1. To activate recovery of substrcture solution, follow these steps:
====================================================================
   a. In the input file tr_parameters.inp set
      recov_yes_no                     = 'yes'
      freq_recov                       =  4 (antiphase) or 3 (for inphase)

   b. In the input file tr_entire_truck_in(anti)_phase.inp uncommnent at all places
      *include,input=tr_substruct_recovery.inp
      *include,input=tr_road_in(anti)phase.inp

2. Sequence of analysis to be followed on generation jobs:
==========================================================
'tr_bed_gen'
'tr_bumper_gen'
'tr_cabin_gen'
'tr_chassis_gen'
'tr_door_left_gen'
'tr_door_right_gen'
'tr_driveshaft_gen'
'tr_engine_gen'
'tr_fuel_tank_gen'
'tr_hood_gen'
'tr_rear_axle_gen'
'tr_seat_gen'
'tr_susp_lower_arm_left_gen'
'tr_susp_lower_arm_right_gen'
'tr_susp_upper_arm_left_gen'
'tr_susp_upper_arm_right_gen'
'tr_wheel_back_left_gen'
'tr_wheel_back_right_gen'
'tr_wheel_front_left_gen'
'tr_wheel_front_right_gen'

3. Then run the usage analysis as follows:
==========================================
abq -j tr_all -input tr_entire_truck_in(anti)_phase -user exa_tr_radial_uel 

Note: Before running generation and usage jobs, all the other inp files attached with the
example must be present in your work directory for the jobs to execute completely.

4. To get a single odb for all the substructure recovered output
================================================================
Use the "substructurecombine" utility as documented in the Execution Procedures section
of Analysis User's manual. 

abaqus substructurecombine <arguments>

required arguments:
-------------------	
baseodb = odb-file-name
copyodb = odb-file-name

optional arguments:
-------------------
[all]
[step=step-name]
[frame=frame-number]
[variable=variable-key]

Note: Refer to Chapter 3, "Execution procedure for combining output from substructures"
      of Analysis User's manual for details on usage. 

=======================================================================
Optional exercise: To get better results, following changes can be made
=======================================================================

1. Initial static step under grav loads (if the model is changed)
    The following three steps have to be taken:
    a.  - obtain a 1st estimate for tire forces, leaf spring stresses, and strut forces
	# in the main input deck use (one increment step)
	  
	  *boundary,amplitude=<load_style>
	  retained_cabin_cg,                    1,2,0
	  retained_cabin_cg,                    6,6,0
	   wheel_spindles,                     3,3,0

	  comment out the *include,input = tr_initial_stresses.inp

	  comment out the UELs

	# in tr_parameters.inp
          lrealPropFlag      =  0.0 -> rigid props for struts and leaf springs

        # this combination ensures very small deformations (desired since
	  the geometry is assumed to correspond to the grav loaded config)

	# the RFs at the "wheel_spindles" will be used as 
	  precompression forces in the tire uels in the next analysis

	# the CEF1 in the struts will be used to compute a strut pre-compression 

	# copy SF for the leaf springs (form .dat) to tr_intial_stresses.inp

    b.	- replace vertical BCs with tires and switch to real props
	# in the main input deck (one increment step)
	  *boundary
	  retained_cabin_cg,                    1,2,0
	  retained_cabin_cg,                    6,6,0
	  **wheel_spindles,                     3,3,0

	  uncommnent  *include,input = tr_initial_stresses.inp

	  uncommnent the UELs

	# tr_parameters.inp: 
	    lrealPropFlag      =  1.0 -> real props for struts and leaf springs
	    place in strut_init_f_* the CEF1s from the prev analysis (from .dat)
	    place in tire_*_*_force_init the RFs from prev analysis

	# this analysis will take 9-10 iterations; RF2s are very small (<1.0); the right 
	  front wheel moves outwards about 3.6 mm

     c.  - iterate to get eventually better results
	# use tire forces from prev analysis into  tire_*_*_force_init
	# use CEF1 from prev analysis into strut_init_f_*
	# use SF from prev analysis into tr_intial_stress.inp
     
	# Not significant improvement in convergence was seen by doing so.

2. To run longer completely over the bump  analyses (the settingss in QA are set for short times over the bumps)
   a. in tr_parameter_anti(in)phase.inp 
      qa_length = 1.0 (or larger)
   b. run the generation jobs + the analysis