Prediction of heating value of straw by proximate data, and near infrared spectroscopy

Exploration of straw resources for energy production has been attracting agricultural scientists and engineers for decades. And the heating value of straw has always been the focus when initiating a straw-based biomass energy project. Nevertheless determination of heating values of straw needs delicate and expensive calorimeter, and is time-consuming. It’s quite desirable to develop quick and easy model predicting heating values of straw. In this study, we proposed three applicable models, first two are multiple linear regression (MLR) equations by contents of moisture, ash, and volatile matter, the other one is based on the near infrared spectroscopy (NIRS) technology. All the models provide satisfactory estimations of heating values of straw samples. The adjusted determination coefficients for MLR models were 0.9049 and 0.9039, and determination coefficients of calibration for NIRS model was 0.9604; When evaluated on independent validation, the determination coefficients were 0.8595, 0.8524 and 0.8946, respectively. The results indicated that both MLR models and NIRS model have the potential to predict the heating values of straw, while the NIRS model presented better accuracy.