Artificial Intelligence for Production Automation (AIPA) - Development of an Machine Learning Platform for Additive Manufacturing

生产自动化人工智能 (AIPA) - 开发增材制造机器学习平台

• 批准号: 77759
• 金额: $ 17.89万
• 依托单位: AUTONOMOUS MANUFACTURING LTD
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2020
• 资助国家: 英国
• 起止时间: 2020 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=77759
• 关键词: Artificial; Intelligence; Production; Automation; AIPA

The crisis caused by COVID-19 has significant economic impact on the UK economy. Over 25% of manufacturing staff have been furloughed and especially R&D Departments face a 17% reduction in R&D budget (IHS Markit, COVID-19 Automotive R&D Impact Survey). Over 50% of respondents postpone deployment of deployment of innovative technologies.Additive Manufacturing (AM) faces a threat due to impact of COVID-19, as innovative technologies are being delayed. As a direct response, AMFG needs to offer the market a solution that offers:1. Process automation for AM \[AMFG offers a solution for this\]2. Rapid deployment of the AMFG solution within a customers environment \[Focus of AIPA, **this innovation project**\]**Role of AMFG:**AM has been in the news, as companies like Dyson and JLR use AM capacity to build face masks and ventilators. AM will also support global supply and value chains, as global companies evaluate reshoring of manufacturing activities to the UK (NatWest, Manufacturing insights: the pros and cons of reshoring).AMFG is a world leader in Process Flow Automation for AM and we have successfully deployed our software into large customers such as ArcelorMittal, Henkel or BMW.Why AMFG needs this:AMFG sees an increase in interest in Additive Manufacturing from global companies. However, our potential customers suffer from significant cuts to R&D budgets and a backlog of work. In order to capture the current trend in the market, AMFG will develop a solution that allows to rapidly deploy our software within a customers environment.The current process is to go through a Proof Of Concept (POC) with potential customers, which typically takes 1 year to complete and costs around £50000\. This is due to an iterative process, where AMFG and customers engage in frequent meetings to elicit requirements, understand our customers environment and develop bespoke concepts. A POC is also a risk for AMFG and our customers, as it depends hugely on the experience of the Project Manager (on customer side). This is a significant drain on resources and results in sub-optimal utilisation of production lines long-term due to potential to miss important requirements early on in the development process (when these issues are easy to fix).**Why AIPA is innovative:**AMFG suggests to take this to the next level and completely automate the initial stage of the POC. AMFG will achieve this by using novel Machine Learning algorithms to determine the best solution for a customer. This will replace an iterative and risky process by capturing best practice in the market from our customers.As a result, AIPA will be an interactive questionnaire that automatically suggests requirements based on our customers self-evaluation, which will be used to develop the required features. Additionally, AIPA can be used as an ongoing self-assessment to confirm the development is on track and according to our customers requirements.AMFG expects that we will be able to triple the number of successfully completed POCs from currentyl 12 to 36 per year.

COVID-19引发的危机对英国经济产生重大经济影响。超过25%的制造业员工已被解雇，尤其是研发部门面临17%的研发预算削减（IHS Markit，COVID-19汽车研发影响调查）。超过50%的受访者推迟了创新技术的部署。由于COVID-19的影响，增材制造（AM）面临威胁，因为创新技术正在被推迟。作为一个直接的回应，AMFG需要为市场提供一个解决方案，提供：1。AM的过程自动化\[AMFG为此提供了解决方案\]2. AMFG解决方案在客户环境中的快速部署\[AIPA的焦点，** 本创新项目 **\]** AMFG的作用：**AM一直在新闻中，因为戴森和捷豹路虎等公司使用AM能力来制造面罩和呼吸器。随着全球公司评估制造活动向英国回流的情况，增材制造还将支持全球供应链和价值链（NatWest，Manufacturing insights：the props and cons of reshoring）。AMFG是增材制造流程自动化领域的全球领导者，我们已成功将我们的软件部署到安赛乐米塔尔（ArcelorMittal）、汉高（Henkel）或宝马（BMW）等大客户中。然而，我们的潜在客户遭受了研发预算的大幅削减和工作积压。为了抓住市场的当前趋势，AMFG将开发一个解决方案，允许在客户环境中快速部署我们的软件。目前的流程是与潜在客户进行概念验证（POC），通常需要1年时间才能完成，成本约为50000英镑。这是由于一个迭代的过程，AMFG和客户经常参加会议，以引出需求，了解我们的客户环境，并开发定制的概念。POC对AMFG和我们的客户来说也是一种风险，因为它在很大程度上取决于项目经理（客户方面）的经验。这是一个巨大的资源消耗，并导致生产线的长期次优利用，因为在开发过程的早期可能会错过重要的要求（当这些问题很容易解决时）。为什么AIPA是创新的：**AMFG建议将其提升到一个新的水平，并完全自动化POC的初始阶段。AMFG将通过使用新的机器学习算法来为客户确定最佳解决方案来实现这一目标。这将通过从客户那里获取市场上的最佳实践来取代迭代和冒险的过程。因此，AIPA将成为一个互动式问卷，根据客户的自我评估自动提出要求，并用于开发所需的功能。此外，AIPA可以作为一个持续的自我评估，以确认发展是在轨道上，并根据我们的客户的要求。AMFG预计，我们将能够三倍的成功完成的POC的数量从目前的12到36每年。