How can I fix 'ValueError: Constraint does not have a proper value' in Pyomo?

ValueError: Constraint does not have a proper value.

ValueError: Constraint 'uvuoncea122222222222[1]' does not have a proper value. Found 'AbstractScalarConstraint'
Expecting a tuple or relational expression. Examples:
   sum(model.costs) == model.income
   (0, model.price[item], 50)

This is the error I encounter for constraint 3. Here is the code and how I defined my sets and decision variables:

## Sets
model.P = pe.Set(initialize = depots, ordered = False) #depots
model.S = pe.Set(initialize = satellites, ordered = False) #satellites
model.I = pe.Set(within=model.P | model.S) #union depots and satellites
model.Z = pe.Set(initialize = customers, ordered = False) #customers
model.Y = pe.Set(within=model.S | model.Z) #union customers and satellites
model.T = pe.Set(initialize=numUV, ordered=False) #FE vehicle
model.F = pe.Set(initialize=numCF, ordered=False) #SE vehicle
model.V = model.P | model.S | model.Z
model.A = model.V*model.V

## Decision variables
model.x = pe.Var(model.T, model.P, model.S, domain = pe.NonNegativeIntegers) #flow FE
model.r = pe.Var(model.T, model.P, model.S, domain = pe.Binary) #binary FE arc used
model.q = pe.Var(model.F, model.S, model.Z, domain = pe.Binary) #binary SE arc used
model.w = pe.Var(model.Z, model.S, domain = pe.Binary) #binary Z gekoppeld aan S
model.u1 = pe.Var(model.T, domain = pe.Binary) #binary FE vehicle used
model.u2 = pe.Var(model.F, domain = pe.Binary) #binary SE vehicle used
model.p1 = pe.Var(model.T, model.S,domain = pe.NonNegativeIntegers) #arrival time at s
model.b1 = pe.Var(model.F, model.Z, domain = pe.NonNegativeIntegers) #arrival time at z
#model.pi = pe.Var(model.F, domain = pe.NonNegativeIntegers) #working duration
model.pprint()

## Objective function
objExpr = (sum(model.r[t,i,j]* c[i,j] * cKm for t in model.T for i in model.I for j in model.I)\
          + sum(model.q[f,i,j]* c[i,j] * cKm for f in model.F for i in model.Y for j in model.Y)\
          + sum(model.u1[t]* ch1 for t in model.T) + sum(model.u2[f] * ch2 for f in model.F))
model.obj = pe.Objective(expr = objExpr, sense = pe.minimize)

## Constraints
#2
model.confcfe = pe.ConstraintList()
for j in model.I:
    for t in model.T:
        expression = sum(model.r[t,l,j] for l in model.I) - sum(model.r[t,j,l] for l in model.I) == 0
        model.confcfe.add(expression)
#3
model.uvuoncea122222222222 = pe.ConstraintList()
for t in model.T:
    expression = sum(model.r[t, i, j] for i in model.I for j in model.P)
    constraint = pe.Constraint(expr=expression <= 1)
    model.uvuoncea122222222222.add(constraint)

The reason for coding constraint 3 in another style then constraint 2 is that I get a trivial/boolean error for constraint3, and it seems like that is fixed while coding the constraint in this manner.

I tried to defining the constraint in different manners (like defining a constraint rule, indirect/direct into the model, defining them as a range (0,sum(...),1). It all didn't work..

The error is telling you that the type of expression needed when making a constraint is a relational expression, so where you have this:

expression = sum(model.r[t, i, j] for i in model.I for j in model.P)
constraint = pe.Constraint(expr=expression <= 1)

bring the RHS (right hand side) up to the expression, like you did in the previous constraint:

expression = sum(model.r[t, i, j] for i in model.I for j in model.P) <= 1
constraint = pe.Constraint(expr=expression)

The constraint above should be able to be constructed several different ways, so there is likely something else wrong/not handled appropriately if it is not working, but there isn't enough information in your question to deduce why.

On a separate note, I think you are setting yourself up for trouble with the way you are indexing/accessing your variables.

You create model.r indexed by [T, P, S] but then in your constraint 2 you are accessing it by [T, I, I] and in constraint 3 you are accessing it by [T, I, P]. That's not going to work as the domain of the sets in the constraints are outside (or supersets) of the variable indices.

If you are stuck there, post a new question with a small instance of the model data that you are initializing all of your variables from and some additional comments about what the purpose of the objectives are and you'll likely get some help. The code in the example should be able to execute model.pprint().