This might be happening because of mechanical issues that the AI crews can't fix.
Find your "papy_ai.ini" in your main NR2003 folder, and create a backup copy first.
Open the file and find the "[ mechanical ]" section. You will see a number of variables under the line "; mechanical problem chances, should total 100".
At the bottom of these variables will be three fields ending with "leak": Fuel, coolant, and oil. Change all of these numbers to 0.0 and add their numbers back into other variables so your total still adds up to 100. Hopefully this resolves the issue.
As an example to show you what this section looks like, here it is from one of the Papy AI files for an offline league I run:
; mechanical problem chances, should total 100
suspension_problem_chance = 2.0 ; % chance any problem involves suspension problem
tire_problem_chance = 10.0 ; % chance any problem involves tire problem
fuel_system_problem_chance = 0.0 ; % chance any problem involves fuel system problem
engine_problem_chance = 9.0 ; % chance any problem involves engine problem
brake_problem_chance = 5.0 ; % chance any problem involves brake problem
suspension_failure_chance = 6.0 ; % chance any problem/failure involves suspension failure
tire_failure_chance = 25.0 ; % chance any problem/failure involves tire failure
fuel_system_failure_chance = 0.0 ; % chance any problem/failure involves fuel system failure
engine_failure_chance = 30.0 ; % chance any problem/failure involves engine failure
brake_failure_chance = 13.0 ; % chance any problem/failure involves brake failure
fuel_leak_chance = 0.0 ; % chance any problem/failure involves fuel leak
coolant_leak_chance = 0.0 ; % chance any problem/failure involves coolant leak
oil_leak_chance = 0.0 ; % chance any problem/failure involves oil leak
As an aside, you can copy and paste the following into your Papy AI file so you can have an offline chat window that keeps track of what mechanical issues AI cars experience, and you can throw arbitrary yellows with it to:
[ debug ]
tire_wear_calibration_enable = 1 ;
update_mechanical_problems = 1 ;
Update: I think this issue is more likely something that the track.ini of Daytona BBMC (for example) has. The original ai_track (yeah I have two copies of the track.ini of Daytona BBMC) the cars pit normally without that glitch. To be more specific is with the [ai_track] which this is what I have (coming from a tutorial to how to get 3-4 wide racing on plate tracks):
[ ai_track ]
pace_merge_from_pit_line_dlong = 320.000
pace_merge_to_pit_line_dlong = 3650.0000
pace_speed_limit_mph = 85
ai_accel_modifier = 1.00 ; acceleration grip efficiency
ai_decel_modifier = 1.00 ; braking grip efficiency
ai_fuel_use = 0.98 ; > 1.0 = more fuel consumed was 0.93
ai_grip_modifier = 1.40 ; > 1.0 = more grip h-l 0.905
ai_drag_modifier = 1.15 ; > 1.0 = more drag, which is slower h-l 1.01
ai_line_modifier = 0.90 ; dangerous hack, don't use unless you are a trained professional
ai_qual_modifier = 1.00 ; > 1.0 = faster during qualifying
ai_wall_offset = 44 ; 100 = full-time AI collision detection
ai_tire_wear_left = 0.35 ; left tire wear adjuster for ai > 1.0 = more wear was 0.30 h-l 0.40
ai_tire_wear_right = 0.44 ; right tire wear adjuster for ai was 0.39 h-l 0.49
ai_arcade_tire_wear = 0.01 ; lower = less wear (used in addition to the above left/right modifiers)
ai_arcade_fuel_use = 0.96 ; > 1.0 = more fuel consumed (used in addition to fuel_use above)
ai_arcade_grip_modifier = 1.0 ; > 1.0 = more grip
ai_arcade_drag_modifier = 1.0 ; > 1.0 = more drag, which is slower
strategy_lap_time_wear_loss = 2.516 ; lap time loss from totally worn tires
strategy_lap_time_fuel_use_gain = 0.118 ; lap time gain from low fuel
strategy_base_pitting_cost = 59.756 ; time lost to a stop & go pitstop
ai_pacing_distance = 1.5 ; car lengths behind while pacing
ai_bunching_distance = 1.4 ; car lengths behind while bunching for start
ai_drafting_distance = 1.17 ; car lengths behind while racing (was 1.45)
ai_panic_decel = 400.0
ai_dlongpad_scale = 0.20 ; scales padding in braking zones (smaller = more scaling)
ai_dlat_pad = 1.0
ai_squeeze_pcnt = 0.00
ai_inverse_slipcurve_k = 0.79 ; how much slip angle AI need to corner. smaller = more, more = increased drag
I guess I have to change something to finally solve that glitch. When I tried with the papy_ai the AI stills the same with Daytona BBMC. Sorry for the interruption but I need to know what should I change to stop the stalling on pit road issue (oh forgot to mention when the cars stall they start shaking a bit which makes the glitch more weird but I guess changing something of the AI track can solve this issue)