1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384
|
# type: ignore
"""Chemical Engineering Design Library (ChEDL). Utilities for process modeling.
Copyright (C) 2018, 2019, 2020, 2021, 2022, 2023 Caleb Bell <Caleb.Andrew.Bell@gmail.com>
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicensse, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
"""
from math import log, sqrt
__all__ = ['polyint', 'polyint_over_x', 'polyder', 'quadratic_from_points',
'deflate_cubic_real_roots', 'exp_poly_ln_tau_coeffs3', 'exp_poly_ln_tau_coeffs2',
'polynomial_offset_scale', 'stable_poly_to_unstable',
'polyint_stable',
'polyint_over_x_stable', 'poly_convert',
]
from fluids.numerics.special import comb
# def stable_poly_to_unstable(coeffs, low, high):
# if len(coeffs) == 0:
# return coeffs
# if high != low:
# from numpy.polynomial import Polynomial
# # Handle the case of no transformation, no limits
# my_poly = Polynomial([-0.5*(high + low)*2.0/(high - low), 2.0/(high - low)])
# def horner(coeffs, x):
# # Keep this copy here
# tot = 0.0
# for c in coeffs:
# tot = tot*x + c
# return tot
# coeffs = horner(coeffs, my_poly).coef[::-1].tolist()
# return coeffs
def poly_add(p1, p2):
# Adds two polynomials p1 and p2
max_len = max(len(p1), len(p2))
result = [0.0] * max_len
for i in range(max_len):
coeff1 = p1[i] if i < len(p1) else 0.0
coeff2 = p2[i] if i < len(p2) else 0.0
result[i] = coeff1 + coeff2
return result
def poly_mul(p1, p2):
# Multiplies two polynomials p1 and p2
result = [0.0] * (len(p1) + len(p2) - 1)
for i in range(len(p1)):
for j in range(len(p2)):
result[i + j] += p1[i] * p2[j]
return result
def stable_poly_to_unstable(coeffs, low, high):
if len(coeffs) == 0:
return coeffs
if high != low:
a = 2.0 / (high - low)
b = - (high + low) / (high - low)
x_poly = [b, a] # Represents the polynomial b + a*x
# Initialize the result with the first coefficient of the polynomial
result = [coeffs[0]]
for c in coeffs[1:]:
# Multiply the current result by x_poly
result = poly_mul(result, x_poly)
# Add the current coefficient c
result = poly_add(result, [c])
coeffs = result[::-1] # Reverse to match the original coefficient order
# Removes leading zeros from a polynomial
for i in range(len(coeffs)):
if coeffs[i] != 0.0:
return coeffs[i:]
return coeffs
def polyint_stable(coeffs, xmin, xmax):
offset, scale = polynomial_offset_scale(xmin, xmax)
scale = 1.0/scale
N = len(coeffs)
out = [0.0]*(N+1)
for i in range(N):
out[i] = scale*coeffs[i]/(N-i)
return out
# from numpy.polynomial import Polynomial
# numpy_to_int = Polynomial(coeffs[::-1], domain=(xmin, xmax))
# new_thing = numpy_to_int.integ()
# coeffs_from_numpy = new_thing.coef
# coeffs_from_numpy = coeffs_from_numpy[::-1].tolist()
# return coeffs_from_numpy
def polynomial_offset_scale(xmin, xmax):
range_inv = 1.0/(xmax - xmin)
offset = (-xmax - xmin)*range_inv
scale = 2.0*range_inv
return offset, scale
def polyint(coeffs):
"""not quite a copy of numpy's version because this was faster to
implement.
Tried out a bunch of optimizations, and this hits a good balance
between CPython and pypy speed.
"""
# return ([0.0] + [c/(i+1) for i, c in enumerate(coeffs[::-1])])[::-1]
N = len(coeffs)
out = [0.0]*(N+1)
for i in range(N):
out[i] = coeffs[i]/(N-i)
return out
def polyint_over_x(coeffs):
N = len(coeffs)
Nm1 = N - 1
poly_terms = [0.0]*N
for i in range(Nm1):
poly_terms[i] = coeffs[i]/(Nm1-i)
if N:
log_coef = coeffs[-1]
return poly_terms, log_coef
else:
return poly_terms, 0.0
# N = len(coeffs)
# log_coef = coeffs[-1]
# Nm1 = N - 1
# poly_terms = [coeffs[Nm1-i]/i for i in range(N-1, 0, -1)]
# poly_terms.append(0.0)
# return poly_terms, log_coef
# coeffs = coeffs[::-1]
# log_coef = coeffs[0]
# poly_terms = [0.0]
# for i in range(1, len(coeffs)):
# poly_terms.append(coeffs[i]/i)
# return list(reversed(poly_terms)), log_coef
def polyder(c, m=1):
"""not quite a copy of numpy's version because this was faster to
implement.
