List index out of range (works fine for list size < 25)

[Wrichik Basu]

I have a program where I am supposed to find out the pressure, given the ideal gas equation with Van der Waal's correction, $$\left(p - \frac{a}{v^2}\right)\left(v - b\right) \ = \ Rt$$I have to take a constant value for $t$ (user input), and $v$ will range from 1 to 20 in steps of 0.01. $a = 0.002$ and $b = 1.2$ are given.

I have written the following code:

[CODE lang="python" highlight="29,34"]import numpy as np
from sympy import Symbol, solve

class P008:
def declarations(self):
global R, a, b
R = 8.314
a = .0002
b = 1.2

def inp_PvsV(self):
"""
Takes input for the P vs V file.
"""
print("Enter the value of T: ", end="")
temp = float(input())
vol = np.linspace(1, 20, num=2001)
vol = np.around(vol, decimals=2)
self.compute_PvsV(vol, temp)

def compute_PvsV(self, vol, temp):
"""
Computes pressure from volume and temperature.
:param vol: The Volume as a numpy array
:param temp: The Temperature as a numpy array
"""
p = Symbol("p")
s = []
for i in range(0, np.size(vol)):
# Equation for p (lhs-rhs)
eq = (p + (a / (vol ** 2))) * (vol - b) - R * temp
# Now solving for p
s1 = solve(eq, p)
s.append(s1[0])
print("i = ", i, " s = ", s)
pres = np.asarray(s)
print(pres)

obj = P008()
obj.declarations()
obj.inp_PvsV()[/CODE]

I found during debugging that in line no. 29, if I change np.size(vol) to anything <= 20, I get a good output:

(Couldn't give the full output screen, but the fact that it finished with exit code 0 means that there was no error.)

When I change the number to anything > 20, I get an error:

This seems more like an overflow problem. My code is unchanged, only that number in the range increases, and I get an error. But I am supposed to accomplish this task for 2001 values.

Any idea on how to do this?

I am working in Python 3.7.4, and using PyCharm IDE.

 

[phyzguy]

Note that the error is in the line that says, "s.append(s1[0])". This means that the line above which is, "s1 = solve(eq, p)" didn't return any solution. You should add a print statement above the failing line to print out s1 and the values of the equation you are trying to solve to see whay there is no solution.

 

[Wrichik Basu]

Note that the error is in the line that says, "s.append(s1[0])". This means that the line above which is, "s1 = solve(eq, p)" didn't return any solution. You should add a print statement above the failing line to print out s1 and the values of the equation you are trying to solve to see whay there is no solution.

Yes, you are right. s1 is empty when i=21. I will have to see how to solve that problem.

 

[Tom.G]

Line 17 might have something to do with it.

 

[Wrichik Basu]

Line 17 might have something to do with it.

Yeah, I know that. Actually there is a problem when $v=b$, where the lhs-rhs becomes 0. I shifted that to the front by changing $b$ to 1.0, and then excluded $v=1.0$ case. With a little bit of extra effort, this can be resolved for any $b=v$ case.

 

[jedishrfu]

I would go through and print your intermediate arrays to make sure they are as you expected.

In particular, the vol array where you use linspace to populate it and then use around to limit the decimal pts which could make duplicates that might somehow get removed shortening the array. I see no evidence of that only its something I would check.

 

[Wrichik Basu]

I would go through and print your intermediate arrays to make sure they are as you expected.

In particular, the vol array where you use linspace to populate it and then use around to limit the decimal pts which could make duplicates that might somehow get removed shortening the array. I see no evidence of that only its something I would check.

I did all that in the debugging phase of the program.

 

[jedishrfu]

If it doesn’t work you’re still in debugging phase.

 

[jedishrfu]

One company i worked for implemented an interesting metric to discover the truthfulness of the project manager based on bug tracking and how a bug was introduced into the code base.

The project phases were:
- design phase
- coding phase
- unit test
- integration test
- system test
...

So while the PM might say confidently that we are in system test ready for ship, the bug stats might say we have some design errors and we are still in coding or unit test phases.

I can’t remember the name of the system but for obvious reasons it wasn't wildly popular with folks looking for promotions.

EDIT: Its called Orthogonal Defect Classification. I think its been superseded by other techniques to get a more real-time feel for what is happening in a software project.

https://en.wikipedia.org/wiki/Orthogonal_Defect_Classification