Industrial scale biofuel pellet production from blends of unbarked softwood and hardwood stems—the effects of raw material composition and moisture content on pellet quality

Abstract A novel process for chipping unbarked tree stems into particles with defined sizes was tested in an industrial experiment conducted in a newly-constructed pellet plant. The factors varied were moisture content of the dried wood particles before pelleting (11–14%) and species composition (Norway spruce, Scots pine and birch; 0–100% of each species). On-line instruments were used to acquire near infrared (NIR) reflectance spectra from the stream of raw and dried particles. The response variables examined were pellet bulk density, mechanical durability, moisture content and ash content. The dryness of the wood particles had a significant effect on all responses (except for ash content). The species composition of the blend only affected mechanical durability and ash content. NIR-based models proved to be excellent for predicting particle dryness and useful for predicting the species composition. Optimal pellet quality was achieved using a blend with 45–75% spruce, 0–55% pine and for 0–25% birch at a moisture content of 12.5–14.0%. Overall, the results obtained demonstrate that it is possible to produce high quality pellets by using fresh stem logs in specified blends with specific moisture contents. NIR spectroscopy was found to be useful for the direct and indirect control of critical pellet quality variables.