Show Summary Details
More options …
New at De Gruyter

# Nordic Pulp & Paper Research Journal

### The international research journal on sustainable utilization of forest bioresources

Editor-in-Chief: Lindström, Tom

IMPACT FACTOR 2017: 1.131
5-year IMPACT FACTOR: 1.262

CiteScore 2017: 1.18

SCImago Journal Rank (SJR) 2017: 0.623
Source Normalized Impact per Paper (SNIP) 2017: 0.555

Online
ISSN
2000-0669
See all formats and pricing
More options …
Volume 33, Issue 1

# Indications of the onset of fiber cutting in low consistency refining using a refiner force sensor: The effect of pulp furnish

R. Harirforoush
• Corresponding author
• University of Victoria, Department of Mechanical Engineering, P.O. Box 3055 STN CSC, V8W 3P6, Victoria, BC, Canada
• Email
• Other articles by this author:
/ J. Olson
• University of British Columbia, Department of Mechanical Engineering, 2054-6250 Applied Science Lane, V6T 1Z4, Vancouver, BC, Canada
• Email
• Other articles by this author:
/ P. Wild
• University of Victoria, Department of Mechanical Engineering, P.O. Box 3055 STN CSC, V8W 3P6, Victoria, BC, Canada
• Email
• Other articles by this author:
Published Online: 2018-05-23 | DOI: https://doi.org/10.1515/npprj-2018-3013

## Abstract

Detection of the onset of fiber cutting is beneficial in low consistency refining as it may prevent reduction of average fiber length, optimize fiber quality improvements by operating at gaps just wider than the critical gap, avoid decreasing the strength properties of paper, and increase energy efficiency. The objective of this study is to understand the effect of pulp furnish on measured bar forces and, more specifically, on the detection of fiber cutting. Bar forces, i. e. forces applied to pulp fibers by the refiner bars, are measured with a custom-designed piezoelectric force sensor. Trials were conducted with an AIKAWA 16-in. single-disc refiner using hemlock/balsam softwood thermomechanical pulp, SPF softwood thermomechanical pulp, northern bleached softwood kraft pulp, and aspen hardwood thermomechanical pulp at 3.0 to 3.5 % consistency at rotational speeds of 1200 and 1400 rpm. The power of the time domain signal of the measured forces is introduced as an indicator of the onset of fiber cutting. Our results show that this new fiber cutting metric is a sensitive and reliable metric for determination of fibre cutting for a range of pulp furnishes. The study suggests that the refiner force sensor has potential to be exploited for in-process detection of fiber cutting.

## Introduction

A crucial parameter in the production of mechanical pulp through refining is energy consumption. The specific energy consumption of refiner mechanical pulp, defined as the total energy expended per unit mass of fiber, is typically between 1600 and 3000 kWh/t (Bajpai 2016), and the majority of this energy is consumed in refining. In British Columbia (BC), 78 refiners consume 5400 GWh/y or 11 % of BC’s total electrical energy production (Luukkonen et al. 2010a). With the increasing cost of energy, the long term survival of mechanical pulping may depend on reducing energy consumption of this process.

One method of reducing energy consumption in mechanical pulping is to decrease the number of refining treatments at high consistency (HC) and increase the number of treatments at low consistency (LC). LC refining of mechanical pulp has been shown to be more energy efficient than conventional HC refining (Muenster et al. 2005, Zha et al. 2008). A reduction of 5–8 % in specific energy has been reported without loss in pulp quality (Musselman et al. 1997). It is hypothesized that the uniformity of treatment of LC refining causes the increased efficiency of this approach (Luukkonen et al. 2010a). Moreover, LC refining is effective for removal of shives, external fibrillation of fibers, removal of latency and improving fiber networking and bonding capabilities (Kappel 1999, Smook 1992). However, the degradation of mechanical properties due to fiber cutting at high refining energies has limited the widespread adoption of LC refining. Understanding the relation between refiner operating conditions and resulting pulp properties is essential to reduce fiber cutting.

In mechanical refining, as the plate gap is closed, fibers are subjected to increasing compressive and shear forces. This leads to increasing net refiner power and Specific Edge Load (SEL), which is defined as the energy expended per unit refiner bar length over one bar-crossing event. SEL, introduced by Brecht (1967), is widely used in industry to characterize refining performance. However, Roux and Joris (2005) showed that SEL cannot be used to predict fiber cutting since the effect of plate geometry parameters (i. e. bar width, groove width, and bar angle) are not considered in this theory.

Modified edge load (MEL) theory, developed by Meltzer (1994), is an improvement over SEL theory that includes plate geometry parameters which are absent from SEL theory. However, MEL theory does not account for some important parameters such as consistency, rotational speed, groove depth, and fiber properties (Senger 1998). The relationship between MEL and the length-weighted fiber length indicates that the theory does not predict changes in fiber length (Elahimehr et al. 2015).

