Performance evaluation of NIR spectrophotometer simulating in-line acquisition for moisture content prediction of woodchips and comparison with hand-held NIR spectrophotometer

The European Green Deal, together with the current political and economic background, aims at improving the use of renewable energy resources rather than fossil fuel dependence. Woodchip could be considered as a valid and widely available alternative, but the high heterogeneity of its properties must be evaluated according to the important influence in each step of the supply chain to combustion process. Moisture content is the main parameter involved in the quality characterization because it influences the energetic (it reduces the biomass calorific value in combustion), the management (during harvesting, handling, and storage steps), and the economic (price based on the load, affected by bulk density) aspects of woodchip utilization. Continuous monitoring of moisture content results essential, but laboratory analyses are time-consuming, destructive, and need economic efforts in terms of instruments and skilled personnel. Near infrared Spectroscopy (NIRS) is a fast, friendly-use, and alternative method to qualitative characterization, based on non-destructive interaction between wavelength and chemical components. So, the aim of this project is to evaluate the prediction performance of moisture content using a portable NIRS adapted to in-line utilization. Then, the results have been compared with the already implemented prediction models of a different portable NIRS, applying statistics to assess the industrial woodchip quality, but also the Tukey's Test method to evaluate the replicates effect. Considering the general consistency of industrial woodchip quality and the use of three different replicates approach, results show good performance ranging from 0.64 to 0.86 as determination coefficient, with errors of prediction from 4.44 to 2.88%, confirmed by a prediction bias below the 2.5% for more than 50% of the samples. Tukey's Test shows that replicates adequately described samples variability thanks to constant and uniform analysis that dilutes the error. Furthermore, accepting a minimum error threshold, no investigation on replicates feasibility is needed, reducing time and economic efforts. In conclusion, the crucial demand of moisture control meets the several advantages of NIRS technique, with the potential in-line application to obtain real-time quality tracing and high-performance energy production.