Is possible a multiperiod optimization using pyomo network?

[juanbaquero4]

Hi, I hope you can help me to solve this optimization problem:

Is it possible to work pyomo network with multiperiod optimization, I am trying to propose a renewable network with process optimization, however it was not possible for me to define correctly the ports. Is possible to declare it on IDAES?

Firstly I declared a block with certain variables for the whole m.date, afterward (here comes the issue) I am trying to declare an extensive port (because I am splitting the stream later in the problem), but it is not possible for me to define correctly the port, here is the code:

This is what I am doing:
m.date = Set(initialize=output.index, ordered = True) # is a 120 periods time

#block of 3 variables:

def  block_fx_CSP (m,t):
    m.area_fxCSP = Var(within = NonNegativeReals, doc = '# area fixed CSP')
    m.fxCSP = Var(within = NonNegativeReals, doc = 'Canadian Solar fixed PV (W)')
    m.y_fxCSP = Var (domain = Binary)
m.fxCSP_block = Block (m.date, rule = block_fx_CSP)

#1st attempt. I declared the port as follows but the output is a list of the same first period date variable:

for t in m.date: 
    m.port_fxCSP = Port(m.date, initialize=[(m.fxCSP_block[t].fxCSP, Port.Extensive)]) 

this is the output and it is not correct:

Key                 : Name  : Size : Variable
2019-12-31 19:00:00 : fxCSP :    1 : fxCSP_block[2019-12-31 19:00:00].fxCSP
2019-12-31 20:00:00 : fxCSP :    1 : fxCSP_block[2019-12-31 19:00:00].fxCSP
2019-12-31 21:00:00 : fxCSP :    1 : fxCSP_block[2019-12-31 19:00:00].fxCSP
2019-12-31 22:00:00 : fxCSP :    1 : fxCSP_block[2019-12-31 19:00:00].fxCSP

The output should be:
2019-12-31 19:00:00 : fxCSP : 1 : fxCSP_block[2019-12-31 19:00:00].fxCSP
2019-12-31 20:00:00 : fxCSP : 1 : fxCSP_block[2019-12-31 20:00:00].fxCSP
2019-12-31 21:00:00 : fxCSP : 1 : fxCSP_block[2019-12-31 21:00:00].fxCSP
2019-12-31 22:00:00 : fxCSP : 1 : fxCSP_block[2019-12-31 22:00:00].fxCSP

#2nd attempt: I declared the port and later the function add, but it replaces all the data therefore the output is a 1 output size port:

m.port_fxCSP = Port()

for t in m.date:
    m.port_fxCSP.add(m.fxCSP_block[t].fxCSP, Port.Extensive)

Output:
port_fxCSP : Size=1, Index=None
Key : Name : Size : Variable
None : <function Port.Extensive at 0x00000165C26A65F0> : 1 : fxCSP_block[2020-01-05 18:00:00].fxCSP

#3rd attempt: I declare a function for the port for each m.date as an iterable, however the error indicates not iterable

def port_fxCSP (m,t):
    return [(m.fxCSP_block[t].fxCSP, Port.Extensive)for t in m.date]

m.fxCSP_port = Port (m.date, initialize=port_fxCSP)

TypeError: 'function' object is not iterable

Any suggestions are highly appreciated!

[andrewlee94]

@juanbaquero4 This question looks like it would be better asked on the Pyomo forums as it appears to be solely related to Pyomo code.

However, I think in this case I think your first attempt was close - instead of Port.Extensive I think you need to provide a name. I think something like the following should work.

m.port_fxCSP = Port()

for t in m.date:
    m.port_fxCSP.add(m.fxCSP_block[t].fxCSP, str(t))

Also note that IDAES models are Pyomo models, so we can only do things that are possible in Pyomo.

[juanbaquero4]

As always, absolutely appreciate your answer! Thank you!