An example of the advances in sensing technology is ABB’s PEMS solution. PEMS stands for Predictive Emissions Monitoring System, which is a software alternative to the traditional emissions measuring solutions, in other words, a virtual measurement solution. It allows users, under the right conditions, to remove any hardware analyzer and device, and instead use data-driven modeling and machine learning to predict the emissions coming from a factory or chimney. This is being done in industrial processes, where analyzers are not installed, but just data is used to assess and report on emissions. This approach is permissible according to regulations in many regions, from the USA to Europe to the Gulf region.
To use virtual measurements, we first get a baseline for the measurement by collecting data with a temporary analyzer for a month or two, to identify correlations between emissions and the plant’s operating conditions. Then we use neural networks to create software models that reproduce the conditions and therefore the predicted emissions. At that point, we remove the temporary sensors and just leave a microprocessor to report the emissions. Of course, models and procedures have to be extremely accurate and reliable; for this reason, an external party must certify them. We have many applications of this solution around the world, all of which are examples of the benefits of digitalization and digital transformation.