Skeleton Unit Model

The IDAES Skeleton Unit model represents a bare bones unit (hence, the name skeleton) that lets the user define a unit model with a custom set of variables and constraints that best suits their needs but that can be connected to other models from the unit model library. This unit can be used to either define a surrogate for a complex unit model or to define a first-principles based model that does not adhere to the typical IDAES modeling hierarchy. To facilitate connection to other models in the unit model library, the user needs to add ports and populate with variables that match the port members of the models this will be connected to. To enable this, a special method is provided to enable adding ports with a specified set of variables.

The two methods available for this unit are:

add_ports: Allows the user to define the inlet and outlet ports for the model and populate the port with a set of variables.
initialize: A default initialize method that will attempt to solve the model as defined or that can be substituted with a custom callback function provided by the user.

Degrees of Freedom

There are no variables or constraints as this is a bare minimum skeleton that will be populated with variables and constraints as defined by the user. Therefore, the number of degrees of freedom will be determined based on the model the user implements.

Model Structure

Given that this unit is only a skeleton, the model structure is whatever the user intends to be. For example, the model has no restrictions on number of inlets and outlets.

Variables

Variables will be defined by the user. Note that when defining variables, they will need to be indexed by the flowsheet time index irrespective if it is steady-state or dynamic. This is to ensure that port members can be connected as there is still a 0 time index even when the flowsheet is steady-state.

Custom Callback

Note that the default value for the initializer config argument is set to trigger the _default_initialize method. To use a custom callback instead, the user needs to first define a function and set the obj.config.initializer to the custom callback function they defined. In the initialize method, after checking that degrees of freedom is zero, the following is triggered:

self.config.initializer(self, opt=opt, init_log=init_log, solve_log=solve_log, initial_guess=initial_guess)

Note that the function signature in the above line expects an instance, opt which is the solver (default is set to ipopt), init_log and solve_log which are the idaes logger objects for initialization and solve, and initial_guess is a dict containing intial values for variables declared in the model. The user can define the custom function in one of the following ways:

def custom_func(unit, **kwargs):

    # custom initialization sequence
    # unit - instance of SkeletonUnitModel

In the above example, the unit refers to the instance and **kwargs will handle the additional function arguments that will be passed in the initialize method. The callback can be written to use these arguments if desired.

def custom_func(unit, opt, init_log, solve_log, initial_guess):

 # custom initialization sequence
 # unit - instance of SkeletonUnitModel

In the above example, the unit refers to the instance and the expected function arguments are listed explicitly and custom values can be passed to the callback when calling the initialize method. This is recommended especially when the user wants to use different values than what is passed in the initialize method, for example, a different solver instead of ipopt.

idaes.models.unit_models.skeleton_model