b'Heavy Industries | Engineer Innovationrecalibrate the controller using a real-time plant model. Finally, the entire test3D animationsetup can be managed by an interactiveSimcenter Amesim Dashboarddashboard.AllowedEffectively developing a controllerEnd of strokethat ensures safe machine operationNot allowedLoizzo and his colleagues successfully CANBUSapplied this methodology in an industrial project where they optimize the control algorithms of a completeControllermaintenance platform train that replaces railway catenary consoles. For safety reasons, such a device has strict constraints in terms of stability. While the operator handles the platform using a joystick, the controls should help prevent the movement of the intervention arm exceeding the safetySimplied Digital Twinmargins as a result of improper manipulation.First, we modeled the machine in Simcenter Amesim to size allFigure 2: Processor-in-the-loop methodology stepscomponents, explains Loizzo. Then after completing the plant model, weweeks of development, and just one could virtually identify potential failureweek of further calibration at the cases and optimize the designcustomer, explains Loizzo. Indeed, our accordingly. Finally, we simplified orprocessor-in-the-loop (PiL) methodology reduced the model to virtually operatelet us do 80% of the testing and it in real time. validation virtually and upfront. That helped us reduce the time we had to In parallel, the engineers developed thespend with the customer fine-tuning control algorithms. The real-timethe control laws for ultimate comfort calculations in Simcenter Amesimimprovement.enable direct exchange of signals between controller and model.Mr. Frederic Lagors, technical director at Simcenter Amesim can provide us, inCerebrum-Ingnierie, put things in a real time, the components position,wider perspective. He compares todays acting as a virtual sensor, explainsachievements with similar projects Loizzo. Those virtual sensor signals areabout five years ago and shares a then introduced in a closed loop withremarkable statistic. We assessed the the real ECU algorithms that adjust theeffectiveness of the PiL methodology by position of the intervention arm in casecomparing this platform train project the platform leaves the authorizedwith a similar one we delivered back in intervention margins. As a precaution,2015, explains Lagors. Five years ago, the software can reduce the platformsit took us four months to develop and movement speed when the boundariesvalidate the machine control are approached, or even stop it whenalgorithms. Now with the PiL method, exceeded. Using a 3D real-timewe are at four weeks. So we can animation of the model, we couldroughly say that the PiL approach allows interactively understand the systemsus to deliver control algorithms four behavior. times faster. This is an enormous reduction of time, cost and risk. nFrontloading up to 80 percent of the control strategy verificationThis virtual testing approach allowed Cerebrum-Ingnierie to dramatically speed up the algorithm design and validation. The catenary train control strategy definition took us only three 81'