Transform Your Machine Maintenance Strategy

Our groundbreaking solution combines the precision of a traditional engineering-based approach with data-driven AI models to identify the impact of wear and tear on complex machines and processes.

This approach is based on a sound mathematical foundation that creates repeatable, and provably accurate results.

Unlock the future with our revolutionary predictive maintenance solution that combines the best of both world: Engineering and AI Models

In the rapidly evolving world of industrial maintenance, the key to staying ahead lies in embracing innovation

ENGINEERING-BASED MODELS DATA-DRIVEN MODELS
Precision for Well-understood Machine Parts and Components:
+ High transparency and effectiveness, even with scarce data
+ Excels in situations where historical data on component failures is limited

Challenges:
— Requires extensive experimentation
— Demands intensive labor for development
— Only feasible for small machine parts or components: Bearings, gears, etc.
Cutting-edge AI at Your Service:
+ Based on processing vast amounts of data to forecast maintenance needs.

Challenges:
— No proof of accuracy due to their "black-box" nature

Our innovative solution: Inverse PID

We use the power of inverse-PID computation to detect changes to controlled systems. PID controllers are ubiquitous. They regulate the behavior of a controlled system using a well-understood mathematical model. 

By reverse-analysis of apparent PID gains,  our approach allows us to identify physical changes to the controlled system and infer from those changes wear and tear.

ADVANTAGES OF OUR HYBRID APPROACH:

  • Predict component failures and inefficiencies by identifying anomalies in PID gains, using Federated Learning for central storage and comparison

  • Ensures utmost data privacy and enables the comparison of hundreds of machines across locations

Federated learning application

 PID gains do not convey machine processing information, thereby safeguarding customer information.