Contribution by Peter Hird
The authors are to be congratulated on addressing the settlement issue that has long been of concern to industry practitioners.28 In the absence of reliable settlement predictions the following sequence of rational single steps is all too common.
A skyline contour is part of the planning permission.
Little or no allowance is made for temporary excess even though it is well-known that waste settles.
The operator surcharges but does so in small increments over a number of years because: (i) nobody notices; (ii) everybody knows waste settles, so a small temporary excess is accepted; or (iii) by the time arguing and appeals are finished the waste has settled back to compliance.
The longer time taken to cap the waste causes gas, leachate and odour controls to suffer.
Eventually capping and restoration soils occur but before settlement is complete.
Final levels are low so it is decided to add more soils and plant trees hiding the skyline for the next 100 years.
Good gas yield predictions are again clearly helpful.
I do have one technical question. Cutting the coarser fraction of the waste into 40 mm pieces tends to make it more single-sized, like filter gravel. Over time each piece will behave differently; some will remain unchanged, others will fragment or crumble and others will disintegrate. This process tends to make the mix more widely graded like a well-graded sand–gravel mix used for concrete. Assuming the same particle SG the density ratio of a single-sized to well-graded gravel is approximately 85%. Have the authors considered how much of this theoretical 15% settlement the artificial process on the test samples contributed towards the total settlement?
Author's reply
The question relates to the processed waste used in this study being more uniform than unprocessed waste in terms of particle size distribution (PSD) which potentially reduces the initial density (due to its higher void ratio) thereby increasing the ultimate settlement potential. This is in principle a possibility, but in practice there are many factors that affect the as-placed density of waste and care was taken to ensure that the compacted dry density in the experiments was representative of that observed in the field and other lysimiter studies. Table 5 shows the waste densities achieved in this study. For comparison, Table 6 shows the waste densities achieved in much larger-scale lysimiter studies using unprocessed household waste14 and which correspond to field measurements. There is generally close agreement.
Waste density achieved in the CAR experiments
| Stage | Duration: days | Applied stress: kPa | Final depth of refuse: mm | Dry density: t/m3 |
|---|---|---|---|---|
| 24 h after loading | n/a | 50 (CAR3)50 (CAR2)150 (CAR1) | ||
| 0.40 (CAR3)0.42 (CAR2)0.34 (CAR1) | ||||
| 0.34 (CAR3)0.36 (CAR2)0.44 (CAR1) |
| Stage | Duration: days | Applied stress: kPa | Final depth of refuse: mm | Dry density: t/m3 |
|---|---|---|---|---|
| 24 h after loading | n/a | 50 (CAR3)50 (CAR2)150 (CAR1) | ||
| 0.40 (CAR3)0.42 (CAR2)0.34 (CAR1) | ||||
| 0.34 (CAR3)0.36 (CAR2)0.44 (CAR1) |
Waste densities achieved in large scale lysimiter studies
| Stage | Duration: days | Applied stress: kPa | Final depth of refuse: m | Dry density:* t/m3 |
|---|---|---|---|---|
| Initial | n/a | 0 | 2.1 | 0.32 |
| 1 | 7 | 40 | 2.0 | 0.39 |
| 2 | 6 | 165 | 1.8 | 0.49 |
| Stage | Duration: days | Applied stress: kPa | Final depth of refuse: m | Dry density:* t/m3 |
|---|---|---|---|---|
| Initial | n/a | 0 | 2.1 | 0.32 |
| 1 | 7 | 40 | 2.0 | 0.39 |
| 2 | 6 | 165 | 1.8 | 0.49 |
Dry density calculated using a water content of 33.7%
The waste used was approximately two-week old household waste and required relatively modest processing of the largest waste fraction. The large waste particles (>40 mm) were shredded using a commercial shredder and the cone-and-quarter method was used to homogenise the original sample. A PSD analysis was then undertaken using the wet sieving method (BS 1377:1975) utilising a standard set of 13 sieves. Beaven et al.29 present the results of this analysis where it is noticeable that the PSD curve indicates a well graded-material.
The authors do acknowledge the difficulties in using these data to extrapolate settlements at field scale and that the degree of preprocessing of the waste may further complicate this process. Watts and Charles30 recorded field settlements in mainly domestic waste (~0·55–0·65 t/m3) at the Brogborough landfill of approximately 25% over a period of six years, which compares favourably with the total settlement observed in the CARs. This may be entirely fortuitous. However, the present study has highlighted the mechanisms by which settlement occurs and the importance of mechanical creep on the overall magnitude of secondary settlement. Furthermore, this study provides an insight into the biologically induced rate of settlement which may now be linked to gas generation rates observed in the field.
In addition, a point discussed in the paper that is also worth reiterating was the important consideration (restriction) in preparing the waste samples, such that the waste particle diameter to cell diameter ratio should not be less than 10 in order that sidewall friction effects are minimised.12 The internal diameter of the reactor was 480 mm. Limiting the maximum particle size to 40 mm (conveniently achieved through the use of a standard set of sieves) led to a ratio of 12.