"""
cnt = m
if cnt == 0:
return c
n = len(c)
if cnt >= n:
c = []
else:
der = [0.0]*n
for i in range(cnt): # normally only happens once
n -= 1
for j in range(n, 0, -1):
der[j - 1] = j*c[j]
c = der[0:n]
return c
def quadratic_from_points(x0, x1, x2, f0, f1, f2):
'''
from sympy import *
f, a, b, c, x, x0, x1, x2, f0, f1, f2 = symbols('f, a, b, c, x, x0, x1, x2, f0, f1, f2')
func = a*x**2 + b*x + c
Eq0 = Eq(func.subs(x, x0), f0)
Eq1 = Eq(func.subs(x, x1), f1)
Eq2 = Eq(func.subs(x, x2), f2)
sln = solve([Eq0, Eq1, Eq2], [a, b, c])
cse([sln[a], sln[b], sln[c]], optimizations='basic', symbols=utilities.iterables.numbered_symbols(prefix='v'))
'''
v0 = -x2
v1 = f0*(v0 + x1)
v2 = f2*(x0 - x1)
v3 = f1*(v0 + x0)
v4 = x2*x2
v5 = x0*x0
v6 = x1*x1
v7 = 1.0/(v4*x0 + v5*x1 + v6*x2 - (v4*x1 + v5*x2 + v6*x0))
v8 = -v4
a = v7*(v1 + v2 - v3)
b = -v7*(f0*(v6 + v8) - f1*(v5 + v8) + f2*(v5 - v6))
c = v7*(v1*x1*x2 + v2*x0*x1 - v3*x0*x2)
return (a, b, c)
def quadratic_from_f_ders(x, v, d1, d2):
'''from sympy import *
f, a, b, c, x, v, d1, d2 = symbols('f, a, b, c, x, v, d1, d2')
f0 = a*x**2 + b*x + c
f1 = diff(f0, x)
f2 = diff(f0, x, 2)
solve([Eq(f0, v), Eq(f1, d1), Eq(f2, d2)], [a, b, c])
'''
a = d2*0.5
b = d1 - d2*x
c = -d1*x + d2*x*x*0.5 + v
return (a, b, c)
def exp_poly_ln_tau_coeffs2(T, Tc, val, der):
'''
from sympy import *
T, Tc, T0, T1, T2, sigma0, sigma1, sigma2 = symbols('T, Tc, T0, T1, T2, sigma0, sigma1, sigma2')
val, der = symbols('val, der')
from sympy.abc import a, b, c
from fluids.numerics import horner
coeffs = [a, b]
lntau = log(1 - T/Tc)
sigma = exp(horner(coeffs, lntau))
d0 = diff(sigma, T)
Eq0 = Eq(sigma,val)
Eq1 = Eq(d0, der)
s = solve([Eq0, Eq1], [a, b])
'''
x0 = 1.0/val
x1 = T - Tc
x2 = der*log(-x1/Tc)
c0 = der*x0*x1
c1 = x0*(-T*x2 + Tc*x2 + val*log(val))
return (c0, c1)
def exp_poly_ln_tau_coeffs3(T, Tc, val, der, der2):
'''
from sympy import *
T, Tc, T0, T1, T2, sigma0, sigma1, sigma2 = symbols('T, Tc, T0, T1, T2, sigma0, sigma1, sigma2')
val, der, der2 = symbols('val, der, der2')
from sympy.abc import a, b, c
from fluids.numerics import horner
coeffs = [a, b, c]
lntau = log(1 - T/Tc)
sigma = exp(horner(coeffs, lntau))
d0 = diff(sigma, T)
Eq0 = Eq(sigma,val)
Eq1 = Eq(d0, der)
Eq2 = Eq(diff(d0, T), der2)
# s = solve([Eq0, Eq1], [a, b])
s = solve([Eq0, Eq1, Eq2], [a, b, c])
'''
x0 = der*val
x1 = Tc*x0
x2 = T*x0
x3 = der2*val
x4 = 2.0*T*Tc
x5 = x3*x4
x6 = T*T
x7 = der*der
x8 = x6*x7
x9 = Tc*Tc
x10 = x7*x9
x11 = x4*x7
x12 = x3*x6
x13 = x3*x9
x14 = val*val
x15 = 1.0/x14
x16 = x15*0.5
x17 = log(-(T - Tc)/Tc)
x18 = x1*x17
x19 = x17*x2
x20 = x17*x17
a = -x16*(x1 + x10 - x11 - x12 - x13 - x2 + x5 + x8)
b = x15*(-x1 + x10*x17 - x11*x17 - x12*x17 - x13*x17 + x17*x5 + x17*x8 + x18 - x19 + x2)
