**1. Introduction**

Over the past few years there has been a significant effort worldwide to increase the utilization of renewable energy sources for electricity and heat production. The European Union Renewables Directive, for example, set a target of 20% of the energy consumption of the European Union to be from renewable sources by 2020 [1] and current decarburization scenarios point at 30% by 2030 and at least 55% by 2050 [2]. In this context, the firing of biomass and the co-firing of biomass with coal in power plants originally designed for coal play a significant role. This approach to the generation of renewable energy has been demonstrated in more than 200 power plants over the recent years taking advantage of the infrastructure already in place within the electricity supply industry, and the associated low capital investment requirements [3].

Besides the expected reduction in both NOx and SOx levels, there are still some difficulties in using raw biomass for co-firing with coal [4]. Amongst the biggest challenges are the diverse chemical and physical properties of biomass that need to be addressed. The ash content as well as volatiles and oxygen content can be higher in biomass than in coal resulting in lower gross calorific value of the

fuel [5]. Additionally, problems may occur if feeding and milling is not adapted. The particle size distribution of ground biomass is different from coal and particles may be considerably larger leading to lower burn-out efficiencies [6]. Overall, co-firing with regular biomass may lead to instability of the combustion process making biomass with coal-like properties the most desirable co-firing fuel.

One intensively researched approach for upgrading regular biomass for co-firing with coal is the thermo-chemical pre-treatment best known as torrefaction. At temperatures in the range of 250–300 ◦C, and in the absence of air, the biomass undergoes a mild form of pyrolysis and is converted into a solid fuel with chemical and mechanical properties similar to coal [7]. This process has been studied on a large variety of feedstock applying different torrefaction settings (e.g., temperature, and time) mainly on lab-scale [8] and most recently also in pilot scale plants [9]. Several technological innovations were made over the past years to bring the process to commercial scale [10] and techno-economic assessments assuming European and Northern American conditions were conducted [11,12]. In both cases, the now more homogenous, brittle and hydrophobic material is compacted to pellets and can be transported, stored for prolonged periods and delivered as a boiler fuel.

Some experts came to the conclusion that the percentage of biomass for co-firing in power plants can be as high as 40% when the torrefied biomass is used [13] while others showed that, according to CFD simulations, coal boilers can handle operation with up to 100% of torrefied biomass without an obvious decrease of energy efficiency and fluctuation of the boiler load [14]. The direct application of torrefied biomass pellets in small-scale pellet boilers may not be advisable at this point unless boiler technology is adapted [15].

For the utilization of torrefied biomass in pulverized coal fired plants it is necessary to assess its fuel properties (chemical and mechanical) beforehand. For most of the parameters standardized analysis methods are available. However, the difference in characteristics between torrefied biomass and coal requires some adaptation of existing methods or even new methods to fully describe their fuels properties, which would then allow comparison with those of coals. A power plant operator will need to, e.g., adjust the mills for achieving the right particle size distribution for optimal plant operation. It is assumed that particle size distribution of the milled material may effect combustion efficiency, the amount of unburned carbon in the ash and the stability of combustion [16]. The Hardgrove Index (HGI), developed in the 1930 by Hardgrove [17], was designed to compare grindability of brittle materials such as black coal. It is commonly used as a simple empirical laboratory procedure for the characterization of their milling behaviour in large coal mills to produce a pulverized fuel with the appropriate fineness. The HGI method described in ISO 5074 [18], however, is not suitable for the characterization of more fibrous material like torrefied biomass and does not consider the actual milling of pellets.

Several studies have determined the grindability of torrefied biomass by applying one of the following methods: the modified Hardgrove Index [16], the Grindability Criterion (GC) [19], the Hybrid Work Index (HWI) [20] and the Impact Grindability Index (PMI) [21]. The HGI was developed for the testing of black coals, which are commonly milled in ball and roller mills. The PMI was developed for low rank coals, which are commonly processed in beater or fan mills. The modified HGI, the HWI and the GC were specifically designed to characterize the grindability of torrefied biomass. A number of boiler makers and mill suppliers also have scaled down coal mills that are employed for test purposes.

All of the current laboratory grindability tests require the preparation of a pre-crushed feed material of a specific particle size. In all cases, the test results provide an assessment of the relative grinding behaviour of the test materials. It may, however, be difficult to extrapolate the data from a laboratory test to provide quantitative information on the behaviour of torrefied biomass pellets in an industrial scale mill.

The aim of the work presented here, which was part of the SECTOR-project (funded under the EU FP7 programme) [9], is to test a grindability method (GM) for torrefied biomass pellets, which is simple, fast and reliable and based on the grinding of pellets rather than pre-crushed material. The method will result in a single parameter: the specific grinding energy (SGE), i.e., the energy consumed when a defined pellet mass (2.5 kg), at a defined mass flow is ground to a particle top size

of 1 mm. This parameter, or the basic principles behind it, may eventually become useful to pellet producers and plant operators if grindability is introduced to a standardized product specification of thermally treated biomass, which is currently being drafted in the relevant ISO TC238 working group (as ISO/PRF TS 17225-8 [22]).

The results of the GM, when done with different mill types (hammer, cutting and impact mill), will be compared and their relation to other pellet stability measures such as the modified HGI, mechanical durability and a hardness test will be discussed. Finally, advantages and disadvantages of the different mill types as proper equipment for a standardized method are discussed.
