Measurement of vibro-acoustic noise of drum brake under various contact conditions

Aditya Chaudhary 1

Indian Institute of Technology Tirupati Tirupati, Andhra Pradesh, India-517506

Akash Yella 2

Indian Institute of Technology Tirupati Tirupati, Andhra Pradesh, India-517506

Yuva Venkat Ajay, Bharinikala 3

Indian Institute of Technology Tirupati Tirupati, Andhra Pradesh, India-517506

Sriram Sundar 4

Indian Institute of Technology Tirupati Tirupati, Andhra Pradesh, India-517506

ABSTRACT Drum brakes are vital safety components, widely used in automobiles due to their greater availability and quick replaceability compared to disc brakes. Since automobiles are used in all weather conditions, the braking performance also changes accordingly. Penetration of water, oil, and dust between the brake lining and the drum directly a ff ects the contact parameters. These critical contact parameters (such as friction coe ffi cient, contact damping, and sti ff ness) in turn a ff ect the dynamic and acoustic responses of the nonlinear system. In this paper, the variation in the vibro-acoustic characteristics of a drum brake due to the presence of water, oil, and dust on the brake lining is studied experimentally by measuring the radiated acoustic signals and the system response. The braking performance under di ff erent conditions is quantified based on the system response while the acoustic behavior of the system is studied using acoustic measurements. This study would help in determining the safety of the automobile even under such abnormal braking conditions. Further, acoustic measurements can be used to assess the contact condition of the braking system. It is envisaged to develop an improved health monitoring system for drum brakes.

1 me20m001@iittp.ac.in

2 me18d506@iittp.ac.in

3 me19b008@iittp.ac.in

4 sriram@iittp.ac.in

a slaty. inter.noise 21-24 AUGUST SCOTTISH EVENT CAMPUS O ¥, ? GLASGOW

1. INTRODUCTION

The braking system is an essential part of an automobile for the safety of passengers and goods. Drum brakes are used in vehicles due to their high durability, economic availability, easy maintenance, and replaceability. Studies have been performed on the dynamic behavior of brakes with variations of thermal conditions [1], operating conditions [2], and friction lining material of brakes [3]. For instance, Mahmoud et al. [4] investigated variation in braking performance of disk brakes due to the contact conditions. The kinetic energy is converted to heat and sound energy in drum brakes, majorly due to friction between brake shoes and rotating drum. Drum brakes produce significant noise in automobiles, irritating to human ears. Engagement of brake shoes with wheel drum has the characteristic sound associated with contact surface conditions of drum brake, which is of interest to many. For example, Ramesh et al. [5] studied acoustic characteristics of drum brakes using a combination of analytical and numerical models, while Yella et al. [6] compared noise generated from simplex and duplex configurations of drum brakes using non-linear vibro-acoustic models. Vehicles are used in all weather (hot and dry seasons to the rainy season), and the brakes are expected to work properly under all conditions. Penetration of water, dust, or brake fluid into braking pad from the master cylinder into the gap between brake shoes and wheel drum a ff ects the performance of the brake resulting in abnormal braking of vehicle. This work aims to experimentally investigate the braking performance of a drum brake with variation of contact conditions (presence of water, dry dust, and oil) by estimating the e ff ective coe ffi cient of friction ( µ e f fective ). The e ff ect of contact conditions on acoustic behaviour during braking is also discussed.

2. METHODOLOGY

The methodology adopted in this work is schematically shown as a flowchart in Figure 1. The experimental setup (described in section 2.1) is used to perform experiment on drum brake with varying contact conditions. Braking torque ( τ braking ), drum velocity ( ω drum ), and acoustic pressure are measured for each experimental run at same operating conditions (initial ω drum and actuation force). These measurements are used to study braking performance of drum brake and estimation of µ e f fective between drum and shoes, and the results are discussed in section 3.

