a drift

Performance for Assets' added value

Hybrid Model: Live + Ideal + Past

Performance for Assets has developed a unique hybrid model that combines Online Monitoring with Physical Models and Historic Data. The hybrid model improves machine behaviour knowledge, increases model robustness and identifies causes more precisely. Its main advantages over traditional online monitoring are:

Early Warning

Detect degradation or performance drift at an early stage to limit impact

Auto Learning

Algorithm learns root cause of failure and adapts itself automatically

Fleet Benefit

Knowledge of one machine’s behavior can be applied to multiple machines to avoid failures

Contact us

For questions concerning our services or pricing.