Shear strength and consolidation properties of Municipal Solid Waste

O. M. Vilar

M. F. Carvalho 

Introduction

In recent years there has been an increasing interest on waste management and disposal. When municipal solid waste (MSW) is concerned, landfills are the most used method of final waste disposal. Apart from environmental aspects, an urban waste landfill must be seen as an engineering structure which must satisfy the usual safety requirements, concerning stability against failure, deformation, stress imposed to foundation soil and the behavior of accessory elements, such as liner and gas and leachate drains. This paper summarizes some aspects of mechanical properties of MSW as measured in laboratory and is a synthesis of the paper by Vilar and Carvalho (2004). These tests were part of a research program performed at the University of São Paulo intended to characterize the behavior of MSW and landfill that encompassed field and laboratory tests, the construction of an experimental landfill and the modeling and analysis of landfill based on test results and results from monitoring.

Characterization of MSW

The tested MSW was recovered from a selected area (Test Area), nearly 450 m2, located at Bandeirantes Sanitary Landfill, in the city of São Paulo (Brazil). The waste in this area is about fifteen years old and it was chosen considering the facilities and the existing instruments (gas and leachate piezometers and superficial markers for settlement control). Field penetration tests (SPT and CPT) indicate some large values of strength that were associated to materials such as stone, rubber and wood. So larger values were associated to more resistant materials and were not considered representative of matrix waste. Excluding the values higher than 20 blows, the average SPT is about 7 blows for superficial layers (0-15m depth). It reaches about 12 blows for deeper layers (between 15 and 30m in depth). Tip and lateral resistance of CPT tended also to increase with depth and typical values varied between 2500 - 7500 kPa and 100 - 400 kPa, respectively.

Infiltration tests showed highly variable rates of flow but typical values of refuse coefficient of Hydraulic conductivity vary between 10-4 cm/s and 10-6 cm/s.
Refuse sampling was done using a continuous auger 40cm diameter. The gradation curve for the collected waste sample was determined through sieving of dry waste and direct measurement of components larger than 50 mm. In the sieving, plastic materials and textiles, which represent almost 20% of the total sample, were excluded. So, it was found that around 35% of the components present grain size larger that 20mm and approximately 50%, between 20 and 2mm. The waste sample used in strength and compression tests presented the following average physical composition: wood (4%), paper (2%), plastic (17%), textile (3%), metal (5%), glass (2%), rubber (2%), stone (10%) and past organic (55%) from which 12% is organic mater and 43%, ash.

Triaxial compression tests

A large triaxial cell was constructed to test samples with 300 or 400mm of height and 150 or 200mm of diameter. Laboratory testing has been conducted using deformed waste samples that were molded to unit weights of 10, 12 and 14 kN/m3. Samples were molded keeping their natural moisture content (in the range between 60 and 68%) and some of them were saturated by back-pressure. Consolidated drained (CD) triaxial tests and undrained tests (CU) were performed at a strain rate of 0.7%/min and effective confining pressures of 100, 200 and 400 kPa. Typical stress- strain curves for CD tests and natural water content waste samples show what can be seen as a typical behavior of urban waste like materials, as reported by many authors. Deviator stress increases with axial strain almost continuously, without any peak in the stress-strain curve or without reaching an ultimate value.

As failure cannot usually be clearly defined when testing MSW and considering the need for an operational shear strength envelope to deal with typical problems such as slope stability, a reliable procedure is to use Mohr-Coulomb criterion related to some value of strain. This is done where shear strength envelopes for axial strains of 10, 20 and 30% are included. As can be seen the mobilized shear strength tends to increase with strain. For instance, for the natural moisture specimens and unit weight of 12kN/m3, cohesion departed from 20 kPa for 10% strain and increased to 71 kPa, for 30% strain. For the same strains the friction angle increased from 22o to 33o. However the mobilization of shear strength parameters does not follow the same pattern. The cohesion has relatively initial low values and tends to increase continuously at an increase rate that is much more pronounced than that of friction angle. The friction angle on the other side has relatively initial higher values and their rate of increase suggest that it will tend to reach an ultimate value as strain increases.

The large cohesion intercepts calculated have been credited to the fibre components (specially plastics) that provides a reinforcement mechanism (König and Jessberger 1997) responsible, for instance, for the existence of vertical cuts as high as 4m in the tested waste in the field. Among the factors studied, it was found that the unit weights used in the test program (10, 12 and 14 kN/m3) showed little influence in shear strength, probably because the corresponding dry unit weight varied little between the samples. Similarly, it was noticed that the results from specimens molded at natural moisture content (degree of saturation of about 70%) did not show appreciable differences from the saturated ones. It is believed that the large volume decrease during consolidation and the high initial degree of saturation of natural moisture samples led both samples to similar conditions during shear justifying the closer results.
 

Confined compression tests

Confined compression tests of waste have been carried out using a oedometer with 385 mm diameter and 365 mm height. The waste sample was placed in the oedometer in 6 layers and lightly pressed to reach the desired initial unit weight. Loading was applied in stages that lasted about 15 days although some stages were prolonged a few days to define adequately the creep characteristics of the waste. Typical curves of deformation against time show an accentuated secondary compression process (creep).

These data allowed to calculate the secondary compression index Cα (Cα =∆e/∆log t) and C'α (C'α = Cα/(1+ eo)) for each stage of loading. Cα ranged between 0.021 and 0.044, with an average value of 0.032, while C'α varied in a narrow range, between 0.012 and 0.016. The overburden stress did not influence both coefficients in the range between 100 and 640kPa, although a slight decrease in Cα was observed as the stress increased.

As an equilibrium void ratio was not attained in the time spans used in the tests, some curves for different times of loading illustrate some compression curves of specimens molded with initial unit weight of 10 kN/m3. The curves were almost parallel but not straight. They were rather slightly concave upwards, however straight curves were fitted to points between 60 and 640 kPa allowing to obtain the primary compression index [Cc= ∆e/∆log(σ)].Cc varied between 0.52 and 0.92, however the normalized compression coefficient C'c [C'c=Cc/(1+eo)] obtained were very close, averaging 0.21.

Conclusions

Typical results of consolidation and triaxial compression tests on large remolded specimens of a 15 year old sample of Municipal Solid Waste (MSW) were shown. Due to space limitations, many of the analysis regarding the whole set of results could not be presented, but the general behavior of the waste can be summarized as follows:

1. Consolidation tests showed the large compressibility of MSW where secondary compression plays a fundamental role.
 

2. Secondary compression index, normalized through eo, C’α (C’α= Cα/(1+ eo)) showed a small range of variation and yielded an average value of 0.013. Primary compression index (Cc) showed to be also dependent on eo and normalized C’c [C'c=Cc/(1+eo)] reached about 0.21.

3. Drained triaxial compression tests showed stress-strain curves that were concave upwards and failure or an ultimate value of deviator stress could not been reached even for large values of strain, above 30%. The shear strength parameters are strain dependent and tended to increase with the deformation.

4. For the same unit weight, the shear strength was little affected by the degree of saturation of the tested specimens. The relatively high initial degree of saturation of the unsaturated specimens (above 70%) and the large volume compression during consolidation probably raised the degree of saturation approaching the saturated condition thus yielding close values of shear strength.

5.The shear strength parameters obtained from specimens with different unit weights presented small variation and an average shear strength envelope could fit the results of the samples in the range of unit weight studied (10, 12, 14kN/m3).

6. Effective stress shear strength parameters from CU tests were misleading and did not agree with the parameters from CD tests. Undrained shear strength was proportional to the effective confining pressure and the relationship between these variables was larger than that usually observed in soils.