“Anything to declare?”: Manufacturers' declared values and Quality Control of CIPP Liners

Ensuring the quality of Cured in Place Pipe Liners

Cured-in-place pipe (CIPP) liners have become the most common method of sewer renovation. Although workmanship is always important, the strength of a CIPP liner is highly dependent on the workmanship. Quality control (QC) is therefore highly important part of installing CIPP liners.

To assist in the better use of QC an ISO Committee is currently developing draft technical specifications ISO/TS 23818 Assessment of conformity of plastics piping systems for the rehabilitation of existing pipelines for various types of lining system. Part 2: Resin-fibre composite (RFC) material will cover CIPP lining systems. The specification gives recommendations for sampling rates and procedures for:

• Type Testing (TT)
• Batch Release Testing (BRT)
• Process Verification Testing (PVT) and
• Audit Testing (AT) by or on behalf of a certification body.

As clients have come to appreciate the importance of ensuring the performance of CIPP liners, they are increasingly requiring independent testing of samples taken from installed liners.

Manufacturer's Declared values for design

ISO 11296-4 Plastics piping systems for renovation of underground non-pressure drainage and sewerage networks Part 4: Lining with cured-in-place pipes, which is the product standard for CIPP liners, no longer specifies values for the key material properties used in design but relies of the manufacturer declaring a value. This is sensible since there are now a huge variety of products with a wide range of material properties.

Design values for different design methods

Design methods in Europe are increasingly using a partial factor method of design which normally use 5% fractile values of key material properties. The French ASTEE 3R and the German DWA method already use this partial factor approach and the design method in the forthcoming WRc SRM Sewerage Renovation Design Guide that will be published in the new year will also use this approach. This is the value which will be exceeded in 95% of test results. If it follows a normal distribution it can however be estimated from the mean value minus a fixed value (k) times the standard deviation (see figure). Where the number of test results is small the value k is increased to take into account uncertainty.

ISO 11296-4 does not give much information on how to determine the declared value, it merely says:

Where requirements are specified in Tables 5, 6 and 7 as declared values, these declarations shall be documented for each CIPP product, with supporting test data or references to such data, in the installation manual for that product.

This begs the question as to whether the manufacturer’s declared value is a mean value, a 5% fractile value or something else. The designer will therefore need to engage with the supplier to obtain appropriate values for the design method being used.

Long and short term properties

As a further complication CIPP materials are subject to creep and so the actual values we want in design are based on long term testing. Long term tests, by their very nature, take time (typically 14 months) and are consequently expensive. So although they can be used in type testing they are not very useful for process verification testing or for site QC testing.

This is covered by establishing the ratio – a conversion factor – between the short and the long-term values for each property and then using only the short-term values for process verification and QC. If the short term values are consistent with the manufacturer’s test results then we can assume that the long term values will be as well.

However if they are not consistent with the manufacturer’s test results then it is not simply a matter of de-rating the long-term values in line with the lower short term values as if the liner has not properly cured we can no longer be sure that our conversion factor is valid.

What should we do with our site QC test results?

We could just compare the 5% fractile value of the sample with the design value, However, we only have a very few results for each liner; typically no more than 5 test results. To estimate the 5% fractile from such a small sample has a high uncertainty. Also, we know that we should expect 1 in 20 samples to be below the 5% fractile value, so just comparing the 5% fractile of our small site sample to the design value is a huge oversimplification.

What we really want to know is whether the material as cured by the installer in the field is the same as the material tested by the manufacturer in the factory during type testing. If it is not, then how can we be sure that the relationship between the long-term material properties and the short-term properties is the same? If the short-term properties are not as good, we cannot simply factor down the long-term properties as such inferior material may be even worse in long-term loading.

However, given the variation in the results how can we be sure they are the same as the material tested by the manufacturer?

A new approach

The only way we can find out if the material cured on site is the same as the material the manufacturer tested is to compare the statistics of our QC test results to the statistics of the results of the original test results that the manufacturer produced in their type testing and on-going process verification testing.

We can do this by checking that the mean and standard deviation of our QC test results are consistent with the mean and standard deviation of the manufacturer’s test data to the required degree of confidence.

To do this we must first know the mean and standard deviation of the manufacturer's short term test data. We can then check this against out own test data using standard statistical test methods.

Details of this new approach will be published in the new edition of the SRM Sewerage Renovation Design Guide which is due out in the new year.