Embed Size (px)
Citation preview
The Clay Research Group Issue 59 – April 2010 – Page 1
K
CONTENTS
Soils Applications on the Web.
Penetrometer –v- Oedometer
Undisturbed Sampling.
Aldenham Headmasters House Update
Synthetic Tree and Polyols.
Weather, Tree and Ground Movement
The corner shop…Insurance in 2010.NEWSFLASH
How do see ourselves in 20 years time? Morecorner shop than Tesco maybe? Measuring,weighing, reporting, talking, losing money. It’seasy to be seduced by the idea that what wedo is “how it is”, sniggering at change. If youridea of innovation is to bring back the brownoverhaul, read no further. It is too late.
Times are changing. We have to catch upwith the rest of the world. Starting now.
The Clay Research Group
Monthly Bulletin
TheClay Research
Group
April 2010
RESEARCH AREAS
Climate Change Data Analysis Electrical Resistivity Tomography
Time Domain Reflectometry BioSciences Ground Movement
Soil Testing Techniques Telemetry Numerical Modelling
Ground Remediation Techniques Risk Analysis
Mapping Software Analysis Tools
The Clay Research Group Issue 59 – April 2010 – Page 2
K
Our thanks to everyone who supplied data forthe Sidcup Road study. Details inside. Thedistribution of claims along this particular roadprovides a snapshot into both past claims andfuture risks, as many houses have trees nearbyof similar proportions and species.
OCA have supplied an analysis of the riskpresented in SE postcodes in terms of thefrequency of notifications. Our risk databaseplaces SE9 at 16,055 of 1,752,000 records. Inthe top 10%.
In next months edition Tony Boobier,Insurance Leader EMEA, Business Analytics atIBM provides an overview of where this worksits in the wider context of BusinessIntelligence.
The Intervention Technique has now beenapplied at sites in Worcester Park, Finsbury,Kenilworth (Mercia Mudstone, right), Islington,Golders Green, Essex, Enfield, Dulwich andNorthampton. Monitoring is ongoing.
CONTENTS
Update and New Projects
“The Sidcup Tree” – Risk by Tree and Claim
Anomalous Results + Intervention Technique
Electrokinesis
Triage
The Clay Research Groupwww.theclayresearchgroup.org
This Edition
An exceptionally busy month with some projects reachingcompletion and others commencing. Richard Rollit andAllan Tew met Dr Ian Jefferson of BirminghamUniversity and John Peterson of Foundation Piling todiscuss electrokinesis.
Birmingham have a PhD student willing to undertake therequired site and laboratory work as part of his originalresearch.
Funding comes from a variety of sources. In addition tothe usual grant funding, contributions are being made byJohn Peterson who is generously topping up any shortfall.
In addition members of the CRG and their clients will beproviding a supplement to meet the cost of site trials asrequired.
The objective is to develop a relatively cheap solution todomestic subsidence that can be installed as a “see andfix” repair, allowing us to retain the tree if possible.
Filter Paper Test Update
MatLab report a significantimprovement in the quality ofthe filter paper soil suctionresults since calibrating eachbatch of filter papers, and arenow detecting anomalies prior toissue – see left. The sampleconsisted of just under 200 tests.
The black, dotted lines representwhat appear to be ‘sensible’results. The odd anomalies arecoloured red and blue.
The Clay Research Group Issue 59 – April 2010 – Page 3
K
SE – Sidcup Road~ a joined up project ~
The street scene is characterised by larger styletraditional semi-detached houses with Plane trees ofsimilar height, age and distance from the buildings.
At first sight, not a particularly risky area. The treesare around 10mtrs high, and 10mtrs from thehouses. They have a fairly open canopy and appearto be relatively young.
And yet, statistically at least, and based onproportions only (not accounting for species), theyare probably a higher than average risk. Youngertrees, 10 – 12mtrs high, 8mtrs from a building areinvolved with more cases of subsidence than anyother.
Cyril Nazareth of the CRG identified that twoneighbouring houses had been damaged, andfollowing investigations (see following page) thecause was proven to be root induced clay shrinkage.
The road offers an interesting opportunity to seewhy some trees (of similar species, age, height anddistance) cause damage, whilst others don’t.
