Attaining greater process efficiency through digital transformation is vital if you want to get ahead in today’s manufacturing environment. In this interview, Dr. Fouad el Khaldi, discusses critical parts of the manufacturing 4.0 transformation which enable predictive maintenance, detection of early signals of deviation and prediction of future incidents with precision.
Today, manufacturing process design, process validation, and actual production are well optimized but remain disconnected. Thanks to recent developments (Big Data, Internet of Things, Artificial Intelligence, etc.), these three phases finally connect! Engineers can benefit from faster iteration loops and better assess the impact of process design decisions.
Data from real life performance offers opportunities for continuous learning, which will benefit the next generation of products by upgrading design assumptions. This becomes key as we speak increasingly of predictive maintenance and managing a product’s performance throughout its lifetime (Product Performance Lifecycle™ - Management or PPL-M), rather than delivering a product that performs on day one (what conventional Product Lifecycle Management or PLM covers).
Virtual Prototyping is a powerful methodology enabling the design and validation of manufacturing processes. It’s at the core of ESI’s Hybrid Twin™ approach, where we combine the virtual prototype with the data coming from industrial plants to measure the real operational performance, to adjust the initial model to real life data and context, and to detect early signals of deviation. A Hybrid Twin™ enables asset managers to get the information necessary to assess cause & effect relationships, and to implement the appropriate corrective measures.
Why hybrid? This is crucial in overcoming the limitations of a digital twin: indeed, if we limit ourselves to data collected from historical and real-life operations, we can only predict behaviors that already took place. Whereas building on a virtual prototype that reproduces the asset as-good-as-real (capturing for instance its material characteristics after manufacturing and assembly) helps us predict almost any kind of future incident with precision, even in the case of changing parameters (materials variations, operating conditions, etc.).
Modern manufacturers rely on simulation and pilot tests to ensure that they’re meeting various time, quality, cost requirements in their manufacturing process design and validation. However, they typically limit the use of simulation to the methods and the validation engineering departments and haven’t deployed it into production for various reasons ─ mainly related to complexity and response time.
Good news for production managers: ESI’s innovative Parametric Reduced Model technology enables the development of a Hybrid Twin™ with real time responses, derived from a predictive detailed 3D model built in the process design and validation phase (see graphic). The Hybrid Twin™ opens new opportunities to augment the PLC (Programmable Logic Controller) capacity for smarter machine control. The Hybrid Twin™ will be loaded on site as edge computing (small processors next to the machine) for obvious performance and security, benefiting from recent IoT advances, such as 5G. Factory production managers will be able to measure and predict production performance more efficiently to detect early signs of deviation and to anticipate troubleshooting - thus maintaining the required quality (reducing scraps) and ensuring optimal performance.
Early pilot projects are already demonstrating the feasibility of such a solution and showing manufacturers how the simulation capabilities will be adapted and streamlined to be implemented right at the heart of the factories ─ with very encouraging outcomes.