Virtual Sensors

Background & Definitions

In Consibio Cloud a virtual sensor is a calculation that happens on top of physical sensor data to produce new insights and knowledge that the physical sensors on their own cannot give.

The term virtual sensor is thus a broad definition for any real-time, automatic calculation that takes a sensor-value and uses a set of analytical tools to produce new insights. The insights could be industry specific Key Performance Indicators (KPI’s) or weekly averages of sensor-values. The possibilities are manifold.

Examples of these virtual sensors are shown below:

Example TitleDescription
1. Biomass EstimationIt can be difficult to take an accurate and representative measurement of biomass weight, whether it is livestock animals (poultry, pigs etc.), insects or microorganisms in general.

However, using advanced mathematical models and the correct sensor feedback, it is possible to accurately estimate and predict the biomass of such living systems.
2. Weekly averagesWhen evaluating what went wrong/right for a specific production/batch, time-based averages can give insight into why something happened.

Whether it is daily, weekly, monthly or yearly averages all depend on the use-case, but all of it is possible in Consibio Cloud through the Virtual Sensors.
3. Metabolic activityMetabolic activity is always interesting when working with living biological systems.
The activity of a living cell can often be correlated well with interesting insights such as diseases, optimal feeding strategies or product concentrations.

However, a “metabolic activity sensor” does not exist but through deep biological knowledge, it is possible to use Virtual Sensors to give a live view of this.
4. EfficienciesOften used to get a measurement of how well a specific production process is performing. Typically shown on a scale of 0-100 %.
5. RatesRates can be seen as either production or the removal of a compound.

A production rate of protein in insect or vertical farming are examples of relevant production rates. In terms of removal rates, this is highly useful in air filtration industries where corrosive and smelly gases such as hydrogen sulphide (H2S), Methane (CH4) and Ammonia (NH3) are removed.

With the help of Virtual Sensors, it is possible to get a live estimate of such production/removal rates, so action can be taken based on data-informed foundation.
A list of example use-cases for Virtual Sensors. Many more exist!

For more examples on specific cases where Virtual Sensors were implement, checkout the projects section at our website!

Physical vs. Virtual Sensors

So why is it that virtual sensors can be justified when compared to a sensor, that physically or chemically measures a parameter in a physical environment?

If you’re not already convinced, here are some of the drawbacks of physical sensors, compared to virtual sensors:

“This sensor is too expensive to acquire”

Sensors come in all different shapes / sizes and so does the price. Some sensors can cost up to several thousands of Euro, depending on the level of precision in the instrument.

With a virtual sensor, the acquisition cost is very low and the the pricing is totally dependent on the value-add.

The environment is too hostile for the sensor to be placed in!

There are processes and productions where the operating environment is so hostile that continuous measurements using sensors are practically impossible to take. Sometimes, the measurement can only be done intermittently such as measuring pH-levels in wastewater or flushing a H2S-sensor with air once every 10 minutes, to increase lifetime of the sensor.

With a virtual sensor, information between readings can be provided and estimated. In other words, virtual sensors can use the sparse data collected to continuously “see” into these environments without a physical presence.

The sensors are just too expensive to maintain

All physical sensors require maintenance at some point and in some cases, they have to be replaced with new ones. In most cases, this requires that a new sensor is ordered and that personal has to be activated once it is ready to install.

Since virtual sensors only rely on calculations happening in the cloud, the level of maintenance is none-existing compared to classical repair of sensors.

It is nice to receive sensor feedback, but what I really want is a prediction of what happens next

Sensors can only provide a historical and present image of the environment that it is placed in. Having this is extremely valuable compared to taking decisions based on “feeling” and not historical, measured data. However, sometimes the operator / manager would like to be able to predict the value of sensor in the near future.

Depending on the complexity, some of the virtual sensors built by Consibio have capabilities to predict and make estimates on what will occur in the future.

So these are some of the drawbacks of physical sensors. However, it should be noted that sensor feedback from physical sensors are a necessity for virtual sensors to work properly. Virtual sensors are not ment to completely replace all sensors, but in certain situations they can be very useful!

How do I get started with Virtual Sensors?

If you want to get started and use Virtual Sensors to get more insight, you have to contact our development team. This can be done via. the email and writing Virtual Sensors as part of the topic field.