Roux and Mayade (1999) developed a comminution model for the kinetics of fiber shortening in LC refining. The model predicts the potential of fiber cutting in the given conditions as a function of the energy per unit mass consumed by the solid phase and the average impact intensity. Olson et al. (2003) also investigated the use of comminution model of LC refining to calculate fiber cutting rate as a function of fiber length for a wide range of operating conditions. They concluded that the probability of fiber cutting during refining is proportional to the applied energy and to the fiber length.

Theoretical net tangential and net normal force per number of bar-crossing have also been proposed to assess refining intensity and to predict changes in fiber length (Roux et al. 2009). The correlation between the average weighted fiber length and these two parameters indicates that the net normal force per number of bar-crossing is adequate to quantify and predict the cutting effect on fibers.

Based on pilot and mill-scale LC refining studies, Luukkonen et al. (2010a) found that the plate gap is a key link between the refiner process variables and resulting changes in pulp properties. Luukkonen et al. (2010b) also proposed a methodology to correlate operating conditions to pulp quality by relating the refiner operating conditions (i. e. power, flow rate, rotational speed and plate gap) to fiber quality (i. e. fiber length and freeness) and then relating fiber quality to pulp handsheet quality (i. e. tensile index, tear index, and bulk). They also showed that pulp quality remains constant as the gap is closed, up to a critical gap which was first introduced by Roux (2001). Beyond the critical gap, fiber cutting commences and paper strength is reduced. Moreover, beyond this point, the relationship between refining power and plate gap deviates from linearity (Luukkonen 2011) which is consistent with the results in Chaussy et al. (2011).

Similar results were observed by Nugroho (Nugroho 2012) for various mixes of softwood (SW) and hardwood (HW) chemical pulps in LC refining. Refining power increases as the plate gap decreased and, at a critical gap between 0.3 to 0.5 mm, the power increases sharply. Moreover, the fiber length and freeness decrease as the plate gap is reduced below the critical gap. This transition occurs for all of the trial conditions.

Elahimehr et al. (2013) studied the effect of plate gap, rotational speed of refiner, and plate pattern on LC refining performance in an AIKAWA 16-inch pilot-scale LC refiner. They found that the non-dimensional power plate gap relationship is a function of refiner plate geometry and pulp furnish. They also found that mean fiber length was constant for gaps larger than a critical gap, near 0.25 mm. The mean fiber length significantly decreased for the plate gaps smaller than the critical gap. Results also show that increasing the rotational speed increases the critical gap. In addition, Elahimehr et al. (2015) defined intensity as the net energy per leading edges of bar-crossing to predict fiber length reduction in LC refining of mechanical pulp. They found that fiber cutting occurs at 0.03 J/m for all plates and rotational speeds.

Berg et al. (2015) found that power increases linearly with decreasing the plate gap over the range of 0.1–0.2 mm in a two-zone TwinFlo72 LC refiner. They calculated average forces applied to fibers using an expression developed by Kerekes and Senger (2006). Fiber cutting was defined as the point where the ratio of outlet to inlet fiber length was equal to 0.95. Plate gap, calculated forces on fibers, and SEL were found to predict fiber cutting.

In all of the studies discussed above, the onset of fiber cutting is detected by post-refining measurement of pulp properties. Typically, this approach does not enable rapid in-process adjustment of refiner operation in response to the onset of fiber cutting. Nor do these studies address the mechanical interactions between fibers and refiner bars. Kerekes (2011) suggests that bar force (i. e. the force applied to pulp fibers by the refiner bars) is the key parameter between bars and fibers in refining that cause important changes in fiber properties.

A piezoelectric sensor, referred to here as the refiner force sensor (RFS), has been developed to measure the forces applied by refiner bars to the pulp. The RFS has a probe which replaces a short length (i. e. 5 mm) of a refiner bar. Forces normal to and tangential to this probe are measured by the RFS, as explained in (Prairie 2005) and in (Olender et al. 2008a). The RFS has been used in a number of studies in both low consistency (Prairie et al. 2007, 2008) and high consistency refining (Olender et al. 2008b, 2007).

Harirforoush et al. (2016) investigated the relation between net power, gap and forces on bars in a LC refining using the RFS in experiments conducted on AIKAWA 16-inch single-disc refiner. As the plate gap is reduced, mean peak normal force and mean peak shear force increase up to threshold values. Here, peak refers to the maximum value that occurs as a bar on the rotor passes over the sensor (Harirforoush et al. 2016, Olender et al. 2007). Beyond these thresholds, the mean peak normal and shear forces continue to increase but the length-weighted fiber length exhibits a negative linear relationship with these forces. In a subsequent study, Harirforoush et al. (2017) found distinct transitions in the parameters characterizing the distributions of bar forces which consistently corresponded to the onset of fiber cutting.

