| Front page | | Contents | | Previous | | Next |

Forureningstransport via utætte boringer

Summary and conclusions

Investigative studies

A study involving the investigation and evaluation of drilling techniques, well construction techniques and well construction materials used in the last century in connection with installation of water supply wells in Denmark was conducted. From this study it can be concluded that there are a number of risks that well leakages will occur and may result in contaminants being transported to the associated groundwater aquifers or to drinking water supply wells.

The wells may be divided into 5 categories, where types I – III are screened wells, while types IV and V are uncased, unscreened wells. Additionally, in type III wells the temporary casing is replaced with a permanent well casing. A more detailed description of the wells is found in section 2.3.

Vertical leaks along the casing

Vertical leakages along the permanent casing are often associated with the use of a boring tube fitted with a socket joint and a casing shoe, as these can create a channel along the casing. In particular, wells with the following characteristics have a high risk of vertical leakages along the casing: wells drilled without drilling fluid (for example using a cable tool or an auger), prior to 1960, where the drilling tube was used as the well casing, and where there was no grouting whatsoever (well types I and II). A similar risk of vertical leakages arises with uncased wells when the work casing is used as the permanent well casing in the upper impermeable layers (well types IV and V). The risk of contaminant influx associated with vertical leakages is greatest when drilling through a compact, cohesive layer. In contrast, the creation of new transport pathways arising from an increase in borehole diameter will only occur to a limited extent in permeable layers.

The risk of leakages in screened wells (type III wells) is closely linked to the presence and quality of the grouting material surrounding the casing. Grout was generally not used in this type of well construction prior to the 1980’s. In cases where grout was used, it was often an unsuitable material. For example, from about 1975 – 1980 various grouting materials were used which, in many cases, resulted in the formation of preferential pathways through the grouting material.

Another risk of contaminant transport along the casing is associated with type III wells where the bentonite sealant has subsided. During auger drilling, cavities can form in the formations along the borehole. While these cavities are later backfilled, subsidence and the formation of vertical preferred pathways often occur. The subsidence associated with auger drilling is the result of improper bore pipe placement during drilling. In the case of rotary drilling, subsidence is primarily associated with insufficient removal of drilling mud and cuttings before well completion. Consequently, a re-bedding of the gravel pack can occur during backwashing causing the sealant to subside.

If the bentonite sealant does not have a suitable height, another problem can arise where the resistance against flow through the sealant is less than the resistance against flow through the surrounding layers. In some cases this error can be attributed to the fact that the standard minimum height of the sealant of 1 m is, under some circumstances, insufficient.

Finally, the sealant can be defective because the permanent casing diameter is too large relative to the borehole diameter making it impossible to seal the well properly.

Leaky permanent casings

Corrosion of permanent steel casings would generally be expected 20 – 30 years after installation. Steel work casing was used as permanent well casing until about 1960, and thus, it can be expected that such wells have a high risk of leakage. In addition, it is possible that holes can occur in other casing materials especially in association with pump installation, etc.

Leaky permanent casing connections

Transport of contaminants through leaky pipe connections is a problem which may potentially occur in all types of wells. Since 1975, connections between PVC pipes have been sealed with teflon tape and in some cases O-rings. However, investigations have shown that teflon tape can have the opposite effect if it is wound around a joint or pipe connection too many times. Socket joints attached with glue and screws, which were used mainly between 1965 – 1975, have an obvious risk of leakage in situations were the screws are screwed through the casing and later rust away. PVC pipe connections in casings are vulnerable to compressive force and thus, often leak. Connections in steel pipes can also leak, although this problem is not considered to be as widespread as that of leaky PVC and PEH connections. The latter problem has not been solved even in newly installed wells.

Leaky casing heads and well covers

Leaky well covers combined with a leaky casing head can lead to direct downflow from the surface into the well. Unintended inflow can also occur via leaky well covers, leaky connections in the concrete manhole ring, and accidental damage to the well construction during burial of underground cables. Inflow of surface water can also occur if the surface housing is placed near ground surface, etc.

The risk of surface water seeping into the well is high in situations where a leaky casing head is found and where the area around the well is backfilled with high permeable materials.

The risk of contamination associated with leaky casing heads and well covers is existing. However, in the majority of cases, only a minor effort is required to minimise this problem. Furthermore, surface water seepage into a well is often easily registered. The problem is believed to be largest in single production wells.

A second risk associated with casing heads arises in connection with sealed casing heads installed without ventilated flanges. Sealed casing heads are numerous because this practice can have a positive influence on the processes in the well so that less well maintenance is required. However, it is crucial that there are no leaks whatsoever, because the vacuum which is created when pumping starts will draw surrounding pore or groundwater through even the smallest holes into the well.