More importantly, they provide an opportunity –over a longer term – to assess which treatmentsoffer most hope.
How can we retain the tree and repair the building,safely and quickly?
The LiDAR image below reveals the fairly regularphysical relationship between the trees and housesalong the road.
The periphery of the modelled root zone (the areadescribed by ground movement sufficient to causemovement to a domestic building, rather than anyidea this is the extent of the actual root zone) just‘touches’ the front house wall.
On the face of it, there appears to be nothing tochoose between them. Are they all at risk? Is itjust a matter of time?
The relationship reinforces our previous studiesdescribing a danger zone. These trees appear topresent a greater risk than others whose rootsextend across the entire building footprint.
How does this fit into the UK Risk Model? Wehave plotted a notional location of a property inSidcup Road on the national scale, taking accountof claims and soil shrink/swell potentials.
The Clay Research Group Issue 59 – April 2010 – Page 4
K
SIDCUP ROAD, SE9
Industry colleagues were approached to see if theywould be willing to share their data with us to plot thelocation of claims.
All but a few replied with actual addresses of claimsthat had been notified over a five year period. Theseare plotted below to show the general distribution.
On the basis of the number that replied, and the factthat the one that did not deals with one of the largerinsurers, we can probably assume that there are at leastanother 5 or so claims in this section of road.
11 claims were reported from a sample ofaround 296 addresses. Of these, some are commercialproperties.
The 11 claims represent 3.7% of the whole sample –including domestic and commercial addresses andhouses with or without trees in influencing distance.
If we exclude properties with no vegetation nearby(we counted 89 properties in this category) – that is,making the sample representative of the risk - thenumber of houses damaged rises to 5.3%. Takingaccount of commercial properties and includingunreported claims doubles this figure to around 10%.
The study was also directed towards understandingwhether concreting the front garden increases ordecreases the risk of subsidence.
Ian Brett-Pitt and James Sloan from Innovationcarried out a walk-over survey, classifying gardens aspervious (lawned), semi-pervious (block or crazypaving) or impervious (concrete finish).
Clearly it is not possible to definitively classifywater flow and run-off etc., but the evidenceappears to suggest that a lawned garden doesn’tprovide any protection.
In fact, a pervious garden provides an idealenvironment for root activity, and the idea thatconcreting in any way increased the risk was notsupported.
The difference appears to reflect the dynamicnature of the problem. Where drives have beenconcreted over, any remaining water is ‘bound’, itsabstraction causes movement very quickly eventhough the environment is hostile.
This is in contrast (but with similar effect) to rootsattracted to a moist environment (lawnedfrontages) that can abstract ‘free’ water rapidlybefore tackling the bound water. The end result isthe same – or so it appears.
Our thanks to all of our colleagues who supplieddata.
Two damaged houses, both with lawned front gardens.
Open red circles represent the likely proportionof ‘missing’ claims and have been placed
arbitrarily where there are domestic properties.
Red dots represent actual claims.
The Clay Research Group Issue 59 – April 2010 – Page 5
K
SE ClaimsThe SE postcode also has a highclaims frequency as can be seen fromthe area plot below.
Is it worth the effort? Yes. There aremany houses with identicalcircumstances and understanding thetrigger (assuming there is one), or whatto do if there is not, is key to much ofour work.
If it is simply ‘a matter of time’ a quickand economic solution that allows usto retain the tree benefits us all.
The CRG “Claims by Postcode” Map
SE Risk by TreeMichael Lawson from OCA has kindly provided the followingmaps and graphs plotting a sample of over 28,000 tree relatedclaims and expressed as a % of the annual total, by postcode.
“SE” produces the highest number, as we see below.Coincidentally, this is the location of the Sidcup Road Study.
The distribution describes the clay belt covering the Midlandsand the South East.
The graphs illustrate the risk clearly (by count – not frequency)with a lot of sectors presenting a fairly moderate risk, and a fewaccounting for most of our claims experience, even when takingfrequency into account.
Michael acknowledges that this won’t be a surprise to anyprofessional in our industry, but it does reinforce the messagethat trees are a problem on clay soils, and the location of thatrisk. This sort of objective analysis is an essential tool tosubsidence practitioners.
Our thanks to OCA for releasing this data and allowing itspublication.
