8 Predicting outcomes with simulation

A digital twin becomes truly valuable when it can do more than reflect the current state of a system. While data integration and visualization allow a twin to describe what is happening, simulation enables it to explore what could happen. By running controlled "what if?" scenarios and changing inputs, conditions, or assumptions, a digital twin transforms from a descriptive mirror of reality into a tool for prediction, optimization, and decision-making.

At its core, simulation is about understanding how a system’s state evolves. Some systems change continuously, for example, temperature rises and falls, pressure fluctuates, and airflow accelerates and decelerates. Others change only at specific moments, such as when a machine starts or stops, a job enters a queue, or a component fails. These two patterns, continuous change and discrete events, form the foundation of nearly all simulation approaches used in digital twins.

8.1 Modeling and simulating physical systems

8.1.1 The digital twin advantage

8.1.2 From static models to dynamic simulation

8.1.3 Continuous versus discrete event simulation

8.2 Continuous systems

8.2.1 Simulating a continuous system

8.2.2 The initial value problem

8.3 Acausal modeling for digital twins

8.3.1 How it works

8.3.2 Modeling tank drainage in OpenModelica

8.3.3 Executing simulations

8.3.4 Integrating sensor data with continuous simulation

8.4 Simulating processes with discrete event simulation

8.4.1 Why DES matters for digital twins

8.4.2 DES with Python using SimPy

8.4.3 Try it out: simulate appliance operation with SimPy

8.4.4 Example application of DES in digital twins

8.5 Finite element and field simulation

8.5.1 Try it out: simulate heat transfer with FEM

8.5.2 Using FEM for virtual sensing in a digital twin

8.5.3 Example application of FEA in digital twins

8.6 Computational fluid dynamics

8.6.1 CFD with Python for lightweight simulation