Our approach for the design of Predictive Maintenance models is in two steps: 

1) Automatic Feature Generation from collected data

2) Design of predictive maintenance models using supervised and unsupervised learning

We allow multiple Implementation Options: On your IoT (proprietary or standard) platform, on our Cloud, or embedded within the equipment.

Predictive Maintenance Model Process

Automatic Features Generation


DiagSign : A Unique Innovation in Automatic Feature Generation for Time Series

DiagSign automatically generates discriminant features out of time series
  • Whatever the nature of the time signal input (electric signal, temperature, humidity, vibration, pressure, …or a combination of these signals)
  • Without the need for a subject matter expert at this stage
  • DiagSign features:
    • Are infinitely rich and discriminant
    • Lead to more performant predictive models with AI and Machine Learning algorithms


Without DiagSign:

discriminant features

With DiagSign:

discriminant features

Predictive Maintenance Solutions

Generic and Bespoke Artificial Intelligence Models

Based on Amiral Technologies innovative models

  • Optimized for your specific equipment
    • Prediction of defects
    • Signs of aging
    • Remaining Useful Life (RUL)
  • Supervised or un-supervised learning depending on the availability of historical data.
  • Intelligent alert thresholds matching your own tolerance to the ratio of prediction rate/false alert rate.

Use cases

Remaining Useful Life Prediction 

Generic Model and its application for Turbofans

Defects and Aging Prediction 

Generic Model and its application for Railway Point Systems and Industrial Printers 

Defects and Aging Prediction

Generic Model and its application for  Wind Turbines and Induction Motors

Energy Supervision Solutions

  • Disaggregation of electric current at the level of the main counter to derive the consumption of each equipment.
  • Detection of equipment based on switch-on signal.

Use cases

Supervision of Industrial Energy Consumption (NIALM)

Non Intrusive Appliance Load Monitoring: Disaggregation of energy consumption to derive individual equipment consumption.

Supervision of Home Appliance

Detection of Appliance in a home environment based on switch-on signals


Our Integration Options

  • In your IoT (standard or proprietary) platform
  • Hosted on Amiral Technologies' cloud
  • Embedded in the equipment
Close Menu