DigiCast:Digital technologies for Intelligent Aluminium Casting Systems

Aluminium er metallet med raskest voksende etterspørsel i verden idag og er et viktig materiale i utviklingen av nye lavutslippsløsninger. For å oppnå en bærekraftig produksjon med lavt karbonavtrykk, vil systemer som gir god prosesskontroll og et godt arbeidsmiljø for operatørene være viktig. Digicast har fokusert på utvikling av smarte systemer for aluminium DC-støping som kombinerer prosess kunnskap i form av prosessmodeller kombinert med datainnsamling og visualiseringsteknologi. Detaljerte modeller for DC-støping av valseblokkmaterialer og ektruderingsbolter i programvaren ALSIM ble benyttet til å generere datasett som utgangspunkt for forenklede og effektive modeller som kan benyttes i digitale prosesstvillinger. Metodikk og numerisk rammeverk for generering av data-sett og forenklede modeller for sentersprekktendens i bolt og oppkrymp under valseblokkstøping ble etablert og demonstrert i prosjektet. Arkitektur for datainnsamling fra de industrielle prosessene, og kameraløsninger til overvåking av støpeprosessen er også utviklet i prosjektet.

Competence on creation of simplified and CPU-efficient models based on syntethic data from accurate, physical based models has been developed in the project. Numerical frameworks for creation syntetic data-sets that is used as a basis for a digital process twin of the DC-casting process has been established. Industrial competence and tools on architecture for data collection from industrial processes, monitoring and visualisation systems has been developed. The competence and tools developed enables the industrial partners to differentiate from competitors in improving process control by the use of digital technologies.

Today casthouse operations must rely on operator experience with casting parameters and recipes. As the requirements on quality such as ingot integrity, homogeneity and geometrical tolerances increase continuously, technical teams necessitate supporting tools to carry out, control and optimize their operations. The project aims at developing intelligent casting systems for improved casthouse operations. The framework will consist of a digital twin as well as visualization solutions to assist operators with information for process planning as well as modification during operations. The digital twin will be self-learning and based on hybrid modelling. The approach takes advantage of strong competences on DC-casting built into process modelling tools and will be combined with data-analytics for increased predictability. The system will be made applicable for multiple industrial alloy systems and DC-casting technologies (billets and sheet ingots). The main R&D challenge will be to establish a fast, accurate and predictive tool. Visualization tools based on augmented reality will be developed for the casthouse by introducing 3D representations of casting equipment. The holographic representations will be linked to sensor data as well as the digital twin to supply operators with process information. By embracing digital twins and AR based control and visualization, several benefits are to be achieved: i) use of real-time data for greater collaboration throughout the process, ii) simulate real-life scenarios before casting, iii) identify casting recipe flaws and make modifications during the preparation phase, iii) improve process and product and entire system function by predicting failures and identifying areas that need modification, iv) increase safety, efficiency and lower production costs, v) competence transfer and training. This will constitute a step change compared to today's operations.