ModelingToolkit.jl cryptic error

Specific Domains Modelling & Simulations
1

I was trying to model a thermo-fluid system with ModelingToolkit.jl, however I encountered this:

julia>     prob = ODAEProblem(sys, Pair[], (0.0, 10.0))
ERROR: AssertionError: !(isalgvar(s, var))

What does it mean?

The connection equations are based on: 15 Stream Connectors‣ Modelica® Language Specification version 3.6-dev

The equations before structural_simplify:

julia> model
Model model with 38 equations
States (42):
  r1₊a₊mdot(t) [defaults to 0.0]
  mixing_valve₊a₊mdot(t) [defaults to 0.0]
  r1₊a₊p(t) [defaults to 100000.0]
⋮
Parameters (7):
  r1₊p [defaults to 200000.0]
  r1₊T [defaults to 353.15]
  r2₊p [defaults to 200000.0]
⋮

julia> equations(model)
38-element Array{Equation,1}:
 mixing_valve₊a₊mdot(t) + r1₊a₊mdot(t) ~ 0.0      
 r1₊a₊p(t) ~ mixing_valve₊a₊p(t)
 r1₊a₊T_in_stream(t) ~ mixing_valve₊a₊T_outflow(t)
 mixing_valve₊a₊T_in_stream(t) ~ r1₊a₊T_outflow(t)
 mixing_valve₊b₊mdot(t) + r2₊a₊mdot(t) ~ 0.0      
 r2₊a₊p(t) ~ mixing_valve₊b₊p(t)
 r2₊a₊T_in_stream(t) ~ mixing_valve₊b₊T_outflow(t)
 mixing_valve₊b₊T_in_stream(t) ~ r2₊a₊T_outflow(t)
 mixing_valve₊c₊mdot(t) + r3₊a₊mdot(t) ~ 0.0      
 mixing_valve₊c₊p(t) ~ r3₊a₊p(t)
 mixing_valve₊c₊T_in_stream(t) ~ r3₊a₊T_outflow(t)
 r3₊a₊T_in_stream(t) ~ mixing_valve₊c₊T_outflow(t)
 r1₊a₊p(t) ~ r1₊p
 r1₊a₊T_outflow(t) ~ r1₊T
 r2₊a₊p(t) ~ r2₊p
 r2₊a₊T_outflow(t) ~ r2₊T
 mixing_valve₊valve_1₊A(t) ~ 7.0e-5mixing_valve₊y 
 mixing_valve₊valve_2₊A(t) ~ 7.0e-5(1 - mixing_valve₊y)
 mixing_valve₊valve_1₊a₊mdot(t) - mixing_valve₊a₊mdot(t) ~ 0.0
 mixing_valve₊a₊p(t) ~ mixing_valve₊valve_1₊a₊p(t)
 mixing_valve₊a₊T_outflow(t) ~ mixing_valve₊valve_1₊a₊T_outflow(t)
 mixing_valve₊valve_1₊a₊T_in_stream(t) ~ mixing_valve₊a₊T_in_stream(t)
 mixing_valve₊valve_2₊a₊mdot(t) - mixing_valve₊b₊mdot(t) ~ 0.0
 mixing_valve₊b₊p(t) ~ mixing_valve₊valve_2₊a₊p(t)
 mixing_valve₊b₊T_outflow(t) ~ mixing_valve₊valve_2₊a₊T_outflow(t)
 mixing_valve₊valve_2₊a₊T_in_stream(t) ~ mixing_valve₊b₊T_in_stream(t)
 mixing_valve₊valve_1₊b₊mdot(t) + mixing_valve₊valve_2₊b₊mdot(t) - mixing_valve₊c₊mdot(t) ~ 0.0
 mixing_valve₊valve_1₊b₊p(t) ~ mixing_valve₊valve_2₊b₊p(t)
 mixing_valve₊valve_1₊b₊p(t) ~ mixing_valve₊c₊p(t)
 mixing_valve₊valve_1₊b₊T_in_stream(t) ~ (mixing_valve₊c₊T_in_stream(t)*max(mixing_valve₊c₊mdot(t), 0) + mixing_valve₊valve_2₊b₊T_outflow(t)*max(-mixing_valve₊valve_2₊b₊mdot(t), 0))*((max(mixing_valve₊c₊mdot(t), 0) + max(-mixing_valve₊valve_2₊b₊mdot(t), 0))^-1)
 mixing_valve₊valve_2₊b₊T_in_stream(t) ~ (mixing_valve₊c₊T_in_stream(t)*max(mixing_valve₊c₊mdot(t), 0) + mixing_valve₊valve_1₊b₊T_outflow(t)*max(-mixing_valve₊valve_1₊b₊mdot(t), 0))*((max(mixing_valve₊c₊mdot(t), 0) + max(-mixing_valve₊valve_1₊b₊mdot(t), 0))^-1)
 mixing_valve₊c₊T_outflow(t) ~ (mixing_valve₊valve_1₊b₊T_outflow(t)*max(-mixing_valve₊valve_1₊b₊mdot(t), 0) + mixing_valve₊valve_2₊b₊T_outflow(t)*max(-mixing_valve₊valve_2₊b₊mdot(t), 0))*((max(-mixing_valve₊valve_1₊b₊mdot(t), 0) + max(-mixing_valve₊valve_2₊b₊mdot(t), 0))^-1)
 mixing_valve₊valve_1₊a₊mdot(t) + mixing_valve₊valve_1₊b₊mdot(t) ~ 0
 mixing_valve₊valve_1₊b₊mdot(t) ~ 44.67661580737735max(min(mixing_valve₊valve_1₊A(t), 7.0e-5), 1.0e-9)*regRoot(mixing_valve₊valve_1₊b₊p(t) - mixing_valve₊valve_1₊a₊p(t), 0.0005000000000000004(202650.0 + mixing_valve₊valve_1₊a₊p(t) + mixing_valve₊valve_1₊b₊p(t)))
 mixing_valve₊valve_2₊a₊mdot(t) + mixing_valve₊valve_2₊b₊mdot(t) ~ 0
 mixing_valve₊valve_2₊b₊mdot(t) ~ 44.67661580737735max(min(mixing_valve₊valve_2₊A(t), 7.0e-5), 1.0e-9)*regRoot(mixing_valve₊valve_2₊b₊p(t) - mixing_valve₊valve_2₊a₊p(t), 0.0005000000000000004(202650.0 + mixing_valve₊valve_2₊a₊p(t) + mixing_valve₊valve_2₊b₊p(t)))
 r3₊a₊p(t) ~ r3₊p
 r3₊a₊T_outflow(t) ~ r3₊T