Figure 1

(a) RFS sensor (b) close-up of the stator plate with the RFS installed.

The objective of this study is to understand the effect of pulp furnish on measured bar forces and, more specifically, on the detection of the onset of fiber cutting using the RFS. Tests are conducted using an AIKAWA 16-inch single-disc refiner using hemlock/balsam SW thermomechanical pulp (TMP), SPF (spruce, pine, fir) SW TMP, NBSK (northern bleached SW kraft) pulp, and aspen HW TMP at 3.0 to 3.5 % consistency at 1200 and 1400 rpm. A parameter used in signal processing, the power of the time domain signal of the measured forces, is shown to be a sensitive and consistent indicator of the onset of fiber cutting. It is also shown that the fiber cutting indicators previously used in (Harirforoush et al. 2017) can be extended to different pulp furnishes and operating conditions.

## Materials and methods

Trials were performed using the AIKAWA 16-inch single-disc refiner at the Pulp and Paper Centre at the University of British Columbia, BC, Canada. The refiner is equipped with a power meter, plate actuation and a variable speed drive. The plate used in these trials is manufactured by AIKAWA FINEBAR®, and has bar edge length of 2.74 km/rev and bar angle of 15° from radial. The bar width, groove width, and groove depth of the plate are 1.6 mm, 3.2 mm, and 4.8 mm, respectively.

Two RFS-type sensors were custom designed and fabricated, based on the design previously used in refiner trials (Olender et al. 2008a). The RFS used in this work is shown in Figure 1a. The sensors are installed in a custom recess that are machined into the back of the stator plate. The sensor wires pass from the back side of the sensor and through a custom gland located in the door of the refiner. The sensor recesses are located so that the tip of the sensor replaces a short length of the refiner bar (i. e. 5 mm), as shown in Figure 1b. The sensor measures force normal to the plate surface (i. e. parallel to the axis of the refiner) and the shear force normal to the bar edge and in the plane of the refiner plate. Two sensors are located on opposite sides of the stator plate. The sensors are positioned in the second bar of a three-bar cluster at a radius of 151 mm. At this radius, the heads of the bolts do not pass over the sensor. 144 bars on the rotor, grouped into 48 three-bar clusters, cross over the sensor in one revolution of the rotor.

Calibration and environmental tests were performed as described in (Harirforoush et al. 2016, Prairie 2005). Modal analyses were performed to ensure that the first natural frequency of the sensor is well above the bar-passing frequency of the refiner so that phase and magnitude distortion due to sensor resonance is minimised. The ratio of the first natural frequency of the sensor to the maximum bar-passing frequency is 8.2 which provides a sufficient margin above the bar passing frequency.

A custom charge amplifier converts the high impedance charge output of the sensors to a low impedance voltage signal which is read by a National Instruments (Austin, TX, USA) PXI 1042 Q high-speed board. National Instrument BNC-2110, Shielded Connector Blocks are used to connect the high-speed DAQ board to the charge amplifier. A custom LabVIEW™ (National Instruments; Austin, TX, USA) interface is used to control data acquisition. A sampling rate of 150 kHz is used which is more than twenty times the maximum bar-passing frequency that occurs during the trials.

Four pulp types, hemlock/balsam SW TMP, SPF SW TMP, NBSK, and aspen HW TMP, at 3 to 3.5 % consistency were used in all trials. Pulp freeness and operating conditions for the trials are shown in Table 1. The refining operating mode was mono-flo with one inlet perpendicular to the outlet. The refiner speed was set with a Variable Frequency Drive. The plate gaps were adjusted from no-load position (3.5 mm) to the smaller plate gaps (0.2 mm). The flow rate was held constant at 250 liter/min. Sensor signals were recorded using the LabVIEW™ interface for approximately 10 s once reaching stable condition at each target point. At each operating condition, the refiner was held steady for about 10 s to give the pulp time to move from the refining zone to the sample collection point. A ten-liter pulp sample was collected over a 5 s interval. The process was repeated until all of the target points were achieved for one rotational speed. Note that the refining machine operated in continuous mode and the pulp was not recirculated.

Table 1

Operating conditions for the pilot-scale refining trials.

All pulp samples were collected and measured for length-weighted fiber length (${\mathit{L}}_{\mathit{w}}$) using a Fiber Quality Analyzer (HiRes FQA, Optest Equipment Inc.; Hawkesbury, ON, Canada), and freeness, defined as the drainage resistance of pulp slurry, (using TAPPI standard T227). Handsheets were made and paper properties of bulk, tear index and tensile index (using TAPPI standard T220) were measured. The calculation of length-weighted fiber length is presented in (Carvalho et al. 1997).

