Configuration Options

This section describes the different configuration options available in TurboFlow.

Turbomachinery

These are the valid turbomachinery options:

• axial_turbine

Loss Models

These are the valid loss model options:

• benner: see Benner-Sjolander-Moustapha (2006) for model details.

• kacker_okapuu: see Kacker-Okapuu (1982) for model details.

• moustapha: see Moustapha-Kacker-Tremblay (1990) for model details.

• isentropic: Ensures that the loss coefficient is zero.

• custom: Implies a constant loss coefficent, specified through custom_value input in configuration file.

Loss Coefficients

These are the valid loss coefficient options:

• stagnation_pressure: see Loss coefficients for details.

• kinetic_energy: see Loss coefficients for details.

Deviation Models

These are the valid deviation model options:

• aungier: see Aungier for model details.

• ainley_mathieson: see Ainley and Mathieson for model details.

• zero_deviation: see Zero deviation for model details.

Choking Criterion

These are the valid choking citerion options:

• critical_mass_flow_rate: see Maximum mass flow rate for model details.

• critical_mach_number: see Critical mach number for model details.

• critical_isentropic_throat: see Isentropic throat for model details.

Note

The package does not currently support supersonic cascades (cascades where both inlet and exit are supersonic).

Objective Functions

The objective function is defined by specifying a variable, type and scale. The objective function is fetched from the results dictionary that stores all calculated variables, and the variable name must be on the form key.column, where key is a key in results, and column is the column header in the DataFrame contained in results[key]. With overall as key, you can access the most common objective functions such as:

• efficiency_ts

• efficiency_tt

• power

The scale is simply a nonzero float or integer, and is used to scale the objective function. This gives the user control to tune the behaviour of the optimization algorithm. The type must be either maximize or minimize. If the type is maxmimize, the sign of the scale changes to ensure that the objective function is in fact maximized.

Constraints

Each constraint is defined by specifying variable name, value, type and normalize. Similar as for objective function, the constraints are fetched from the results dictionary that stores all calculated variables, and the variable name must be on the form key.column, where key is a key in results, and column is the column header in the DataFrame contained in results[key]. Here are some example of valid keys:

• overall to access variables such as mass flow rate (overall.mass_flow_rate)

• geometry to access geometrical variables such as flaring angle (geometry.flaring_angle).

• plane to access variables such as relative mach number (plane.Ma_rel)

• additional_constraints to access additional variables such as interspace area ratio (additional_constraints.interspace_area_ratio)

Type specifies if the constraint is an equaility or inequality constraint, and must be either <, > or =. The constraints are defined such that the parameter must be < or > than the given value. Value specifies the value that the variables should be equal to, less than or greater then, and normalize specifies if the constraint should be normalized with the value or not.

Radius Types

These are the valid radius types:

• constant_hub: Generates turbine with constant hub radius.

• constant_mean: Generates turbine with constant mean radius.

• constant_tip: Generates turbine with constant tip radius.