Test Results and Discussion
The 83 diesel samples collected with the normal FQM programme were sent to Inspectorate for testing in June 1998. There was insufficient fuel in one sample to test for both CN and CI, so for this sample only CN was measured. A further sample was a duplicate of another within the batch. These two samples were omitted from the analysis.
Time Distribution
Graph 1 shows the values of CN and CI976 plotted against date of sampling. The striking feature of this is the variability of diesel quality during the period. For the first three months, May to July 1997, quality was reasonably consistent. CN varied between 48 and 52 and CI976 varied between 50 and 54. From August 1997 average CN and CI976 both rose slowly. The average increase in CN was smaller than the increase in CI976. Also the quality of diesel, as indicated by the range of values of CN (between 47 and 58) and CI976 (between 51 and 59), varied more than in the first three months.
Graph 1: Time Distribution
The difference between CN and CI976 of six out of the eight samples collected between mid August and mid September was less than for the rest of the period. Four of these samples had CN higher than CI976. The difference between CN and CI4737 for these same samples does show some effect was also unusually small. According to Marsden Point Refinery, cetane improver was added to the diesel produced by the Refinery between 11 and 21 August 19979, which explains this observation.
CN is shown to be lower than CI976 on average over all samples. This is expected. The limits in the Regulations are a minimum CN of 45 and a minimum CI976 of 47, the fuel having to meet at least one of these criteria. All samples met both of these requirements and also the density requirements. This indicates either that suppliers are good at tailoring fuel to meet the specifications, or that the limits set in the Regulations are in the right relationship with each other.
From the point of view of ability to comply with the Regulations, this study indicates that there is no need to change the specified limits of CN and CI. Any increase in specification limits (for CN and CI) results in better engine performance and would have environmental benefits but it is likely that changed manufacturing process or the addition of a cetane improver would be required to achieve this.
Frequency Distribution
Graph 2 shows the frequency distribution of CN and CI976. About 88% (71 samples out of 81) have CN between 49 and 54 and about the same percentage (90%) of sample have CN between 51 and 58. This reinforces the observation that an average CN of New Zealand diesel is 2 less than its CI976.
Graph 2: Frequency Distribution
Measured Properties
Averages of the measured properties: CN, density and distillation temperatures are given in Table 1.
Density of these diesel samples varies between 0.820 and 0.858 g/mL with average of 0.842 g/mL. All samples fell within with the limits specified in the Regulations (0.81 g/mL minimum and 0.86 g/mL maximum).
Density can provide useful indications of the fuel's composition and performance-related characteristics such as ignition quality (CN), power, economy, low-temperature properties and smoking tendency. Fuel injection equipment operates on a volume-metering basis, so a change in density will influence engine output due to the different mass of fuel injected. A higher density fuel will tend to produce more smoke as well as more power.
The distillation, or volatility, characteristics of diesel are expressed in terms of the temperature at which successive portions are distilled from a sample under controlled heating. The distillation range of a fuel influences properties such as viscosity, flash point, auto-ignition temperature, CN and density. The low distillation temperature indicates the ease with which the fuel will start to vaporise, while the high distillation temperature10 gives an indication of the extent to which complete vaporisation of the fuel may be expected. High boiling components may not burn completely; forming engine deposits and increasing smoke emissions. The average distillation temperatures as well as the extremes are given in Table 1.
Five samples appear to have come from the same batch of diesel with a higher than usual distillation temperature. Full distillations of four of these samples were done. The fifth was not fully tested because there was insufficient sample. Although the 90% distillation temperature for each of these samples was above 350°C, they met the regulatory requirements at 85% distillation.
Number of samples | | CN | Density
[g/mL] | % Distillation temperature [°C] | CI976 | CI4737 |
|---|
10% | 50% | 90% |
|---|
| 81 | Lowest | 47.1 | 0.8201 | | 263 | | 48.6 | |
| Highest | 57.7 | 0.8584 | | 302 | | 58.6 | |
| Average | 51.3 | 0.8423 | | 283.3 | | 53.0 | |
| 79 | Lowest | 47.1 | 0.8201 | 197 | 263 | 319 | 48.6 | 48.3 |
| Highest | 57.7 | 0.8584 | 250 | 302 | 358 | 58.6 | 59.5 |
| Average | 51.4 | 0.8423 | 226.8 | 283.1 | 338.5 | 53.0 | 53.7 |
27 (NZR)
(Auckland) | Lowest | 48.1 | 0.8242 | 208 | 272 | 327 | 50.3 | 51.0 |
| Highest | 57.2 | 0.8555 | 247 | 297 | 349 | 58.0 | 59.4 |
| Average | 52.1 | 0.8419 | 232.4 | 286.8 | 340.0 | 53.8 | 55.0 |
Table 1: Summary of Test Results
Cetane Index vs. Cetane Number
Graph 3 plots CI976, against CN for 81 samples and CI4737 against CN for 79 samples. A line of CI976 = CN is drawn. These samples with CN equal to CI976 are found on this line. Those with CI976 greater than CN are found above the line and those with CI976 less than CN are found below the line. Most points are found above the line which reinforces the observation that CI976 is generally higher than CN. A very small number of samples, 7 out of 81 was found below or on the line. Four of these between the period of mid-August to mid-September when the cetane improver product was supplied. These samples are different from the majority of samples either because the batch they are drawn from was processed by different refinery using a different method or because there was a major change in processing system, such as the addition of cetane enhancers.
