John Deere engineers lower noise emission levels at their equipment with LMS Noise and Vibration tools

Construction equipment is designed primarily to fulfill operational requirements, although operator comfort and noise concerns are important design considerations. Quietness is becoming a critical product attribute in construction equipment, and manufacturers that cannot measure up will have a tough time competing in the international market. Traditionally, companies try to reduce noise emissions with sound-damping add-ons and structural modifications when noise problems crop up during prototype testing near the end of product development. But these last-ditch efforts involve lots of guesswork and are costly and time-consuming, and often create more problems than they solve. John Deere is breaking new ground by designing quietness into the equipment, creating strong brand values and strengthening its leadership position.

With growing concern over noise emissions and operator comfort, both from a government regulation and customer demand perspective, sound performance is becoming a major product differentiator in the construction equipment industry. Often, noise and vibration issues are not addressed until late in product development when unexpected noise levels during prototype testing raise concerns. By this time, designs are all but finalized and major changes are out of the question.

The addition of acoustic trim and other sound-absorbing materials dampen noise somewhat but fail to solve the problem at its source. Structural modifications to fix problems often add needless material and weight, and may set up unintended resonance in other parts of the equipment. The process is often frustrating and involves considerable hit-or-miss changes that may or may not work.

As John Deere sought to learn how to more efficiently lower noise emissions on its products, they looked for new methods that would enable their engineers to effectively predict noise and vibration in the early stages of product development when designs can be more readily optimized. The strategy was to establish this method by applying predictive noise and vibration technology to one particular product line.

The product Deere targeted for establishing this early noise reduction work was their line of skid steer loaders, general-purpose machines with sets of wheels driven by a set of hydraulic motors and control valves on each side. The direction of the wheels is fixed, and turning is accomplished by applying hydraulic drive power on one side of the vehicle while the other side is braked and skids through the turn. Because of this elegantly simple hydraulic drive and steering concept, the equipment is highly reliable and versatile.

The acoustic and vibration work to develop predictive technologies on the skid steer loader was a collaborative effort between John Deere engineers from multiple locations and the LMS Engineering Services group, which provided technical assistance in establishing the acoustic simulation process.

LMS technology was selected based on capabilities that are well suited to the requirements of acoustic work in these types of applications. LMS Test.Lab has extensive test control and measurement capabilities for gathering experimental data through its LMS SCADAS III front-end acquisition system and performing analysis on the results, including Transfer Path Analysis (TPA) for structure-bone noise and Acoustic Source Qualification (ASQ) for airborne noise for tracing sound paths back to their sources. LMS Virtual.Lab Noise and Vibration offers advanced capabilities for building full-system simulation models, combining test-derived models of existing components with Finite-Element (FE) models of newly designed parts. Based on this hybrid model, LMS Virtual.Lab predicts vibration responses and sound levels for any proposed change to the product configuration.

“One of the primary advantages of using the LMS software in this approach is that the entire process is performed in a single environment, which eliminates problems with data transfer and conversion,” says Loren DeVries, Senior NVH Engineer at the John Deere Technical Center in Moline, Illinois. “Moreover, the process is captured, making it easy to do multiple analyses in studying the impact of design changes, and in applying the process to other products.”

DeVries explains that for the skid steer loader, the process was completed in four phases: a baseline series of initial sound tests to identify noise sources, a second series of tests to quantify the strength of each sound source and determine sound transfer paths, then building a hybrid acoustic model based on these test results, and finally using that model for acoustic simulation to predict sound levels in studying the impact of design modifications.

For baseline testing, the LMS testing and analysis system was used to make standard sound measurements through a series of microphones set up in and around a skid steer loader, which was run through its normal operating modes. Analysis of test data identified primary noise sources according to frequency: a 3rd order peak from the vehicle engine and 9th, 13th, and 18th order peaks from the hydraulic motors and pumps. In addition to these peaks, the engine structure, air intake/exhaust, and fan also contributed to overall noise levels.

To quantify the contribution of each of these noise sources, a series of focused tests was performed: running mode tests measuring operating accelerations at more than a thousand points on the structure, experimental modal analysis using a force input from shakers, and reciprocal FRF (Frequency Response Function) measurements of sound source transmission to the operator’s ear. These data were analyzed to establish the paths taken from the noise sources: TPA for structure-borne noise (through the frame, mounts, valve block, etc.) and ASQ for noise transmitted through the air (from the engine, body panels, and intake/exhaust outlet nozzles). TPA and ASQ analysis revealed that emitted airborne noise was less of a contributor than structure-borne.

For building the hybrid model, forcing functions determined through TPA in the previous phase were used as input loads to an FE model to establish surface vibration data needed in generating a hybrid acoustic model of the skid steer loader in LMS Virtual.Lab. The structural FE model was correlated with vibration response data taken in the second phase. In this hybrid approach, empirical load data becomes part of the virtual simulation model to provide a noise prediction of sound level at the operator’s ear. Predictive capabilities within LMS Virtual.Lab include Acoustic Transfer Vectors (ATV) that relate surface velocity to sound pressure at a target location for a range of frequencies, and modal ATVs (MATV) that superimpose structural modes of vibration on the transfer vectors. Predicted results from this acoustic model were then correlated with measured data from baseline tests and were found to agree well.

With the validity of the acoustic model verified, Deere engineers could complete the process in fully evaluating noise sources and sound transfer paths for the skid steer loader to improve its sound performance. The simulation focused on the largest noise contributor, which was hydraulic noise that followed a structure-borne path to outer body panels and the floor, where it radiated into the air. The LMS Virtual.Lab model was then used to study the impact of several design modifications on overall sound levels. Fundamental changes to the configuration of the machine were more effective than subtle changes to the frame structure. In this way, acoustic simulation allowed engineers to study the impact of these changes and identify some of the best ideas to guide future designs without having to go through multiple physical prototypes.

DeVries explains that the benefit of virtual simulation is that noise and vibration sources are clearly identified before prototype testing and engineers can immediately study the sound-level impact of proposed changes on the design. “As a result, trial and error testing on multiple physical prototypes is reduced significantly, allowing us to reduce the time and cost of product development and refine the design more effectively than is practical otherwise,” he says. “Testing is still an important part of development, but the purpose is shifted from that of last-minute troubleshooting to a dual role of data capture for building the hybrid model in the early stages of development as well as verification that design goals have been met near the end of development.”

Time and cost savings for Deere through acoustic simulation are significant. According to DeVries, reducing noise using test-based methods exclusively required three or four physical prototyping cycles, each taking weeks to complete and costing tens of thousands of dollars for redesign and assembly of new casting, sheet metal panels, noise barriers and other components. By reducing the number of prototype cycles to only one or two, predictive noise and vibration simulation has the potential to shorten development time and reduce expenses considerably.

Gordon Miller, Engineering Manager for the Skid Steer line at John Deere, explains that the benefits of acoustic simulation go far beyond time and cost saving. “Noise reduction is becoming increasingly important in the global construction equipment market,” says Miller. “The effective use of analytical tools is critical in reducing prototype iterations to improve our speed to market.”