In the recent past, as more farm power is being demanded on farms, due to increased farm sizes and operating speeds, larger and heavier farm machines are deployed in various farming operations. Their cumulative negative effects have become more apparent with increased incidences of soil compaction problems. This has forced many farmers to practice deep tilling, using subsoilers to break up compacted subsoil layers. In some maize growing regions of South Africa, conventional subsoilers are used in a tandem configuration. The farmers believe that the use of subsoilers in this mode reduces the draft force per unit area tilled. This probably happens because the critical depth for the rear subsoiler is increased beyond its working depth of 600 mm. Operating in this mode necessitated this study, with the ultimate goal of testing an appropriate existing force model for a single tine in predicting the force requirements of the front subsoiler in a tandem configuration. Secondly, to develop an alternative model for the rear subsoiler based on the three-dimensional failed soil-profile and to determine the relative position of the front subsoiler at which energy utilization is optimized. To develop the proposed model, an analytical approach based on limit equilibrium analysis was used and a Matlab-based computer program was coded to solve it. Its verification was conducted through field experiments in sandy clay loam soil. The experiments consisted of a continuous measurement of the horizontal and vertical forces acting on each subsoiler by a two-dimensional force transducer system. At the same time, the three-dimensional and thus the cross-sectional areas of the disturbed soil-profiles at different sections were measured, as well as the soil characteristics. A manual method employing a pin-profile meter was used to measure the vertical cross-sectional areas of the failed soil-profiles at 100 mm intervals. Further more, a technique using an automatic penetrometer and a computer program was developed to identify and map the three-dimensional failed soil-profiles. This technique indicated that the subsoiler failed the soil beyond its maximum operating depth and width. The results also indicated that the soil-failure pattern at close spacing is in phase at both subsoilers, leading to reduced total draft force requirements. At a wider spacing, the soil-failure pattern was out of phase, thus resulting in increased total draft force requirements. At the same time, the cross-sectional area tilled per unit draft force increased with increased spacing. This was because the failed maximum cross-sectional area increased in size faster than the total draft force as the spacing was increased. The proposed model verification results show that the predicted and recorded forces at the rear subsoiler correlated reasonably well at a wider spacing. When the front subsoiler was shallow working and close to the rear subsoiler, the model under- predicted the measured forces on the rear subsoiler, whilst the Swick-Perumpral model over predicted the applied forces to the front subsoiler and this was generally the case at wider spacings. Furthermore the efficiency of the subsoilers was maximized when the longitudinal spacing was such that it allowed the soil failed by the front subsoiler to stabilize before the rear subsoiler reached it. The maximum cross-sectional area failed per unit draft force was recorded when the depth of the front subsoiler was equal to about 80% of the rear subsoiler-operating depth. The knowledge contributed by this research will not only facilitate qualitative field operations and optimize energy use, but also promote better management decisions. / Thesis (PhD (Engineering))--University of Pretoria, 2004. / Civil Engineering / unrestricted