Tree Data Supplied by OCA
The Clay Research Group Issue 59 – April 2010 – Page 6
K
ELECTROKINESIS
Richard Rollit and Allan Tew met BirminghamUniversity academics and John Peterson fromFoundation Piling to discuss the electrokinesisproject.
The proposal involves cation exchange using amodification of a well-established technique thathas been in use for de-watering for probably 60years or so. The approach involves moving waterfrom one location to another, and flocculating thearea from which the water was removed to renderit stable – and possibly hostile to root activity.
An electrical charge is passed through the soilbetween two nodes drawing ions from one pole tothe other. The ions effectively ‘drag’ the negativepole of the water molecule to the new location.
The change is permanent and leads to a reductionin hydraulic conductivity at the positive node. Theend-game is to retain the tree, but reduce groundactivity beneath the building.
The proposal would stabilise the soil by thebuilding, leaving sufficient moisture on the ‘treeside’ to satisfy demand.
If this process stresses the roots locally, resulting inan increase in ABA production, then so much thebetter.
Timetable
The current approach will be (a) confirmation offunding for our PhD student, (b) laboratory testingto understand the reaction in a controlenvironment, (c) with the permission of AldenhamSchool, undertake field trials at the site of theWillow and if successful, (d) identify around 6actual claims involving root induced clay shrinkageto determine its efficacy.
Assuming the laboratory work is successful (i.e. thetreatment doesn’t result in greater shrinkage thanoccurs due to root activity) applying this at thetreatment zone of the Willow (see below) andcomparing pre- and post-treatment groundmovement profiles should yield some useful data.
We will be working with arboricultural colleaguesto assess the environmental issues of the treatment.Although NaCl is toxic, the fact the ions are boundto the clay (i.e. not free to be absorbed by roots)may overcome any concerns.
The Clay Research Group Issue 59 – April 2010 – Page 7
K
SOIL CUBE
TREE CUBE
NOT SEASONAL SEASONAL
High Risk
Lower Risk
Many Claims
COMPOSITE
A Beta version of the updated Triage applicationwhich takes account of uncertainty using Bayesianand Heuristic analysis has just been handed over toone of our members.
Triage – handling a claim at the point of the initialcall and ‘putting it on course’ for an appropriateprocess - isn’t a precise science, and the applicationresolves this using statistical techniques.
Conditional probabilities take account of the variouspermutations of soil, climate and historicexperience. For example, whilst around 40% ofclaims might be valid in the winter months, thatmight change to say 70% in the summer, but alldepending on the soil type, climate and history.
The degree of certainty in correctly assessing theprobability increases in a hot, dry year on a clay soilfor example and reduces in the winter.
Further layers of uncertainty lie in how the operatorhandles the call and the description of damage bythe homeowner. Small cracks can appear to beevidence of major structural distress to a nervouspolicyholder.
The application also has intelligence. It adjusts tolearn from its experience and, as far as we are aware,is an industry first. By constantly reviewing closedclaims, the system will modify its decision-making‘on the fly’.
The combined probability cube (above) accounts for all of theelements, and the portion that resolves the tree component isshown below, left. It acts as a layer superimposed ontopostcodes with a shrink/swell potential, amplifying the signal,but only at a certain time of year.
The geological element of the probability cube is shownbelow, right, representing the postcodes on a shrinkable soilthat are influenced by seasonal activity.
Each layer has a weighting, and some contribute to risk thanothers. We can see that claims experience amplifies anunderlying risk, if one exists, and suppresses it in thealternative.
The cubes illustrates the three-dimensional nature of theproblem and above we show our ‘risk signature’ of the UKwith a value attributed to every single postcode.
Elements of the cube above,accounting for vegetation and soil. The
cubes account for location, risk but moreimportantly the dynamic nature of the
combination seasonally.
Right, clay soils only pose a risk in thesummer months, and that risk varies withthe soil PI but is amplified by vegetation,which contains a sub-set of risks. “How
tall is the tree, how far away, and of whatspecies”.
The composite cube accounts for allpermutations to deliver a ‘probability ofvalid claim’ and ‘likely peril’ from verylimited data. It also drives scripting to
ensure only relevant questions are askedat the time of claim notification.
