Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

The Bnancial bene6t &own a site rate and may not be apparent in the cash flows.

may not be apparent in the cash flows. The risks are


primarily taken by the banker or investor.
investigation strategy ~Insurable risks, eg technical risks connected with project
by IL Whyte, Department of civil and construction and operation. These can be covered by a
premium payable to cover the risks. The premium is normally
structural engineering, UMIST, UK considered to be insurance but it can equally be the cost of
necessary planning and investigations.
Introduction ~ Specific risks. These reflect the uncertainty in the
In general construction, including land reclamation, the site assumptions used to estimate the cash flow.
investigation component often suffers &om an ill-defined Ideally, the risks should be borne by the partly most able to
management strategy and a lack of resources in terms of time control the event. In traditional forms of contract the employer
and money. In a proper strategy adequate resources would be does not have a strong motive to do enough site investigation to
provided. In practice this is likely to require an increase in establish the effect of ground conditions upon construction cost
finance and time allowed beyond the amounts traditionally (the contractor is assumed to have inspected the site and to
allocated. For such increases to be sanctioned, then it essential have carried out his own site assessment). Thus if only a
that the benefits be realised, particularly cost benefit. minimal investigation has been conducted, the prices may be
The paper presents data &om risk modelling of two sites based on a wrong view of the subsurface conditions and the
where the ground hazards are 'high'nd 'low', a reclamation risk to the employer of later delays and cost escalation
development is included. The modelling produces information increases. The employer should be more motivated to do
which is indicative of the range of financial effects rather than enough investigations, thus ensuring the contractor does not
absolute and accurate data. This is due to a combination of the carry a large risk for which he can do little to reduce.
uncertainties inherent in the figures and elements of subjective The perception of what constitutes a substantial loss is not
judgement. easy. The evidence suggests that variations in cost and time
Historic data on the cost uncertainties associated with the beyond about 10% to 15% are not acceptable. It would be of
ground is given. When these are combined with the results of great interest to know in greater detail what level of loss or cost
the risk models, then the financial argument for a proper site escalation is considered to be unacceptable.
investigation strategy is clear. Investment in good planning, The available evidence suggests the ground risk to be one of
management and communications produces sound economic the largest uncertainties to the construction process. Ground
benefit and return; this is particularly true for estimates and for hazards often influence activities on the critical path and are a
the levels of financial risks to which a developer becomes significant feature of cost over-runs and delays. Over the past
exposed. 30 years, geotechnical engineers have made significant efforts
to improve their work. Often this is to little avail. Under market
Examining the risks conditions, the retail price index has increased to about 240%
The perception of risk varies according to viewpoint —one since 1979. The ground investigation price index has gradually
man's loss may be another's gain. Engineers and professionals fallen by about 25% to 30% over the same time period (taking
in construction tend to view hazards and risk &om the RPI and GIPI to be 100% in 1979)'. Insufficient resources
technological and environmental viewpoints. Developers and produce inadequate investigations. A coherent strategy is
financiers view events in terms of their economic power and the necessary, this will inevitably lead to higher cost for
market. There is a problem in relation to ground conditions in investigation. Is this cost increase worthwhile?
that these have both technical and financial consequences.
These are exacerbated by the often low standards used, a Ground risks
recent survey'eports that 80% of specialists are not satisfied The most serious effects of risk events to construction are:
with the general standards of SI in construction. OFailure to keep to cost estimate.
The main concerns of developers and financiers in terms of OFailure to achieve the completion date.
project security and finance include: OFailure to meet quality and operational requirements.
~ Capital over-runs: mainly due to inadequate preliminary
engineering and poor estimating procedures. Identi6cation of ground risk
~ Completion delays: almost as common as capital over-runs. Ground investigation is a low cost activity. Data over many
~Cash flow variations: operating costs, productivity, currency years suggest costs to rarely exceed 1% of contract value and
fluctuations, etc. typically the cost may range &om 0.2% to 0.5%. (The full SI
~Market difficulties. cost is rarely reported but may be realistically estimated to be
~Resource risks. no more than double the ground investigation cost). The cost
~Political actions. of 'unforeseen'round events can exceed 10% of contract value
~Project inefficiency. and on occasion can exceed 100%. Data is sparse but has been
~ Environmental risks. summarised in an ICE report'. More detailed information is
~ Casualties. available &om tunnelling contracts in America4 and for
~ Insolvency. highways in the UK'.
Proposals for development are subject to risk analysis. The The American report on 87 tunnel projects showed a direct
principles are: correlation between the accuracy of cost estimate and the level
~Risks attributable to investment issues, eg political and of SI cost. The average cost of investigation (probably the
general risks. These can be covered by an increased discount ground investigation) was 0.44% of estimate cost; it was
33'ROUND

