A method of predicting changes of soil dry bulk density beneath agricultural wheels

Abstract

Existing research into factors influencing soil compaction and methods of modelling compaction processes have been examined.As the relationships between soil stresses and strains were very complex, simpler interpretations of soil mechanical theories were required for the modelling of soil compaction.

Proposals were made for examination of stress prediction equations developed by Sohne (1953» 1958). Use was also made of the Critical State theory of soil mechanics to describe rela­tionships between soil stresses and strains. Interpretations ofthis theory were made to derive soil mechanical functions from soil packing state before and after the application of certain levels of stresses. These functions were the apparent virgin compres­sion line (’VCL*) and the primary function; soil packing state being expressed as dry bulk density.

Experiments were made to test Sohne’s stress prediction equa­tions and identify the 'VCL1 and primary function from in situ and laboratory measurements of stress and strains of field soils. Stresses were applied to soils in the field by single tractor rear wheels with controlled loads (approx. 1.0 to 2.0 tonnes) and con­trolled inflation pressures (approx. 80 to 170 kPa). Stresses were measured in situ beneath the experimental wheels by deformable spherical transducers. These transducers had been especiallydeveloped for this research and could measure combinations of hydrostatic and deviator stresses up to three bar. Two other field experiments, one on a loam, the other on a sandy loam soil, examined the responses of soil of different initial dry bulk densities (approx. 1 .0 to 1 .4 g/cm^) at different soil moisture contents (approx, 14 to 28 per cent, w/w) to tractor rear wheels of three different loadings and inflation pressures. Dry hulk density was measured in situ by gamma ray transmission equipment. Measurements of soil strength, soil moisture content and soil moisture tension were also made in situ as well as descriptions of the transverse profiles of the wheel ruts. A soil tank experiment using a sandy loam soil examined relationships between the horizon­tal projection of the tyre/soil contact area and wheel sinkage. ’Triaxial’ equipment was used in the laboratory to measure changes of dry bulk density by different levels of spherical pressure up to five bar. Loose and ’undisturbed' samples of field soils were used at different moisture contents.

Apparent virgin compression lines ('VCL') and primary functions could be derived from the laboratory data. These could be com­pared to estimations of the 'VCLs' from the field data. Therelationships between wheel sinkage and tyre/soil contact area enabled estimation of contact areas by the wheels used in the exper-I!iment. These contact areas were used for Sohne's stress prediction equations. Comparison of measured and predicted values of first principal stresses (for soils of different strengths) gave quanti­tative soil strength limits forthe use of each prediction equation. The predicted soil stresses, and measurements of dry bulk density after wheel passage were used to estimate apparent virgin compres­sion lines ('VCL') from the field data.

The differences between the 'VCL* derived from laboratory measurements and those estimated from field data were very similar to those expected from the differences between the method of stress application used in the field and that used in the laboratory. Therefore it was suggested that the method of identifying 'VCL1 had been successful, but further evidence was required for more conclusive proof.

Primary functions were derived from the laboratory tests on loose and 'undisturbed' soil. These were combined with the estimates of apparent virgin compression lines from the field data andtlSohne's stress prediction methods to construct a simplified model of soil compaction beneath the centre line of a moving wheel. The applicability of the model was confined to loam or sandy loam soils of cone resistance greater than five bar and initial dry bulk density greater than 1 . 1 g/cm(3) and to beneath the centre line of wheels with less than about five per cent slip or ski

The compaction model was tested against field measurements of dry bulk density and other soil physical properties, made before and after the passage of suitable agricultural wheels with differ­ent loads and contact areas. The field measurements had been made during separate and independent experiments carried out on loam and sandy loam soil. The model was considered a sufficiently accurate method of predicting soil compaction, within the range of appli­cability mentioned above, since most of the predicted values fell within five per cent of the observed values.

Simulation of soil compaction by the model, and examination of the data from the field experiments suggested that, for soils before ploughing and after sowing operations, increasing tyre/soil contact area and reducing wheel load is the most suitable means of reducing(iii) compaction of 'topsoil'. However reduction of wheel load and not reduction of contact area appeared to reduce compaction of 'subsoil'.

It also seemed that loose soil, below the strength limit of the compaction model, would compact to higher bulk density than firmer soil, with strength above the limit, when run over by the same wheel. Observations of the variation of the apparent virgin compression line with soil moisture content also suggested that soil became more susceptible to compaction at moisture contents above the lower plastic limit, determined by the drop-cone test on soil aggregates similar to those in the field.