The EU Tyre Noise Label: The problem with measuring the noise level of only a few of all tyre variants Ulf Sandberg 1 Swedish National Road and Transport Research Institute (VTI) SE-58195 Linköping, Sweden Piotr Mioduszewski 2 Gdańsk University of Technology (GUT) ul. Narutowicza 11/12 PL-80-233 Gdańsk, Poland

ABSTRACT The STEER project, described in another Inter-Noise 2022 paper, has evaluated the performance of the noise label of the European tyre label. The major finding was that uncertainties of the tyre/road noise measurements were higher than should be accepted. One of the worst uncertainty sources was found to be the common practice to measure only some tyres of all sizes or variants within a tyre line, to save money. Generally, only the noisiest tyre(s) is/are measured individually and other tyres in that line get the same level, which means that many if not most tyres are labelled with too high noise levels. Then consumers cannot find the quietest tyres. STEER lacked resources to study this problem, but a Swedish project supported STEER in this respect, by making it possible to measure noise of 53 tyres selected from tyre lines of three major tyre manufacturers. The results showed that even though tyres in each line were labelled with the same noise level, in practice they differed up to 6 dB in noise emission. To avoid this very serious source of uncertainty, a special simplified test is suggested in order to be able to label tyres correctly without too much extra effort.

1. INTRODUCTION – THE EU TYRE LABEL

In 2009, the EU introduced a tyre label for wet grip, energy and tyre/road noise, which was imple- mented from 2012 and revised in 2021 [1]. The label of 2021 is shown in Figure 1. Other countries or regions have produced similar tyre labels, although not all of them include the noise parameter [2]. By establishing a mandatory label for noise levels of tyres, this in principle allows consumers to select a tyre based on its noise emission, in addition to the two other important properties: wet grip and rolling resistance (energy). The impact of the EU noise label for tyres on road traffic noise emissions has not yet been evaluated quantitively, but it is clear that it has substantially raised the interest in this environmental property of tyres.

This paper focuses on the noise property of the label. This has been studied in a special project named STEER, with the main purpose to analyze and improve the label, which is reported in another paper at this conference [3]. The STEER project made an extensive analysis of the uncertainties in the label value both in its present state and after implementing all improvements suggested by STEER. This uncertainty analysis is fundamental to the work and is reported in another conference

1 ulf.sandberg@vti.se

2 pmiodusz@pg.edu.pl

a 2022

paper [4]. In this analysis, it was concluded that two sources of uncertainty dominate the total uncer- tainty of determining the noise label value. These are:

1. The variability of the test surface on which noise measurements are made (surface according to ISO 10844). 2. The fact that, generally, only a few of all tyres are actually measured according to the standard measurement procedure (coast-by of a test vehicle over an ISO test track). Instead, the labelling is often simplified by using the same levels for (usually quieter) tyre variants as measured when determining if the tyre line meets the type approval (noise) limit, according to Regulation (EC) No 661/2009 [5] and ECE Regulation 117 [6]. The STEER paper [4] deals with the first uncertainty source as well as a number of other less influ- ential sources; while the second uncertainty source is the subject of this paper. A much more com- prehensive report of this work is found in [7].

a 2022

Figure 1: The EU tyre label implemented from May 2021 [1]. Note that the noise level is shown in both A-weighted dB (integers) and in noise class A, B or C. 2. TYRE TRADE-NAMES, TYRE LINES AND VARIANTS

Tyres are defined by its manufacturer and a certain trade or product name; for example, Michelin Primacy 4 and Goodyear EfficientGrip Performance 2. However, there are normally several different variants having the same trade-name; with similar construction, materials and tread pattern, but dif- ferent width, rim size, aspect ratio, load index and speed rating. Such variants sharing the same prod- uct name are often referred to as a certain “ tyre line ”. Within a popular tyre line there may be 50-100 variants having different dimensions, load index, and speed ratings. When labelling all these variants for tyre noise it would be an extensive job to test each individual variant to assign a certain noise level to each variant.

code ring to he European {is 4 Bt rect egy ree ENERG' Pra in oe ling resistance (At) = + Wet rp (Ato) ‘Optana pictogram to cover = SaaS crenumsenwo— gl) A A sana wer yes

Therefore, the labelling regulation does not require that all variants of tyres within a certain tyre line are tested. Instead, it is common and accepted that tyre manufacturers save money by testing only the noisiest tyre(s) within such a line, and then giving the other tyres the same label. It is also common that a few tyre variants are noise tested and chosen to represent all variants in a certain range within a tyre line.