ENGINEERING ~ OCTOBER 1995 GROUND ENGINEERING ~ OCTOBER ~ 1995


recommended that this be increased to 3%. This was based on but this was not on the critical path. There was a three week
an analysis of cost increases and claims. About 60% of the delay in the activity at an extra cost of $22,000. SI costs were
claims were for large amounts with the level and number of not reported but are estimated at about $ 10,000 (some 0.5%
claims rising exponentially as the amount of investigation of civil tender).
decreased. It was concluded that the more that was known OA commercial development on a 'brownfield'ite. The
about the ground in advance of construction reduced the level estimate costs were $4.3M. Serious obstructions to piling (on
of uncertainty. This leads to better planning and estimating the critical path) caused an eight week delay and costs
with a reduction in the number and degree of severity of increased by $0.9M. SI costs are not known in detail but
claims. probably were less than 0.5% of the estimate. It is believed that
The UK report'n 58 highway projects (of which 21 were subsequent industrial archive researches indicated the probable
subject to detailed analysis) showed the average SI cost to be existence of the obstructions.
1.2% of tender value and 0.9% of out-turn cost (ground The analysis consisted of two components: calculation of the
investigation cost is probably about 50% of these figures). The influence of ground uncertainty on estimated costs and
cost increase above tender on the 21 projects ranged &om 7% durations for each project; and to quantify the efFects of costs
to 61% with an average of 30%. About 54% of the cost and delays &om unforeseen ground conditions when these are
increase was due to geotechnical origins (ie. about 16% of the encountered during construction.
total cost increase). The range of uncertainty over the total Details of the analysis are presented elsewhere".
increase in construction cost related to SI cost is broadly as
follows (58 project analysis): Ground uncertainty and estimates
The influence of uncertainties in the knowledge of ground
li liiio;:u i 5I I I I S ltl I,'II' II II conditions on project estimates were modelled for five levels of
»»i; II I " Ik ilIIIlrii Irinmili@

"%!i!~III"-'=P.-:::-'::~-:
iI r;
i i i
site investigation: very good, good, fair, poor and very poor.
-'--': Results are summarised below (the analysis for delays is not
presented):
0-100 26 <1
TilT'HiiF, I iWWiiii
SIL'Z:--::::;:;:---'-:::-:-:.':"'":-:'F~-'-":
i:-----':::!'--:-:-- ''ii B I Is l,li Igl I I

0-30 15 3-4 , Bi 'li III

II
'
i RYi ~ ~ Is i! li%

Note average geotechnical cost increase is 54% of the total.


Table 1. Good 5.8 0.8 - 9.9 3.2 0.6 - 8.0

While there is no direct relationship between SI and


increase in construction costs, the data clearly shows that Poor 11.6 2.5 - 22.0 8.8 1.4 - 15.6
low levels of investigation result in a large range of
uncertainty. This uncertainty reduces with investment in site
investigation. Table 2.
The limited published data for roads confirms the cost of site
investigation to be low and the level of investment (an The following can be noted:
'insurance'remium) relates to final construction cost. Low (i) The greater is the uncertainty over ground conditions then
premiums (<3%) are taken by gamblers since the final out-turn the more likely is the risk of cost escalation (and delay) to a
is uncertain —a few contracts will meet estimates and a few will project.
experience capital over-runs of perhaps 100%. The most likely (ii) It is significant that the range of uncertainty on an
out-turn is likely then to be a cost escalation of about 25% to estimated cost decreases with improved site investigation. The
30%. factual data on tunnels and highways confirm such ranges of
cost uncertainty do exist in construction.
Evaluating the ground risk (iii) For a project to have a cost estimate which is unlikely to be
Ground risks have been evaluated in financial terms using the exceeded by 10% due to ground conditions then at least a good
CASPAR program (Computer Aided Simulation for Project site investigation is required. A very good investigation
Appraisal and Review). This is based on a precedence network produces the most reliable estimate.
with time, resources, costs and revenues attached to each (iv) The benefit (in cost terms) &om a site investigation can be
activity in the network and time related charges are handled by quantified, the return &om good practice is the probability that
means of a hammock. The cost model allows risks and cost estimates are more reliable: financial uncertainty is
uncertainties to be assigned with results in terms of parameters reduced.
such as NPV, IRR etc. The results are derived in the form of
probability distributions. Effect of 'unforeseen'round in
Two projects analysed are outlined below: construction
OIndustrial development on a 'greenfield'ite. The It was assumed in the model that a 'good'ite investigation has
tendered target cost for the civil works was $ 1.8M and this was been commissioned. Designs are completed and construction is
34 effectively met. Ground problems did occur during earthworks, in progress. During groundworks activities (eg piling,

