Monte Carlo Simulations

Download Report

Transcript Monte Carlo Simulations

Experiments 8 and 9
Monte Carlo Simulation
 Monte Carlo simulations in PSpice can be run as either:
 a worst case analysis where the maximum deviation from
the nominal values of each component are used in the
calculations
 a statistically-driven variation from the nominal values of
the components using a bell shaped curve for the
distribution of deviations.
Prior to Monte Carlo
• Set up a Simulation Run using Analysis Setup
– You must select one of the other types of analysis first
–
Click on DC Sweep
–
Pick either of the voltage sources to sweep
• Make both the start and end
values equal to the nominal
value of the voltage source
• Pick any value for the
increment.
• Click OK.
Analysis Setup
 Select Monte Carlo/Worst Case
 In the pop-up window, enter the number of runs in MC Run


At least 500 is recommended for statically reasons
Not more than 2000 due to run time issues
Do not unselect the Bias Point Detail
Monte Carlo or Worst Case
• In Output Var: box, enter V(node
name) or V(component:node)
• This is the advantage of labeling
the nodes. So, you will enter V(X)
and V(Y) for the two Monte Carlo
simulations that are required.
–
Note that you will also have to put a
voltage marker on to the circuit
schematic no matter what you put as
the output variable
– If you have selected a DC Sweep,
select DC as the Analysis Type.
Monte Carlo or Worst Case
 Select All on the MC Options.
 Then, click OK.
 Once the schematic is saved,
you are ready to run the
simulation.
Open the Plot Window
Click OK.
Output Plot is a Histogram
Histogram
 The shape of the histogram and the values of the mean,
maximum, and minimum voltages at node X will differ
every time you run the simulation because a random
number generator is used to determine the values of
each of the five resistors in every one of the runs.
 The differences in the mean, maximum, and minimum
values are negligible if sufficient number of runs are
entered in the MC Run box when you set up the simulation.
 The shape of the histogram becomes more symmetrical as
the number of runs approaches 2000.
Your Experimental Results
 If your voltage supplies are exactly +5 V and +9 V, then
the voltages that you measure at nodes X and Y should
fall somewhere within the range from the minimum to
maximum values of the voltages calculated during the
Monte Carlo simulation.
 If you measure the actual values of the +5 V and +9 V
supplies and enter them into the simulation, then there is a
100% likelihood that your measurements will fall within this
range.