What If You Could Measure Things That Don’t Have Sensors?
What If You Could Measure Things That Don’t Have Sensors?
Industrial facilities have never had more access to data.
Modern machines are equipped with pressure sensors, temperature sensors, flow meters, vibration monitors, encoders, and countless other measurement devices designed to provide visibility into system performance. Yet despite all of this instrumentation, engineering teams still encounter an unavoidable reality.
Many of the variables they would most like to measure are often the hardest to access.
Internal component forces, localized pressures, mechanical stresses, leakage rates, actuator loads, and system efficiencies frequently remain hidden behind physical, technical, or economic limitations. Installing additional sensors may be impractical, expensive, or simply impossible.
This raises an interesting question.
What if you could measure values that do not have sensors?
The Limits of Physical Instrumentation
Every sensor provides information, but every sensor also comes with constraints.
Sensors must be installed, calibrated, maintained, and integrated into existing control architectures. They introduce additional cost and complexity while only providing visibility into specific locations within the system.
For many industrial applications, engineering teams quickly reach a point where additional instrumentation becomes difficult to justify.
This is particularly common in hydraulic and mechanical systems where the variables of greatest interest are often buried deep within the machine.
An engineer may want to understand:
- Internal cylinder forces
- Pressure distributions throughout the system
- Component loading under specific operating conditions
- Energy losses within a hydraulic circuit
- Mechanical stresses at locations where sensors cannot realistically be installed
Obtaining this information through physical measurement alone can become prohibitively difficult.
A Different Approach: Virtual Sensors
One of the more practical applications of digital twins is the use of virtual sensors.
Rather than physically measuring every parameter within a machine, engineers use a physics-based model operating alongside the real equipment. The model receives the same inputs as the physical machine and continuously calculates system behavior based on those operating conditions.
The result is a virtual representation of the machine that provides visibility into quantities that would otherwise remain inaccessible.
Instead of adding another physical sensor, engineers can estimate values using the physics already embedded within the model.
A Servo Press Example
During a recent digital twin implementation, a manufacturer of servo presses faced a common challenge.
The company wanted greater visibility into machine behavior but recognized that adding physical sensors throughout the equipment would significantly increase complexity and cost. There were numerous variables they would like to monitor, but instrumenting every location simply was not practical.
Their solution was to run a physics-based model alongside the operating machine.
At the end of each operating cycle, the machine inputs were passed to the model, which calculated how the system behaved during that cycle. This effectively created a set of virtual sensors capable of providing information that did not exist within the physical instrumentation package.
Instead of being limited to measured values alone, engineers gained visibility into pressures, flow rates, forces, torques, and other system characteristics throughout the machine.
Seeing Inside the Machine
This is where the real value begins to emerge.
- Most physical sensors can only tell you what is happening at a specific point.
- A pressure transducer measures pressure at its installation location.
- A flow meter measures flow through a particular section of the circuit.
- A load cell measures force at a single location.
Virtual sensors allow engineers to move beyond those physical limitations.
Because the model represents the entire system, it becomes possible to estimate conditions throughout the machine rather than only where instrumentation exists. This creates opportunities to investigate questions that would otherwise require significant testing effort or additional hardware investment.
For development teams, maintenance organizations, and machine builders, this often results in a much deeper understanding of how equipment is actually operating.
Supporting Better Engineering Decisions
Virtual sensors are not simply about generating more data.
They are about generating more useful data.
Engineering teams are constantly making decisions related to machine performance, maintenance strategies, system upgrades, and operational improvements. The quality of those decisions depends on the quality of the information available.
By providing visibility into quantities that are difficult to measure directly, virtual sensors help fill gaps that often exist within conventional monitoring systems.
This additional context can support:
- Root cause analysis
- Performance investigations
- Condition monitoring
- Design validation
- Operational optimization
- Predictive maintenance initiatives
The objective is not to replace physical instrumentation. It is to extend what that instrumentation can tell us.
The Connection to Digital Twins
Virtual sensors are often one of the first practical applications organizations adopt when implementing digital twins.
They deliver immediate value because they leverage data that already exists while providing insights that would otherwise require additional hardware investment.
As digital twin programs mature, these same models can support trend analysis, fault detection, operational forecasting, and future AI initiatives. The virtual sensor becomes one component of a larger strategy focused on understanding system behavior and improving decision-making across the asset lifecycle.
For many organizations, the journey begins with a simple objective.
Understanding more about the machine than physical sensors alone can provide.
Discuss Your Application
Every industrial system contains variables that are difficult or impossible to measure directly.
Whether you’re developing new equipment, improving existing assets, or exploring digital twin initiatives, virtual sensors can provide additional visibility into system behavior without significantly increasing instrumentation costs.
If you’re evaluating how physics-based models can support monitoring, diagnostics, condition assessment, or operational decision-making, Agito can help you explore practical approaches based on your specific application.