To measure only the (estimated) noisiest tyre in a line is perfectly legal, and practical and econom- ical as well, since the measurements for labelling are allowed to be the same as the measurements for type approval. Type approval intends to make sure that a certain noise level is not exceeded for a certain tyre line. The label will then show a noise level which is conservative for most of the tyre variants, but as the real noise level (at labelling conditions) would only be better than or equal to the labelled value, one cannot say that the system is violated from a legal aspect.

This simplification in the measuring system unfortunately means that when consumers select tyres, they will generally not have the full potential of the labelling system to select the quietest tyres. When such a simplified procedure is implemented, the consumers might in the best case only be able to compare the noisiest tyres in tyre lines, which seriously limits the value of the labelling system. 3. HOW THE TYRE NOISE LABEL IS DETERMINED

The labelling regulation specifies the use of the noise measurements method described in UN ECE Regulation 117 [6], here referred to as R117. It is the same method as specified in the type approval for tyre lines. In summary, it requires coast-by of the test vehicle equipped with the tested tyres at 8 speeds between 70 and 90 km/h, determined by noise-speed regression as the noise level at the refer- ence speed of 80 km/h. This is for passenger car tyres (designated C1 tyres) and for van tyres (desig- nated C2 tyres), while for heavy vehicle tyres (designated C3 tyres) the speeds are 10 km/h lower. During the coast-by, the maximum noise level is recorded in microphones located 7.5 m beside the centre of the test track, the average of which will be the measured noise level. This level is then given a “political discount” of 1 dB and is further truncated to the nearest integer value.

In this paper, only C1 (passenger car) tyres are considered. The STEER project is also focused on C1 tyres. 4. PROJECT OUTLINE

A substantial number of new tyres were selected for noise testing. The testing was made by rolling the tyres on a laboratory drum fitted with a replica of an ISO test track at speeds between 70 and 90 km/h. Two microphones were located close to the test tyre as in the CPX method of ISO 11819-2: 2017 and each tyre was loaded and inflated in accordance with R117.

This was of course a different measuring method than the coast-by method run on an ISO test track, but the conditions were similar except for the closer location of the microphones and the some- what curved drum surface. It allowed the tests to be made within a very narrow temperature range and compared to coast-by tests, only ¼ as many tyres were needed. Earlier tests had shown that such measurements are well correlated with coast-by measurements if the same ISO test track is used. In this project, only the differences between tyres within the same line were evaluated; thus, the absolute levels were unimportant.

The measurement results were compared within each tyre line and with the labelled values. Since it was found that tyres within a tyre line could differ up to 6 dB, despite they had the same label value, proposals were made how to avoid such high deviations in the labelling procedure. This required more testing, but to reduce the testing workload, proposals were made how the required more com- prehensive testing could be made with a minimum of extra work.

a 2022

5. THE MEASUREMENT METHOD

The measurement method is similar to the CPX method (ISO 11819-2:2017) but adapted for indoor drum measurements. Two microphones are located beside the tyre sidewall, at a distance of 0.2 m from the sidewall (counted perpendicular to the sidewall); exactly like the CPX method. Since the microphones were located close to the tyres, combined with sound absorptive materials on sound reflective surfaces, the effect of acoustic room reflections were considered negligible.

The noise measurements were made as time averages over 32 s at each of the speeds of 70, 73, 76, 79, 81, 84, 87, 90 km/h. Ambient temperatures were within 19.0 and 21.0 °C. Average tyre tread temperatures were within the range 31.0 to 36.5 °C. Tyres were warmed up by running on the drum at high speed before any measurements took place. Loading and inflating tyres was according to requirements in R117 for each tyre variant. Tyres were new when delivered and were run-in before the measurements as is required in the R117 method.

