Pyomo 动态规划与离散输入集的目标函数

Question

我刚开始学习Python和Pyomo，下面这段代码是用来根据一个简单的微分方程来计算电容器的大小。这里有一个离散数组，经过插值处理后用在问题的公式中。为了验证我的结果，我使用了这一行代码，但它报错了。

Pref = [pyo.value(m.intPload[i]) for i in m.t]

错误信息: ValueError: 评估参数值时出错 (intPload[0.03569])：当前参数值设置为无效值。这通常是因为标量参数或可变索引参数没有初始值或默认值。

**完整代码：

import numpy as np
from scipy import integrate
import math
import matplotlib.pyplot as plt
import pyomo.environ as pyo
from pyomo.dae import ContinuousSet, DerivativeVar
model = m = pyo.ConcreteModel()


# power and time data
t1 = np.array([0, 1, 2, 3, 4, 5]) #s
P1 = np.array([0, 10, 20, 30, 40, 0]) # W
# resample the data
dt      = 1 # delta time, resampled every one second
tFinal  = math.floor(t1[t1.size-1])
numSims = math.ceil(tFinal/dt)
tsecs   = np.linspace(0, tFinal, numSims)
Pload   = np.interp(tsecs, t1, P1)
# Capacitor rating
Cc      = 1
Rsc     = 2.4e-3 # capacitor internal resistance
VcMax   = 3 # maximum capacitor cell voltage
IcMax   = 50

# Pyomo problem formulation
m.t     =  ContinuousSet(initialize=tsecs)
m.Vc    = pyo.Var(m.t, bounds =(0.5*VcMax, VcMax))
m.VcT   = pyo.Var(m.t, bounds =(0.5*VcMax, VcMax))
m.Ic    = pyo.Var(m.t, bounds=(-IcMax, IcMax))
m.N_sc  = pyo.Var(bounds=(0, 100))
m.N_pc  = pyo.Var(bounds=(0, 100))
m.dVcdt = DerivativeVar(m.Vc, wrt=m.t)
m.intPload = pyo.Param(m.t, mutable=True)

timepoints  = list(m.t)
Pload_Int   = np.interp(timepoints, t1, P1)
for i,t in enumerate(timepoints):
     m.intPload[t] = Pload_Int[i]

m.obj = pyo.Objective(expr=(m.N_sc+m.N_pc)**2 + sum( (m.VcT[i]*m.N_sc*m.Ic[i]*m.N_pc - m.intPload[i])**2 for i in m.t), sense=pyo.minimize)

def _Vcdot(m, t):
    return m.dVcdt[t] == -m.Ic[t]/Cc
m.Vcdot = pyo.Constraint(m.t,rule=_Vcdot)

def _VcTR(m, t):
    return m.VcT[t] == m.Vc[t] -m.Ic[t]*Rsc
m.VcTR = pyo.Constraint(m.t,rule=_VcTR)

def _init_conditions(m):
    yield m.Vc[0] == VcMax
    #yield m.VcT[0] == VcMax
m.init_conditions = pyo.ConstraintList(rule=_init_conditions)

# Discretize model using Orthogonal Collocation
discretizer = pyo.TransformationFactory('dae.collocation')
discretizer.apply_to(model, nfe=8, ncp=5)


solver=pyo.SolverFactory('ipopt')
results = solver.solve(model,tee=True)


x1 = [pyo.value(m.Ic[i]) for i in m.t]
Pl = [pyo.value(m.Ic[i]*m.VcT[i]*m.N_sc*m.N_pc) for i in m.t]
Pref = [pyo.value(m.intPload[i]) for i in m.t]

#Pl2 = [pyo.value(m.intPload[i]) for i in m.t]


diffPl = [pyo.value(m.Ic[i]*m.VcT[i]*m.N_sc*m.N_pc - m.intPload[i]) for i in m.t]


plt.plot(x1)
plt.show( )

错误信息： ValueError: 评估参数值时出错 (intPload[0.03569])：当前参数值设置为无效值。这通常是因为标量参数或可变索引参数没有初始值或默认值。

在使用Pyomo DAE时，如何在某个时间段内给变量分配一个常量值？我想让m.intPload[i]的值呈阶梯状，但用上面的代码却只得到了采样实例。总之，我想把代码执行的设计空间离散化，或者至少让一些变量有一组离散的值。

目标函数参数设置微分方程动态规划离散输入插值处理变量分配设计空间

Pyomo 动态规划与离散输入集的目标函数

1 个回答

撰写回答