| OCR Text |
Show NONLINEAR REGRESSION MODEL BASED PEMS The general procedure for building, installing and certifying a nonlinear regression model based PEMS is as follows: 1. Collect emissions data with a highly-accurate reference method or CEM as depicted in Figure 1. The pollutants of interest could be NOx' CO, O2 , CO2 , S02' VOC, or other compounds. 2. Operate the process and/or pollution control device over the range of anticipated operations for 3 to 5 days and concurrently collect process and/or pollution control device data with the emissions data. 3. When the data gathering period is finished, use a modeling tool such as Pavilion's Process InsightS® to build emissions and sensor validation models as a function of the process variables. 4. Install the PEMS in the computer which receives the process data. 5. Perform a relative accuracy test audit (RAT A) of the PEMS to demonstrate conformance with accuracy requirements. Environmental Monitoring Van Reference~ethod onlinear Regression Modeling Tool and PEMS Figure 1: Process and emissions data are collected simultaneously from the process data historian and by a hardware CEMS or a reference method monitoring van, respectively The emissions model is built with a modeling tool such as Process Insights by correlating these two sets of data. Then the PEMS is installed on the plant computer which receives the process data and the PEMS runs real-time, continuously. 7 |
