## Conclusions

The variability of the coating process, the number of coats of paint, the complexity of the surface and the use of a minimum DFT rule can all result in a mean DFT far greater than that which is specified. The shift in the mean can be close to or exceed the x2 DFT maximum value that paint companies generally recommend as good practice.

Clearly, it is easier for a shipyard to apply more paint to make up for low DFT than to remove paint in the event of excessive DFT. While is important to achieve a certain minimum DFT for the coating to perform, there is a real danger that the use of minimum DFT rules will lead to higher than expected DFT readings, and this can also lead to performance drop off or even failure of the coatings.

The problem for the yard is that the application of this extra paint not only increases man-hours and coating costs, but also extends over-coating and drying times, and increases VOC emissions.

For the owner, the problem is that the DFT provided may be in excess of that recommended by the paint supplier as good practice, and the impact of excessive DFT on coating performance may not be well understood.

The reality is that unless current coating application techniques are improved, the range of readings that will be obtained in practice for any given specification will depend on:

- The number of coats
- Structural design complexity
- The skill of the applicator
- The condition of the equipment used

The Technical Data Sheets (TDS) provided by paint suppliers need to be very specific as to the DFT value that is being quoted. It is likely to be preferable to simply quote a range from the minimum to the maximum acceptable for each coat, rather than some vague value that could be interpreted as a minimum, a mean or some other measure such as nominal.

Paint suppliers would be prudent to test their products at expected DFTs that may be achieved in the field and provide data on the TDS for the elevated thickness expected.

Thus the IMO PSPC specification may be better written as a range of 288 µm–640 µm; this would imply a mean of about 464 µm. The only problem, given the complexity of some aspects of ship structures, is that for the range to be practically achievable, it must have a greater maximum. This would probably be more like x3 the nominal value of 320 µm, thus giving a range of 288–960 µm, implying a mean of 624 µm (assuming a normal distribution).

In another example for ballast tank coating, the maximum limit set in South Korean yards is 2,000 µm. While this seems high, it is clearly an attempt to push the upper specification limit well beyond the capability limit of the application process to ensure that there is never any need for re-work in the form of removing excessively thick paint.

The use of minimum value rules in a specification, such as 90:10 or 80:20, will tend to increase the mean DFT excessively. It will tend to move the whole distribution to the right. The process variables tend to also lead to a non-normal distribution and result in a skewed distribution. The most significant of those variables is the touch-up process that takes the lower DFT values and moves them significantly to the right by the addition of more paint through brush or spray application.

## DFT Gauge Readings

DFT gauges are set up to assume that readings represent a normal distribution and provide the statistical results associated with such a distribution. However, as demonstrated, the DFT readings for a ship tend to fall into a skewed distribution, which would potentially raise concern about how the data from a DFT gauge is presented.

However, not all is lost because more statistics can come to our aid. But that requires a careful look at how we collect DFT readings.

If, instead of taking individual readings, the requirements of SSPC PA 2 are considered, then there is a requirement to take three readings for a spot measurement. This “grouping” of readings will tend to result in the data being forced to form a normal distribution. This occurs as a result of the Central Limit Theory (CLT) and is why the SSPC PA2 method results in a requirement for fewer DFT readings.

If you do not group the readings in that way, you require more readings to invoke the CLT to generate a normal distribution. This is the basis of Control Chart Theory as advocated by Shewhart in the 1960s (Statistical Process Control – Grant and Leavenworth, McGraw Hill).

Therefore—if in doubt—the more readings you can collect, the better the overall picture you will obtain of the coating of a particular area. However, where time is limited, a smaller set of data collected correctly can give you a reasonable overview.

## Implications for Paint Specification

What should be taken away from this study is that the way that coatings are currently specified is inadequate and how the DFT is provided on the TDS can be quite misleading.

It is recommended that the TDS should simply contain a maximum and minimum value for DFT rather than a single ambiguous value. This would leave each paint supplier to determine the DFT range over which their products will provide the claimed performance. Of course, this would add some complications in that drying times, cure times and other data that may be affected by DFT (such as time to service) will need to reflect the range that is provided.

Coating specification writers should also consider the DFT range as more important than a specific DFT value (nominal, mean or otherwise). The range would then reflect any minimum/maximum values recommended by the paint supplier.

The challenge is to achieve a range that is achievable by the application process.

## Other Sources of Variability

It would be nice to be able to conclude that the problem stops there. However, there are other sources of variability that also have an impact on results and even these can be significant when measuring the DFT achieved. These can be summarized as follows:

**Sampling Regime:** How many samples are taken and where they are taken? Does everyone follow the same sampling regime? If not, there can be significant differences in the results.

**Accuracy of DFT Gauge / Repeatability of Readings:** DFT gauge variability can cause different results even for people using the same sampling regime, as highlighted by Ault (Coating inspection data reproducibility, J. Peter Ault, Mega Rust 2008).

**Operator Capability: **The skill of the operator is critical and can also affect the readings taken.

**Location of Readings and Substrate: **Are the readings taken across a representative sample of the structure, as per IMO PSP requirements? Is the substrate condition known to enable allowances for profile to be made?

It is a well-known adage that the more you inspect, the more you will find wrong and so in part, the inspection process needs to be well defined to assure that it is not punitive, while achieving the required assurance.

**What if some few points are outside the range either way, what is acceptable?**

The authors are of the view that the range on the TDS should define the absolute minimum and maximum for various in-service uses for any given product. This range should assure the required in-service performance—assuming drying/cure times, over-coat intervals, etc. are adhered to.

The problem with ships, in particular, is that even if you accept 1 or 2 % of the readings to be outside the range provided, the total area equating to such a small percentage can still be quite large. For example, 5% (equating to ±2s) of 30,000 sqm of cargo hold would mean that 1,500m^{2} would be outside the range. This is not an insignificant area.

It is clear that variability should be minimized and that technical documentation needs to better reflect the in-service practicalities.

*****This article was co-written with John Fletcher.**

*With more than 45 years’ experience in the corrosion, protective coatings and electronic **inspection technology fields, John Fletcher serves as technical support manager at Elcometer Ltd. in Manchester, England. He is the current president of the Institute of Corrosion (iCorr), and chairman of ASTM International Committee D01 on Paint and Related Coatings, Materials and Applications. As a top international expert in paint testing and inspection methods, Fletcher also leads Subcommittee D01.23 on Physical Properties of Applied Paint Film.*

*****

__More in the "The Problems with Meeting Marine Paint Specifications"__

__s__**eries:**Working to Overcome Poor Quality Coating Practices in International Shipbuilding

Ambiguities in Determining if 'Specified DFT' has been Achieved

The Impact of Minimum & Maximum DFT Values on Coating Performance