GROUND ENGINEERING OCTOBER ~ 1995


IS )rid) I) tl rI I I I Wlliilt I )I gl'I I I I

'. I'1
I

I'I a l ' II III iii


II I) ii al I: ' its~ ~ II '' Ill ~ lI iljTI I| I I)RIIII)w a; ~ ill FIN )Mil ' llll'Kji
l ll li,
III~—
i i l I

li
1 6.6 1.1- 10.5 2 8.3 5.3 - 13.9

9.9 4.1 - 14.9 19.0 13.7- 23.8

Table 3. Results for site investigation 'good'.


muck-shifting), 'unforeseen'onditions are encountered which Controlling the ground risk
lead to delay. The financial effects of the delays are calculated Ground risks can be controlled by reducing the uncertainties
using probability analysis. Results are summarised in Table 3. associated with knowledge on the ground hazards. Knowledge
The following can be noted: is obtained &om investigation: the better the investigation the
(i) The cost variations for no delay reproduce the data Rom less the uncertainty. Site investigation consists of the following
Table 2 and represent the baseline uncertainty in cost estimate essential components:
as apparent at the planning stage for the projects. ~ Desk top study and walk-over survey.
(ii) Should delay occur due to 'unforeseen'round during ~ Ground investigations.
construction then the cost efi'ects are significant and of a ~ Construction monitoring and feedback.
greater order than cost escalations envisaged at planning. This ~Performance observations and feedback.
is particularly the case for groundworks on the critical path of The main difficulties are the development of an effective
an activity schedule. management system, strategies for investigation and provision
(iii) In practice, most 'unforeseen'round events are capable of resources. An outline strategy considering levels of ground
of detection by best SI practice and management strategy. risk and development complexity has been suggested'.
Planning estimates are then more reliable and cost uncertainty For good practice in SI, resource has to be available. These
is reduced by up to a factor of 10. Best practice in SI has consist of time and money. Time is required to carry out full
measurable and significant financial benefit to developers and desk top studies, plan the explorations, to specify reliable and
financiers. representative levels of drilling, sampling and testing and to
(iv) The cost to these projects for a delay of one week is about report the data. Money is necessary to fund works of the right
0.8% for the industrial development (on a low risk 'greenfield'ite)
extent and quality. Time and money should be considered as
and about 2.7% for the commercial development (on a equals, time does not necessarily relate directly to cost. It is to
high risk 'brownfield* site). Such costs would be likely to be noted that additional time risks relating to the programme
purchase a 'very good'ite investigation for both these can occur where a site investigation is subject to a form of
developments. regulatory appraisal, for example on contaminated land sites
For planning and estimating, therefore, a realistic guide and highway schemes (particularly in Scotland).
figure for costing a SI is likely to be the cost of a one week Management structures and strategies have to be planned
delay. This produces the benefit that such a delay would have a and co-ordinated, preferable by an appointed geotechnical
low probability of occurrence and thus be avoided at a net advisor'. Experience shows two major management problems
financial balance. More importantly, the risk of severe cost have to be addressed:
escalation due to ground uncertainty is reduced to low and ~A lack of appreciation or awareness of the potential ground
acceptable levels. hazards leads to wrong questions being asked leading to
(v) The model analysis can be compared to events at the incorrect or inappropriate answers. This is particularly true in
prototype works. The industrial development had the planning investigations, defining objectives and preparing
earthworks delay of three weeks at a reported $22,000 cost. specifications. The consequences of a lack of awareness can be
This activity was not on the critical path and the development severe, not only in technical terms but also financially.
was completed almost to cost and time. The additional cost ~ If ground hazards are suspected, or known, then the problems
and delays were within the range Rom the model but at a are not communicated to other stages in a development,
calculated low probability. The site investigation was 'good't a eg design, construction and monitoring. Poor communication
low ground risk greenfield development. can lead to hazards being overlooked or neglected as projects
The commercial development experienced eight weeks delay progress.
on a critical path activity with a reported $900,000 cost. This A problem exists within the site investigation industry in
cost is within the range &om the model and is close to the most relation to the quantity of information and data produced.
likely value (note $900,000 = 21% of estimate). The site When conventional methods are used the quality of the data
investigation was 'good'ut it has been reported that can be poor and can be erroneous or contradictory. For
subsequent investigation of the site history may have identified example, on a recent contract a borehole log described 'Firm
the extent of the obstructions, these may have been Peat'et the soil description from test samples reported 'StifF
predicted. Clay'or the same horizon. Out of 81 strength tests, 40 tests 35
GROUND ENGINEERING ~ OCTOBER '995 GROUND ENGINEERING ~ OCTOBER ~
1995
were described as being stronger than the borehole log stated References
(eg stiff clay on the test sheet, soft clay on the log), five tests 'non. 'SI Psyche'. Ground Engineering, Vol28, No2, pp22-28 March 1995.
described samples as being weaker than the log record and in a 'ott MacDonald & Soil Mechanics Ltd. 'Study of the efficiency of site
investigation practices'. Proj Rep 60, TRI Crowthorne, 54pp, 1994.
further six cases there were direct contradictions between 'CE. 'Without site investigation ground is a hazard'. Report 1, Site Inv Steering
sample description and log record —a total of 51 anomalies Gp, Ground Eng Board. Thomas Telford Ltd, London 1994.
fiom 81 results, only 27% were consistent! Under commercial 'ational Research CounciL 'Geotechnical site investigations for underground
pressures of tendering and fee bidding it is difficult to maintain projects'. Vols 1, 2. Nat Ac Press, Washington DC, 1984.
'lhalaby NMH & Whyte IL 'The impact of ground risks in construction on
standards and detect and correct quality deficiencies when project finance'. Risk and reliability in eng. (BO Skipp Ed). Proceedings Conf
these occur. Major ground risks may not be detected under ICE, Thomas Telford Ltd, London, 54.67, 1994.
such circumstances. 'hyte IL. 'Site investigation and construction project risks'. Final report, SERC
The high level of dissatisfaction with standards amongst site Bldg Des, Tech and Management, Ref GR/G55682, 1994.
'onks D & Whyte IL. 'Site investigation strategy for contaminated land'. Proc
investigation companies is of concern. They blame clients, or Conf Polluted & Marginal Land '94: Re-use of contaminated land and land fills,
those procuring services, for a reluctance to spend a realistic Brunel Univ, ECS Publications, Edinburgh, June 1994.
amount on SI services and the acceptance of lowest tenders
without checking quality of service or product. Equally,
perhaps, the SI companies can be blamed for deficiencies in
marketing their services and product, not surprising when they
appear to have little faith in their standards. Clients have to be
convinced of the need for a planned and managed SI strategy
and the geotechnical engineer is in the only position to do this
in an informed manner. Argument has to be not just on
technical and quality grounds but also has to be based on
sound financial and commercial sense.