and after

julia>     sys = structural_simplify(model)
Model model with 1 equations
States (3):
  mixing_valve₊valve_1₊b₊T_outflow(t) [defaults to 293.15]
  mixing_valve₊valve_2₊b₊T_outflow(t) [defaults to 293.15]
  mixing_valve₊valve_2₊b₊T_in_stream(t) [defaults to 293.15]
Parameters (7):
  r1₊p [defaults to 200000.0]
  r1₊T [defaults to 353.15]
  r2₊p [defaults to 200000.0]
⋮
Incidence matrix:
  [1, 1]  =  ×
  [1, 3]  =  ×

julia> equations(sys)
1-element Array{Equation,1}:
 0 ~ (r3₊T*max(44.67661580737735regRoot(r3₊p - r1₊p, 0.0005000000000000004(202650.0 + r1₊p + r3₊p))*max(min(7.0e-5mixing_valve₊y, 7.0e-5), 1.0e-9) + 44.67661580737735regRoot(r3₊p - r2₊p, 0.0005000000000000004(202650.0 + r2₊p + r3₊p))*max(min(7.0e-5(1 - mixing_valve₊y), 7.0e-5), 1.0e-9), 0) + mixing_valve₊valve_1₊b₊T_outflow(t)*max(-44.67661580737735regRoot(r3₊p - r1₊p, 0.0005000000000000004(202650.0 + r1₊p + r3₊p))*max(min(7.0e-5mixing_valve₊y, 7.0e-5), 1.0e-9), 0))*((max(44.67661580737735regRoot(r3₊p - r1₊p, 0.0005000000000000004(202650.0 + r1₊p + r3₊p))*max(min(7.0e-5mixing_valve₊y, 7.0e-5), 1.0e-9) + 44.67661580737735regRoot(r3₊p - r2₊p, 0.0005000000000000004(202650.0 + r2₊p + r3₊p))*max(min(7.0e-5(1 - mixing_valve₊y), 7.0e-5), 1.0e-9), 0) + max(-44.67661580737735regRoot(r3₊p - r1₊p, 0.0005000000000000004(202650.0 + r1₊p + r3₊p))*max(min(7.0e-5mixing_valve₊y, 7.0e-5), 
1.0e-9), 0))^-1) - mixing_valve₊valve_2₊b₊T_in_stream(t)
2

Open an issue for this. Could be a bug. Please share how to recreate it.

3

It is not, the model was missing four equations.
Is there a way to check the consistency of the model?
First thing might be if the number of equations and number of states are the same. But is this true in general.

4

We’re adding a check_consistency fairly soon. See the issue on it.

2 Likes