| dc.description.abstract | The overall objective of the study was to investigate how batch log processing based on a recovery optimizing criterion could be used to improve recovery of band sawmills. Cutting patterns were developed to provide a support system for curbing the limited decision making capacity associated with portable band sawmills. Data were collected on 150 logs selected randomly at Albertine Timbernet sawmill in Hoima district and for each log, top diameter, taper, eccentricity, sweep, number and size of knots were evaluated. Timber recovery of logs under existing practices at the sawmill was determined. Cutting patterns for each log were then developed through mathematical analysis and the maximum attainable recovery from such a log determined. Cluster analysis was used to assign logs to three classes using a 20% dissimilarity in top diameter. The difference in recovery of the formed log classes was then tested using a one way ANOVA while the difference in the existing and maximum recovery of the logs was tested using an independent student t-test. There was a significant difference in recovery between log classes with large diameter classes yielding more recovery than small diameter classes. There was also a significant difference in the existing recovery and maximum recovery that would have been attained if the logs had been sawn using their corresponding optimal cutting patterns. The mean existing recovery was 27.27 per cent while the mean maximum recovery was 41.52 per cent indicating therefore that there is a potential for significant increase in recovery. To ease the implementation of the proposed patterns, logs were assigned into four classes and it was recommended that the sawmill adopts the patterns. Other recommendations for further studies included determining the applicability of the proposed patterns at other mill types, doing analysis based on revenue maximization and development of patterns that would maximize recovery while meeting customer demand for the different timber sizes | en_US |