c = x16*(-x1*x20 - x10*x20 + x11*x20 + x12*x20 + x13*x20 + 2*x14*log(val) + 2.0*x18 - 2.0*x19 + x2*x20 - x20*x5 - x20*x8)
return (a, b, c)
def deflate_cubic_real_roots(b, c, d, x0):
F = b + x0
G = -d/x0
D = F*F - 4.0*G
# if D < 0.0:
# D = (-D)**0.5
# x1 = (-F + D*1.0j)*0.5
# x2 = (-F - D*1.0j)*0.5
# else:
if D < 0.0:
return (0.0, 0.0)
D = sqrt(D)
x1 = 0.5*(D - F)#(D - c)*0.5
x2 = 0.5*(-F - D) #-(c + D)*0.5
return (x1, x2)
def polyint_over_x_stable_helper(coeffs, i, n, scale, offset, scale_powers, offset_powers):
# term = scale**(i)
term = scale_powers[i]
inner_term = 0.0
for j in range(n):
multiplier = comb(j, i)
# delta = multiplier*coeffs[-j-1]*offset**(j-i)
delta = multiplier*coeffs[-j-1]*offset_powers[j-i]
inner_term += delta
return term*inner_term/i
def polyint_over_x_stable(coeffs, xmin, xmax):
'''Take a stable polynomial coefficient series as
evaluated by horner_stable and the limits e.g. Tmin, Tmax
and transform them into the integral over x.
This has the unfortunate property of breaking the stability
of the series. The impact of this is bad but nothing
catastropic has been found yet.
The output int_over_x_coeffs, log_coeff should
be evaulated with horner_log.
I tried using math.fsum for power accuracy in the coefficients but it
did not help.
The coefficients from this function can be converted
to stable form (goes directly into horner_stable_log) as follows:
from numpy.polynomial.polynomial import Polynomial
stable_coeffs = Polynomial(terms[::-1]).convert(domain=(Tmin, Tmax)).coef.tolist()[::-1]
However, the precision of the conversion is worse.
'''
offset, scale = polynomial_offset_scale(xmin, xmax)
n = len(coeffs)
scale_iter = 1.0
scale_powers = [scale_iter]
for i in range(n):
scale_iter *= scale
scale_powers.append(scale_iter)
offset_iter = 1.0
offset_powers = [offset_iter]
for i in range(n):
offset_iter *= offset
offset_powers.append(offset_iter)
log_coeff = 0.
for i, coeff in enumerate(coeffs[::-1]):
log_coeff += coeff*offset_powers[i]
terms = [0.0]
for i in range(1, n):
term = polyint_over_x_stable_helper(coeffs, i, n, scale, offset, scale_powers, offset_powers)
terms.append(term)
terms.reverse()
return terms, log_coeff
def poly_convert(coeffs, Tmin, Tmax):
# from numpy.polynomial.polynomial import Polynomial
# return Polynomial(coeffs).convert(domain=(Tmin, Tmax)).coef.tolist()
off = 0.5*(Tmin + Tmax)
scl = 0.5*(Tmax - Tmin)
degree_P = len(coeffs) - 1
Q_coeffs = [0.0] * (degree_P + 1)
# Precompute powers of off and scl up to degree_P
off_powers = [1.0]
scl_powers = [1.0]
for _ in range(degree_P):
off_powers.append(off_powers[-1] * off)
scl_powers.append(scl_powers[-1] * scl)
for i in range(len(coeffs)):
coeff_i = coeffs[i]
binom = 1.0
for k in range(i + 1):
term = coeff_i * binom * off_powers[i - k] * scl_powers[k]
Q_coeffs[k] += term
binom *= (i - k) / (k + 1.0)
return Q_coeffs
|