University uses Siemens Digital Industries Software’s solutions for acoustic analysis and noise reduction study

A game-changing study

Kettering University in Flint, Michigan, is home to the NVH and Experimental Mechanics (NVHEM) Laboratory. The lab is used by faculty, undergraduate and graduate students in various engineering degree programs.

With electric vehicles (EV) gaining in popularity, the motivation to study and optimize noise from the powertrain in hybrid and electric vehicles has never been greater. The noise generated by an electric vehicle is perceivably much lower compared to a vehicle powered by an internal combustion (IC) engine. However, prolonged high-frequency sounds from motors can be unpleasant for passengers. Owing to this desire for a lower sound level, noises from components such as tires, rotating fans and other vehicle accessories that would usually be masked by the engine’s sound become more prominent.

Kettering University studied the sound generated by the magnetic force using an electromagnetic model of a three-phase induction motor to understand the nature of forces (radial or tangential) acting on the stator core, followed by a thorough structural and acoustic analysis using a finite element (FE) model.

In addition to the electromagnetic and FE model, a modal analysis was carried out on an Ironhorse 3-phase general purpose motor. The modal analysis helped find resonant frequencies of the stator and motor housing. In addition to performing a modal analysis, operational deflection shapes of the motor housing and a sound intensity test of the motor in an operating condition of 1,770 revolutions per minute (rpm) were also carried out.

“We expected this project would give us an accurate understanding of the acoustical behavior of electric motors and set a new feasible method to reduce stator housing vibrations,” says Javad Baqersad, director, NVH and Experimental Mechanics Laboratory, Kettering University.

The project’s scope was limited to academic research, with the goal to find a cost-effective way to tackle noise from electric motors. Reducing a motor’s sound by using modified end brackets could simplify the process of isolating frequencies that amplify perceived sound. Furthermore, this is one of the first studies that proposes the use of modified end brackets with vibration damping properties to reduce the sound pressure level of a motor.

Kettering University carried out an acoustic test in a semi-anechoic chamber on a smallscale general purpose motor to validate their simulation results. Modifying end brackets to reduce sound was challenging given the materials at their disposal. Nevertheless, if successful, Kettering University’s study could provide electric motor manufacturing companies a new way of reducing inherent motor noise.

Figure 1

Kettering University employs Siemens’ Simcenter solutions

Many companies are investing heavily in EVs because they are environment friendly and considered the future of mobility. The heart of an EV is its motor. Modern EVs are quiet, and with the lack of an IC engine to mask most sounds from other components, the sound from the electric motor and other auxiliary parts becomes more prominent. Therefore, a method to reduce the sound from the electric motor is crucial to a pleasant driving experience.

The motor whine is a type of sound from an electric motor and accounts for the majority of the motor’s sound. The masking of this electric whine noise by the road/tires mainly comes into play at speeds over 45 miles per hour (mph). Depending on the design of the motor, the electromagnetic pulses and corresponding torque pulses can be very powerful. These pulses can be radiated as noise directly from the motor housing and can also be transmitted structurally to the support structure through the motor mounts. Rubber isolation systems used to mount the motor can be tuned more efficiently, and conventional materials can be used to block and absorb airborne noise energy. The primary challenge is to block and isolate these noises without adding additional weight and cost to the vehicle. By modifying the motor’s end bracket, Kettering University proposed a feasible solution to tackle the sound radiated by the motor.

One of the technical challenges of this study was integrating different software elements. The simulation was divided into three parts: electromagnetic (EM) analysis, structural analysis and acoustic analysis. Siemens Digital Industries Software’s Simcenter Testlab™ software and Simcenter 3D were used for the structural and acoustic analysis, respectively.

“Siemens has helped us a great deal, especially by extending the Simcenter 3D license to us,” says Anand Krishnasarma, graduate research assistant, NVH and Experimental Mechanics Laboratory, Kettering University. “A wide range of training material has helped the NVH and Experimental Mechanics lab here at Kettering get accustomed to different types of acoustic simulation techniques.”

Simcenter Testlab is used in several courses in the mechanical engineering and physics departments at Kettering universities. Both undergraduate and graduate students use this product to conduct academic projects, providing students with a unique opportunity to learn a software package that is used in the industry. Many of Kettering’s co-ops and alumni use Siemens software packages at their companies. Students who learned these skills at Kettering University successfully found internships and jobs in the industry.

Kettering University has used Simcenter Testlab for over three years and finds it superior to competitive products. Also, Simcenter 3D is an acoustic simulation software that performs a variety of analyses, in particular tests specific to the automotive industry like tailpipe noise vibro-acoustic response and panel acoustic contribution analysis (PACA).

Kettering University used Simcenter 3D to carry out an acoustic response analysis of their three-phase asynchronous alternating current (AC) induction motor and are preparing to use Simcenter Testlab for testing the motor in a semi-anechoic chamber.

