Simpson Park, Harworth by WYG
Shortlisted for Brownfield Awards Category 1 - Best Project Preparatory Work
The former Harworth Colliery site, now known as Simpson Park, is being regenerated by Harworth Group. It is located 12 km south of Doncaster and covers an area of approximately 73 hectares. WYG designed enabling and infrastructure works on development plots in the north of the site were completed in 2018 and housebuilding is ongoing. Redevelopment of the rest of Simpson Park is at master planning stage with engineering design optimising areas for development to residential end use.
In the west of the site Phase 1 Geo-environmental work identified historical evidence of a large pit. From discussion with a former employee of the colliery and review of historical photos and available ground investigation data, it was established that the pit had been infilled through end tipping with coal fines from coal washing and colliery spoil mostly during the 1960s.
In early 2019 WYG produced a Foundation and Ground Improvement Options Review for the area of the site where the historical pit was located. This document contained a review of existing ground investigation data, an assessment of the properties of made ground within the pit, a review of options to improve the ground and alternatives to piled foundations. A ground model was produced from the existing data. The made ground associated with the pit was known to extend to at least 7 metres below ground level (mbgl), underlain by till deposits (cohesive and granular) with bedrock comprising sandstone. Groundwater levels within the sandstone were approximately 20 mbgl.
The review specifically identified a potential risk of collapse compression from inundation. BRE (2015) describes this process as ‘collapse compression on wetting’ and it occurs where poorly compacted, partially saturated fills undergo a reduction in volume when inundated or submerged for the first time. If it occurs subsequent to construction upon the fill, buildings can suffer serious damage. BRE (2015) and also ICE (1996) state that it often represents the most serious hazard for buildings on fill and is recognised as a major risk during construction on colliery fill.
 BRE, Building on Fill: Geotechnical Aspects, 2015
 ICE, Geotechnical Engineering Panel Paper: The assessment of the collapse potential of fills and its significance for building on fill., 1996.
Inundation Test Design, Innovation and Best Practice
Ground investigation was required to further characterise the infill of the pit with one of the objectives being to assess the risk of collapse compression on inundation. Whilst laboratory analysis of disturbed samples could provide some information on the properties of the soil there was no laboratory test that could definitively establish if in-situ collapse compression on inundation was a risk.
Best practice was exemplified by focus on the ground model and adopting a lines of evidence approach to assessing the potential for the risk of collapse compression to be realised through a combination of collection of field data, laboratory data and in-situ field testing.
Measurement of the potential for collapse compression is challenging due to the difficulties associated with either taking undisturbed samples of uncompacted fill for lab testing or in conducting in situ testing. The construction of boreholes often involves the introduction of water which in itself could induce collapse compression before in situ testing and monitoring begins.
BRE (2015) provides a case study from the 1970s where water was added to a trial pit and settlement measured at various depths using magnet extensometers installed around the trench. ICE (1996) describes a test pit also constructed by BRE, made from concrete with an outlet to control water egress to the soils beneath to simulate inundation of two material types, glacial till and colliery spoil. Using these case studies as a starting point, WYG designed a methodology that was practicable and would provide in-situ data that would contribute to the lines of evidence approach. At the forefront of the design was the constraint that each in-situ test location could only be used once to provide representative data.
The test involved the excavation of a trial trench to c. 4 mbgl which was backfilled with cobble sized material. A windowless sampler borehole was drilled on the long side of the trench and installed with a groundwater monitoring well to monitor changes in groundwater level in the surrounding made ground once water was added to the trench. Windowless sampler boreholes were drilled approximately 0.5m away from each corner of the trench. Magnet extensometers were installed at a variety of depths targeting made ground. Two deep extensometers were installed, one at c. 0.5m below the base of the trial pit and one at c. 0.3m above the base of the trial pit at opposite corners. Shallow extensometers were installed between 1.0m and 1.75 mbgl at all four corners. Baseline readings were obtained for all measuring points and then water was introduced to the trial trench.
Water levels were measured at the groundwater monitoring well and ground movement was monitored both at surface and within the made ground at the extensometer points. Measurements were obtained every 15 minutes for 3 hours on the day of the test. To assess the effects of inundation over a longer period, two further topographic level survey visits were made targeting the top of the extensometer pipes, whilst the in-ground extensometers were monitored weekly over a 6-week period.
Inundation Test Results
Inundation tests were completed at three locations. The initial test was treated as a trial of the proposed methodology and was used to refine the procedure. The test locations were selected to provide coverage across the area of infilling and to provide test data for different types of made ground.
The results of the initial test provided good data relating to the amount of saturation of the ground. Water measured within the monitoring well rose by 0.6m indicating the surrounding ground, in particular at the depth of the deeper extensometers, had been saturated. Measurements of the extensometers showed relatively limited movements of up to 8mm over the 4-hour period. The trial test identified that the GPS based topographical surveying of the level of the top of the extensometer pipes was not sufficiently accurate and the surveying technique was changed to use a total station with an accuracy of ±2mm. This was the only variation from the initial method.
