![[BioMatNet Database - FAIR Program]](../images/Fair.gif) |
FAIR-CT95-0286
Blowline Blending and Fibre Drying in the Wood Fibre Processing Industries |
 |
Contract No: | FAIR-CT95-0286 |
| Date Prepared: | March 2001 | |
| Source: | Final Report Executive Summary |
Introduction
The overall aim of this project was to achieve an improved understanding of the blowline resin blending and flash drying processes used in the wood fibre processing industries. This would enable the resultant products to be manufactured both at lower cost and with more consistent properties, and that the processes themselves can be made more environmentally friendly. The project demonstrated at pilot scale the effect of blowline blending, flash drying and resin variables on resin distribution, resin chemistry, drier emissions and product properties. This should enable the potential for refiner blending to reduce resin consumption and drier emissions as well as to improve product properties and conduct preliminary full scale industrial trials.
Objectives
One of the key reasons for the lack of published work on MDF process variables has undoubtedly been the lack of available research facilities with the capacity to emulate industrial practice. The objectives of this work were therefore to use a pilot scale MDF facility to study and determine the relative importance of a range of selected process variables. Key objectives of the work were as follows:
Results
Predictions of blowline flow conditions and resin drop size and residence time were produced, based models developed to calculate the steam flow generated in the pilot scale MSD/Digester/Refiner system and the flow conditions in the pilot scale blowline. Blowline wall temperature measurements made along the pilot scale blowline showed excellent agreement with model predictions. It was also demonstrated that the model could be used to calculate wall shear stresses in the blowline. From evidence gathered during the project, it was possible to estimate a value for wall shear stress above which fibre/resin adherence to the blowline wall would be eliminated.
Additional non-standard resins were manufactured for evaluation and further work was carried out aimed at investigating the effect of resin variables on the MDF process and panel properties. Results suggested that as the solids content of applied resin was decreased, resin distribution on fibres was improved. Again, though, it was apparent that board properties were relatively unaffected.
One experimental resin (EXP 2150), containing additives which reduced surface tension and increased water tolerance, showed improved performance (in terms of board properties) over the reference resin. This was particularly true in the case of internal bond strength and thickness swelling. This resin was therefore selected for investigation in industrial scale trials.
A number of additional trials aimed at determining the influence of blowline design / resin injection position on resin distribution and panel properties gave inconclusive results, while no firm evidence could be found to suggest that either resin distribution or panel properties were influenced by blowline design.
It was evident that the addition of wax emulsion had a significant influence on the thickness swelling of boards hot pressed for both 150 and 180 seconds. The optimum addition level appeared to be between 0.4 and 0.8%; above this level addition had no significant effect (reduction) on thickness swelling. Internal bond strength appeared to be progressively reduced with increasing wax addition level. Wax addition level appeared to have relatively little significant effect on the bending properties of boards.
There was little firm evidence that either wax type (emulsion / molten) or injection position (blowline / refiner) had any significant influence on board properties. This was certainly true for internal bond strength and bending properties. There was some suggestion that wax emulsion might impart improved thickness swelling to boards compared with molten wax. However, in general the significance of apparent differences imparted by the two wax types was not supported statistically.
It was evident that board properties were generally reduced when the drier inlet temperature was increased from 155 to 200 ·C. However, only the increase in thickness swelling and reduction in bending properties observed in boards made from furnish dried at 200 ·C were statistically significant.
Boards could not be successfully manufactured from 100% wheat straw and the standard reference resin. This was the case even when high levels (1.5% on resin solids) of aluminium sulphate catalyst were used coupled with target board densities of 650 kg/M3 and long (300 seconds) hot pressing times. Boards were successfully manufactured from 50:50 mixtures of wheat straw and softwood. This was accomplished using the standard hot pressing time of 180 seconds and the addition of 0.5% aluminium sulphate catalyst to the standard resin. The bending properties of boards containing 50% straw were acceptable. However, the internal bond strength and thickness swelling properties fell short of what would be considered acceptable. This was particularly the case with thickness swelling.
Furnish moisture content had a significant influence on internal bond strength and thickness swelling. An increase in furnish moisture content from around 8.5 to 11 % resulted in a 40% reduction in thickness swelling and a 50% increase in internal bond strength.