The critical gap is identified based on fiber length data and Equations 12: $(Lw)i+1−(Lw)i(Lw)i<−2%i=1,…,n−1$(1)$(Lw)i+2−(Lw)i+1<0(Lw)i+3−(Lw)i+2<0(Lw)i+4−(Lw)i+3<0$(2) In these equations, ${\mathit{L}}_{\mathit{w}}$, is length-weighted fiber length, i, is the index of data point, and n is the total number of data points at each trial. The data point index,i, ascends as the gap is closed. When the conditions presented in Equations 1 and 2 are both satisfied, then fibre cutting is determined to have begun.

The bar force data is processed to identity the passage of rotor bars over the sensor. A bar passing event is defined to have occurred when a maximum or peak value in the force data exceeds a predefined threshold value. The algorithm and its verification are described in (Harirforoush et al. 2016).

For each bar-passing event, peak force is defined as the difference between the force at the base of the local valley that precedes the peak and the force at the peak (Prairie et al. 2007). The mean peak normal force and mean peak shear force are calculated for each operating condition. The peak coefficient of friction is calculated by dividing the peak shear force by the peak normal force. The peak normal and shear force data are assessed in the form of distributions to which the two-parameter Weibull distribution function is fit. The Weibull distribution function has been used in previous HC and LC refining studies (Olender et al. 2007, Prairie et al. 2008).

The average power (P) of a time domain signal $\mathit{f}\left(\mathit{t}\right)$ can be defined as the sum of absolute squares of time-domain samples divided by the signal length, and defined by Equation 3. (Papoulis 1962): $P=limT→∞12T∫−TTft2dt$(3) where T is the signal length. Note that this term is not equal to the mechanical power that is dissipated on the sensor probe. Rather this is a signal processing term which is equivalent to the square of RMS (root-mean-square) magnitude of a signal.

In this study, the power of the normal force signals and the shear force signals are calculated at each operating condition. As the gap is closed, a transition in the relation between the power of the normal force signal and the plate gap is identified when the conditions shown in Equations 45 are satisfied: $Δj=Pnormj+1−Pnormj>0.01;j=1,…,n−1$(4)$Δj+1>0.01Δj+2>0.01$(5) where ${\left({\mathit{P}}_{\mathit{n}\mathit{o}\mathit{r}\mathit{m}}\right)}_{\mathit{j}}$ is the normalized power of normal force and is defined by Equation 6: $Pnormj=Pj−minP1…PimaxP1…Pi−minP1…Pii=1,…,n−1$(6) In these equations, ${\mathit{P}}_{\mathit{j}}$ is the power of normal force signal, i and j are the indexes of data point, and n is the total number of data points at each trial.

## Results

The relation between net power and the inverse of plate gap for each pulp furnish is shown in Figure 2. The net power increases as the plate gap is reduced. A considerable increase in net power is seen for the plate gaps of less than 0.5 mm ($2\phantom{\rule{0.1667em}{0ex}}{\text{mm}}^{-1}$). At a constant gap, increasing the rotational speed increases the net power. This trend is similar to the results presented in (Elahimehr et al. 2013). Moreover, for all plate gaps, more energy is transferred to SW pulp than to HW pulp.

Figure 2

Net power versus the inverse of plate gap for hemlock/balsam SW TMP (diamond), SPF SW TMP (circle), NBSK (triangle), and aspen HW TMP (hexagram).

Figure 3

${\mathit{L}}_{\mathit{w}}$ versus the inverse of plate gap for (a) hemlock/balsam SW TMP (b) SPF SW TMP at 1200 rpm (c) SPF SW TMP at 1400 rpm (d) NBSK (e) aspen HW TMP at 1200 rpm, and (f) aspen HW TMP at 1400 rpm.

The relation between ${\mathit{L}}_{\mathit{w}}$ and the inverse of plate gap for all pulp furnishes is shown in Figure 3. In each case, ${\mathit{L}}_{\mathit{w}}$ remains relatively unchanged as the plate gap is reduced until the plate gap reaches the critical values, indicated by red dashed lines. The critical gaps, for all but the NBSK pulp, are determined based on an algorithm expressed in Equations 12.

For the NBSK pulp (Figure 3d), the critical gap defined by this algorithm is indicated by the circled data point. However, inspection of the data suggests that the data point indicated by the red dashed line is more likely to be the point of transition. During the NBSK pulp trials, the flow rate was unstable and this is thought to be the cause of this anomalous data.

${\mathit{L}}_{\mathit{w}}$ decreases in an approximately linear manner as a function of the inverse of plate gap for plate gaps smaller than the critical gap. This trend is in agreement with the results of (Harirforoush et al. 2016, 2017). In addition, our results show that the critical gap depends on both pulp furnish and rotational speed and it occurs at different SEL values. For example, the SEL at the onset of fiber cutting is 0.31 J/m for hemlock/balsam SW TMP while it is 0.42 J/m for SPF SW TMP.