Figure 1: Methodology of the work

2.1. EXPERIMENTAL SETUP A sub-scaled experiment was developed to imitate two-wheeler drum brakes, as shown in Figure 2. The experiment consists of a drum brake on one end of the shaft and a flywheel is mounted on other end. The shaft is supported by four stands with ball bearings for shaft rotation and fifth stand supports backing plate of drum brake. Keys are used to lock the flywheel, wheel drum to the shaft, and the backing plate to fifth stand (ground). An electric motor (not shown in the figure) was used to provide initial angular velocity to flywheel. Brake shoes mounted on the backing plate were actuated

using a mechanical cam which was controlled by a lever. The lever was attached with dead weight to provide constant load during braking. Load applied on the lever forces cam to rotate which pushes both brake shoes towards the rotating drum. After removal of applied load, the shoes were pulled back to initial position by retaining springs.

Figure 2: Drum brake experimental setup

The laser tachometer [Model: EILas2 Make: VISPIRON ROTEC] was used to capture angular velocity of the drum. The telemetry type torque sensor [Make: manner sensor telimetric] was mounted on shaft for determining τ braking . These sensors were connected to 6 - channel torsional vibrator recorder / analyzer [Model: RASdelta8 Make: VISPIRON ROTEC] data acquisition system.Two free field microphones [Model: 378B02 Make: PCB PIZOTRONICS] were placed in plane of the drum brake to capture acoustic response of the system as shown in Figure 3.The microphones were connected to 16 channel vibration recorder / analyzer [Model: OR38 Make: OROS] data acquisition system.

Figure 3: Acoustic sensors in the experimental setup

The experiment was performed for the following four conditions associated with real life situations of the brake shoes: (1) dry condition or normal condition of brake shoes (2) with water on the brake shoe to imitate situation of rainy season (3) applying silver sand (150-250 µ m grain size) on the brake shoe to emulate situation of dust accumulation on brake shoes (4) applying SAE 20W40 engine oil on the brake shoe to replicate situation of oil leakage from master cylinder. The brake shoes with di ff erent contact conditions are depicted in Figure 4. The flywheel was rotated at initial ω drum of 200 rpm and constant actuation force ( F actuation ) of 30 N was applied on the lever using dead weight. The vibro-acoustic signals obtained from data acquisition system were post-processed. The torque data was used to estimate µ e f fective value between brake shoes and drum.

Figure 4: Contact conditions of brake shoe: (1) dry, (2) water, (3) dry dust, (4) oil.

2.2. ESTIMATION OF THE COEFFICIENT OF FRICTION The braking performance of the drum brake depends on various factors like thermal conditions, operating conditions, drum geometry and contact conditions. Braking torque, a key indicator of brake performance is function of coe ffi cient of friction and normal force. The µ e f fective was evaluated using moment balance of forces acting on brake shoes and τ braking as shown in Figure 5.

Figure 5: Free body diagram of the shoes

Here, F l , F r are forces transmitted from cam to left and right brake shoes respectively and l 1 , l 2 are their respective perpendicular distance from pivot p 1 and p 2 . F s 1 , F s 2 are retaining spring

Brake shoe —_ Fa _-> Brake shoe —~ Diver.

forces acting on both shoes at perpendicular distance of d 1 and d 2 from pivot respectively. Forces transmitted from cam to both shoes were assumed to be equal. Rigid contact was assumed between cam and shoe tip. Braking shoes were assumed to be in line contact and normal force to be acting at centre of respective shoes to reduce computation complexity.

Coe ffi cient of friction was estimated using eq. 1.

µ = τ braking ( N l + N r ) ∗ r c (1)

Here, µ is coe ffi cient of friction between Brake shoes and drum, N l and N r are assumed to be normal forces acting at centre (perpendicular distance r c from pivot) of left and right brake shoes respectively.

