Tools for ML models of thermohydraulics#

The help provides tools for preparing additional descriptions of ML thermophysical models for use by the Approximation model (TL) block when loading the model.

class repeat_onnx.ModelTH(nodes: list[NodeTH])#

The class is intended for additional description of machine learning models stored in the neutral ONNX format.

The class allows for a unique mapping of the input and output parameters of a machine learning model with the nodes of the Approximation model (TL) block of the Thermohydraulics library of the REPEAT platform.

The names of the input parameters for pressures and temperatures, and the output parameters for flow rates and enthalpy fluxes, are specified using the :py:class:NodeTH class of the model node.

Parameters:

nodeslist[NodeTH]: A set of nodes with names of pressures, temperatures, flow rates, and enthalpy flows.

Attributes:

flows: An attribute that allows you to get the names of the streams of all model nodes
potentials: An attribute that allows you to obtain the names of the potentials of all model nodes

Methods

to_onnx()

The method returns a string representation of the JSON structure for subsequent saving to a neutral ONNX format file.

Returns:

None

Examples

>>> from pprint import pp
>>> import repeat_onnx  as rx
>>> 
>>> 
>>> model_description = rx.ModelTH(
>>>   [
>>>    rx.NodeTH(
>>>      "Input unit", "pressure_1", "temperature_1", 
>>>      "flowrate_1", "enthalpy_flux_1"
>>>      ),
>>>    rx.NodeTH(
>>>      "Output unit", "pressure_2", "temperature_2", 
>>>      "flowrate_2", "enthalpy_flux_2"
>>>      ),
>>>    ]
>>>   )

property flows: list#

An attribute that allows you to get the names of the streams of all model nodes

Returns:

list: A set of stream names for all model nodes.

property potentials: list#

An attribute that allows you to obtain the names of the potentials of all model nodes

Returns:

list: A set of potential names for all model nodes.

to_onnx()#

The method returns a string representation of the JSON structure for subsequent saving to a neutral ONNX format file.

An additional description of the ML model should be located in the metadata_props structure under the "repeat_contract" key.

Returns:

str: JSON describing the model for the ONNX neutral format file.

Examples

>>> import onnx
>>> import repeat_onnx  as rx
>>> 
>>> 
>>> model_description = rx.ModelTH(
>>>   [
>>>    rx.NodeTH(
>>>      "Input unit", "pressure_1", "temperature_1", 
>>>      "flowrate_1", "enthalpy_flux_1"
>>>      ),
>>>    ]
>>>   )
>>> properties = onnx.StringStringEntryProto()
>>> properties.key = "repeat_contract"
>>> properties.value = model_description.to_onnx()
>>> model = onnx.load(path)
>>> model.metadata_props.append(properties)
>>> onnx.save(model, path)

class repeat_onnx.NodeTH(name: str, pressure: str, temperature: str, flowrate: str, enthalpy_flux: str)#

This class is designed to describe a node in a thermal-hydraulic model. The node is a mandatory component of the thermal-hydraulic model, which is described using the :py:class:ModelTH class.

The node class provides a one-to-one mapping in the model of pressure and temperature to the corresponding names of the input parameters of the user’s machine learning model, stored in the neutral ONNX format. The NodeTH class also provides an unambiguous comparison of the flow rate and enthalpy flux with the calculation results on the output parameters of the machine learning model.

Parameters:

namestr: Model node name. The node name will be displayed in the properties of the Approximation Model (TL) block in the Thermohydraulics library of the REPEAT platform.
pressurestr: The name of the pressure input parameter in the model node
temperaturestr: Name of the temperature input parameter in the model node
flowratestr: The name of the output parameter with the result of the mass flow calculation in the model node
enthalpy_fluxstr: The name of the output parameter containing the result of calculating the enthalpy flow in the model node

Attributes:

flows: An attribute that allows you to get the names of the node’s threads
potentials: An attribute that allows you to get the names of node potentials

Returns:

None

Examples

>>> from pprint import pp
>>> import repeat_onnx  as rx
>>> 
>>> 
>>> node = rx.NodeTH(
>>>   "Input unit", "pressure_1", "temperature_1", 
>>>   "flowrate_1", "enthalpy_flux_1"
>>>   )
>>> pp(node)
>>> # NodeP(
>>> #   name='Input unit',
>>> #   pressure='pressure_1',
>>> #   temperature='temperature_1',
>>> #   flowrate='flowrate_1',
>>> #   enthalpy_flux='enthalpy_flux_1'
>>> #   )

property flows: list#

An attribute that allows you to get the names of the node’s threads

Returns:

list: Node stream names.

property potentials: list#

An attribute that allows you to get the names of node potentials

Returns:

list: Names of node potentials.