Ineffective well closures

Only wells which were closed after 1980 by means of filling them with a suitable material are believed to be effectively closed. Effective sealing materials were introduced first in 1980. Thus, the problem of unsealed abandoned wells has only been solved in wells closed after 1980 using suitable materials.

The risk of contamination via ineffectively closed wells is believed to be common. Many abandoned wells are located on water works sites, often close to in-use water supply wells, which is an unfortunate situation in terms of hydraulic gradients. Furthermore, in some cases an improperly closed well can act as a huge preferential pathway with the potential to contribute significant contamination to the near-well region of an aquifer.

In addition to the above descriptions of potential leakage situations, there are examples of short circuiting between various aquifers. This can occur when a single well is constructed with several screened intervals which are separated only by a casing with or without grout. This practice of multiple screened intervals within a single well was common in the past, but ought to be avoided due to the obvious risk for short circuiting between aquifers.

It should be noted that the types of leakages described can occur alone or in various combinations. The potential risk of contamination is increased in the latter case.

Database studies

Field investigations

The following section presents conclusions from selected field studies and investigations which were reviewed as part of this project. The discussion focuses on well construction. Based on the field studies and fild investigations the following conclusions can be made:

Generally, the contamination situation is complex, although the hypotheses regarding risk of contamination at the selected locations appear to be clear. In many cases the well construction is flawed, but there are also contaminant contributions via other pathways.

Field investigations indicate that numerous, relatively new wells are flawed, particularly by leaky connections (where there are connections) and improperly placed sealant.

Prefix has not been used consistently at all water works, based on concentrations in soil samples taken at water works sites. It is not possible to make general comments regarding the delivery routine. The source of the contamination may in these cases be the water works itself. This theory ought to be investigated in specific field studies.

Tests performed in wells with leakage from casing connections reveal a decreasing inflow compared to what normally could be extracted during pumping of a water supply well. Estimates of the effects of preferential pathways, leaky casing heads as well as of improper well closures were determined and are summarised in table 1.

Tabel 1.
Probable leakage to the aquifer from leaky wells.

Table 1 illustrates that the leakages in a defective well will often be minimal relative to contamination of an aquifer with BAM. A contribution via a well-related leakage often occurs, but generally, under normal pumping rates, the BAM concentration in water supply wells is under the maximum permissible value. The explanation for the low concentrations is the low BAM concentration in the secondary groundwater as well as limited infiltration. During this project, BAM concentrations in the secondary groundwater seldom exceeded 1 µg/l and concentrations over 10 µg/l have never been seen in investigations. Thus, the above discussion demonstrates that BAM concentrations will be under the maximum permissible value regardless of the type of leakage, as long as normal pumping rates from a water supply well can be assumed.

At small water works, with lower rates of water extraction, and where operation may be discontinuous, leaky wells, particularly in association with preferential pathways, can lead to drinking water quality problems, if high BAM concentrations are present in the upper secondary groundwater. Experiments with continuous alternating operation indicate no apparent difference in the BAM concentrations in the contaminated wells. In these cases, the aquifer may be contaminated with BAM.

Leakages resulting from improper well closure may be the cause of significant BAM transport, especially considering the often unfortunate location of abandoned wells close to in-use water supply wells.

Water quality samples from a test well drilled into the secondary or primary aquifer and located directly upgradient from a contaminated well will often (if contamination is found) provide evidence as to whether the contaminant transport has been via leakages in a well or via the aquifer.

Unsuccessful attempts to renovate a defective boring may be attributable to an incorrect hypothesis regarding the source of contamination or its pathway(s).

Summarised results from water works well survey

The majority of the wells which were investigated were found to have some type of flaw. Evidence of defective well construction was registered in 144, or approximately 84% of the wells.

The majority of the leaky connections were found in type III wells (screened wells) (57%) and in PVC-cased wells (61%). The majority of holes were observed in type I and type II wells (wells were the temporary casing remains becomes permanent) (46%) and in wells where a steel pipe is used as the permanent casing (42%).

Conclusions derived from BAM investigations conducted as part of this project

Of the 27 completed investigations there is definite evidence of aquifer contamination at 13 locations. There are 8-9 locations with contaminant transport via both the well and the aquifer. Transport contamination which can be attributed exclusively to defective well construction was observed in only 2-3 investigations. It was not possible to determine the transport pathway in 3 investigations.

