Elsevier

Powder Technology

Volume 378, Part A, 22 January 2021, Pages 704-715
Powder Technology

Analysis of wood pellet degradation characteristics based on single particle impact tests

https://doi.org/10.1016/j.powtec.2020.10.017Get rights and content

Highlights

  • Design of impact test facility for cylindrical particles

  • Development of empirical prediction model of wood pellet breakage behaviour

  • Selection and breakage function for cylindrical biomass particles

  • Degradation characteristics of cylindrical wood pellets

  • Analysis of influence of pellet quality and target material on degradation behaviour

Abstract

The knowledge of size reduction of wood pellets during pneumatic conveying is important to achieve failure-free system operation and to set-up adequate quality assurance processes. In the present study, experiments are performed with a single particle impact test facility. A stereoscopic high-speed camera set allows 3D-particle tracking and visual analysis of the particles' degradation properties. Based on the data obtained, empirical correlations for the statistical description of the pellets' breakage behaviour are presented depending on particle length, impact velocity and collision angle. These relationships are expressed mathematically by two key functions: the so-called selection function (breakage probability) and the breakage function (fragment size distribution). As expected, higher impact velocities lead to more damage, especially at normal collisions (90°) due to the maximum change of momentum. Furthermore, smaller particles tend to be more breakage resistant as they contain less impurities and cracks. Finally, an outlook on the influence of pellet quality and target material on the particles' degradation behaviour is given. Here, pellets with higher durability tend to break at higher impact loads and into larger fragments. In addition, a softer target material (e.g. HDPE) causes less particle breakage than e.g. steel.

Introduction

Wood pellets are among the most popular and widely used commercial types of biomass fuels [1]. Thanks to their advantageous storage and handling properties, high energetic density and comparably homogeneous particle shapes and sizes, they are suitable for industrial heat and power generation as well as for decentralized domestic heating systems [2].

Wood pellets are typically delivered by blowing trucks, pneumatically conveyed into local storages and thus exposed to various mechanical loads. These mechanical impacts cause attrition and abrasion and lead to unintended pellet breakage and an increase of fines [3]. The fines produced lead to operational problems and system failures during further transportation steps and storage, may increase the risk of dust explosions and result in higher pollutant emissions during combustion [4,5].

The level of particle fragmentation during pneumatic conveying is affected by multiple parameters such as particle's material properties, the occurring inter-particle and particle-wall collisions and the particles' motion behaviour during the conveying process [6]. Identifying the influence of these individual parameters is quite challenging when considering realistic conveying line situations.

Single particle impact studies can help to resolve this problem by separating the influence of the relevant factors. Recent studies [[7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20]] examine the fragmentation behaviour of individual spherical particles. The intention of these studies is to develop models to predict breakage probability and characteristics, based on empirical correlations.

Salman et al. [7,8] carried out single particle impact studies investigating the influence of impact velocity, collision angle, the mother particle size and the target material using the example of spherical Al2O3-particles of different sizes already in 2002. These tests served to develop an empirical model based on Weibull distributions to predict both breakage probability (selection function) and size distribution (breakage function) of the resulting fragments.

A more comprehensive modelling approach supported by single particle breakage experiments was presented by the research group led by Kalman (e.g. [[9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20]]). Their method is generally based on two types of functions: the process functions (also known as machine functions) and the comminution functions [21]. The concept is sketched in Fig. 1. The machine functions describe the particle dynamics in a specific system. They provide the statistics of particle-particle or particle-wall collisions including velocity and angle of the impact. These data can be obtained from preceding DEM-CFD simulations of a certain system [9,10,22]. Comminution functions define, for example, the probability of breakage (selection function), the size distribution of the daughter particles (breakage function) or the strength deterioration due to a preceding particle contact (fatigue function). These functions can then be combined with, for example, a one-dimensional two-phase model of the flow field to predict particle breakage for a given system as demonstrated in [20]. This approach avoids conducting full 3D-CFD-DEM breakage simulations, which is quite expensive.

Based on this concept, Portnikov et al. [15] recently predicted the size reduction due to elbows during pneumatic conveying by machine functions employing additional parameters such as distributions for collision angle and particle position inside bends without DEM-CFD just on the basis of empirical models.

