Virtual Analysers are a cost efficient solution because instead of physically measuring a parameter like a hardware analyser does e.g. CO, NOx etc., our VirtuaLyser (Virtual Analyser) calculates the parameter to be monitored using other process inputs & parameters. The VirtuaLyser is a part of our Predictive Solutions.
If you are responsible for upkeep and maintenance of hardware analysers at your site then you must have asked these questions to yourself:
- Is the analyser working perfectly and giving accurate readings
- If you think it is working perfectly, is it because you calibrated it yesterday
- If you have a preventive maintenance program in place, can you really be sure if the interval between two calibrations is adequate just because the vendor told you so
- All of this decision making is based on subjective inputs i.e. there is nothing tangible or measurable which can confirm that:
- The analyser is working perfectly (unless you calibrate it)
- It has not drifted and is still accurate
Virtual Analysers take the guess work, the subjectivity, out of the equation and replaces it with Objectivity i.e. solid data to prove whether an analyser is accurate or not. Here is how the VA does it:
The VA takes data from different inputs to confirm the validity of the hardware analyser. This example will explain it best:
Suppose you have an O2 (oxygen) analyser in the exhaust stream of a furnace, it is crucial for you that it gives accurate readings as the entire combustion process is hinged on it. That makes the O2 analyser a safety critical item. A Virtual Analyser takes inputs from other process analysers and data points like air flow analyser, power consumption of the air blower, damper position, fuel flow AND the O2 analyser to “predict” the actual O2. After a one time calibration, the software of the Virtual Analyser can “learn” to distinguish between drift, span or calibration error.
Virtual Analysers use machine learning models which can accurately predict the actual analyser reading and inform you if the health of the actual analyser is good or requires calibration. Just like any other Machine Learning device using Artificial Intelligence, the Virtual Analysers improves in accuracy as it gathers more & more data and it is not uncommon to have an accuracy of upto 98%, all the time. This predictive engine is US EPA Certified for use in continuous emissions monitoring, also known as PEMS (Predictive Emissions Monitoring).
It makes perfect logical sense when you reach a conclusion based on the inputs of 4 or 5 independent parameters instead of just one.
A VA doesn’t:
- Require Calibration
- Require Plant shut down for any maintenance on them
- Drift (hence always stay accurate)
VA can be:
- Cheaper to install & maintain
- Used in hazardous locations
- More accurate as they rely on more than one input to give the results
Most locations where hardware analysers are installed can have VA, either as the main or back up analyser.
You can download this short Power Point Presentation Virtual Analyser to fully understand this product.
If you are interested, we can provide a free demo installation at your site.
A picture is worth a thousand words. Below are the pictures of hardware required for a typical O2 analyser. All of this equipment is NOT required for a VA.
Please feel free to Contact Us for any further details.