Quantic Defect Analytics

Context

The growing global awareness of climate change has promoted a tighter focus in the manufacturing sector around sustainable work practices and environmentally friendly initiatives such as the so-called circular economy. Such initiatives are encouraging companies to seek out greater efficiencies around manufacturing yields, primarily to reduce wastage through tighter control of the manufacturing scrap rate.

For this reason, there is a significant demand for a generic and low-cost analytics solution to assist small to medium scale enterprises in identifying sources of manufacturing defects and facilitating the implementation of targeted quality improvements. This represents the target market for the Quantic defect analytics suite.

Idea

The goal of the Quantic project is to enable and empower companies to capture, understand and generate business value from their manufacturing process defect data.

The primary project objective is to research an analytics suite that takes defect data, such as from the Quantic defect data capturing app or from existing internal systems and carries out some generalised and advanced analytics to support quality improvements.

Through the use of these analytical techniques (e.g. artificial intelligence, machine learning) it is envisioned the project can highlight fundamental process defect information and identify the most critical signals that are showing anomalous trends or revealing details never observed before. It is further envisioned that this approach will lead to better manufacturing yield outcomes, increased digitisation and overall increased efficiency of member company operations.

Aim

Specific project objectives are as follows:

• To engage in research around an analytics suite that takes manufacturing process defect data and carries out a generalised analytics framework to support quality improvements.
• To devise and implement a standardised framework to allow for the manual collection of defect data (defect type, location, image) across various stages of a generic manufacturing process through a mobile device or similar.
• To enable a means for clear and easy visualisation of defect data across multiple end-users.
• To allow the user to identify and select relevant defect data and present back to the user the most likely cause of a certain defect location/type based on numerical analysis .
• To enable comparison of defect data by shift/machine/operator or other variables.
• To allow users to understand their scrap rate, what is impacting it the most, how to reduce it and enable the development of actionable insights.
• To provide a report on the state of the art in the analysis of numerical defect data to root cause quality problems.
• To provide a benchmarking report on the use of numerical defect data across multiple sectors of industry.

Challenge

The manufacturing scrap rate is a major concern for plant managers because of the waste produced and the resources that are consumed during the production of defective parts (i.e. energy, raw material and equipment uptime). While many companies collect scrap information, few if any collect this information in an automated manner with the intention of transforming it into intelligence. As a result, companies are often left reliant on paper-based data collection methods and/or minimal digital records, resulting in a limited view of the scrap rate without having visibility on why, how, when and where the scrappages occurred.

Quantic Consortium

The project consortium consists of:
• Grant Engineering
• Combilift
• McHale Engineering
• Decotek Automotive
• iQuTech

Quantic is an 18-month project funded through an Enterprise Ireland (EI) Innovation Partnership Programme.