The studies mentioned above reveal that higher impact velocities lead to more particle degradation and smaller particles have a lower tendency to break due to their lower number of impurities and cracks. Furthermore, the normal collision (impact angle 90°) causes the greatest damage due to the maximum change of momentum caused by the impact. The data indicate that the damage is approximately equal between 50 and 90°, while at angles below 50° the particle damage decreases significantly. This behaviour could be predicted in all studies with satisfactory accuracy by the empirical models developed, so that this procedure appears to be suitable for prediction of particle breakage.

Note that apart from the studies of Salman et al. [7,8] and Portnikov et al. [14,15], the influence of the collision angle has not been considered so far. Additionally, only the fragmentation behaviour of (roughly) spherical and mostly homogeneous particles has been examined. Until now, non-spherical, e.g. cylindrical particles such as wood pellets with additional comparatively inhomogeneous structure have not been considered at all.

Therefore, this paper concentrates on the fragmentation behaviour of cylindrical wood pellets including the effect of collision angle. It covers the influence of impact velocity, initial particle size, target material and pellet quality. Using a test rig for single particle impact tests, wood pellets of different lengths were accelerated towards a target plate at various velocities and angles. Particle trajectory and velocity were recorded for each particle using a stereoscopic pair of high-speed cameras. For the current paper, we concentrate on deriving a) the selection function (Breakage probability = BP) and b) the breakage function (BF) for wood pellets as the basis for further model development.

Section snippets

Experimental setup

The working principle of the designed single particle impact test facility is sketched in Fig. 2.

Up to 100 pellets (maximum length and diameter 50 and 6 mm, respectively) can be stored in the vertical 100 bores of the rotating pellet depot. By rotating the depot bore by bore, every single pellet drops through a short tube into the rotation unit mounted below (Fig. 2, green circle). Here, the pellet is rotated from vertical to horizontal position (shaft diameter 50 mm, bore diameter 7 mm). In

Results and discussion

The aim of this contribution is the statistical description of the degradation by using the concept of breakage probability and a breakage function. The necessary (mostly material-specific) model parameters are determined by model fitting using the data obtained experimentally. The accuracy of the fitting is evaluated by the coefficient of determination (R2), a benchmark how accurately the least squares fit curve represents the experimental data (values close to 1 correspond to high accuracy).

Conclusions

The current study deals with the breakage behaviour of wood pellets under impact loads. It includes the development or extension of an empirical model for predicting the particles breakage probability and resulting fragment size distribution in dependence on particle length, impact velocity and collision angle. Existing models suitable for spherical and homogeneous particles were adopted and modified to be also valid for cylindrical inhomogeneous wood pellets. For experimental analysis, a test

Declaration of Competing Interest

None.

Acknowledgements

The IGF-Project 20565 N of the DVV e.V. was funded by the German Federation of Industrial Research Association (AiF) on behalf of the German Federal Ministry of Economics and Energy.

References (41)

Cited by (4)

  • A novel oblique impact model for unified particle breakage master curve

    2023, Chemical Engineering Science
    Citation Excerpt :

    The experimental database is composed of breakage results from three publications (Jägers et al., 2021; Portnikov et al., 2018; Salman et al., 2002). The test materials under oblique impact include 2.36-3.35 mm salt particles (Portnikov et al., 2018), 3.2 mm fertiliser particles (Salman et al., 2002), and 15 mm wood pellet (Jägers et al., 2021). Note that the breakage probability in the database from (Salman et al., 2002) is defined by the number of unbroken particles.

  • Numerical simulation of feed pellet breakage in pneumatic conveying

    2022, Biosystems Engineering
    Citation Excerpt :

    Fulchini et al. (2019) also used the CFD-DEM method to analyse the breakage of manganese oxide particles in cyclone. Jägers et al. (2021a, 2021b) investigated the degradation of wood pellets based on a single particle impact test. The degradation of wood pellet in pipeline was analysed through a CFD-DEM approach.

  • DEM-CFD simulation of wood pellet degradation by particle-wall impact during pneumatic conveying

    2021, Powder Technology
    Citation Excerpt :

    Based on these experiments, they determined the degradation-describing statistical material functions (selection and breakage function) for cylindrical wood pellets. In the current contribution, the degradation model of Jägers et al. [32], in terms of selection and breakage functions for particle-wall impacts, is implemented in a coupled DEM-CFD approach. Direct breakage simulation is performed.

View full text