Significance of well leakages in relation to BAM transport

In the majority of cases, BAM concentrations in drinking water supply wells with any type of leak will be well under the maximum permissible value, assuming that the BAM concentrations in the secondary aquifer are normal and the pumping rate is normal. However, at small water works, where the pumping rates are lower or discontinuous, preferential pathways, etc. may cause problems, if high contaminant concentrations are present in the secondary groundwater. For example, it would be possible to find BAM concentrations exceeding the maximum permissible value (0.1 µg/l) in water extracted from a well under the following conditions: the well has an unsealed casing; the pumping rate is about 1 m³/h); due to the effect of preferential pathways along the outside of the casing, the hydraulic conductivity is equivalent to that of a medium-grained sand; and there is a minimum concentration of 8 - 10 µg BAM/l in the near-surface groundwater which is in physical contact with the well. If, however, the hydraulic conductivity of the preferential pathways is equivalent to that of gravel, then a BAM concentration of only 0.5 – 1.0 µg BAM/l in the secondary groundwater in contact with the well is required before the maximum permissible value is exceeded in the well water. A situation like the first example would be rare because BAM is seldom present at these concentrations in the upper groundwater. While the BAM concentration levels are more realistic in the second example, the effect of preferential pathways of this magnitude is likely uncommon.

Evaluation of investigative methods as diagnostic tools

In the process of collecting data regarding well characteristics invaluable information was obtained which can be used to organise the subsequent well examination strategy.

TV inspection is normally a good tool to detect leaks. In cases of poor visibility, for example under the water table, only limited information can be retrieved. Furthermore, this method provides no indication of the volume of inflow. Pressure tests of the casing can test for leaks, but these tests are associated with practical problems and tests in old wells may lead to collapse of the casing.

Packers may be used to seal off an interval within a well and thereby obtain an estimate of the inflow via leakages. However, it is often difficult to maintain the pressure on the packers during testing. This is essential to ensure that any inflowing water during a test comes from leakages into the well, and is not well water flowing past poorly sealed packers.

While gamma logs are often used to verify the placement of the sealant, this type of borehole logging provides no information regarding the watertightness of the seal or the volume of inflow (if any). Conductivity logs can be used to supplement gamma logs, but the latter are not so widely used because they can give false signals in connections with iron screws, etc. Other types of borehole logging can be conducted, for example a caliber log (which logs the borehole dimensions), a flow log (which measures inflow), a resistivity log (which measures the geology in the screen section), a sonic log (attached to the casing). All these logging methods have the common characteristic that they are indirect and thus under normal circumstances they cannot be used alone, but must be considered as a supplement to water quality samples.

Multi-level water quality sampling is the most effective and precise method to determine from which level contaminated water is entering the well. Great care is required with this type of testing to obtain representative samples and thus, a correct interpretation of the hydraulic and contaminant situation. This requires the use of 2 or more pumps. Additional tools such as flow logs, packers and heat pulse probes can be used to ascertain the correct levels for sampling. Water samples of the annulus water in the casing can be used to determine the inflow contribution from leakages in the casing. In addition to BAM analyses, it would be expedient to analyse water samples for other parameters which are characteristic of surface water. Leakages on the outside of the casing can be determined by taking a water sample at the top of the screen pipe. There can be uncertainty as to the source of contamination when samples are taken at the top of the screen pipe. Contaminants could be transported down along the casing via vertical preferred pathways to the well or the source can be near-well aquifer contamination, or a combination of the two. The source can be definitively determined by long term separation pumping where numerous water samples from specific levels within the well are taken over a longer period. If representative water samples from the middle or bottom of the screened interval are taken during the same pumping period, information can be obtained regarding the inflow from the formation from more distant contaminant sources.

CFC dating can be used to estimate the risk of BAM contamination of a groundwater aquifer via natural infiltration or via well leakages where groundwater is mixed in the well with younger, near-surface groundwater. In cases of short circuiting involving younger groundwater, the estimated age of groundwater at the depth of the screen will normally not be representative of the known groundwater age at that depth. There are, however, process and factors which can affect dating results, and thus there are a number of uncertainties associated with this diagnostic tool. These reservations must be considered and overanalysis of CFC dating results should be avoided.

Information obtained from investigations inside the well can be verified by conducting investigations outside the well. Test wells drilled into the secondary or primary aquifer and located upgradient from the contaminated well provide water samples which can determine whether the source of contamination is related to leakages in the well, or to contamination in the aquifer. Analysis of near-surface soil samples for dichlobenil and BAM are other types of analyses which ought to have a high priority.

Reclamation of wells and remediation techniques

If well improvements are to have an effect on the water quality it is necessary to definitively establish the true cause(s) of the water quality problem before any attempt is made to reclaim the well. In many cases a contaminated well will continue to be so if remediation attempts are based on flawed or incorrect investigation conclusions. Clearly, renovation of a well will have no effect unless the most important contaminant transport pathways are closed.

To ensure that other wells do not become contaminated, it is often necessary to continue to control the hydraulic behaviour at the site via pump and treat and to continue to monitor the site after wells are reconditioned. Experiments with various pumping scenarios in areas with contaminated aquifers were unable to demonstrate a change in BAM concentrations regardless of whether pumping was continuous or alternating.

If a decision is made to seal a well, it is necessary to seal all leakages, as well as any preferential pathways. This will often require that the well is re-drilled or over-drilled at least to the same dimensions as the original well.

Conclusions