readme.txt

Benchmark Control Problems for Seismically Excited Nonlinear Building

+------------------------------------------------------------------+
|******************************************************************|
|*                                                                *|
|*               Benchmark Control Problems                       *|
|*       for Seismically Excited Nonlinear Building               *|
|*                                                                *|
|*           Produced by University of Notre Dame                 *|
|*                                                                *|
|*        Y. Ohtori ([e-mail] or [e-mail]) *|
|*        R. Christenson ([e-mail])                        *|
|*        B.F. Spencer, Jr. ([e-mail])                      *|
|*                                                                *|                             
|******************************************************************|
+------------------------------------------------------------------+
-------------------
--- 0. Contents ---
-------------------
   1. Description of Files
   2. Important Parameters
   2. Required Software
   3. Getting Started
   4. Post Processing
   5. Notes

-------------------------------
--- 1. DESCRIPTION OF FILES ---
-------------------------------
1.1 Index of Files Used
  The following programs are required for this benchmark study: 
  - Run_NLBM.m
  - Bld_NLBM.m
  - struct#.m
  - inout_#.m
  - ctrlr_#.m
  - red_mod.mat (required to run sample controller)
  - Sim_NLBM.mdl
  - Nonlin12.mexsol (Work Station) or Nonlin12.dll (PC Windows)
  - elcentro.mat, hachinhe.mat, kobe.mat, northrdge.mat
  - EvalNLBM.m
  - unctrl#.mat
    ----------        ------------         ----------          -----------
   | Run_NLBM | ---> | Bld_NLBM | --->  | Sim_NLBM |  --->  | EvalNLBM |
    ----------        ------------         ----------          -----------
                     - struct#          - Nonlin12          - unctrl#
                     - inout_#      - elcentro, hachinhe,
                     - ctrlr_#       kobe, northrdge
                        * red_mod

1.2 Explanation of the Files

<< Run_NLBM.m >>
 This program will run the 10 required simulations of the Benchmark 
control strategy for the earthquake time histories. 

<< Bld_NLBM.m >>
 This program makes the structural data and sets the parameters for the
SIMULINK model. DO NOT MODIFY THIS PROGRAM.

<< struct3.m, struct9.m or struct20.m >>
  These files define the structural configuration, member properties and 
calculation conditions. DO NOT MODIFY THESE PROGRAMS (except to change
calculation time steps).

<< in_output_3.m, in_output_9.m inout_20.m >>
  These files define order and required input and output results. 
Designers/Researchers should define both the location and the type of 
structural responses required for evaluation, sensors and connected 
measurements. The locations of control device forces should be defined 
here also. Note that a "inout_20.m" file has been created for the 
sample control design and to provide an example as to how one might
write this file.  

<< controller_3.m, controller_9.m or ctrlr_20.m >>
 This file defines control devices and control algorithms. Designers/
Researchers should implement their own control devices and algorithms. 
Note that a "ctrlr_20.m" file has been created for the 
sample control design and to provide an example as to how one might
write this file.

<< red_mod.mat >>
  This data file of a reduced order model for the 20-story benchmark
building is included for use in the development of the sample control 
design. This data file is not required for Designer's/Researcher's
own control designs.

<< Sim_NLBM.mdl >>
  This MATLAB SIMULINK program is the nonlinear benchmark simulator.
Designers/Researchers should modify the Sensors and Controller blocks
in accordance with their own control strategies.

<< Nonlin12.mexsol or '.dll' >>
  This S-file performs the nonlinear calculations for the simulator.
DO NOT MODIFY THIS.

<< elcentro.mat, hachinhe.mat, kobe.mat, northrdge.mat >>
  These data files are the time histories of the earthquakes used for
evaluation purposes. See section 1.3. DO NOT MODIFY THESE DATA FILES.

<< EvalNLBM.m >>
  This program determines the evaluation criteria for the 20 story 
benchmark building and control strategy presented here. Designers/
Researchers should modify this program to determine the evaluation 
criteria of their own control strategies. This is provided as a template.

<< unctrl3.m, unctrl9.m, unctrl20.m >>
  These are data files of the uncontrolled values required to calculate the
evaluation criteria. DO NOT MODIFY THESE DATA FILES.