2.3. SOUND PRESSURE LEVEL (SPL) During a braking event the sound is radiated due to the contact between the brake shoes to drum surface whose parameters vary with contact conditions. The acoustic pressure of microphones in two transverse directions to axis are depicted in Figure 6 for overall breaking event under dry condition of brake shoes. The initial portion [0s - 2s] sec of braking event was used to obtain the SPL as given in eq. 2.

S PL = 20 log P rms

P re f (2)

where, P re f = 20 µ Pa .

Figure 6: Acoustic pressure at initial ω drum = 200 rpm and F actuation = 30 N for dry brake shoe condition. (a) From Mic 1 in radial X-direction to drum (b) From Mic 2 in radial Y-direction to drum

(a ‘Acoustic pressure [Pa] ‘Acoustic pressure (Pa) = . oe ; _ IMHO nnn considered for overall PL during braking event —— Microphone 2

3. RESULTS AND DISCUSSION

3.1. Variation of SPL with contact conditions of drum brake The variation of SPL for various contact conditions during braking event is shown in Figure 7. SPL was found to be more in case of dry contact conditions while it reduced in presence of any material (water, sand, oil) between the shoes and drum. The reduction in SPL can be associated with reduced friction or increased damping between surfaces. SPL was found almost same for the system with water, dry dust, or oil.

Figure 7: Variation of SPL under various contact condition at initial ω drum = 200 rpm and F actuation = 30 N. Key: - - ◦ - - Mic 1 in radial, x-direction - - □ - - Mic 2 in radial, y-direction

The noise for di ff erent contact condition is compared in frequency domain in Figure 8. For contact condition of dry shoes and presence of water, the acoustic pressure is found to be high in low- frequency spectrum whereas in high-frequency spectrum, the acoustic pressure is found to be higher for contact condition with dust and oil for both microphones. The frequencies around 363 Hz, 1818 Hz and 7608 Hz (spectrum A, B, C) are appearing in all contact conditions for microphone 1. Almost same frequencies around 373 Hz, 1809 Hz and 7671 Hz (spectrum D, E, F) appear for all conditions for microphone 2. These might be natural frequencies of brake system. From spectrum B and E it is found that frequencies corresponding to non-dry contact conditions (water, dust, oil) is slightly lesser than of dry contact condition. The reduction in frequency associated with peaks might be due to increase in damping in non-dry conditions. The variation in amplitude between dry, water condition and dust, oil condition is high in high-frequency region for microphone 1 whereas this trend is at relatively lower frequency region for microphone 2.

SPL (dB re 20Pa) 85 nt a 3 60 Ory Water Contact condition Dry dust Oil

(a)

(b)

Acoustic pressure [Pa] ° 2000 4000 8000 10000 12000 6000 Frequency [Hz]

Figure 8: Comparision of acoustic spectrum for various contact conditions at initial ω drum = 200 rpm and F actuation = 30 N: (a) Mic 1 (b) Mic 2. Key: — Dry — Water — Dry dust — Oil

3.2. Variation of braking performance with contact conditions of drum brake The variation of ω drum during braking event for various contact conditions is shown in Figure 9. The ω drum was found to retard rapidly in presence of water on brake shoes compared to dry brake shoes. The rate of reduction of ω drum was found to be less in presence of dry dust and even lesser with application of oil on brake shoes. The ω drum seems to decrease linearly with time which is representation of a coulomb friction.

Figure 9: Variation of ω drum with time during braking for various drum brake contact conditions at initial ω drum = 200 rpm and F actuation = 30 N. Key: — Dry — Water — Dry dust — Oil

The variation of mean and max τ braking with various contact conditions is depicted in Figure 10. The τ braking surged in presence of water as compared to dry contact condition. The presence of dust and oil on contact surface between shoes and drum resulted into lower τ braking since it is highly dependent on µ e f fective .