The laboratory of the Gdansk University of Technology (GUT) in Poland, includes a suitable drum facility with replica road surfaces and where temperature variation and other uncertainty factors can be controlled, see Figure 2. A drum with a diameter of 2.0 m was used for testing the selected tyres. The drum surface was a replica from a real ISO test track in Sweden, meeting the latest requirements of ISO 10844 (having a texture measured as MPD of 0.40 mm).

Thus, except for the drum curvature and the closer microphones, the measurement conditions were similar to those of R117. Due to the laboratory conditions, the measurements were made with much less environmental influence, with very high reproducibility and with no test vehicle influence.

a 2022

Figure 2: The laboratory drum facility at the GUT in Gdansk. The drum (diameter 2.0 m) has three drum surfaces (left to right: ISO replica, smooth steel, sandpaper) but only the ISO was used. Note the position of the microphones, visible as “black balls” 0.1 m from the drum surface.

6 TESTED TYRES

As test objects in the project, a major criterion was that the line should include many variants covering a wide range of dimensions, loads and speeds, and yet being labelled with the same noise level. The project budget allowed a maximum purchase of around 50 tyres. After initial spot-checks over a wide selection of tyre brands and tyre lines, three major tyre brands were chosen by VTI for this study: one premium European one, one premium Asian one and a budget brand from China. For two of these lines, all tyres were labelled with the same noise level, despite a wide range of dimensions and ratings, but for one of the lines, two noise labels were used. In total 53 car tyres (and suitable rims) were purchased of very different dimensions (e.g., test loads vary from 270 to 670 kg and tyre width vary from 145 to 245 mm); see Figure 3. The three tyre lines chosen were (see details in [2] and [7]):

• Pirelli Cinturato P1 Verde: representing a European (and international) quality brand • Yokohama BluEarth-ES (ES32): representing an international (non-European) quality brand • LingLong Greenmax HP010: representing an Asian (and international) budget brand

a 2022

Figure 3: The test tyres stored before transported to GUT for testing.

Table 1: Selected condensed data about the test tyres.

Rim range [inches]

Width range

Profile ra-

Test load

Max speed

Tyre line Number

tio range

range

range [km/h]

of tyres

[mm]

[%]

[kg]

Pirelli Cintur a to P1 Verde 22 14 - 17 165 - 215 50 - 70 294 - 524 190 - 240

Yokohama BluE a rth-ES (ES32) 16 13 - 18 145 - 245 40 - 70 270 - 570 190 - 270

LingLong Gre en max HP010 15 14 - 17 185 - 235 50 - 65 361 - 665 210 - 240

7 RESULTS

7.1 Measured noise levels versus labelled levels The noise measurement results are summarized in Table 2 and Figure 4. The noise levels measured by the two microphones at each speed are averaged to represent an average speed of 80 km/h. Third- octave band frequency spectra were also measured but are not reported here, since the labelling deals only with A-weighted overall noise levels. Table 2: Summary of labelled and measured noise levels for the selected tyres. The speed range is 70-90 km/h but the level represents the average at 80 km/h. The noise level range is the difference between the maximum and minimum noise levels for the selected tyres. The colours show how seri- ous the deviations are (red = too high, green = acceptable, orange = not too bad).

Noise level measured on drum

Labelled noise level

Tyre line No. of

Average

Stand. dev.

Range

tyres

[dB]

[dB]

[dB]

[dB]

LingLong Greenmax HP010 15 71 96.6 0.64 2.4

Pirelli Cinturato P1 Verde 5 70 99.4 0.60 1.5

Pirelli Cinturato P1 Verde 17 69 98.3 0.62 1.9

Yokohama BluEarth-ES (ES32) 16 68 98.6 1.55 5.8

a 2022

Figure 4: Distribution of measured A-weighted noise levels for each of the tyre brands (with each tyre’s labelled level on the ordinate). One symbol for each tyre variant. The Pirelli tyre line had la- belled the tyre variants with two levels: 69 dB for some and 70 dB for others.

