Discrete manufacturing: model-based process optimization and control

Comments (0) Blog

The ways in which process manufacturing and discrete manufacturing are respectively controlled differ a lot. Dealing with discrete manufacturing, STREAM-0D requires the implementation of sophisticated analysis within the process. This activity is managed by STAM, which in this article for the Bits Of STREAM-0D inititative, outlines the features of such approach.

STAM is an Italian private engineering company whose mission is to provide engineering services to industries, specializing in design and manufacturing of innovative mechanical systems, based on conventional and non-conventional robotics and mechatronics.

The main role of STAM in STREAM-0D is to provide the model-based control system (interface, programmed algorithms, software, hardware, connections) that will be tested and checked under a HIL simulated environment. STAM also participates in the design of some of the in-line measurement stations, support the development of data-based models and coordinate the implementation of the system in one of the end users’ production lines.

Discrete manufacturing v. Process manufacturing

Discrete manufacturing is an industrial term for the manufacturing and assembly of finished products that are distinct items (e.g. cars, smartphones and mechanical components). It differs from process manufacturing, where the product is created as a controlled mixture of raw ingredients and the final product cannot be broken down to its basic components.

This difference is also visible in the way the two types of processes are controlled. In process manufacturing sophisticated process tracking and real-time control are implemented to keep operations running at peak efficiency.

In discrete manufacturing, where feedback control is often difficult to implement, since each component has its own specific characteristics and the result of a critical operation might be visible only a number of operations later – too late for an effective correction –, manufacturers make use of after-the-fact statistical analysis to continuously improve their processes.

STREAM-0D approach

The STREAM-0D approach, taking advantage of a combination of both reduced order physic-based and data-driven models, allows for real-time control of all manufacturing processes, resulting in a drastic reduction of defected products. STREAM-0D control module run the models at key points of the manufacturing process, producing a real-time prediction of the process outcome that is compared against the product KPIs, and through advanced optimization techniques is able to dynamically adjust the process parameters in order to prevent defect generation rather than correcting them once they are measured.

Through the combination of precise models predictions and real-time process parameters optimization, STREAM-0D strategy can represent an innovative paradigm for process control in discrete manufacturing, bringing benefits to potentially all industrial sectors.


This article by STAM is part of the Bits of STREAM-0D series of contributions from the project’s consortium partners. Periodically, on STREAM-0D channels, a different topic will be given prominence, along with videos featuring members involved in the project.

Check Bits of STREAM-0D’s previous contributions by STREAM-0D partners:

Follow STREAM-0D on: