Effect of Refiner Plate Pattern Design on Refined Fibre Size Distribution – a Time Series Study

In a multi-step batch or continuous operating process, quantitative & qualitative optimization of every single step is always aimed at. Even the smallest aberration in processing parameters affects the quality of the fi nal product. Therefore, it is necessary to study time series quality of products delivered by every single step, so as to ensure product quality in a long run. Refi ning is one of the most crucial steps in fi breboard manufacturing, as it has a major impact on fi bre quality and energy consumption. Refi ning plate pattern is key to good results, which have to be optimized based on experience or onsite testing. We tested three different refi ning plate patterns: straight bar design pattern (TYPE 1), spiral bars design pattern (TYPE 2) and bar with groove pattern (TYPE 3), and their impact on fi bre size quality in a time series. Reported results include fi bre size distribution of three different types of plate patterns observed for 1500 hrs and relative quantity of fi ne fi bres observed in time series for three variants of disc type refi ners. Our research showed that fi bre quality within desired limits was best delivered by TYPE 2 plates; however, fi bre quality also varied with time series for three of them, so different time periods of desired fi bre quality were analysed.


UVOD
Hydrothermal refi ning is the process of converting steamed chips into fi bre bundles, widely utilized in fi bre board and pulping industries (Lumiainen, 2010).Factors affecting the choice of a refi ner can be divided into two categories, fi rst refi ner considerations and second system considerations.Refi ner considerations include plate diameter, pattern design detail, surface and sub-surface dams, refi ning intensity (fi bre type) and alloy.System considerations include refi ner throughput, system stability, stock contamination, available horsepower, operating scheme, and control method.Refi ning plate pattern is the criteria affecting energy consumption as well as fi bre size quality.That is why refi ner plate pattern is continuously researched and improved.Plate pattern has changed from classical constant angle type and bar with groove type to more advanced and improved types, for example curved refi ning bars with jagged leading sidewalls (Gingras, 2011), plates with logarithmic spiral type bars (Antensteiner, 2008), damless refi ner plate for wood fi bres (Savujz'irvi and Liifgren,1999), rough edged refi ner plate cutter bars (Wasikowski,1996), tooth refi ner plates having V-shaped teeth (Gingras, 2012).Despite the efforts to fi nd a computer model to design plates, in the end the selection is based on experience and onsite testing of the process (Rowell, 2012).
The refi ner plate design is characterized by bar width (b), groove width (g), groove depth (gd), sector angle (q), and bar angle (f); plate designs are typically characterized by the length of intersecting edges between the opposing plates, termed as the "bar edge length" (BEL) (Figure 1).
BEL is a standardized measure in the industry estimated using TAPPI standard TIP 0508-05 (1994):Eq.(1) Where n r (r) and n s (r) are the number of bars on the rotor and stator plate at a given radius, R 1 and R 2 are the inner and outer radii of the refi ner plates, respectively.
To approximate this integral, the number of bars on each plate can be roughly estimated by setting (2): (2) With this, Equation 1 can be integrated directly to yield Eq. 3 (3) At this point, attention must be directed to defi ne an operating parameter used through the research, i.e. the no-load power.No-load power refers to the power used by the refi ner for purposes other than changes in fi bre morphology.The LC (low consistency) refi ner no-load power has been estimated to 20-35 % of the total refi ner motor power, (Dietermann and Roux, 2005).During the process of defi bration, radial compression collapses the lumen, the cross section bound by fi bre cell wall, which enhances paper smoothness and sheet uniformity in paper making (Page and Grace, 1967).In this work, effect was observed of plate pattern type on the size quality of fi bres produced.The purpose of this research was: (1) to compare on-site effi ciency of three types of plate pattern, (2) to understand differences in time series performance of these types of plate pattern in a disc type refi ner.

MATERIJAL I METODE
In this research refi ner plates with different pattern design (TYPE 1, TYPE 2 and TYPE 3) were used and the fi bre-size i.e. fi bre-quality distribution was evaluated.Wooden raw material composed of 80 % softwood and 20 % hardwood was processed at 1500 rpm.The distance between refi ner plates was automatically optimized according to tool wear.Samples of fibres were taken from belt conveyer at the interval of 2, 10, 30, 200, 350, 500, 650, 800, 950, 1100, 1250 and 1378 hrs.Samples were stored in laboratory plastic bags to provide constant relative humidity and marked by number, time and date.The weight of specimens were scaled for 30 g and homogenized at 20 °C and 60 % relative humidity for 24 hours.For size analysis, the cascade laboratory sieving machine was employed (Figure 1).The sieves with different sizes (2; 1; 0.5; 0.25; 0.16; 0.08 mm) were used to determine the volume of fi bres on sieve of each type of sieve opening.The fi bres were sieved for 50 minutes and the weight of fi bres from each section was determined.This laboratory process was used for all samples separately.Data were mathematically evaluated and listed in graphs.Although the size range was very wide, the solution of size distribution in time was included as well as the quantity of particles by means of descriptive statistics of the total produced amount of each fi bre type by each tested plate.The trends were evaluated using software MS Excel and then software STATISTICA.10 to obtain the total amount of fi bre type produced by each tested refi ner plate.The ANOVA test with Duncan´s post-hoc test was employed to determine statistically signifi cant differences between the effects of refi ner plates.
Quantity of fi bres with specifi ed size (Q) was calculated using equation (4), where m n is the weight of fi bres on n th sieve in a cascade starting from the top, and m 0 is the total weight of the fi bres. (4)