Board properties were generally improved as the resin addition level was increased. However, it was evident that the relationship between resin addition level and both internal bond strength and thickness swelling was not as strong as that observed for furnish moisture content. Furthermore, the results suggested that the critical addition level might be around 10%; above this level, in many cases apparent improvements in board properties were not statistically significant.
The free formaldehyde content of boards produced on the pilot plant was around 8mg/l00g, which is within the limits specified for El panels. There was no significant difference in the measured free formaldehyde content of boards made with resin addition levels of between 10 and 14%. Boards made without any added resin contained significant and measurable levels of free formaldehyde (as much as half that in boards made with resin).
Injection of resin into the stator side of the refiner blender resulted in some improvements in board properties compared with resin injection into the rotor side of the blender. However, the properties of boards made with refiner blended furnish were still worse (or at best, no better) than boards made with blowline blended furnish. The use of modified plates did not appear to have any advantages in terms of board properties. Indeed, on average, board properties were typically poorer than those made with the non-modified plates. The modified plates did, however, reduce the incidence of blockages in the refiner blender.
There was clear evidence of significant resin spotting on the surfaces of boards made with refiner blended furnish. This was indicative of relatively poor resin distribution, certainly in comparison to furnish blended with resin in the blowline. Boards also tended to contain dark spots / flecks. These were identified as charred resin/fibre fragments, their source being traced to the refiner blender. It was evident that during running there was a significant build up of resin and fibre on the surfaces of the plates within the refiner blender; over time these deposits became darkened / charred, and fragments of the deposits were breaking off regularly and contaminating blended furnish.
A new design of industrial blowline was evaluated. This consisted of a reduction in the internal diameter of a section of the blowline at the point of resin / catalyst injection. This design was predicted to reduce resin drop size in the blowline from 124 to 92 microns. Panels of various thickness were manufactured with the new blowline design. It was evident that the properties of boards made with the re-designed blowline were not significantly different from those of boards made with the standard blowline.
Industrial scale resin trials led to the production of 4mm thick high-density panels produced with the experimental resin. The results suggested that in general there was little difference in the properties of boards made with the experimental resin and the standard resin used on the plant. Boards made with the experimental resin exhibited slightly higher thickness swelling (this was in direct contrast to the results found at pilot scale). The most promising feature of boards made with the experimental resin was their lower free formaldehyde content (around 30% lower).
Discussion
The extremely good agreement obtained between the predictions of the blowline flow model and measured values provides good supporting evidence for the validity of the model. As such, it should be a useful tool for characterising industrial blowlines, and thereby determining whether improvements can / should be made. The model could also be used to suggest improved designs / configurations.
The results obtained in this work for the influence of wax addition level on board properties are in line with typical industrial practice i.e. industrial addition levels are typically between 0.6 and 1 %. Although it was evident that both thickness swelling and internal bond strength were influenced by level of additions, bending properties were less influenced. This is possibly due to the fact that bending properties are more heavily dependent upon the density of board surface regions rather than inter-fibre bonding.
The finding that neither the type of wax used nor the injection position significantly influenced panel properties to some extent fits with known industrial practice (all combinations can be found in full-scale production plants). Wax suppliers have attempted to promote the potential benefits of wax emulsion (improved properties / lower addition rates), by claiming its use results in improved distribution. Whilst there was some tentative evidence from this work that the use of wax emulsions caused improvements in thickness swelling, this was not supported statistically. Further detailed work, with considerably higher levels of replication, would be required in order to confirm whether the apparent improvements were real.
The finding that increased drying temperatures resulted in reduced panel properties was in line with expectations. This was explained in terms of increased resin pre-cure and hence reduced bonding efficiency. Given that drier inlet temperatures can be heavily influenced by (primarily seasonal) variations in ambient conditions, the results suggest that panel manufacturers need to ensure that resin properties / addition levels are adjusted appropriately in order to maintain quality targets.