Figure 4 shows the relation between the power of normal force and plate gap for all pulp furnishes. Note that the x-axis of these plots is the plate gap not the inverse of plate gap. This format is consistent with the format used in Harirforoush et al. (2017) to show the transitions that indicate fiber cutting.

For all pulp furnishes, as the plate gap is reduced, the power of normal force is relatively constant down to a gap at which there is a transition below which the power increases sharply. This transition gap, based on the algorithm expressed in Equations 45, is indicated by the blue dotted line. The critical gap, based on the fiber length data (see Equations 12 and Figure 3) is indicated by the red dashed line. For each pulp type, except the NBSK pulp (Figure 4d), there is an agreement between the transition gap and the critical gap. The difference between the critical gap and the transition gap for the NBSK pulp, shown in Figure 4d, is believed to be due to instability of the flow rate during these tests.

The relation between freeness, tear index, and tensile index at 1200 rpm is depicted in Figure 5. As it is shown in Figureure 5a and 5b, freeness changes with decreasing plate gap. This is consistent with the results of (Elahimehr et al. 2015, Luukkonen et al. 2010b) who showed the changes of freeness with specific energy. A substantial reduction of freeness occurs for the plate gaps smaller than the critical gap highlighted by dashed lines.

In Figure 5a, the tear index is relatively constant as the plate gap is decreased until the critical gap. After the critical gap, fibers start to cut and tear index drops considerably. This is in agreement with (Elahimehr et al. 2015). However, the trend is not apparent for aspen HW TMP, Figure 5b. In addition, the trend of decreasing tear index with increasing tensile index, Figure 5a, is in agreement with the results of (Luukkonen et al. 2010a). Similar trends to those shown in Figure 5a were observed for SPF SW TMP and NBSK pulp but the results are not presented here in the interest of brevity.

The relation between ${\mathit{L}}_{\mathit{w}}$ and mean peak normal force for hemlock/balsam SW and aspen HW at 1200 rpm is shown in Figure 6. As shown in Figure 6a, ${\mathit{L}}_{\mathit{w}}$ remains constant as mean peak normal force increases (i. e. as the gap is closed) up to a threshold value, indicated by the dashed circle, which corresponds to the critical gap, as determined by Equations 12. ${\mathit{L}}_{\mathit{w}}$ decreases in an approximately linear manner beyond the critical gap. This trend is consistent with the results of (Harirforoush et al. 2016, 2017). The mean peak normal force at the critical gap is approximately 9 N. This threshold normal force is significantly higher than the equivalent force (i. e. 3 N) for the aspen HW pulp, Figure 6b.

Figure 4

Power of normal force versus plate gap for (a) hemlock/balsam SW TMP (b) SPF SW TMP at 1200 rpm (c) SPF SW TMP at 1400 rpm (d) NBSK (e) aspen HW TMP at 1200 rpm, and (e) aspen HW TMP at 1400 rpm.

Figure 5

Freeness (circle), tear index (rectangle), and tensile index (square) versus the inverse of plate gap for (a) hemlock/balsam SW TMP and (b) aspen HW TMP at 1200 rpm.

Figure 6

${\mathit{L}}_{\mathit{w}}$ versus mean peak normal force for (a) hemlock/balsam SW TMP, and (b) aspen HW TMP at 1200 rpm.

A similar trend is seen in the relationship between ${\mathit{L}}_{\mathit{w}}$ and mean peak shear force (Figure 7). However, the shear force at the critical gap is significantly lower than the threshold of mean peak normal force. Comparing the threshold values of Figure 7a, and 7b, the mean peak shear force at the critical gap is of similar magnitude for both HW and SW pulps, approximately 1.4 N. If the average shear force is assumed to be one-half of the peak shear force, the average shear force per length of sensor tip is in relatively good agreement with Kerekes and Meltzer (2017) who showed that for zero length reduction, the threshold of bar shear force is 200 N/m.

The trends shown in Figure 6 and 7 also appear in the relation between ${\mathit{L}}_{\mathit{w}}$ and mean peak normal and shear forces for SPF SW TMP, and NBSK pulp. These results are not presented here due to the interest of brevity. However, as tabulated in Table 2, the mean peak force at the onset of fiber cutting depends on pulp furnishes.

Table 2

Mean peak normal force, mean peak shear force and coefficient of friction at the onset of fiber cutting for all pulp furnishes at 1200 rpm.

