Consultings

Scope:

Traditional control systems are reactive, and do not include capabilities of prediction to prevent defects that could improve the control system behaviour / performance. On the contrary, near real-time control system can correct the deviation of the process, by modifying process parameters based on predictions from the combination of simulation and AI in order to avoid defective parts. 

Approach:

Process Control Manager (PCM) is composed of:

• a generic functionality of “Monitoring process parameters” based on OPC UA
• a custom functionality for “Detection of defective parts”, fully dependent on the specific manufacturing process, based on specific customized “Reduced Order Model” (ROM) and “Data Driven Model" (DDM) which are invoked from the PCM during the manufacturing process of the parts

PCM, along with ROM and DDM, leads to major improvements (innovation):

• near real-time predictive control
• incorporation of simulation capabilities in production
• data-driven and hybrid approaches (physical models, machine learning, real sensor data)

The benefits of ROM and DDM-based process control are:

• higher product quality
• increased production throughput
• lower operational costs
• reduced dependence in trial-and-error tuning

Differences with current solutions:

• ROM and DDM-based predictive control supports prevention of defects in near real time while the traditional process control is reactive (corrects after defect occurs)
• it has adaptive capabilities while traditional solutions are rigid (fixed PID or rule-based control)

Outcome:

The customer obtains:

• reduced defects rates
• more reliable and adaptive process control
• improved process visibility and decision-making