To effectively construct models and operate predictive analytics and optimization systems, the acquisition of sensor data becomes imperative. Frequently, a shortfall in data quantity or data accuracy is encountered. In some cases, there may even be a complete absence of sensors, despite their critical importance for model development. Such a scenario unfolded at a chemical plant.
Our client presented us with a clear objective: to boost product output using our optimization tool. During our project execution, a significant challenge surfaced — the measurement of production only occurred after the material had been processed in the machinery. This crucial parameter was not integrated into any existing data systems, and manual data entry posed an elevated risk of human error. So, what was the solution?
Following an intensive brainstorming session, our team devised a virtual tag. This innovative approach enabled us to compute production figures in real-time based on available data, such as flow rates and tank levels, directly within the data connector.
The outcome was remarkable. Real-time production figures were now seamlessly integrated into our optimization system. Our technologists and engineers had truly worked wonders!