Results for work on alternative fibrous raw materials suggested that the manufacture of NMF from wheat straw alone and E1 UF resins is not likely to be economically viable in the EC, since extremely long press times would probably be required coupled with significant levels of added catalyst. The use of more reactive resins could be one technical solution, though the free formaldehyde content of resultant panels would in all likelihood be unacceptable. The results did suggest, however, that there may be scope for incorporation of at least some straw in MDF. Although boards made with 50:50 mixtures of softwood and straw still exhibited unacceptable thickness swelling, these contained no additional added wax; there may in fact be a requirement for the addition of some wax. The use of alternative raw materials is an area that warrants further investigation.
Perhaps the most important finding of the work conducted on the influence of process variables was that pertaining to furnish moisture content. Whilst the results can be explained in terms of increased rates of heat transfer to the core of the mattress (and hence increased resin cure) and increased stress relaxation of fibres, it was nevertheless surprising that moisture content could have such a significant effect on board properties. It is believed that the results indicate that furnish moisture content could be one of the most critical factors in terms of process control and board properties. The results certainly suggest that panel manufacturers should strive at all times to control moisture content within very strict limits. It is an area that warrants further investigation.
Conclusions
The results on resin addition levels were broadly in line with expectations. The fact that board properties did not appear to be as heavily dependent on levels of resin addition as on furnish moisture level provided further support for the conclusions from the work on furnish moisture content. That is that that careful control of the furnish moisture content could potentially realise 4 significant savings in resin consumption in an industrial process. The fact that as much as half of the free formaldehyde present in finished MDF panels could be attributable to the wood alone suggests that in order to reduce formaldehyde levels further, research might be better directed towards process variables other than the resin itself. For example, it is likely that the formaldehyde generated by the wood is caused by the high temperatures used in the manufacturing process. Efforts directed towards reducing processing temperatures might therefore prove to be of benefit.
The results obtained from refiner blending studies were disappointing. It was evident that, compared with blowline blending, the method resulted in a very uneven distribution of resin on fibres, resulting in a high incidence of resin spotting on board surfaces. Such boards would certainly be considered as unacceptable by end-users of MDF. The fact that resin was so poorly distributed indicates that there was insufficient inter-particle contact within the blender. It was postulated that this might be improved with larger / full scale refiners. However, validation of this was outside the scope of this work, and would certainly be extremely expensive as a research undertaking.
The fact that, as well as resin 'spotting', it was observed that refiner blended boards contained dark flecks and spots on their surfaces was also detrimental, since these features would again render such boards unacceptable to end users. The heavy deposits of resin and fibre within the refiner blender which were identified as the source of the dark spots in boards would be difficult to eliminate without increasing shear stresses (friction) at the plate surfaces. The circumstantial evidence from this work suggested that such an approach would either result in blockages or worse, fires within the blender.
Leaving the spotting issues aside, since refiner blending appeared to offer no advantages in terms of board properties for a given resin addition level, it was concluded that it would also be more costly in terms of both capital investment and maintenance compared with blowline blending.
The use of the modified design of blowline on an industrial scale did not appear to offer any advantages in terms of either reduced resin consumption or improved panel properties. Although disappointing, this was in line with findings at pilot scale. It was believed that the results may simply indicate that the industrial blowline used may already be near optimum and that there are no advantages in either reducing resin drop size or increasing velocity at the resin injection point. It is worth noting that prior to the scheduled industrial trials, the design of the industrial blowline used had already been altered since the start of the project. This design change resulted in an increase in velocity and reduction in resin drop size, and was shown to reduce the incidence of resin spotting on board surfaces. This was related to elimination of adherence of resin / fibres to the inner wall of the blowline.
The relatively short duration of the industrial scale resin trial precluded the drawing of any definitive conclusions regarding the experimental resin. The increased thickness swelling observed with the resin was in direct contrast to that found at pilot scale and is therefore difficult to explain. The reduced free formaldehyde content of boards made with the experimental resin would be significant and highly desirable if realised for production runs of much longer duration.
Results were published as:
Hague, J.R.B., Robson, D.J. & Riepen, M. (1999). NMF Process Variables - An overview of their Relative Importance. In: Proc. 33rd International Particleboard / Composite Materials Symposium, WSU, Pullman, WA. 13-15 April 1999. pp 79-87.
© Copyright 2006 Policy Statements
Updated
by CPL Press:
03/07/2007
- biomatnet@biomatnet.org
 |
 |
 |
News |
 |
Events |
|
|