Two further tests were completed, with the groundwater monitoring recording a water level rise of 1.4m in one well at the location where the made ground was more cohesive and remaining dry in the other, where the made ground was mostly granular. The initial testing period (3 hours) recorded minimal movements (<5mm) at the extensometers and the top of the extensometer pipes. Movements were both up and down suggesting a combination of both settlement and heave.
The 6-week period of monitoring of the extensometers did not show any significant movements at any of the test locations with the exception of one location where both the top of the pipe and the extensometer targeting the shallow ground showed settlement of 55mm. This occurred at the test location where the made ground was predominantly cohesive and the results indicate that movement occurred as the inundation water rest level dropped. This data suggests slow infiltration rates within the shallower cohesive material appeared to have the most influence on collapse compression.
In addition to the in-situ test other aspects of the properties of the fill and ground conditions were assessed as part of the lines of evidence approach as follows:
25 No. laboratory compaction tests were undertaken on samples of made ground to determine the optimum moisture content which was compared to the initial (natural) moisture content. Most tests recorded initial (natural) moisture contents at or above the optimum moisture content. Only four samples recorded initial moisture content lower than optimum, three of which were associated with the area of historical infill. This suggests a reasonable proportion of the fill had been subject to some wetting in the past.
To attempt to allow a visual assessment of made ground compaction, two boreholes were drilled utilising continuous undisturbed sampling to try to facilitate visual assessment of the soil compaction state. The material sampled was found to be generally cohesive with occasional granular elements. Visual assessments could not provide any indication of compaction; however, the technique demonstrated the variable composition of the colliery spoil.
The presence of groundwater within made ground was assessed through groundwater level monitoring of traditional wells with response zones targeting the made ground. Groundwater level monitoring was carried out on six occasions over a period of four months. Perched groundwater was present in most boreholes providing a further indication that collapse compression due to water inundation is unlikely at these locations since wetting of the ground has already occurred.
Cost was a key driver in the design of this relatively untested methodology. WYG selected inexpensive intrusive investigation techniques that were already being utilised as part of the wider ground investigation of the site. The extensometers used are a cost effective and established piece of geotechnical monitoring equipment specifically designed for the purpose of recording ground movement.
The cost effectiveness of the technique allowed three tests to be undertaken. This approach provided better confidence in the data collected compared to a single more expensive test.
Compliance with Codes and Guidance
BS5930 does not make reference to collapse compression on inundation. Section 4.1 of NHBC Standards refers to collapse compression being potentially associated with infill and made ground but does not provide any methods of assessment. BRE (2015) provides the most comprehensive assessment of the mechanisms behind collapse compression on inundation but it does not provide a methodology for in-situ investigation.
WYG’s approach of investigating several lines of evidence improved confidence in the conclusions reached and allowed stakeholders to view the evidence and interpretation behind the conclusions drawn. WYG’s liaison with NHBC is detailed below.
NHBC was identified as a key stakeholder. Harworth Group are using NHBC’s Land Quality Endorsement service for the site and consequently contamination assessments, geotechnical assessments, earthworks and remediation designs and specifications are subject to review and comment by NHBC. In advance of the ground investigation WYG met with the NHBC geotechnical engineer for the site to discuss the objectives and the test methodology. NHBC was kept up to date with plans for ground investigation and the same representative from NHBC attended site during the first of the three tests. This allowed NHBC to see the practicalities of the test methodology in the field and allowed them to be party to the settlement data obtained on the day.
The ground investigation report produced includes a section specifically addressing the data collected relating to collapse compression on inundation. Prior to issuing the report WYG again met with NHBC to present and discuss the findings. The report was then submitted to NHBC as part of the LQE process.
WYG kept Harworth Group up to date with progress. On completion of the ground investigation works the key information was communicated to the project design team via the regular progress meetings that are arranged by WYG. In addition, the Foundation and Ground Improvement Options Review report was updated to take account of the findings of the ground investigation.
Environmental, Economic and Social Benefits
The ultimate economic benefit of the test relates to the foundation solution for the site redevelopment. The demonstration of a low or negligible risk for collapse compression on wetting reduces the development cost. Piled foundations or time-consuming surcharging are no longer required and a more sustainable and cost-effective design using turnover and re-engineering of the made ground to make it suitable for shallow foundation solutions can be employed. This in turn improves the viability of the development of this part of the site, thereby contributing positively to the housing requirements for the local area.
A Robust, Sustainable, and Defensible Solution
The risk associated with collapse compression on inundation has been assessed by an appropriately planned, designed and executed ground investigation. The lines of evidence approach provided data that were reliable and reasonably representative of a poor case scenario allowing a robust conclusion to be made. The findings will be subject to scrutiny and review by NHBC who have been engaged as a stakeholder from the planning stage onwards.