Writing a JSON configuration file¶

This guide will help you write a JSON configuration file to define FMUs, their properties, the connections between FMUs and connections with external data.

Structure of the configuration file¶

The configuration file consists of three main sections:

FMUs (fmus): defines individual FMUs and their properties (id, path, step size, supplier, etc.)
Connections (connections): describes the connection graph.
Global settings: specifies global configuration options.

1. Defining FMUs¶

The fmus section contains the list of FMUs. Each FMU is represented by a dictionary with the following attributes:

id: a unique identifier for the FMU (e.g., "FMU1" or "rotor").
path: the file path of the FMU (see below).
stepsize: the step size for the FMU simulation (e.g., "0.3 sec").
name (optional): a human-readable name for the FMU (e.g., "load").
initialization (optional): key-value pairs to initialize variables in the FMU. An initialization value could be overridden if a connection is defined for this variable.
supplier (optional): the supplier or creator of the FMU.

Here is an example of an FMU description:

{
    "id": "load",
    "path": "load.fmu",
    "stepsize": "0.2 sec",
    "name": "variable load",
    "initialization": {"V_in": 800, "R": 2.5},
    "supplier": "My company"
}

Paths to FMU files¶

Paths to FMU files can be specified as:

Relative paths. The configuration parser will search for files using path value, by considering the following:
Files are placed at the same level than the configuration JSON file or in nested folders.
Each FMU file at the previously mentioned location has a unique name.
Absolute paths. In this case, the path field will remain untouched and the FMUs will be loaded from the specified locations.

2. Defining Connections¶

The connections section is a list of connections. A connection links a source to a target. The generic JSON structure for a connection is:

{
    "source": {...},
    "target": {...}
}

Three types of connections can be defined in the JSON file:

a connection between two FMUS
a connection from a data source to an FMU
a connection from an FMU to a data "sink"

Each connection type is described in the following subsections.

Between two FMUs¶

A connection between two FMUs represents the interaction between the subsystems in the co-simulation. The connection links the output variable of an FMU to the input variable of another FMU. A source/target for an FMU is defined with:

id: the FMU id as defined in the fmus section.
variable: the variable name in the FMU
unit: (optional) the unit of the variable

Example¶

For a connection from variable I_out of FMU battery to variable Iload_in of FMU load, the JSON description is:

{
    "source": {"id": "battery", "variable": "I1_out", "unit": "A"},
    "target": {"id": "load", "variable": "Iload_in"}
}

Note

A connection source/target has a type which is fmu by default. It is possible to add this field to be more explicit in the JSON: "type": "fmu".

From a data source to an FMU¶

A connection from an external data source to an FMU is required when one input variable of a FMU is controlled by external data. The data source will be used as input values for the FMU. In CoFMPy, we provide multiple external data sources: from single values, from a CSV file, or from a Kafka stream. You can also create your own data source (see documentation).

The information required varies depending on the data source. For each type of source, a field type is required. Note that for all types, the field unit is optional.

From single values (literal)¶

type: literal
values: a key-value dictionary with the values of the variable at different timestamps, e.g. {"0.1": 12.2, "0.5": 8.6}. Note that keys in JSON must be strings, timestamps "0.1" and "0.5" above are then represented as strings.
interpolation: (optional) interpolation method (see interpolation reference). Defaults to previous.

Example¶

The variable R of a FMU resistor is defined as "literal". The corresponding JSON description is:

{
    "source": {
        "type": "literal",
        "values": {"0": 100, "12.5": 250},
        "interpolation": "linear"
    },
    "target": {"id": "resistor", "variable": "R", "unit": "Ohm"}
}

From a CSV file¶

type: csv
path: the path to the CSV file
variable: the variable name to read in the CSV. It must correspond to the column header name.
interpolation: (optional) interpolation method (see interpolation reference). Defaults to previous.

If you want to know how CSV files must be formatted to be correctly read by CoFMPy, refer to the section on CSV data source.

