Q: Where does the data come from for Profit SensorPro models?A: Data for Profit SensorPro can come from a variety of sources, just like data for Profit Controller or Profit Optimizer. These data sources include: Experion PKS, PHD, Loop Scout, Excel, text file, clipboard, etc.
Q: Can Profit SensorPro model non-linear relationships?A: Yes. The polynomial Partial Least Squares (PLS) supports model orders of 2 and higher. The User Entered Structure models support nonlinear modeling based on the form of the entered equation.
Q: Do users need to understand all of the modeling techniques to use Profit SensorPro?A: No. The Profit SensorPro user interface has been designed to make it easy to use for those who are less familiar with the statistical modeling techniques used, while at the same time providing additional information for more “expert” users.
Q: How can the user tell if the model is working properly on-line?A: Profit SensorPro uses the embedded version of Uniformance PHD that is part of Profit Suite as a data repository for its calculations. As a result, the model predictions as well as bias values can be easily trended and tracked using the Uniformance desktop tools. Additionally, Profit SensorPro model performance can be tracked, problems diagnosed and recommendations made to improve the performance using Honeywell’s Profit Expert.
Q: How does Profit Controller or Profit Optimizer use Profit SensorPro model outputs?A: During development of the Profit SensorPro model, Profit Design Studio enables Profit SensorPro outputs to be “connected” to Profit Controller or Profit Optimizer inputs by simply dragging and dropping the outputs into the Profit Controller or Profit Optimizer workspace
Q: What are some example applications where Profit SensorPro has been used?A: Some examples include: naphtha 90% point prediction, kerosene flash point prediction, and Light Gas Oil cloud point prediction, to name a few.