Labelled noise level [dB] n n 70 69 68 67 © LingLong 71.48 © Pirelli 69470 dB 0c cmm 0 00 0 © Yokohama 68 48 ce 00 woecce 95 96 7 98 99 100 101 102 Measured noise level [dB] 103

For each tyre line, one variant was tested with four tyre samples; the reason be-ing to study how large the spread would be in noise levels between different tyres of nominally identical construction and performance. The results are presented in Table 3. Note that uncertainty according to the section about test method is less than 0.10 dB [7]. Table 3: Results of noise measurements on four tyre samples nominally identical, for each of the three tyre brands. Within each tyre brand/line all tyres were produced the same week, except for the Pirelli tyres which were produced eight weeks apart.

Tyre code

Tyre brand/line designation Dimension, load and speed index

Label value

Measured noise level

Notes

L7 LingLong Greenmax HP010 205/50 R16 87 V 71 96.14

L8 LingLong Greenmax HP010 205/50 R16 87 V 71 96.24

L9 LingLong Greenmax HP010 205/50 R16 87 V 71 96.53

L10 LingLong Greenmax HP010 205/50 R16 87 V 71 96.45

Average level: 96.34

Standard deviation: 0.18

Range (max – min) 0.39

P10 Pirelli Cinturato P1 Verde 195/65 R15 91 V 69 99.02 Produced 5118

P11 Pirelli Cinturato P1 Verde 195/65 R15 91 V 69 98.51 Produced 5118

P12 Pirelli Cinturato P1 Verde 195/65 R15 91 V 69 98.03 Produced 0719

P13 Pirelli Cinturato P1 Verde 195/65 R15 91 V 69 98.64 Produced 0719

Average level: 98.55

Standard deviation: 0.41

Range (max – min) 0.99

Y8 Yokohama BluEarth-ES (ES32) 205/55 R16 91 V 68 98.02

Y9 Yokohama BluEarth-ES (ES32) 205/55 R16 91 V 68 97.81

Y10 Yokohama BluEarth-ES (ES32) 205/55 R16 91 V 68 98.08

Y11 Yokohama BluEarth-ES (ES32) 205/55 R16 91 V 68 98.65

Average level: 98.14

Standard deviation: 0.36

Range (max – min) 0.84

Average standard deviation: 0.32

Average range (max-min): 0.74

It appears that for the LingLong tyre while the standard deviation was only 0.18 dB, for the Pirelli tyre it was 0.41 dB and for Yokohama it was 0.36 dB. For the Pirelli tyre it appears that the tyres were from two different batches (years), with about 0.5 dB between each one, and had all four tyres been from the same batch it is probable that the standard deviation would have been less than 0.3 dB.

a 2022

For the Yokohama tyre, it is one of the tyres which is 0.5 dB noisier than the other three, despite being produced in the same week. However, it may still be from a different batch.

The results suggest that for tyres produced in the same week, the standard deviation between tyre samples is around 0.3 dB, but for different batches, larger differences may occur. These results are in-line with estimations from ETRTO which suggest that the standard deviation due to tyre sample differences is around 0.26 dB [2].

7.2 Relations between noise and the other tyre parameters In total, the database from the experiment contains about 30 000 data cells. This allows more analyses than just the primary data needed for this project. For example, we have third-octave frequency spec- tra for each measurement. An attempt to analyze these have not been made here as frequency spectra are not considered in the labelling system.

However, some analyses of how the various parameters in the database are correlated with each other have been made. For example, a multiple regression analysis was made with noise level as the dependent variable and tyre width, load index and speed rating as independent variables. To avoid tyre brand “disturbing” the result, the tyre brand average noise level difference was compensated for. It appeared that only the tyre width had an influence on noise level which was statistically significant (since tyre width and load are rather closely correlated). Therefore, the relation between noise level and tyre width is shown in Figure 5.

The Yokohama tyres, which cover the widest range of tyre widths shows that noise increases with 4 dB over the 145-245 mm range. This fits well with the data shown in [8] which suggest an effect of 4 dB for that tyre width range. Also, a regression of measured noise level on purchase cost showed no significant correlation.