Example¶

The variable V of FMU battery is controlled according to the values in the CSV file path/to/battery_data.csv (column called voltage). The interpolation is set to previous. The corresponding JSON connection is:

{
    "source": {
        "type": "csv",
        "path": "path/to/battery_data.csv",
        "variable": "voltage",
        "unit": "V",
        "interpolation": "previous"
    },
    "target": {"id": "battery", "variable": "V", "unit": "V"}
}

From a Kafka stream¶

type: kafka
uri: the server URL (with port), e.g. 127.0.0.1:9092
group_id: the consumer group id.
topic: the topic to listen to.
variable: the variable name to read in the Kafka topic. It must correspond to the variable name used in the Kafka message.
interpolation: (optional) interpolation method (see interpolation reference). Defaults to previous.
timeout: (optional) delay, in seconds, to wait (blocking) after reception of the first message before proceeding with the rest of the cosimulation. Defaults to 2.

If you want to know how Kafka messages must be formatted to be correctly read by CoFMPy, refer to the section on Kafka data source.

Example¶

The variable V of FMU battery is controlled by the values from a Kafka server. A single message is formatted as follows: {"t":, "":, "":, "":, }. The corresponding JSON connection is:

{
    "source": {
        "type": "kafka",
        "uri": "172.17.0.1:5000",
        "group_id": "my_group",
        "topic": "cofmpy_topic",
        "variable": "voltage",
        "interpolation": "linear",
        "timeout": 1,

    },
    "target": {"id": "battery", "variable": "V", "unit": "V"}
}

From an FMU to a data sink¶

A connection from an FMU to a data "sink" represents a connection towards the "outside". A sink can be a data storage system (e.g. a file or database system) or an outbound data stream (e.g. a Kafka data stream). The JSON description is equivalent to the previous cases.

By default, CoFMPy will save all output variables of the co-simulation in a single CSV file for each simulation step. The location of this file is storage/results.csv.

To a CSV file¶

type: csv
output_dir: the directory where the CSV file will be created.
variable: the name of the variable, used as the CSV filename and the header column name in the CSV file.
overwrite: (optional) overwrite an existing file if true. Defaults to true.
unit: (optional) the unit of the variable.

Example¶

{
    "source": {"id": "FMU1", "variable": "V_out", "unit": "V"},
    "target": {
        "type": "csv",
        "id": "csv_out_V",
        "output_dir": "path/to/folder/",
        "variable": "FMU1.V",
        "overwrite": true
    }
}

To a Kafka stream¶

id: an ID to name the connection
uri: the Kafka server URL (with the port number)
variable: the name of the variable, used to send messages on the Kafka topic.
group_id: the consumer group id.
topic: the topic to listen to.
unit: (optional) the unit of the variable.

Example¶

A JSON example for Kafka data stream:

{
    "source": {"id": "FMU2", "variable": "I_out", "unit": "A"},
    "target": {
        "type": "kafka",
        "id": "kafka_out_1",
        "uri": "172.17.0.1:5050",
        "variable": "FMU2.I",
        "config": {"group_id": "my_group", "topic": "FMU2.I"},
        "unit": "A"
    }
}

3. Global settings¶

Besides the description of the FMUs and the connections, the JSON configuration file also specifies global settings used by CoFMPy.

cosim_method: specifies the algorithm for solving system loops (e.g., jacobi, gauss-seidel).
edge_sep: defines the separator for connection naming. Default is ->.
root: specifies the root directory for relative paths.

Tips for writing a configuration file¶

Use unique IDs: ensure all FMUs and data sources have unique IDs.
Verify paths: check that all path attributes are valid and accessible.
Specify units: use consistent units across all variables to avoid mismatches.
Test connections: verify connections are logical and complete.
Validate JSON: Use a JSON validator to ensure the configuration file is error-free.

This guide should help users create and manage their JSON configuration files effectively. A full JSON example is provided in the "Examples" section.

Interpolation reference¶

Interpolation is used when the system needs a value between two time points when data external to FMUs is used (literal, csv, kafka, etc.).

The interpolation uses cofmpy.utils.Interpolator class which supports a wide variety of interpolation methods. The currently tested methods are: - previous, which returns the value at the previous time point (default value) - linear, which calculates a linearly interpolated value between the previous and next points.

The other methods will log the following warning: Method '{method}' is in beta version. We recommend using 'linear' or 'previous'

For more details, please refer to: ./notebooks/interpolator.ipynb