Conclusions
OGround hazards give rise to risks of cost escalation and delay
to construction projects. Evidence &om tunnel contracts in
USA and highway schemes in Britain shows the levels of
financial uncertainty on out-turn costs to decrease rapidly with
increased investment in site investigation, ie. better knowledge
leads to better planning. Low levels of site investigation can
lead to cost over-runs of between 0% and 100% (or more) of
the tender amount.
OSite investigation is a low cost activity and good practice
requires increased levels of resources compared to general past
practice. It is to be noted that the GIPI has fallen by about
25% to 30% since 1979.
OTwo development projects, one on a 'greenfield'ite and the
other on 'brownfield'and, provided data for a probability risk
analysis of the potential ground hazards. At the
planning/estimating stage, good SI practice can be expected to
produce estimates with an acceptable range of uncertainty.
Should development proceed with a lack of knowledge on
ground conditions then 'unforeseen'vents can cause delay
leading to unacceptable levels of cost escalation. It is suggested
that a suitable measure for a site investigation could be the cost
of a one week delay to a development.
OGround risks can be at acceptable levels with effective site
investigation. This requires management, communication and
resources of time and money.

Acknowledgement
The author acknowledges the work of Dr WS Peacock and Dr
NMH Alhalaby. Research was funded by SERG and thanks are
given to those individuals and companies that provided
information for the case studies.

36
GROUND ENGINEERING ~ OCTOBER ~ 1995

You might also like