103

y = 0,0304x + 92,602 R² = 0,3545

102

101

100

Noise level

99

98

97

96

95

130 140 150 160 170 180 190 200 210 220 230 240 250

Tyre width [mm]

Figure 5: Correlation between A-weighted noise levels and tyre width, for all the 53 tyre variants.

a 2022

8 DISCUSSION OF THE RESULTS

8.1 Notes about a few of the extreme noise levels The following observations are noteworthy with regard to the noisiest tyres measured in our study:

• The four noisiest Yokohama tyres have widths 225-245 mm, rim diameters 16-18 inches, and load indices of 91 to 98. The quietest Yokohama tyre has the dimension 145/65 R15 and load index 72. • The noisiest Pirelli tyre has the dimension 215/50 R17 and load index 95. It is the largest of the Pirelli tyres both in width and rim. • It is interesting that it seems that the five Pirelli tyres labelled at 70 dB are about 1 dB noisier than the ones labelled at 69 dB, which is logical; especially for the noisiest for each label level. • The noisiest LingLong tyre has the dimension 225/65 R17 and load index 102. It is the second largest of the LingLong tyres both in width and load index, and the largest in rim size.

Consequently, as one could expect; the larger tyres rather consistently give the higher noise levels.

8.2 Observations regarding not selected tyres Apart from the three selected tyre lines, also several other lines were studied as potential candidates. However, many of the desired tyre variants on those lines were not available for purchase in Sweden, at least not from the major outlets, which gave us too few tyres to test.

When looking at other tyre lines, the picture is more mixed regarding different or same label levels with a tyre line: some have different (noise) label values, while some (the majority it seems to us) have the same label for the entire line or maybe two levels for two parts of the line (the latter similar to our Pirelli line used here).

One major European tyre line that was considered was possible to purchase in 27 different variants, from 14 to 19 inches (rims) and from 165 to 225 mm (width). However, these were no useful for our experiment as they were labelled with noise levels 66, 67, 68, 69, 70 and 71 dB. This was a typical example for a case when it appeared that most (if not all?) variants with the tyre line had been tested (?) for noise. It also illustrates the effect of tyre dimensions (5 dB from the quietest to the noisiest variant). Most probably, that tyre line was available (not for us but in other markets) in more dimen- sions and variants than the 27 which we were able to identify as possible to purchase.

Another European brand offered one tyre line with more than 30 variants, all labelled with the same noise level, according to its website. This would have been ideal to include in this study but for budget reasons this was not selected. But it illustrates that the policy of choosing to label the entire or most of the tyre line with only the highest noise level is indeed common.

8.3 Implications for tyre consumers From a tyre consumer point of view, this flaw in the labelling system means that the possibility to select a tyre based on noise level is very limited. For many tyre lines, only the noisiest tyre(s) is/are appropriately labelled, and this depends on the dimensions of the noisiest tyre in the line. Thus, for one tyre line the tyre variants may be labelled as 71 dB, measured based on the widest tyre (say 255/50 R20) while for another tyre line the variants may be labelled as 69 dB, measured based on the widest tyre in that line (say 205/65 R18). It may then appear as the latter is a 2 dB quieter tyre line, but it is only because of dimension of the “worst” (generally widest) tyres in each line, and not the case for tyres of the same dimension.

The fact that tyres which have the same noise label, in reality can be up to 6 dB different, in addition to the reasoning in the previous paragraph, implies that the value of the noise label for the consumer as it is determined currently is very limited. The noise labelling system is in principle a

a 2022

very useful idea for making tyre selections based on technical performance, rather than (say) price and visual appearance, but it is not currently designed to work in this way.

8.4 The effect of variability of ISO test surfaces In Figure 4 it appears that the tyre line with the highest noise label (71 dB) surprisingly include the quietest measured levels. One reason for this, the authors believe it is the dominating reason, may be the effect of the test track surface (the surface according to ISO 10844). In [2] and [3] it is shown that there may be 3-4 dB of difference in the noise properties of ISO surfaces despite they meet the stand- ard. In the case in Figure 4, it may be that the LingLong tyres have been tested on an unusually “noisy” test track. If that test track actually was 3 dB “noisier” than those used by Pirelli or Yoko- hama, the relation between measured and labelled levels would overall look much better.

