The digital twin is a virtual representation of a product or process. Whether it is based on physics or is data-driven, the comprehensive digital twin enables design optimization and provides various types of services when the asset is being operated. From a design perspective, the digital twin created in 1D, 3D or computational fluid dynamics (CFD) simulation tools is primarily used for design analysis and optimization. The digital twin is usually specialized to predict accurately a limited set of attributes.
By conveying their core attributes only, reduced-order models valorize digital twins in a much broader context. Building and exporting ROM is now made available for all with Simcenter Amesim ROM Builder. This tool provides in an intuitive interface the best reduction technique from machine learning, linear algebra and statistics.
Learn step by step, the workflow to build reduced-order models.
The Simcenter Amesim ROM Builder grabs data that represents a model or a process, and in a few clicks, provides reduced versions of it. These reduced models come with unique features: They have a small memory footprint, are tool-agnostic (a fixed step Euler solver is enough) and can be operated in real time. For instance, complex 3D models can be simplified into a ROM that can be used in a system simulation model or on an edge device.
Further, the Simcenter Amesim ROM Builder provides to any simulation, testing, production or service professional a fast and easy way to create, evaluate and export ROM. With advanced automation, no expertise is needed to convert data into ROM.
The Simcenter Amesim ROM Builder unleashes the power of existing models and data to provide extra value for design and for operations.
Reducing complexity to relevant system-level expectations promises more affordable runtime. However, simplifying the model often requires expertise outside of the application domain: applied mathematics, statistics, machine learning, etc.
The digital twin applies to operations as well as design. Maximizing throughput and performance of mechatronics systems can be achieved by leveraging the system’s knowledge with smart controllers. Such controllers, empowered by real-time models for predicting behavior, can define setpoints that balance the system, environment and economic constraints. The Simcenter Amesim ROM Builder exports real-time models with small memory size that are tailored for this purpose.