1.3 Earthquake Data Files used for the simulation
   The simulation uses four earthquake records, such as El Centro, Hachinohe,
Kobe and Northridge earthquakes. Each file contains the time vector [sec] 
and the ground acceleration vector [m/sec/sec]. The first row of the file 
is time and second is acceleration. The last point in each record has been 
set to time=600, acceleration = 0 in order to take enough margin. Following 
files are required.
  elcentro.mat
 - El Centro earthquake record (e).
  hachinhe.mat
 - Hachinohe earthquake record (h).
  kobe.mat
 - Kobe earthquake record (k).
  northrdg.mat
 - Northridge earthquake record (n).

-------------------------------
--- 2. Important Parameters ---
-------------------------------

No_bld: Target Building
      3: LA  3 Story Building
      9: LA  9 Story Building
     20: LA 20 Story Building

Idx_linear (1: Linear, Otherwise: Nonlinear)

Idx_Band   (1: Banded Solver, Others: Symmetrical Solver)

--------------------------
--- 3. GETTING STARTED ---
--------------------------

3.1  Required Software
   - MATLAB 5.2 or greater (must be allocated ~32 MB of memory) 
   - MATLAB SIMULINK Toolbox
   - MATLAB Control Toolbox

3.2 The Benchmark Control Problem
  The essence of the benchmark control problem lies here. As a Designer/
Researcher it is your task to define a candidate control strategy. This 
is done in the following steps.
  1) Select a structure that you wish to control from the three provided
     within this study: the 3-story LA SAC building; the 9-story LA SAC
     building; or the 20-story LA SAC building. Indicate your choice by
     defining <No_bld=3>, <No_bld=9>, or <No_bld=20>.
  2) Define the number and location (as well as local dof) of the structural
     responses required by your control strategy for evaluation, sensor and
     connectivity purposes in the file inout_#.m. Include models for 
     your sensors. See section 5.2 for further information.
  3) Define the number and location (as well as local dof) of the control
     forces to be used in your control strategy also in the file 
     inout_#.m.
  4) Define your controller, either passive, semi-active or active, in 
     the file ctrlr_#.m.
  5) Modify the 'Sensor' and 'Controller' blocks in the SIMULINK model
     Sim_NLBM.mdl to accommodate your control strategy.
  You are now ready to execute the simulations.

3.3 Executing the Simulations
  There are two ways to execute the simulations. One is to use the 'Run_NLBM.m' 
program and the other is to run a particular simulation directly from the
command line and Simulation menu in the SIMULINK window. 

<<Run_NLBM>>
  1) Type <Run_NLBM> at the MATLAB prompt. This command runs the file 
     'Run_NLBM.m' which execute simulations.

     Note: The program is designed to execute the simulation in which El Centro 
           earthquake is first used. Sim_NLBM must be set up initially with
           the Ground Accelerations port of El Centro connected. 

<<Command Line>>
  1) Type <No_bld=3>, <No_bld=9>, or <No_bld=20> at the MATLAB prompt to
     indicate which target building you wish to simulate.
  2) Type <Bld_NLBM> at the MATLAB prompt. This command runs the 
     file 'Bld_NLBM.m' which makes structural data and set parameters 
     to the SIMULINK model.
  3) Type <Sim_NLBM> to open the SIMULINK simulator.
  4) Select input earthquake and define the proper stop time in simulation
     parameters (100 sec for El Centro, Hachinohe and Northridge, 180 sec
     for Kobe).
  5) Type <intensity=0.5>, <intensity=1.0>, or <intensity=1.5> at the MATLAB
     prompt to indicate which level of the earthquake you wish to simulate.
  6) Select 'Start' in the SIMULINK Simulation.

3.4 Changing the Integration Time Step
  The time step for integration can be set to any value between 0.001 sec 
and 0.01 sec. This should correspond to the sampling time of the digitally
implemented controller for active and semi-active strategies, and should
be set to the maximum (0.01 sec) for passive strategies, for optimal 
simulation speed. The integration time step can be changed by changing 
the parameter 'dt_cal' in the data file 'struct#.m'
The default time step is 0.01 sec.

--------------------------
--- 4. POST PROCESSING ---
--------------------------
   Once the simulation has been performed for each of the earthquakes and 
   the data is stored, the sample control system can be evaluated using the 
   MATLAB function 'EvalNLBM.m'. This program is provided as a template 
   for calculating the evaluation criteria. Modifications to this 
   program may be necessary for different choices of control devices, 
   configurations, and algorithms.