The acoustic simulation carried out using Simcenter 3D required meshing the structure before the acoustic response case is simulated. The wrapper mesh is surrounded by an acoustic layer (air in this case). Finally, a data transfer case was created in order to map the structural displacements to the inner layer of the acoustic tetrahedron filler mesh, and a spherical field point mesh of one micrometer (m) radius was created around the structure to calculate the sound pressure levels at different points around the motor.

“The initial stage of the project gave us a clear understanding of how the variable electromagnetic flux in the air gap between the stator and motor translates into forces that affect the performance of a motor for a range of operating speeds,” says Krishnasarma. “In the second stage of the project, we saw how the motor housing deforms at different operating speeds and frequencies in order to try and isolate frequencies at which the electromagnetic forces result in unpleasant vibrations.”

Kettering University anticipated that most resonant frequencies would be from 1,000 to 5,000 hertz (Hz). The acoustic analysis using Simcenter 3D helped them identify the frequencies at which the sound pressure level peaked and offered insight into the steps that needed to be taken to isolate those frequencies from being excited to the maximum extent. Kettering University expected the modified end brackets to give more support to the main housing around the stator and reduce deformations in areas that are subject to maximum strain.

The structural response on the actual motor and sound intensity analysis revealed frequencies and sound pressure levels from the end bracket and motor housing that could potentially contribute to a major component of sound emitted by the motor.

Identifying sources of motor vibrations and sounds

To clearly determine the parts of the motor that vibrate more and contribute to a major proportion of the sound that is heard, an acoustic simulation of the motor, an operational deflection shapes test and a sound intensity test was performed. The acoustic simulation involved creating a spherical mesh with microphones around the vibrating meshed motor. Virtual microphones could then be placed anywhere on this spherical mesh. The microphones are used to compare sound pressure levels between different parts of the motor and at multiple distances from the motor

The operational modal analysis helps visualize actual motor vibrations at certain points on the motor that are being analyzed (also called operating deflection shapes). A sound intensity test helps find frequencies at which the sound levels peak and helps validate the operational deflecting shapes.

The acoustic simulation included a field point mesh with microphones at a distance of 1m from the motor. The pink spots on the field point mesh represent microphones to capture the sound. The sound captured by the microphones can be analyzed as seen in figure 1. For an example where only one frequency, 50 Hz, has been analyzed at a motor speed of 2,000 rpm, we see that the sound pressure levels around the stator housing are approximately one decibel higher than the sound at points closest to the end brackets. This behavior is not true for all frequencies. The operational modal analysis and sound intensity test will highlight other frequencies that contribute to higher levels of vibration and noise.

The acoustic response simulation was followed by a modal analysis on the actual motor’s stator in free-free condition and motor housing fixed at the base. From the modal analysis of the stator core, the first four modes occurring at 1,403, 2,545, 2,833 and 3,264 Hz can be extracted from the sum of the measured frequency response function (FRF). The modal impact test on the assembled motor shows the resonant frequencies of the motor housing. The first five mode shapes of the assembled motor with the base fixed are shown in table 1.

The operational modal analysis test on the motor housing in operating condition reveal operating shapes of the motor that can’t be extracted using the impact hammer. Multiple peaks can be seen, the first of which is at 29 Hz, which is very close to 29.5 Hz which is the rotating frequency of the motor. Peaks are seen at multiples of 29.5 Hz. Curve-fitting the sum of the cross spectrum of operational response in the frequency range of 1,000 Hz to 1,500 Hz figure 2 identifies some of the operational deflection shapes of the end bracket.

The Fast Fourier Transform (FFT) from the operational deflection shapes for the motor running at a speed of 1,770 rpm at no load show peaks at multiples of 29.5 Hz. A closer look at the modes that are excited between 1,000 Hz to 1,500 Hz, where most of the dominant resonant frequencies lie, reveals that the deformation corresponding to these frequencies are predominantly end bracket contributions. The harmonics of the fundamental rotation frequency are also evident, at multiples of the rotating frequency fundamental peak. The sound pressure level from the intensity analysis show peaks that very closely correlate with the spectrum from the operational deflection shapes.

The operational deflection shapes reveal some resonant frequencies that did not exist in the impact hammer modal test. From the correlation between the operational deflection shapes and the sound pressure measurements, it is seen that the resonant frequencies that cause the end bracket to vibrate strongly relative to the rest of the motor housing majorly contribute to the motor sound that is perceived by the human ear.

“The results of this study will enable us to provide a methodology to identify and isolate resonant frequencies in an electric motor,” says Krishnasarma. “Applying this methodology to small-scale motors used in auxiliary components, industrial motors and motors used in a variety of electric drives, including those used to power vehicles like golf carts, motorcycles, boats and mining vehicles, is very desirable and will be the subject of the proposals that are going to be sent to funding agencies.”

Table 1

Figure 2