The comparison between the CI976 and CI4737 is interesting. Overall only 17 of 79 samples (22%) had CI4737 higher than CI976. Of the 30 samples collected in the Auckland area, 27 were collected from service stations supplied by the four established oil companies. Of these 27, only one (4%) had CI4737 higher than CI976 (see graph 3 and 4).
The other three samples were of diesel supplied by the new entrant in the market.11 It is known that all this fuel is sourced from overseas. For all three of these samples CI4737 was greater than CI976.
There is a strong probability that apart from diesel supplied by Challenge, all diesels supplied to the Auckland area originates from the Marsden Point refinery. It is therefore reasonable to assume, on the basis of the observed differences in the relationship between CI976 and CI4737, that some of the diesel supplied through terminals other than Auckland is sourced from overseas refineries.12
Graph 3: Comparison of Cetane Number and Index
An investigation of the four-variable equation (ASTM D 4737) was carried out by a panel of the Institute of Petroleum using 167 European distillate fuels.13 The CN range this study was between 31 and 62.14 Results show a very large variation in the relationship between CI and CN. The variation is much larger than found in this test because the range of CN for diesel fuel used in New Zealand is narrow (47-58) and close to the majority diesel made in world.
A line of best fit is drawn on graph 3. The equation for this line is CI976 = 0.951922 * CN + 4.12858, which has a coefficient of correlation of 0.7 (one is for a perfect fit relationship). This equation has limited use to estimate CI976 from CN or vice versa. The main reasons for a relatively weak relationship are the random error in the test methods and variations in the fuel samples which may come from a variety of refineries or change in process system by same refinery, including addition of cetane improvers.
Diesel from Marsden Point Refinery
Detailed information about the fuel from which the samples represented in this study, such as the refinery the fuel came from, its make-up and levels of cetane improver, is not available. Therefore it is generally impossible to divide results into groups for further analysis to determine the relationship between CN and CI more closely.
The exception is that all four established oil companies use the direct pipeline between the local refinery at Marsden Point near Whangarei and Wiri in Auckland to transport petrol and diesel. Therefore we may assume that most of the diesel supplied from the Auckland terminal by the four established suppliers comes from Marsden Point refinery. There are 27 such samples.
Diesel Supplied to Auckland
Graph 4 plots CI976 and CI4737 versus CN as measured by Inspectorate for the 27 Auckland samples.
On this graph, all but two CI976 points and all CI4737 points were above the line of CI = CN. In other words, for fuel from Marsden Point CI976 is almost always higher than CN and CI4737 is always higher than CN. The overall indication is that for diesel from Marsden Point CI976 is about 1 to 2 higher than CN, and that CI4737 is about 3 to 4 higher than CN. Similarly, CI4737 is about 2 higher than CI976.
Thus, for diesel sourced from Marsden Point Refinery, CI976 gives a closer approximation to CN than CI4737 does.
A line of best fit is drawn on graph 4. The equation is CI976 = 0.9849 * CN + 2.41224. The coefficient of correlation for the relationship is 0.75, a stronger relationship than for the whole sample.
Graph 4: Comparison of Cetane Number and Index for Auckland Samples
Comparison between Laboratories
As has already been described, the samples used in the study were collected during the routine FQM programme. Instead of the usual one litre can, two one litre cans were collected at each sampling site. Throughout the year, the first can of each sample was tested for CI by BP Lab within 50 days of sampling using ASTM D 976. Testing samples in batches helps to keep costs down. The second cans of each sample were retained until shipped to Inspectorate for testing.
The sampling method involves putting the fuel samples into cans which are then immediately sealed. Both can and seal are suitable for the purpose.
Graph 5: Comparison of Test Laboratories
Graphs 5 and 6 show a comparison between the test results of the two laboratories, Inspectorate (CIIS) and BP (CIBP), both using ASTM D 976. In graph 5 CIBP is plotted against CIIS. In graph 6 (CIBP - CIIS) is plotted against date of sampling.
Graph 5 shows good agreement between CI as determined by each of the two laboratories. For only 8 of the 81 samples was CIBP - CIIS more than 1. The coefficient of correlation between the results from the two laboratories is 0.93.
Graph 6: Comparison of Test Laboratories
By the time tests were done by Inspectorate (Singapore) the first batch of samples was over a year old, and so the time between the tests done by BP Lab and the tests done by Inspectorate is up to a year. Graph 6 shows no discernible trend in CI as a function of the age.
This analysis demonstrates two things. The first is that both laboratories are in good agreement, given the uncertainty of testing. This gives us greater confidence in the routine test results.