12 Time [s] 220/- 200 180+ 160 - 140+ 120- 100 - ump, [wid]

Figure 10: Variation of τ braking with various drum brake contact conditions at initial ω drum = 200 rpm and F actuation = 30 N. Key: - - ◦ - - mean τ braking - - □ - - maximum τ braking

T raking IN.m] 12 Dy Water Contact condition Dry dust oil

The µ e f fective between brake shoes and drum was computed using τ braking and system parameters depicted in section 2.2. The variation of µ e f fective for di ff erent contact condition is depicted in Figure 11. The µ e f fective of dry contact condition is found to be 0.20. The µ e f fective was found to be relatively higher in presence of water between shoes and drum. Thermal cooling of surface with water could a ff ect the friction coe ffi cient. The presence of tiny dust particles on brake shoes reduced the µ e f fective . The sliding friction is assumed to be acting in dry contact condition of brake shoes. Rolling friction might act between contact patch of brake shoes and drum due to dust particles, and friction in rolling is much lower than sliding friction causing reduction in µ e f fective . Oil film on brake shoes was found to reduce µ e f fective to larger extent.

Figure 11: Variation of µ ef fective with various drum brake contact conditions at initial ω drum = 200 rpm and F actuation = 30 N. Key: - - ◦ - - µ ef fective

4. CONCLUSIONS

A sub-scaled model of drum brake was developed to study the vibro-acoustic behavior of the drum brake under various contact conditions (dry, water, dust, oil) of brake shoes and to estimate e ff ective coe ffi cient of friction under these conditions. The experiment was conducted with fixed initial ω drum and constant braking load. The presence of water resulted in increase in τ braking and µ e f fective . The presence of dry dust and oil film on brake shoes reduced the τ braking and µ e f fective . The presence of water, dust, and oil resulted in reduced SPL in microphones with respect to dry brake shoes. The acoustic pressure with dry shoes and in presence of water was found to be dominating in low- frequency spectrum whereas acoustic pressure in presence of dust and oil dominates in high-frequency spectrum. It is envisioned that, this work can be used to detect abnormal contact conditions based on acoustic characteristics during braking and compute µ e f fective between brake shoes and drum. In future, this work can be extended to develop predictive maintenance and fault detection system for the drum brakes.

04 T T T T 0.35 03> 0.25 e N ont 0.05 - ° 1 1 1 1 Dry Water Dry dust oil Contact condition

ACKNOWLEDGEMENT

This research was funded by Start-up Research Grant from Science and Engineering Research Board (SERB, Govt. of India) under project number: SRG / 2019 / 001172.

REFERENCES

[1] Daanvir Karan Dhir. Thermo-mechanical performance of automotive disc brakes. Materials Today: Proceedings , 5(1):1864–1871, 2018. [2] Xingming Xiao, Yan Yin, Jiusheng Bao, Lijian Lu, and Xuejun Feng. Review on the friction and wear of brake materials. Advances in Mechanical Engineering , 8(5):1687814016647300, 2016. [3] Ho Jang, Kang-hee Ko, Seong-jin Kim, Rena Hecht Basch, and James W. Fash. The e ff ect of metal fibers on the friction performance of automotive brake friction materials. Wear , 256(3- 4):406–414, 2004. [4] Khaled Mahmoud and Mohamed Mourad. Influence of water, oil and dust on the performance of conventional and wedge disc brakes. International Journal of Vehicle Structures and Systems , 6(3):71–75, 2014. [5] Ananthapadmanabhan Ramesh and Sriram Sundar. Estimation and study of drum brake noise using a comprehensive nonlinear vibroacoustic model. In INTER-NOISE and NOISE-CON Congress and Conference Proceedings , volume 261, pages 5531–5540. Institute of Noise Control Engineering, 2020. [6] Akash Yella and Sriram Sundar. Comparison of noise generated from simplex and duplex configurations of drum brake using non-linear vibro-acoustic models. In INTER-NOISE and NOISE-CON Congress and Conference Proceedings , volume 263, pages 1415–1425. Institute of Noise Control Engineering, 2021.