The mean coefficient of friction is plotted in Figure 8 as a function of the inverse of plate gap. As shown in Figure 8a, the mean coefficient of friction decreases as the plate gap is reduced. This finding is in accordance with Roux (2001), Senger et al. (2004), and Harirforoush et al. (2017). The trend indicates that by decreasing the plate gap, the changes in mean peak normal force are more significant than the changes in mean peak shear force. It can be hypothesized that the changes in pulp and paper properties that are associated with closure of the gap, in particular those properties affected by fiber cutting, are more strongly related to an increase in compressive force than to an increase in shear force. As an aside, the results show that the mean coefficient of friction at the critical gap depends on the pulp furnish. This value, as tabulated in Table 2, is 0.22, 0.54, and 0.41 for hemlock/balsam SW TMP, SPF SW TMP, and NBSK, respectively.

Similar trends are observed for aspen HW TMP, Figure 8b. However, the mean coefficient of friction increases up to the plate gap of 0.65 mm, and decreases beyond this point. For this pulp furnish, the onset of fiber cutting occurs when the mean coefficient of frictions equals to 0.49. At the same plate gap, the mean coefficient of friction is relatively constant at two different rotational speeds, Figure 8b.

Figure 7

${\mathit{L}}_{\mathit{w}}$ versus mean peak shear force for (a) hemlock/balsam SW TMP, and (b) aspen HW TMP at 1200 rpm.

Figure 8

Mean coefficient of friction versus the inverse of plate gap for (a) hemlock/balsam SW TMP, SPF SWTMP, NBSK, and (b) aspen HW TMP.

Our results also show that the scale parameter of Weibull distribution of peak forces, and the magnitude of dominant frequency, previously used as the indications of the onset of fiber cutting (Harirforoush et al. 2017), can be employed for different pulp furnishes. This data is not shown in of the interest of brevity.

## Discussion

In a previous work (Harirforoush et al. 2017), we observed distinct transitions in parameters that characterize the distributions of peak normal and shear forces, namely mean peak force and the Weibull scale parameter, as measured by the RFS during LC refining. These transitions consistently correspond to the onset of fiber cutting. In addition, the analysis of the power spectrum of the sensor data shows that the magnitude of the dominant frequency can be used as an indicator of fiber cutting.

In the current study, we investigate the generality of these findings by performing further trials measuring forces during LC refining of a range of pulp furnishes (i. e. SW, HW, TMP and chemical pulp). The results show that the RFS-based indications of the onset of fiber cutting apply to all of the tested pulp furnishes.

The power of the time domain signal of the normal force is shown to be a reliable and consistent indication of the onset of fiber cutting. This parameter consistently identifies the critical gap, as determined by fibre length data, for all pulp furnishes except the NBSK pulp. Flow instability during the NBSK trials is believed to account for this anomalous result.

This study also shows that the mean peak normal and shear force at the onset of fiber cutting depend on pulp furnish. At the onset of fiber cutting, the mean peak normal force of the SW pulp is much higher than for the HW pulp. The level of compressive strain in fibers imposed by fiber forces depends on the properties of the fibers (i. e. modulus of elasticity and fiber geometries such as length, diameter and wall thickness) as well as applied forces (Kerekes and Senger 2006). HW and SW have quite different geometries (i. e. SW fibers are longer than HW fibers) which result in different normal forces at the onset of fiber cutting. Kerekes and Senger (2006) estimate that, for low consistency refining, the normal force on a fiber is proportional to fiber length taken to the power of 3/2 and uncompressed fiber diameter.

However, the mean peak shear force for HW and SW pulps at the onset of fiber cutting are approximately equivalent. The shear force that is generated as opposed bars cross is comprised of two components (Batchelor et al. 1997): a corner or ploughing force at the bar edge and a friction force on axial-facing bar surface. Due to the dissimilar geometries of HW and SW fibres, we expect that the magnitude of the corner force for HW pulp will be different from the corner force of SW pulp. Similarly, we expect that the magnitude of the friction force for HW pulp will be different from the friction force for SW pulp. Our results suggest, however, that these two forces vary such that their sum is approximately equivalent for HW and SW pulps, at the onset of fiber cutting.

The mean coefficient of friction at the critical gap also depends on the pulp furnish. At 1200 rpm, the mean coefficient of friction is 0.22, 0.54, 0.41, and 0.49 for hemlock/balsam SW TMP, SPF SW TMP, NBSK, and aspen HW TMP, respectively. For all pulps, the coefficient of friction decreases as the plate gap is closed. Since the coefficient of friction is the ratio of peak shear force to peak normal force, it is hypothesized that the changes in pulp and paper properties that are associated with closure of the gap and, in particular, those properties affected by fiber cutting, may be caused by an increase in compressive forces rather than by an increase in shear forces on fibers.