8.5 The case for the rolling resistance label? When we studied the label for energy (i.e., rolling resistance) we noticed that for some tyre lines all or at least most of the tyre variants were labelled with the same class, despite being of very different dimension. This makes us suspect that the same problem as for noise may exist for the energy label. We suggest that this matter is studied in a future project. 9 SUGGESTED IMPROVEMENT OF THE LABELLING PROCEDURE

To eliminate the insufficient noise labelling, due to not measuring all tyres in each tyre line, three options are suggested:

1. Measure and report the noise levels of all tyre variants within a tyre line, using the method of R117 (coast-by on ISO test track). Note that this may already be implemented by some tyre manufacturers for some tyre lines. In this option, the workload to produce the noise labels will be high or very high compared to the present situation (needs four test tyres, test vehicles, driver, access to ISO test track and proper weather). It may also be difficult to have access to test tracks for such extensive measurements.

2. Use a simplified laboratory measuring method to determine differences between tyre variants within a tyre line and use this difference to assign noise labels to all tyre variants, with the type approval level as a reference, and (optionally) with a few more tyres tested by the coast-by method for additional noise references. A practical and relatively fast method is proposed, namely measurements indoor in a laboratory, utilizing a drum with appropriate surface and with microphones close to the tyre/drum contact patch as in the so-called CPX method (CPX = Close- ProXimity) adapted from trailer to laboratory drum conditions. The drum must have a replica surface based on an ISO test track (which may be produced globally from the same original). This option is described in some detail in [7]. It is essentially the method used in this study.

3. Use a noise modelling or simulation procedure to determine differences between tyre variants within a tyre line and use this difference to assign noise labels to all (or at least most) tyre variants with really measured tyres as references (using the coast-by procedure). This procedure should be limited to labelling tyres which are only slightly different in dimension or materials to the closest measured tyre(s). It relies on the availability of a reasonably accurate simulation model for noise emission of tyres. Such models for general use do not yet exist, with an acceptable accuracy. But in this case, only relatively limited noise differences need to be modelled, only on the same ISO surface and same speeds, between tyres within the same line (i.e., essentially the same construction, tread patterns and materials) This should be possible to model with reasonable

a 2022

accuracy. This third option has the potential to reduce the measurement workload substantially compared to the first option

These options are described in more detail in [7]. Applying one of them is necessary to make the noise labelling system reliable and meaningful for the possibility of selecting quieter tyres.

10. CONCLUSIONS

First it must be noted that there is no evaluation by the authors of the tyre lines or brands in this study, other than how these particular selections are labelled. That the labelling is more or less accurate does not mean that one tyre line or brand is better than the others, since they have all done what is required and allowed in the regulation. It is not fair to qualitatively compare the tyre brands here.

The extensive measurements in this project indicated that the comparison between labelled and measured noise levels came out as follows:

• The labelling could be better but is reasonably fine for the Pirelli tyre line selected here. • Labelling is not acceptable for the LingLong tyre line selected here. • Labelled values are highly incorrect for the Yokohama tyre line selected here. Tyres that are almost 6 dB quieter than the noisiest in the tyre line are labelled with the same noise level (68 dB). • However, given the existing labelling regulation, there is nothing illegal in this; it just shows that the regulation is insufficient.

In conclusion, the results show that the practice of measuring only some of the tyres within a tyre line makes the labelling regulation questionable and practically makes it very difficult for tyre buyers to make choices based on the noise label, since it can be substantially wrong. Commonly, the quietest tyres are not possible to identify – only the noisiest in the particular tyre line, which often are the widest tyres. This problem is considered common for many tyre lines, but there are also tyre lines in which most, if not all, tyres in the line are appropriately noise labelled based on measurements. Which alternative that apply may well be due to ambition and/or available testing resources for the tyre line.

The results of this experiment show that by picking some arbitrary tyres from one or more tyre lines, noting the noise labels of these tyres and measuring noise emission of them by the R117 method (or whatever standard method), would almost automatically fail to show a reasonable correlation between label and measured levels, since the label values may be so different from reality. This has been indicated in many studies, and the result of our study shows why it will happen.

The study also suggests solutions to the problem, namely that all tyres within a tyre line shall be labelled individually with noise levels that are based on either noise measurements or calculations of noise levels based on simulation tools (tyre noise models). The following three options are suggested:

1. The default option is to measure and report the noise levels of all tyre variants within a tyre line, using the method of ECE Regulation R117 (coast-by on ISO test track). Note that this may al- ready be implemented by some tyre manufacturers for some tyre lines. This option requires sub- stantial testing resources.

