Yesterday, a lightning struck the SLS' umbilical tower during a scrubbed wet dress rehearsal.
Is it normal for a lightning to strike the umbilical tower instead of the 3 lightning arresters around the pad?
What would have happened if the lightning struck the Orion Launch Escape System instead of the umbilical tower?
If the highly-improbable event of lightning striking the launch abort system of the SLS-Orion stack occurred, it still is unlikely to ignite the rocket motors or trigger an abort scenario. According to this image: 
The LAS seems to be a metal casing, which probably would conduct electricity pretty well. As the entire SLS is made of metal, it would conduct the electricity through the umbilicals and be safely drained. However, if a spark were to ignite the LAS' propellant, it certainly would fire but it is unlikely that the capsule would be separated from the rest of the stack.
Again, this is a speculative answer so feel free to leave comments and/or add in extra information that you find.
The efficiency of the 3 lightning arresters around the rocket may be checked using the rolling sphere method.
You need a scaled model of the arresters, the rocket and the umbilical tower. A sphere with a radius of 20, 30, 45 or 60 m is rolled all over the model. A smaller radius is used for critical and sensitive facilities, a larger for uncritical and robust facilities.
The arresters are efficient if the rolling sphere only touches the arresters and never the rocket or the tower. If not, there is a certain probability for hits of the rocket or the launch tower.
If the distance between the aresters is too large, the rolling sphere may touch the tower without touching the arresters.
See
https://www.nasa.gov/feature/lightning-towers-stand-tall-at-nasa-kennedy-s-launch-pad-39b
Several years ago, while working on a contract as Kennedy’s Advanced Technology and Development Laboratory technical lead, Mata used software he designed to run simulations for a new lightning protection system at the pad. During design, he positioned the towers using this software and was able to obtain full coverage of the mobile launcher by using three lightning towers, rather than the single tower that was part of the previous pad infrastructure during the space shuttle era and only partially protected the orbiter.
“He did a tremendous amount of field work to ensure we had the most robust system while still saving a significant amount of money for NASA,” Perez-Morales said.
At nearly 600 feet, the taller towers are positioned to provide more protection to flight hardware. The height of the towers is a function of the height of the mobile launcher and the SLS rocket when at the pad.
Mata’s design is based on a rolling sphere theory. While the old system used during the Apollo and shuttle programs featured a 45-degree cone of protection, which left a portion of the hardware exposed, Mata’s design completely protects all hardware – including the mobile launcher – under catenary wires. Supported by long insulation masts at the top of the towers, the wires run to the ground almost diagonally, steering the lightning current away from the rocket.
https://file.scirp.org/pdf/EPE_2013051415384450.pdf
https://images.nasa.gov/album/Lightning_Strikes_at_Pad_39B
https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-smt.2018.5002
See how NASA is using a rolling spheres lightning protection system to expand the cone of safety currently used on Launch Complex 39. Find out how the towers and catenary wires create an easier path for the lightning to get to the ground.
Some more editing to be done later.
