This paper discusses some improvements to ultrasonic synthetic aperture imaging in solids with primary applications to in-situ non-destructive testing of railroad tracks. Specifically, the study examines three different subarray emission modalities that enhance the signal-to-noise ratio in ultrasound inspection using a phased array with the addition of a transducer wedge. The subarray emission approach effectively compensates for the loss of synthetic transmit power in sparse transmission. Numerical analysis and experimental testing of a point scatterer evaluate the focusing ability of the three subarray emission methods with different subaperture configurations. The Delay-and-Sum (DAS) algorithm builds synthetic aperture images by tracing the refracted ray paths in the multi-layered structure. The transmit subarrays are modeled as virtual elements in the DAS algorithm to accommodate the phase coherence of the synthetic wavefronts. The best subaperture configuration is found by evaluating the resolution and contrast of synthetic aperture images of a subwavelength scatterer in an aluminum block. The SAF imaging approach proposed in this paper have been tested in a portable system aimed at the imaging of internal flaws in rails. The ability to obtain 3D images of rail flaws is badly needed in the railroad maintenance community to enable the most informed remedial decisions following the detection of a defect. The portable system utilizes GPU parallel computation to realize quasi real-time DAS beamforming, and defects are visualized in 3D with calibration of the rail profile. Results are shown from a series of tests conducted on a variety of flaws present in rail sections. These results show the improvements in both imaging speed and signal strength that can be achieved by the subarray emission methods over the conventional synthetic aperture imaging methods.