The second is that the current sample storage and transport systems appear to be appropriate for diesel and that at least one property, CI, does not deteriorate over a period of up to a year.
Comparison with Diesel of Other Countries
Because Inspectorate (Singapore) is a part of a worldwide group it was able to provide CN and CI data for diesel supplied in eight other countries. This information is given in Table 2.
Overall, it seems that New Zealand's diesel quality is slightly better and less variable than most countries. Except for Hong Kong and Sweden the CN of New Zealand diesel is on average 1 to 5 CN higher than that of the other countries listed in Table 2. Hong Kong and Sweden have average CN about 3.5 CN higher than New Zealand. However it should be noted that there are considerable uncertainties in an ad hoc comparison of CN like this because the effects of additives, climate variation in production in the different countries have not been taken into account. The average CI4737 of 2059 production samples of diesel supplied by Mobil Oil Australia is 51.1.15 The Australian fuel has a much larger range than in New Zealand (44.7-62.1). Whilst Australia doesn't have government regulations they do work to a minimum specification of 45 CI.
Generally the difference between the average CI976 and average CN is small - in most instances well below 2. In some cases average CN is higher than average CI976; in other cases it can be lower than CI976. New Zealand fuel is at one end of the range with CI - CN = 1.5 while Sweden is at the other end with CI - CN = -2.6.
| Country | Number of Samples | CN615 | CI976 | CI4737 |
|---|
Lowest | Highest | Average | Lowest | Highest | Average | Lowest | Highest | Average |
|---|
| New Zealand | 79 | 47.1 | 57.7 | 51.4 | 48.6 | 58.6 | 53.0 | 48.3 | 59.5 | 53.8 |
| Austria | 4 | 48.4 | 51.4 | 50.0 | 48.9 | 50.1 | 49.4 | 48.9 | 50.0 | 49.3 |
| Denmark | 4 | 47.9 | 50.4 | 49.2 | 47.4 | 50.2 | 48.6 | 46.5 | 49.1 | 47.7 |
| France | 13 | 46.3 | 51.3 | 49.0 | 45.9 | 54.8 | 50.0 | 44.8 | 54.3 | 49.3 |
| Germany | 24 | 46.7 | 54.8 | 50.2 | 47.2 | 53.0 | 50.3 | 47.2 | 53.6 | 50.2 |
| Russia | 3 | 44.6 | 47.4 | 46.1 | 45.4 | 47.6 | 46.5 | 50.4 | 53.6 | 52.5 |
| Hong Kong | 5 | 53.3 | 59.8 | 56.9 | 53.0 | 57.2 | 55.3 | 54.3 | 60.4 | 57.0 |
| India | 6 | 45.7 | 53.4 | 51.1 | 47.2 | 55.3 | 51.6 | 45.5 | 55.2 | 50.0 |
| Sweden | 5 | 53.4 | 55.9 | 54.9 | 50.7 | 53.3 | 52.3 | 52.0 | 55.2 | 54.0 |
Table 2: Cetane Number and Cetane Index of Diesel from Various Countries
Differences between average CI4737 and average CN are less consistent. For some countries, CI4737 and CN are very close but in other cases the difference between them is much greater. The largest difference was found for Russian diesel. The average CI976 is very close to the average CN, but CI4737 is different by about 6.5. However it should be noted that data from most of these countries is based on very small number of samples.
The information on diesel from other countries needs to be treated with caution, and should not be used to make definitive statements about quality of diesel in these countries. No information is known about the samples other than tabulated test results. Quality of diesel varies from one batch to another, from one refinery to another and also seasonally. Sample sizes are small, generally less than 10. With this data alone it is difficult to compare the quality of our diesel with that of other countries to any great depth.
Implications for the FQM Programme
When the specifications in the Regulations were first made in 1988 diesel supplied throughout New Zealand was more likely to have come from one source, Marsden Point Refinery, than is the case now. The relationship between CN and CI976 that appears to have been assumed in setting the Regulations would have been a good match with the diesel from Marsden Point Refinery.
The petroleum supply industry has much more diversity than when the specifications were first drawn up. Companies other than the four established suppliers have entered, or are about to enter, the market. These new entrants source their entire diesel from overseas and the established companies also source some of their fuel from overseas. In addition, as has already been discussed, there has been at least one occasion when the local refinery has added cetane improver.
Because the relationship between CN and CI varies with the refining process and any additives used, the use of CI instead of CN has to be treated with caution. Despite this using CI is appropriate for FQM monitoring on a routine basis because of the expense and logistical difficulties of continually sending diesel samples overseas for CN testing.
However it is recommended that the routine FQM programme be reinforced by further CN testing on an ongoing basis. Options include:
- Repeating this study every three to five years;
- Sending up to half the number of samples for CN testing every second year to monitor significant changes;
- No further CN testing and placing reliance on CI results alone.
The second option is most likely to balance costs and benefits, provided sufficient samples are tested to provide a statistically valid result.
Back to Top