Skip to main content

Knowledge-Based Real-Time Diagnosis and Repair for a Complete Robotised Handling and Storage System

Objective

The objective of this project is to design and implement a new generation of knowledge-based diagnostic tools that continuously acquire data from programmable logic controller (PLC) systems, identify possible malfunctions, search for their causes and suggest corrective actions. The main innovation will lie in the use of diagnostic models that integrate the classical heuristic and associative approaches.
The objective of this project is to design and implement a new generation of knowledge based diagnostic tools that continuously acquire data from programmable logic controller (PLC) systems, identify possible malfunctions, search for their causes and suggest corrective actions. The main innovation will lie in the use of diagnostic models that integrate the classical heuristic and associative approaches.
The project is investigating the use of deep models that reproduce the functional and operational logic embedded in the ladder logic of a PLC. A study has shown that these models contain design and engineering knowledge in a more usable form compared to that held in the minds of the designers themselves. 2 software tools will be developed at a prototype stage: the specifications and a demonstrator of an automatic modelling module, which will be a component of a possible PLC programming support environment (with specification and code generation support facilities) and be able to send the models produced to the diagnostic system; and a development environment for multimodel diagnostic systems, implementing general structures, which will be an extension of an existing diagnostic system.
The tools will be tested on a synthetic fibre spinning plant. A knowledge based multimodel diagnostic tool (MMDT) has been installed and is running in the plant. Its main innovation is the extensive use of deep models of the plant, together with the hollow models generally used by previous diagnostic tools. The specification of a prototype knowledge acquisition tool (automatic modelling module (AMM)) has been developed to extract automatically the behavioural model of the plant from the step 5 code running the plant PLC. The development of the toolkits Toros and C Max, provided by Siemens, has allowed the implementation of the knowledge base and the inference mechanism.
The final results of the project will be the refinement of the MMDT prototype, the implementation of the AMM and th e refinement of Toros and C Max.
The project is investigating the use of deep models that reproduce the functional and operational logic embedded in the ladder logic of a PLC. A study has shown that these models contain design and engineering knowledge in a more usable form compared to that held in the minds of the designers themselves.

Two software tools will be developed at a prototype stage:

- the specifications and a demonstrator of an automatic modelling module, which will be a component of a possible PLC programming support environment (with specification and code-generation support facilities) and be able to send the models produced to the diagnostic system
- a development environment for multi-model diagnostic systems, implementing general structures, which will be an extension of an existing diagnostic system.

The tools will be tested on a synthetic fibre spinning plant.

Coordinator

Snia Bpd Fiat Group
Address
Via Borgonuovo 14
20121 Milano
Italy

Participants (3)

BMT Fluid Mechanics Ltd
United Kingdom
Address
1 Waldegrave Road
TW11 8LZ Teddington
Pirelli SpA
Italy
Address
Viale Sarca 222
20126 Milano
Siemens AG
Germany
Address
Paul-gossen-straße 100
91052 Erlangen