REZULTATI
The research provided solid ground for size-evaluation in time series, and size distribution of fi bres produced by each refi ner plate type was reported.Results in fi gures represent trends of change in fi bre size with time and also relative quantity of fi ne fi bre produced by different refi ner plates.For plate Type 1 (Figure 3), in the time range of 500-800 hrs, the relative amount of dust fi bres (undesirable) was found to be minimum, whereas other desired fi bre sizes of 0.16 to 1 were found to be at a maximum relative amount.So, the range between 500 and 900 hrs can be considered as optimum performance time for the refi ner plate.
Fibres processed by TYPE 2 (Figure 4) were found to be in the desired size in time ranging between 300 and 900 hours, the relative amount of dust fi bres (bottom) shows a steep fall in the fi rst section of operation and rise after 900 hours.The amount of fi bres with dimension of 0.08 to 1 mm showed a decrease during the whole duration of operation.Larger particles (above 1 mm) showed a steady tendency with the same percentage as at the beginning.Their share during the production is up to 8 %.
Type 3 plate pattern (Figure 5) is represented by decreasing quantity of dust fi bres whose size is smaller than the last sieve i.e. 0.08 mm up to 600 hrs.The quantity of fi bres with the size of 0.08 to 0.016 mm is increased or constant in the second half of the production, as well as the quantity of particles with dimensions greater than 0.5 mm.Although these facts are not convincing for mill effi ciency, the positive results are achieved for fi bres concerned.It can be concluded that the quantity of high quality size fi bres are used in a steady range throughout the production time up to 750 hours.
Total performance of the refi ner plates were also evaluated as mean values of the total produced fi bres along the refi ner life span (see Table 1).Considering  fi bres on 1 mm sieve, TYPE 1 and TYPE 2 produced the same amount of fi bres (ANOVA; p> 0.05).At the same time, TYPE 3 produced a higher amount of 1 mm fi bres compared to TYPE 1 (ANOVA; p<0.05).The plates delivered the same amount of fi bres considering 0.5 mm sieve.On the other hand, the dimension of 0.25 showed insignifi cant differences between the total amount of produced fi bres considering all types of refi ner plates.All tested Types performed with fi bres amount to around 15 % (ANOVA, p>0.05).With the use of 0.16 mm sieve, different amounts were produced by TYPE 3 and TYPE 1 (ANOVA; p<0.05).Fibre production of TYPE 3 was lower by 1.5 %.Fibre production on 0.08 sieve was higher by 3.5 % with TYPE 1 compared to TYPE 3 (ANOVA; p<0.05).Interestingly, the dust fi bres on the last sieve (Bottom) were produced in the same amount of 29 % by TYPE 3 and TYPE 1, while TYPE 2 showed a signifi cantly lower production of fi bres (ANOVA; p<0.05).
According to results, the plate pattern has significant infl uence on fi bre-quality during its lifetime.With TYPE 1 and TYPE 2, longer life span and higher volume of high quality fi bres were obtained, however low quality dust fi bres were also produced in considerable amount.TYPE 2 and TYPE 1 have the highest amount of high quality fi bres (sieve range 0.08 to 0.25 mm) with no signifi cant change until the end of the tool life span.Nevertheless, the percentage share of the high quality fi bres is not constant during the whole processing time.Results with TYPE 1 showed a decreasing tendency of high quality fi bres during the whole operation.On the other hand, the lifetime of refi ner plates was longer than in case of the last Type.Although TYPE 3 has the shortest production time, the quality of fi bres is high.The share of the fi bres with high quality is constant and high from the beginning to the end of lifetime.On the other hand, the lifetime limits its recommendation for future use.With respect to the presented facts, TYPE 2 and TYPE 1 are considered as the best option for the use in the process, the lifetime is long and the volume of high quality fi bres is high throughout the production.Despite the short lifetime, TYPE 3 is then recommended due to constant fi bre quality, which was also high at the start of production, compared to TYPE 1 and TYPE 2, and did not significantly fall during the production.Considering TYPE 1 compared to TYPE 2, high quality fi bres were produced at the beginning, however with decreasing tendency during the production.The quantity of highquality fi bres reached the peak in time ranging between 200 and 500 hours and after that its quantity declined.Since plate pattern design includes numerous mathematical and technical niceties, more research is required to understand the behaviour of each plate.

CONCLUSIONS 4. ZAKLJUČAK
1.In this study, three Types of refi ner plates were successfully compared in terms of mean values of fi bre sizes produced along the life span.Various trends and different plate life spans were presented.2. The results show that TYPE 1 and TYPE 2 plates provide very similar total amount of fi bres since ANOVA shows difference only in the amount of fi bres on the bottom of cascade sieve i.e. dust particles.3. TYPE 3 is found to provide constant fi bre quality, which does not signifi cantly fall during the life span of the plate.On the other hand, the total life span of the plate is lower than that of TYPE 1 and TYPE 2. p,q,r,s,t,u,v,w,x,y,z values having the same letter are not signifi cantly different (Duncan test) / Vrijednosti označene istim slovom značajno se ne razlikuju