Genevieve Rodrigue^{1} and Chris K. Mechefske^{1}

A persistent issue with MRI scanners is inadequate vibration isolation. The objective of this work is to implement vibration isolation in a prototype MRI scanner base to minimize stringent and costly structural requirements for MRI suites. To first use a computational model of an MRI base as an analysis tool, modal analysis has been conducted to validate the model. Furthermore, analysis of a pneumatic isolator has been completed to assess its suitability to the MRI application. Testing demonstrated adequate performance of the isolator under most expected loading cases. Computational analysis of an idealized isolator model supports the experimental results.

**Introduction: MRI Vibration Problem and Project Description**

A persistent issue with MRI machines is inadequate vibration isolation. This problem is two-faceted. First, the scanner produces significant noise and vibrations during operation. These vibrations, when transmitted to the machine’s environment, have negative repercussions such as the production of uncomfortable and even harmful acoustic noise. Second, MRI scanners take relatively long to perform a scan. Any motion of the scanner will lead to suboptimal imaging. Therefore, it is necessary to isolate the machine from its environment as much as possible. As a result, current MRI suites require significant and costly physical modifications to make them suitable for MRI operation.

Since the 1990s, numerous
relatively successful passive and active methods have been implemented to
address this issue. A recurring design element makes use of a compliant passive
component coupled with piezoelectric actuators^{1}. Unfortunately,
these patents are challenging to implement as they require changes to the
internal structure of the MRI machine, as suggested by Lee et al^{1}. As
such, there is a need to develop a means of isolating an MRI machine from its
environment that does not require modifications to the internal structure of
the machine. The objective of the work presented herein was to implement
passive vibration isolation in the base of a low weight prototype MRI scanner. The
first step was the creation and validation of a finite element (FE) model of
the MRI base, which requires computational and experimental modal analysis. The
second step was the analysis of a load-dependent natural frequency, high
damping pneumatic isolator to assess its suitability to this application.

Modal analysis is used to describe a system in terms of its
dynamic characteristics: natural frequency, mode shapes, and damping ratios^{2}.

A simplified model of the MRI scanner base was imported into the FE modelling software ANSYS for modal analysis. Free-free boundary conditions were used. The natural frequencies and mode shapes of the model up to 600 Hz were calculated.

Subsequently, a modal test was performed on a prototype of the MRI support structure. A shaker provided burst random excitation for 50% of the sample window while accelerometers were roved to 310 points on the surface of the structure. The base was set on rubber mounts to simulate free-free boundary conditions.

Seven mode pairs were successfully visually correlated. An example
of a mode pair is shown in Figure 1. The natural frequencies of correlated mode pairs are
within a 4% difference of each other. One way of evaluating the degree of
correlation between the mode pairs is to plot the experimental frequencies
versus the computational frequencies and fit them with a linear regression. A
slope of 0.968 indicates good correlation. Thus far, good subjective
correlation between experimental and computational results has been found. Conclusions
can now be drawn from the FE model with confidence in their accuracy.

**Isolator Analysis Method and Results**

Testing was performed to give insight into the isolators subjected to their operating conditions since substantial simplifications were necessary in prior analyses. For these tests, accelerometers were mounted above and below the isolators and the floor excited with a 3Hz (near isolator resonant frequency) impact excitation. The isolators were shown to perform suitably well in a typical environment.

In conjunction to experimental testing, the isolators were modelled computationally with the previously validated model of the base. The isolators were represented as a spring damper system with hysteretic damping as a function of excitation frequency. A harmonic analysis was performed at 3.25Hz, the natural frequency of the isolator, and at 10Hz, the estimated resonance frequency of the floor. The isolator performs very well at 10Hz, attenuating most vibrations (see Table 1). At resonance, it is not causing amplification indicating adequate damping is present.

**Conclusions**

1. Lee, J., Rudd, B., Li, M., & Lim, T. C. Sound Reduction Technologies for MRI Scanners. Recent Patents on Engineering. 2008; 2(2), 1-8.

2. Avitabile, P. Modal Testing: A Practitioner’s Guide. 2017. Hoboken, NJ: John Wiley & Sons.

Correlated Mode Pair. The top figure shows the computational result while the bottom figure shows the experimental result.

Figure 2 Linear regression of Correlated Mode Pairs shows a slope above 0.9 indicating good correlation.

Figure 3 Isolator testing
results show good attenuation between floor (orange) and scanner (blue).

Figure 4 Computational Analysis Results