## Conclusion

This study investigates the effect of pulp furnish on measured bar forces and, more specifically, on the detection of fiber cutting using a custom designed piezoelectric force sensor measuring bar forces in an AIKAWA 16-in. single-disc refiner. Trials were performed using hemlock/balsam softwood thermomechanical pulp, SPF softwood thermomechanical pulp, northern bleached softwood kraft pulp, and aspen hardwood thermomechanical pulp at 3.0 to 3.5 % consistency at rotational speeds of 1200 and 1400 rpm. The pulp was sampled at regular intervals, and the length-weighted fiber length, freeness, tear index, and tensile index were measured for each sample. Distinct transitions occur in the plot of the power of time domain signal of measured normal force versus plate gap which consistently identifies the critical gap, as determined by fibre length data. This indication is a sensitive and reliable metric for in-process detection of fiber cutting. Moreover, our results show that at the onset of fiber cutting, the mean peak normal force, mean peak shear force, and mean coefficient of friction depend on pulp furnish. The results of this study suggest RFS as a potential sensor to detect the onset of fiber cutting.

## Acknowledgments

The authors wish to thank M. Miller, E. Jahangir, R. Seifert, and G. Soong for their assistance during the preparation and execution of the refining trials, and for conducting the sample characterisations at the Pulp and Paper Centre at the University of British Columbia.

## References

• Bajpai, P. (2016) Pulp and Paper Industry: Energy Conservation, Elsevier. Google Scholar

• Batchelor, W.J., Martinez, D.M., Kerekes, R.J., Ouellet, D. (1997) Forces on fibres in low-consistency refining: Shear force. J. Pulp Pap. Sci. 23(1):J40–J45. Google Scholar

• Berg, J.-E., Sandberg, C., Engberg, B.A., Engstrand, P. (2015) Low-consistency refining of mechanical pulp in the light of forces on fibres. Nord. Pulp Pap. Res. J. 30(2):225–229.

• Brecht, W. (1967) A method for the comparative evaluation of bar-equipped beating devices. Tappi J. 50(8):40A–44A. Google Scholar

• Carvalho, M.G., Ferreira, P.J., Martins, A.A., Figueiredo, M.M. (1997) A comparative study of two automated techniques for measuring fiber length. Tappi J. 80(2):137–142. Google Scholar

• Chaussy, D., Martin, C., Roux, J.-C. (2011) Rheological behavior of cellulose fiber suspensions: Application to paper-making processing. Ind. Eng. Chem. Res. 50(6):3524–3533.

• Elahimehr, A., Olson, J.A., Martinez, D.M. (2013) Understanding LC refining: The effect of plate pattern and refiner operation. Nord. Pulp Pap. Res. J. 28(3):386–391.

• Elahimehr, A., Olson, J.A., Martinez, D.M. (2015) Low consistency refining of mechanical pulp: how plate pattern and refiner operating conditions change the final properties of pulp. Nord. Pulp Pap. Res. J. 30(4):609–616.

• Harirforoush, R., Olson, J., Wild, P. (2017) In-process detection of fiber cutting in low consistency refining based on measurement of forces on refiner bars. Tappi J. 16(4):189–199. Google Scholar

• Harirforoush, R., Wild, P., and Olson, J. (2016) The relation between net power, gap, and forces on bars in low consistency refining. Nord. Pulp Pap. Res. J. 31(1):71–78.

• Kappel, J. (1999) Mechanical Pulps: From Wood to Bleached Pulp, Tappi press, Atlanta. Google Scholar

• Kerekes, R.J. (2011) Force-based characterization of refining intensity. Nord. Pulp Pap. Res. J. 26(1):14–20.

• Kerekes, R.J., and Meltzer, F.P. (2017) The influence of bar width on bar forces and fibre shortening in pulp refining. In: PTS Faserstoff Symposium, Dresden, Germany. Google Scholar

• Kerekes, R.J., Senger, J.J. (2006) Characterizing refining action in loconsistency refiner by forces in fibres. J. Pulp Pap. Sci. 32(1):1–8. Google Scholar

• Luukkonen, A. (2011) Development of a methodology to optimize low consistency refining of mechanical pulp. Ph.D. thesis, University of British Columbia. Google Scholar

• Luukkonen, A., Olson, J.A., and Martinez, D.M. (2010a) Low consistency refining of mechanical pulp, effect of gap, speed and power. J. Pulp Pap. Sci. 36:28–34. Google Scholar

• Luukkonen, A., Olson, J.A., and Martinez, D.M. (2010b) Low consistency refining of mechanical pulp: A methodology to relate operating conditions to paper properties. J. Pulp Pap. Sci. 36:107–111. Google Scholar

• Meltzer, F.P. (1994) Neue Möglichkeiten zur Vorherbestimmung des technologischen Mahlergebnisses. Papier 48:578–583. Google Scholar

• Muenster, H., Ferritsius, O., Lecourt, M., and Petit-Conil, M. (2005) Energy savings in TMP by high temperature LC/MC refining. In: International Mechanical Pulping Conference, pp. 213–223. Google Scholar

• Musselman, R., Letarte, D., and Simard, R. (1997) Third stage low consistency refining of TMP for energy savings and quality enhancement. In: 4th Int. Refining Conf., pp. 141–147. Google Scholar

• Nugroho, D.D.P. (2012) Low consistency refining of mixtures of softwood & hardwood bleached kraft pulp: Effects of refining power. M.Sc. thesis, Asian Institute of Technology School of Environment, resources and development, Thailand. Google Scholar

• Olender, D., Wild, P., and Byrnes, P. (2008a) A piezoelectric force sensor for mill-scale chip refiners. Proc. Inst. Mech. Eng. Part E-J. Process Mech. Eng. 222 (E2):115–122.

