Industry 4.0 Case Study: Automated Inspection Software offers £24 million saving
How Leveraging Big Data Technology can help manufacturers obtain the Ultimate Smart Factory.
Inviso Technologies have recently completed a project for a global gas turbine engine manufacturer to improve their inspection analysis and concession creation capability.
The project involved identifying areas for improvement within a manufacturing facility producing complex aerospace vanes. A significant rise in load has led to increased inventory and a need to eliminate bottlenecks.
A shop floor walk identified work in progress (WIP) queuing at inspection areas so a Value Stream Map (VSM) was created to analyse the work flow around the inspection and concession process. This identified that not only was the process taking an excessive 7 days but there were also multiple opportunities for potential non-conformance in the defect reporting due to the manual nature of the method. The frequencies of mistakes were also contributing to the inspection lead time due to the process having to be repeated to correct the errors, in some instances this loop was being repeated 5 times. This was in addition to the 7 days that the process was taking and was delaying deliveries out of the factory.
The proposal involved utilising big data technology to quickly analyse the inspection output from the Co-ordinate Measuring Machines (CMM) and report any deviations. It also applied the same logic that the quality procedure and inspectors use to assess them. It would also provide Feature analysis reports and Statistical macros for the Manufacturing Engineers so that they could access Statistical Process Control (SPC) data necessary to make data driven process decisions.
The platform needed to be accessible by everyone but restrictive IT policies and the need to keep the software regularly updated meant that the system couldn’t be provided as a standalone software installation, so a Platform as a Service model (PaaS) was decided upon.
A demonstration system that used email and a cloud based containerised Python system to process the data was created. This allowed all users to use the latest version of the software and using container’s meant that a continuous deployment approach was employed with updates taking minutes to deploy. As access was required 24 hours a day, 7 days a week disruption had to be kept to a minimum.
The inspectors and Manufacturing Engineers had to attach the CMM output files to an email and send it to a dedicated secure inbox. The setup of the system ensured that the data was encrypted both in flight and at rest and once the data was processed it was securely erased so that there was no risk of a data breach. It was also critical that the data did not leave the European Union (EU) due to export control limitations, so the platform and email was restricted to UK based servers.
The subject line of the email could also be used to provide different options for the output reports and analysis i.e. an SPC capability report instead of the standard six pack analysis. The automated response detailed everything that had been sent by unique serial number and grouped it by batch for easy concession generation. The format for the reports was produced in Microsoft Excel format so that all users could view the data and it was easily interchangeable with other software.
The average response time for the reports was less than 1 minute with the actual computing processor time taking only milliseconds (the remaining time was due to getting through the corporate email system). During the development and trial period the system analysed and converted in excess of 39 million lines of data with a total computing time of 68 minutes. This was a reduction in time for the Value Stream Map of 99.99% and the data’s Right First Time was 100% instead of an estimated 46%.
The initial calculations from the demonstration software showed that there was an opportunity to save £2.7 million in the first year and then an ongoing saving of at least £0.5 million a year for the following 9 years (The estimated life of the system), giving a total saving of £7.2 million. This was based on an initial inventory reduction once the system was initialised and there was also a reduction in data reporting for the Manufacturing Engineers. They could now get their data reports in near real-time instead of several days to produce Statistical Process Control reports and run charts to analyse processes. This would also free up capacity in the inspection area’s allowing the facility to eliminate the bottleneck and meet the required load.
Based on this, a pilot project was launched to validate 25% of the load going through the manufacturing facility. This demonstrated that the actual system would perform better than anticipated providing in excess of £50k savings per month. This would result in a saving of £24 million based on the current facility load (which was forecast to increase further) over the life of the system.
This case study clearly demonstrates the power of leveraging the latest technologies with manufacturing processes to realise significant benefits in quality, cost and delivery performance for larger manufacturers which can also be beneficial when applied to SME’s.
About Manufactory
Manufactory is a Blockchain technology company combining artificial intelligence (AI), smart contracts and the Internet of Things (IoT) into an advanced de-centralised platform for the manufacturing industry.
White paper and more information coming soon.
Intelligent Scalable Connected Manufacturing.
Join us on Telegram https://t.me/joinchat/H7ilAkk1Km-RlN90VJegeQ