4.1 MEM_DAMAGE DATA FILE

   In addition to the output data specified by each participant (in the 
   in_output.m file) to determine the evaluation criteria for the building
   responses, control devices and control strategy, data is automatically
   saved to the data file Mem_damage.out for determination of the building
   damage evaluation criteria.

   This file contains the variable Mem_damage and the following information:

   Phi_j       - curvature at the end of the j-th connection divided by 
                 the yeild curvature: Mem_damage(j,5)
   Phi_j_norm  - normed curvature at the end of the j-th connection divided by 
                 the yeild curvature: Mem_damage(j,6)
   dEj         - dissipated energy at the j-th connection divided by the product
                 of the yeild moment and yeild curvature: Mem_damage(j,7)
   Nc          - number of damaged connections: as indicated by a actual to
                 yeild curvature ratio of greater than 1.0 
                 (e.g. Mem_damage(:,5)>1.0)

----------------
--- 5. Notes ---
----------------
5.1 UNITS

   In the model, the units of the inputs and outputs of the structure are:

   Inputs:
      Ground Acceleration [m/sec^2]      Control Forces [N]
   Responses:
      Relative Displacements [m]
      Relative Velocities [m/sec]
      Absolute Accelerations [m/sec^2]

5.2 SENSORS

   The designer should describe and model the sensors employed in each 
   proposed control system, including the sensitivity of the device. In 
   the simulation 'Sim_NLBM.mdl', a {Sensors} block is included, which 
   converts the physical measurements (e.g. accelerations) to sensor outputs 
   (voltages) before any feedback signals pass through the 'Saturation' 
   and {Quantizer} blocks (simulating the effects of the A/D converter). 

   For the sample control design, feedback is provided by accelerometers 
   with a sensitivity of 1g = 10V where 1g = 9.81 m/sec^2. It is left to 
   the researcher/designer to define the units for his/her sensor model(s)
   and control device model(s) and ensure that they are consistent with 
   those of the structure. 

5.3 Structural Data File

   Structural data file defines structural data for simulation. In the 
   benchmark study the LA 3-, 9-, and 20-story building structures are 
   already defined. Researchers/Designers can change calculation parameters 
   in this file. The dt_cal and dt_out and output options can also be changed
   within the constraints specified in Section 3.4 and in the paper.

5.4 User Defined Input/Output data file

   In the user defined file, such as 'inout_#.m' researchers/designers 
   should define the input-ports and out-ports from the Newmark-beta block.
   Researchers must define the location (node number), direction (horizontal, 
   vertical or rotational) and the response (absolute acceleration, relative 
   velcoity or relative displacement). The variable of 'obs' defines the location, 
   direction and response of ye, used for the evaluation. The variable of 'snr'
   defines the location, direction and response of ym, which is for sensors. Note 
   that only absolute acceleration responses are available for feedback measurement 
   in the nonlinear benchmark study. The variable 'cps' defines the location, 
   direction and responses of yc, which is for connection points. The variable of
   'cps' defines the location of control forces, which are calculated and 
   generated in the control block. 

   Also note that for the sample control strategy, the output data 
   is specified to produce a vector of evaluation responses for the 
   displacement, velocity and acceleration of each level (and not all
   displacements, all velocities, all accelerations). This can be
   changed for different proposed control stategies if desired, but 
   calculation of the evaluation criteria for the sample control strategy 
   is based on this format.

           +----------------------+
           |                      |---> ye
   ag  --->|                      |
           |                      |---> ym
    f  --->|                      |
           |                      |---> yc
           +----------------------+

    obs: Defines Observations for Evaluations  (ye)
    snr: Defines Sensor Positions for Acquisition (ym)
    cps: Defines Connection Positions of Passive Devices (yc)
    cf : Defines Location of Control Forces (f)


5.5 User Defined Controller data file

   The 'ctrlr_#' data file defines controller parameters which 
   are required in the SIMULINK model during the simulation. 

   Data for Sensors (for Structure: acceleration)
   Data for Control Signals

   Researcher/designers should define each algorithm and device 
   in this file.

---------------------------------------------------------
*********************************************************
   Questions can be directed to Prof. B.F. Spencer, Jr. 
     via e-mail at: [e-mail]
*********************************************************
---------------------------------------------------------