#	Pagina
attuale pagina	/open-h2020/projects/211479/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - CARL-PdM (Next Generation Holistic Predictive Maintenance Software)

Teaser

Summary

\"\"\"Industry 4.0 preaches a complete revolution of industrial process and promises huge efficiency gains by a complete virtualization of the factory, numerical design tools, automation of the logistics and the routing of the parts, smart machines, 3D printing, cyber-physical systems, predictive maintenance and control of the whole factory by an intelligent system.
In the next 10 years, industry 4.0 is expected to change the way we operate our factories and to create 1250 Billion â‚¬ of additional value added in Europe.
Also , according to ARC Advisory Group, the predictive maintenance market is estimated to grow from 1,404.3Mâ‚¬ in 2016 to 4,904.0Mâ‚¬ by 2021.
CARL-PdM is a innovative IIoT data powered predictive maintenance platform encompass the core of \"\"Industry 4.0\"\" with a new maintenance paradigm : maintenance is a production function whose aim should be to optimize production output and quality.
We will leverage the IoT revolution to achieve these goal.
\"

Work performed

\"This software solution, CARL-PdM, provides many core capabilities in industrial scenarios, including edge analytics who provide a way to pre-process the data so that only the pertinent information is sent to the predictive layer (Auto Classification and Machine learning).
The predictive layer will categorize data into abstract class which represent technical assets behavior. It is a reliable and reproducible approach.
Competitive advantages:
- Reduce failure by 50%, maintenance cost by 30%, production stops by 70%, energetic consumption by 20%, Time To Repair by 30%
- Increase production flexibility
- System agnostic to machines
- Machine-learning algorithm that compares the fault prediction and sensor data with historical data, predicting best maintenance activity regarding to production and quality objectivesWith this digital platform which is capable of analyzing real time data provided by communicating sensors, technical services will be able to improve equipment operation (equipment configuration, predictive maintenance, prediction of failure, etc.) thanks to statistical analysis and generation of predictive models.

The development of this prototype dedicated to technical services places CARL Software among the pioneers of tools dedicated to the management of equipment for the industry of the future, Smart Building and Smart Cities.

CARL Software focuses its R&D on studying new maintenance uses created by the Internet of Things and by artificial intelligence to integrate them into its CMMS.
In the short term, we plans to connect its CARL Source CMMS to all types of IoT to have necessary planning data for an optimal triggering of preventive maintenance and optimization of equipment operation.
In the medium term, the equipment digital data will be processed and analyzed by the platform which will generate behavioral and predictive models that will enrich the characteristics of the equipment managed on CARL Source and create a \"\"digital twin\"\" of the equipment.
\"

Final results

\"We experiment the prototype of CARL-PdM (Predictive maintenance Platform powered by IIOT) with technical equipments and building energy management processes in concrete situations.
It is our own headquarters, a building of 2500 M2, which served as a test base.
It includes:
- An air handling unit for the production of hot / cold air,
- A low energy data center (with free cooling and free heating system) with two air conditioning systems
- Storage and energy transformation devices etc.

The platform protoype developed in this project allowed data analysis from ours equipment and to measure the energy performance of heating and air-conditioning systems of our building and Data Center.
The project has provided the opportunity to experiment with different technologies of communicating sensors (LoraWAN, SigFox, Ethernet, WIFI, Bluetooth ...), thanks to the different environmental sensors installed in the building, as well as machine learning techniques based on the data analysis and modeling of systems behavior.

During the first months of experimentation, the platform recorded and analyzed over a million of measurements / day.
In a few weeks, the platform has created \"\"digital twins\"\" of the free cooling / free heating system, identified behavioral models, observed singular modes of operation and highlighted a number of anomalies in the building energy management system that were so far undetectable in the supervision system.
CARL Software technical services have already been able to optimize the operation of the data center cooling system based on recommendations provided by the platform, and also saved 15% of electrical powers and reduced our energy footprint and improved our ecological balance sheet !
Today they rely on forecasting guidelines to anticipate the adjustments to make on these systems.
\"