Skip to content

Python program to generate SMV language for Signal Interpreted Petri Net (SIPN) model checking

Notifications You must be signed in to change notification settings

Faaizz/smv_generator

Repository files navigation

unit_test_workflow

SMV and ST Generator

Python program to generate SMV language for model checking and IEC 61131 Structured Text (ST) PLC code from JSON representation of Signal Interpreted Petri Nets (SIPN).

How It Works

The program takes a json specification of the SIPN as input, and produces a symbolic model checker in the NuSMV language. For the program to function properly, the input json must follow some laid out syntax which is outlined below.

The program translates the json input into NuSMV code by performing the following activities:

  • Declare all places as variables of type "boolean"
  • Declare all inputs as variables of the specified types
  • Define transition firing conditions based on specified pre-places, inputs combinations (which may be specified as a raw string), and post-places for each transition
  • Define variable stab to check stability of markings
  • Define the set and reset conditions for each boolean output based on information obtained from the places
  • Define the present value of each output at every stable marking
  • Assign initial values to inputs and construct how they obtain next values accordingly (automatically for boolean inputs, but relies on provided inofrmation for inputs of other types)
  • Assign initial values to places accordingly and construct how they obtain next values based on the firing of transitions

The program translates the json input into ST code by performing the following activities:

  • Declare all places as variables of type "BOOL" and assign initial values based on the specified initial markings
  • Declare all inputs and outputs as variables of the specified types and PLC addresses
  • Define transition firing conditions based on specified pre-places, inputs combinations (which may be specified as a raw string), and post-places for each transition
  • Fire transitions in a loop in order to set outputs only at stable markings
  • Set the value of each output at every stable marking

How to Run

You can generate NuSMV for model checking and ST PLC code by running:

python ./launch.py /path/to/json/file

Note that the output files are stored in the parent directory of the specified JSON file.

JSON Syntax

It should be noted that terms like inputs, places, transitions, and outputs in this section are used to represent the names of such entities.

1- Places

The root json object must have a places attribute, which holds an array. This places array in turn contains an attribute to represent each place in the SIPN. Subsequently, each SIPN place attribute is an array with the following elements:

  1. Element 0: An array with 2 elements. The first being an array of boolean outputs that should be set at the place, and the second an array of boolean outputs that should be reset at such place. Required
  2. Element 1: An array of manually specified output assignments for non-boolean outputs. Optional
  3. Element 2: A comment string. Optional

In addition, there's an initial attribute in the places object which holds an array of places that should be marked initially. Each place array can contain an optional 3rd element which holds comments about the place, this element is not used in processing.

Illustration:

{
    "places": {
        "PLACE1": [
            [
                ["AL1_GRAB"],
                ["RP_AL1_CLAMP"],
            ],
            ["AL1_X_SET=80"],
            "Textual description of the significance of this place"
        ],
        "PLACE2":[
            [
                ["RP_AL1_CLAMP"],
                ["AL1_GRAB"]
            ],
            "Comment text."            
        ],
        "initial": [ "P1", "P3" ]
    }
}

In the above example, PLACE1 when marked, sets AL1_GRAB as TRUE, RP_AL1_CLAMP as FALSE, and AL1_X_SET as 80. While when PLACE2 is marked, RP_AL1_CLAMP is set TRUE and AL1_GRAB FALSE.

2- Inputs

The inputs into the SIPN are registered in the inputs attribute of the root json object.

Illustration:

{
    "inputs": {

        "S_AL1_ST1": ["boolean", "%IX100.0"],
        "S_AL1_ST2": ["boolean", "%IX100.1"],
        "S_AL1_ST3": ["boolean", "%IX100.2"],
        "RP_AL1_ST_CLAMPED": ["boolean", "%IX100.3"],
        "AL1_ST_DETECTED": ["boolean", "%IX100.4"],
        "RESET": ["boolean", "%IX100.5"],
        "AL1_ST_X_POS": [
            [
                "{0, 400, 860}",
                0,
                "{0, 400, 860}"
            ],
            "%IW100"
        ],
        "AL1_ST_Y_POS": [
            [
                "{0, 300, 560}",
                0,
                "{0, 300, 560}"
            ],
            "%IW101"
        ],
        "AL1_ST_Z_POS": [
            [
                "{0, 200, 410, 750}",
                0,
                "{0, 200, 410, 750}"
            ],
            "%IW102"
        ]
        
    }

3- Transitions

A transitions attribute must be present in the root json object. This attribute holds an object which contains named attributes representing each transition.

Each such attribute should be an array with at least 3 elements:

  • Element 0: An array of places that must be marked for the transition to fire;

  • Element 1: An array of arrays indicating the required transition conditions. The transition conditions may be specified as raw text by specifying a string in the array. Otherwise, it can be specified with as a combination of inputs that should be HIGH and inputs that should be LOW. The first element is an array of inputs that should be set for the transition to fire, while the second is an array of inputs that should be reset for the transition to fire; and

  • Element 2: An array of places that must be left unmarked for the transition to fire.

For example, if transition T1 has pre-places P1 and P2, requires inputs I1 and I2 to be set and I4 not set or input I3 to be set and I4 not set, and has post-place P3, we have in NuSMV:

Then the entry for transitions in the transitions object should look like:

{
    "transitions": {
        "TS3": [
            ["PS2"],
            [
                [
                    ["RP_AL1_ST_CLAMPED"], []
                ]
            ],
            ["PS3"]
        ],
        "TS4": [
            ["PS3"],
            [
                [
                    "(AL1_ST_Y_POS>540) & (AL1_ST_Y_POS<580)"
                ]
            ],
            ["PS4"]
        ]
    }
}

About

Python program to generate SMV language for Signal Interpreted Petri Net (SIPN) model checking

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published