Mass manufactured products are used by both consumers (public) and industrial users for work activities. Latent defects introduced at the design or production phase are pernicious since many users can be adversely affected by a
single fault. If these issues result in ‘safe’ failures, then user operability issues arise. However, ‘dangerous’ failures can result in harm to many people and product recall initiatives are costly, time consuming and rarely result in the recall of a substantial proportion of the affected products. Within theUK there is increasing scrutiny and interest in product safety matters. There are also specific concerns in some industry sectors around counterfeit products.
This project aims to carry out a feasibility study to determine whether the Health & Safety Executive’s (HSE’s) product safety data could be systematised to provide insights into the root causes of product safety issues. HSE’s regulatory focus is upon the product safety of industrial products used for work, but this project also aims to assess whether there is value in combining intelligence from both consumer and industrial applications.
HSE holds a large regulatory dataset which includes the outcomes from many investigations into product safety issues. These often arise from complaints from equipment users and following accidents to people at work using industrial products. Historical analysis of the multiple factors which lead to an individual product safety issue has been challenging due to the volume of intelligence stored as free text including in document attachments within the database established.
More Information
Industry drivers
Manufacturers of quality products realise the importance of brand reputation. Product safety issues, including multiple accidents and poorly managed product recalls can heavily damage reputation, sometimes irreparably. At the same time purchasers of industrial products, particularly high-value or high-volume items, need to ensure that they procure safe products. Product safety issues attract international media coverage and well-established brands are especially vulnerable to such scrutiny. At the same time, sharing of product safety learning is challenging, particularly given the commercial confidentiality around product design
Aims and objectives
The study aims to deploy text mining techniques, alongside suitable subject matter experts to extract useful categorised intelligence which will ultimately allow stakeholders to search through anonymised data whilst filtering by, for example:
- Type of product
- Country of manufacture
- Age of product at time of failure.
- Failure mode(s)
- Initial technical cause of failure.
- Root causes existing in design or manufacture
- Human factors issues and user foreseeable misuse
Within later parts of the project more detailed feedback will be sought from potential users of the intelligence resource to guide the future focus of the project.
Key Benefits
If the feasibility study suggests that the data is conducive to producing valuable insights, then in later projects these insights could be used to:
Empower ‘intelligent customers’ – i.e. those procuring industrial products as to how to better select products and seek assurance they are safe
Improve standards of technical design and manufacture leading to less harm to the public and better economic progress around the globe
Provide designers and manufacturers with insight and case studies to improve their design and manufacturing assurance processes to avoid costly product safety issues
Improve product safety performance for industrial products
Related Content