Parameter sensitivity study of a Field II multilayer transducer model on a convex transducer

A multilayer transducer model for predicting a transducer impulse response has in earlier works been developed and combined with the Field II software. This development was tested on current, voltage, and intensity measurements on piezoceramics discs (Bæk et al. IUS 2008) and a convex 128 element ultrasound imaging transducer (Bæk et al.

ICU 2009). The model benefits from its 1D simplicity and hasshown to give an amplitude error around 1.7‐2 dB. However, any prediction of amplitude, phase, and attenuation of pulses relies on the accuracy of manufacturer supplied material characteristics, which may be inaccurate estimates. The previous test cases have assumed the simulation parameters to be exact as received from the manufacturer.

In this paper the influence of a deviation in the accuracy of the different parameters is studied by comparing simulation and measurement. The long term objective is a quantitative calibrated model for a complete ultrasound system. This includes a sensitivity study aspresented here.Statement of Contribution/MethodsThe study alters 35 different model parameters which describe a 128 element convex transducer from BK Medical Aps.

The changes are within ±20 % of the values supplied by the manufacturer, which are considered the zero reference (ZR). Simulations of a system consisting of a transmit unit, a five material layer transducer, and the FIELD II predicted pressure are performed by altering in turn the value of a single parameter in steps of 2 %.

The remaining simulation parameters are held fixed at the ZR. The influence of the parameter change is determined by calculating the pressure and the intensity at adistance of 112 mm on an element’s center axis and comparing it with hydrophone measurements. These are performed with a water bath hydrophone setup using an Agilent MSO6014A oscilloscope that is set to average consecutive pulses 48 times for noise reduction of the hydrophone output.

A commercial transmitter unit is used to drive the transducer with a 10 cycle tone burst at a frequency of 4.0 MHz and a maximum excitation amplitude of 31 volt.ResultsPredictions using the ZR give a pressure pulse error (PPE) and an intensity error (IE) of 32 % and 23 %, respectively, relative to the measured.

Altering the piezoelectric permittivity +12 % from ZR decreases the PPE to 30 % and the IE to 2 % relative to the measured. Changing the stiffness constant of the lens -4 % from ZR increases the PPE and the IE with 6 % and 1 %, respectively. Performing the same with the ceramic stiffness the PPE is lowered 1.5 % and the IE is lowered 12 %.Discussion and ConclusionsPPEs are found mainly to be sensitive to lens properties and piezoceramic properties, but minor sensitive to changes in matching layers.

IEs are mainly sensitive to the piezoceramic properties. The study shows that minor changes can improve predictions significantly.