A New Design Parameter: Feedback

Industry 4.0 has opened a world of feedback data to engineers, and they are actively using it to take motion control and machine design to the next level. “In Industry 4.0, operating data closes the loop between design assumptions and real performance,” explains Rodrigo Toro Olmedo, global offering leader for Mining, Minerals & Metals at Honeywell Process Solutions. “Continuous digital oversight allows designers to observe how torque, vibration, temperature and duty cycles evolve in the field, transforming empirical knowledge into design inputs.”

Data related to predictive maintenance such as bearing wear or thermal drift also allows engineers to identify patterns of mechanical stress or control instability that can be corrected in the next design iteration. “This feedback reduces oversizing, improves system efficiency and drives reliability-centered design,” Olmedo notes. “In essence, data turns ‘best guesses’ into measured truths, ensuring new machines are designed based on evidence rather than assumptions.”

In the view of Sameer Kher, senior director of Product Development, Systems and Digital Twins at Synopsys, data from predictive maintenance and machine operation helps engineers address all three of their main ongoing goals: lowering cost, improving time to market and adding new capabilities to products.

READ MORE: From CAD to Co-Design: Mastering AI, Material Science, Digital Twins and MBSE

With costs, incorporating data into existing design processes means models of system or machine function are more accurate than idealized models, which in turn reduces costly physical testing processes. With time to market, Kher explains that integrating feedback data with less computationally demanding models should result in rapid creation of more-accurate models, reducing product development time. “Finally, realistic models during design will reduce the need for over-engineering,” he says, “allowing for more product features.”

Part of how the integration of real-life operational data aids design engineers to reduce costs and optimize the performance of the next iteration is the way it aids engineers in optimal component and material selection, says Hugues Therrien, division manager for Drives and Motors (Motion Business) at ABB Canada. “By examining past failure events and wear data, designers can pinpoint parts and components most likely to fail or underperform,” he explains. “This in turn assists in selecting more-reliable components, addressing weak points and strengthening commonly failing electrical and mechanical systems.”

Which Data Matter Most?  

The most valuable feedback data depends on the specific use case/application, but according to Kher, parameters that correlate strongly with failure modes and performance optimization are generally the most useful. “Temperatures, torques, vibrations/stress, etc. all have a role to play, depending on the application,” he says. 

Looking specifically at motion systems, some operational and maintenance feedback parameters are significantly more impactful for design enhancement than others, says Dragos Dobreanu, service manager (Motion Business) at ABB Canada. 

“For instance, in a powertrain system featuring a drive and motor, current consumption and voltage stability can indicate stress on the drive or motor, revealing potential design inefficiencies,” he says. “Additionally, monitored parameters such as vibration offer clear insights into mechanical integration, component choices and material selections by signaling potential issues like misalignment, looseness, imbalance or bearing failure.” Dobreanu adds that temperature data is equally critical, as thermal patterns can expose inadequate cooling design, overloading conditions or component degradation over time. 

READ MORE: Physical AI in Motion—How Machine Learning Drives Next-Gen Industrial Automation

In addition to power consumption, vibration and temperature data, Olmeda identifies a couple of other metrics that directly connect real-life operational behavior with mechanical integrity and lifecycle cost, allowing engineers to design for actual conditions. He points to torque and load spectra (which define structural fatigue and optimize sizing) and Mean Time Between Failures/To Repair and failure modes (which help direct improvements in materials, seals and serviceability).

At the same time, Kher says a key challenge faced by customers is that sometimes parameters of interest are difficult to measure—for example, the temperature inside a furnace. “In these cases, they might need to measure other parameters,” he says, “and create virtual sensor models to get the information they need.” 

Design Effects

As to which aspects of motion control and machine design are more affected by using feedback in the design process than others, Therrien puts primary motion components at the top. 

“Particularly, drives and motors are significantly influenced by digital performance and maintenance feedback,” he says. “This impact is most evident in the controls, processors, sensors, diagnostics and communication architecture, as these system elements are where algorithms and analytics provide critical insights into performance and reliability.”

Drilling down a little further in the context of motion systems, he differentiates that digital performance and capabilities determine achievable outcomes regarding speed, diagnostics and precision, while maintenance feedback indicates reliability, robustness and life expectancy. “Together,” he says, “these factors contribute to the development of more intelligent systems equipped with advanced monitoring and self-diagnostic functionalities that enable proactive intervention and continuous optimization.”

READ MORE: Modeling Machine Designs that Seal Deals 

Looking at the design process itself, Olmedo says feedback has the greatest impact on the early design and verification phases. That’s the point where decisions have the most influence on lifecycle cost. Real-world operating data at this point in the design process refines system requirements, improves component selection (motors, gearboxes and actuators), and guides structural and thermal modeling.

However, control strategy design also benefits, says Olmeda, from the use of feedback. “PID tuning and motion profiles are validated against the actual mechanical response captured in field data,” he explains. “During validation, test cases can be based on recorded transient events rather than theoretical load cases.” In addition, he says design for serviceability and maintainability are also impacted by insights from predictive maintenance trends, ensuring the next generation of equipment is smarter, more efficient and easier to sustain.

The Role of Feedback in MBSE

Feedback is also useful in MBSE. Olmedo says that in MBSE, feedback transforms static models into living digital assets, where data collected from the field (e.g., dynamic loads, temperature gradients or vibration spectra) is used to calibrate simulation models, ensuring that virtual representations match as-operated conditions. “This allows engineers to validate control algorithms, mechanical designs and maintenance intervals,” he explains, “within a single systems model.” 

However, he notes that operational data also supports requirements for traceability within MBSE platforms. “Field performance metrics such as Mean Time Between Failures or energy consumption could directly update the design’s SysML or simulation parameters,” says Olmeda. “As a result, the ‘digital twin’ becomes continuously synchronized with the physical asset, enabling design decisions based on verified operating behavior.”

In their view of the role of feedback in MBSE, Kher and his colleagues first explain that they believe the eventual goal of MBSE is to enable the concept of Continuous Integration/Continuous Delivery (CI/CD) in software development for engineered products. 

“Key within this is the need to flow data and models continuously across the CI/CD loop, all the way from design to operations and back,” says Kher. “Data can be used to calibrate models to improve accuracy or it can also be encapsulated into executable models that can be incorporated into the rest of the design flow.”