• Olender, D., Wild, R., Byrnes, R., Ouellet, D., and Sabourin, M. (2007) Forces on bars in high-consistency mill-scale refiners: Trends in primary and rejects stage refiners. J. Pulp Pap. Sci. 33(3):163–171. Google Scholar

• Olender, D., Wild, P., Byrnes, R., Ouellet, D., and Sabourin, M. (2008b) Forces on bars in high-consistency mill-scale refiners: Effect of consistency. J. Pulp Pap. Sci. 23(2):218–223. Google Scholar

• Olson, J.A., Drozdiak, J., Martinez, M., Garner, R., Robertson, A.G., Kerekes, R. (2003) Characterizing fibre shortening in low-consistency refining using a comminution model. Powder Technol. 129(1):122–129.

• Papoulis, A. (1962) The Fourier Integral and Its Applications. McGraw-Hill, New York. Google Scholar

• Prairie, B. (2005) Measurement of forces in a low consistency refiner. M.Sc. thesis, University of Victoria, British Columbia, Canada. Google Scholar

• Prairie, B., Wild, P., Byrnes, P., Olender, D., Francis, D.W., and Ouellet, D. (2007) Forces during bar-passing events in low consistency refining: Effects of refiner tram. Pulp Pap. Can. 108(9):34–37. Google Scholar

• Prairie, B., Wild, P., Byrnes, P., Olender, D., Francis, W., and Ouellet, D. (2008) Forces during bar-passing events in low-consistency refining: Distributions and relationships to specific edge load. J. Pulp Pap. Sci. 34(1):1–8. Google Scholar

• Roux, J.C. (2001) Review paper on pulp treatment processes. In: Trans. 12th Fundamental Research Symposium, Oxford. Ed. Baker, C.F. The Pulp and Paper Fundamentals Research Society, Bury, U.K. pp. 19–80. Google Scholar

• Roux, J.-C., Bloch, J.-F., Bordin, R., and Nortier, P. (2009) The net normal force per crossing point: A unified concept for the low consistency refining of pulp suspensions. In: 14th Fundamental Research Symposium, Oxford. pp. 51–83. Google Scholar

• Roux, J.C., and Joris, J. (2005) Angular parameters beyond specific edge load. TAPPSA J. Tech. Asso. Pulp Pap. Ind. South Afr. 1–10. Google Scholar

• Roux, J.C., and Mayade, T.L. (1999) Modeling of the particle breakage kinetics in the wet mills for the paper industry. Powder Technol. 105:237–242.

• Senger, J.J. (1998) The forces on pulp fibres during refining. M.Sc. thesis, The University of British Columbia. Google Scholar

• Senger, J., Siadat, A., Ouellet, D., and Wild, R. (2004) Measurement of normal and shear forces during refining using a piezoelectric force sensor. J. Pulp Pap. Sci. 30(9):247–251. Google Scholar

• Smook, G.A. (1992) Handbook for Pulp and Paper Technologists, 2nd Edition. Angus Wilde Publications, Vancouver, BC. Google Scholar

• Zha, Q., Lanouette, R., Law, K.N., Bousquet, J.-P., and Bussières, S. (2008) Refining of long fibre fractions after fractionation. In: Prepr. 94th Ann. Mtg., PAPTAC, Montreal, Canada, 2008, p. B481–7. Google Scholar

Accepted: 2017-12-12

Published Online: 2018-05-23

Published in Print: 2018-05-23

This work is supported by a Collaborative Research and Development grant, of \$4 million over 5 years, provided by Natural Sciences and Engineering Research Council of Canada (NSERC) and the following partners: AB Enzymes, Alberta Newsprint Company, Andritz, BC Hydro, Canfor, Catalyst Paper, FPInnovations, Holmen Paper, Meadow Lake Pulp (Paper Excellence), Millar Western, NORPAC, West Fraser, Westcan Engineering, and Winstone Pulp International.

Conflict of interest: The authors do not have any conflicts of interest to declare.

Citation Information: Nordic Pulp & Paper Research Journal, Volume 33, Issue 1, Pages 58–68, ISSN (Online) 2000-0669, ISSN (Print) 0283-2631,

Export Citation

© 2018 Walter de Gruyter GmbH, Berlin/Boston.