If the first two carriers are going to be STOBAR, why manufacture anything less than 50+ J-15s, the number needed to fully equip both carriers, with a few left over? CATOBAR-capable J-15s would not be needed for either of those carriers and mixing them with the STOBAR J-15s would be less than ideal from a logistics standpoint.
The idea would be that all carriers (STOBAR, CATOBAR) in the medium term will field an all catapult compatible J-15A fleet, and the small number non catapult compatible and lower capability J-15s would be phased out (or at least relegated to other less front line tasks). The rationale would be to streamline logistics to a single type of J-15 in the naval aviation fleet, but also to increase flexibility -- so say if a J-15A took off from Liaoning or 001A, it could still land on an 002 carrier and be able to take off from it. Or putting it another way, all J-15As in the country will be able to be assigned to either Liaoning, or 001A or 002 (or any other carrier).
For example, if a naval air base on land with J-15As gets attacked, J-15As from another naval air base can still operate from 002. OTOH, having a mixed fleet of J-15As and J-15s will mean if one base of J-15As gets wiped out, then the Navy is may be left with only J-15s that can take off from Liaoning and 001A and 002 thus is unable to launch J-15s of any kind (or 002 is only able to launch a smaller number of J-15As -- this of course is just one scenario of the risks of a lack of standardization, not meant to be exhaustive)
There would be some costs -- specifically, the costs of additional structural enhancements between J-15 and J-15A to accommodate the stress of catapult launches -- but I think this can be vastly offset by reducing the number of standard J-15s purchased and increasing the total number of J-15As purchased, not to mention the significant benefits of providing increased flexibility and cross decking of an all J-15A fleet.
There is also the avionics factor to consider -- current J-15s do not field very competitive subsystems, such as radar, and likely its cockpit can undergo a spruce up along with other avionics like ESM, EW. Part of this is because the current J-15s were designed in a time when the domestic industry couldn't reliably supply systems like AESA radars, and I suspect they cut costs and complexity deliberately by basing many of J-15's subsystems on J-11B.
The catapult compatible J-15A OTOH should be able to leverage a much more advanced subsystem supply chain, and field AESA, improved avionics, and newer weapons systems etc, alongside the structural reinforcements.
So, if the Navy does choose to manufacture 50+ non catapult J-15s like you suggest, then they will eventually have to eventually develop an upgrade package for these standard J-15s, while also fielding the J-15A which should come with these subsystems as standard. Or, they can try to develop both the J-15A and develop an "improved" non catapult compatible J-15 simultaneously, where both field the same newer subsystems but the only difference being the J-15A is structurally reinforced for catapult launches.
But, I think it could be much cheaper to purchase a smaller number of non-catapult, older type J-15s as an interim capability for the two STOBAR carriers, and invest fully in a single standardized (and more capable) naval fighter for all carriers (STOBAR and CATOBAR) with a single avionics and subsystems chain and development pathway, without having to create an upgrade package for older J-15s or develop an improved non catapult J-15 alongside J-15A.
This obviously is in addition to the long term benefits of cross decking/flexibility I mentioned above.