2. Use a simplified measuring method to determine differences between tyre variants within a tyre line and use this difference to assign noise labels to all (or at least most) tyre variants with a really measured tyre as a reference (measured according to the ECE R117 method). For example, the reference noise level could be the type approval noise level for the tyre line.

3. Use a noise modelling procedure to determine differences between tyre variants within a tyre line and use this difference to assign noise labels to all (or at least most) tyre variants, with noise levels

a 2022

of really measured tyres as references (measured according to the ECE R117 method). To avoid too high uncertainties, this procedure should be limited to labelling tyres which are only slightly different to the nearest measured tyre(s). Yet, it may reduce the testing effort substantially.

In the second option, a significantly simplified procedure to measure many more tyres than pres- ently is suggested, without causing too much extra work. The new method is a simplified method conducted on a drum in a laboratory, a drum having a replica of an ISO surface.

If the suggestions in this report are implemented, the uncertainty in the noise labelling system will be significantly reduced. Currently the uncertainty contribution of tyre properties was calculated in STEER to be between 0.59 and 1.2 dB, but the proposal based on this study would reduce that to only 0.13 dB [4]. Together with improvements in the ISO test track specifications or calibrations, the total uncertainty of the noise levels in the labelling system can become halved [3]. This will make the labelling system much more reliable and useful for the customers and eliminate the worst flaws in the current system. 11. ACKNOWLEDGEMENTS

This project was conducted in support to the project STEER (Strengthening The Effect of quieter tyres on European Roads) sponsored by CEDR (Conference of European Directors of Roads). It was financed by the Swedish Transport Administration, conducted by VTI with GUT as subcontractor.

The authors are very grateful to Mr Lars Dahlbom at the Swedish Transport Administration for his support during the STEER project in the role as project responsible officer on behalf of CEDR, and in particular for sponsoring the Swedish project reported here. The project plan for this project was made in cooperation between the authors and him. 12. REFERENCES

1. European Commission (2020): “Labelling of tyres with respect to fuel efficiency and other parameters”. Regulation (EU) 2020/740 of the European Parliament and of the Council on the, amending regulation (EU) 2017/1369 and repealing regulation (EC) No 1222/2009. 2. Bühlmann, E.; Sandberg, U.; Berge, T.; Goubert, L.; Schlatter, F. (2022): “STEER: Final Report”. Report for CEDR (https://www.cedr.eu/peb-research-programme-2018-noise-and-nuisance). 3. Schlatter, F.; Sandberg, U.; Bühlmann, E.; Berge, T.; Goubert, L. (2022): “Project STEER: Improving the EU Tyre Noise Label”. Proc. of Inter-Noise 2022, Glasgow, UK. 4. Goubert, L.; Berge, T. (2022): “Project STEER : The Effect of Uncertainties in Determining the EU Tyre Noise Label”. Proc. of the 28th International Congress on Sound and Vibration (ICSV 28), 24-28 July 2022, Singapore. 5. European Commission (2009): “Regulation (EC) No 661/2009 of the European Parliament and of the Council of 13 July 2009 concerning type-approval requirements for the general safety of motor vehicles, their trailers and systems, components and separate technical units intended therefor" (https://eur-lex.eu- ropa.eu/eli/reg/2009/661/oj/eng). 6. ECE R117 (2011): “Uniform provisions concerning the approval of tyres with regard to rolling sound emissions and to adhesion on wet surfaces and/or to rolling resistance”, Regulation No. 117 of the Eco- nomic Commission for Europe of the United Nations (UN/ECE). 7. Sandberg, Ulf (2022): “Noise labelling of tyres: The effect of not measuring all tyres”. Technical Report TRV 2019/119986, Swedish Road and Transport Research Institute (VTI), Linköping, Sweden. 8. Sandberg, U., Ejsmont, J., (2002): Tyre/Road Noise Reference Book. Informex HB, Kisa, Sweden, (http://www.informex.info/html/book__tyre_road_noise_.html).

a 2022