Predicting exhaust emissions of a liquefied petroleum gas engine

Quantifying nonroad engine emission levels under different ambient and operating conditions can complement current manufacturer's certifications. An initial data set consisting of CAN, weather, and emissions information was obtained from a 104 hp LPG irrigation engine using different RPM and torque settings over an ambient temperature range of 75°F to 103°F. Polynomial models were developed to determine what variables affect emission levels of oxides of nitrogen (NOx) and Carbon Monoxide (CO) at different ambient and operating conditions. Variables shown to have a direct affect on predicting NOx emission levels at 90 percent confidence levels were fuel rate, rpm, and humidity. Variables shown to have a direct affect on predicting CO emission levels at 76 percent confidence levels were rpm, percent torque, and manifold temperature. Using modeling software to predict gaseous emissions of a LPG engine is possible.