ARDI ModelHost
ModelHost is a small Python library that allows you to create and host models of your system(s) that can be run, modified and tested on-the-fly.
In this context, a model is a system that takes one or more input values, and produces one or more output values - they often calculate the physical limits on machines or perform analytics and classification on your data.
ModelHost helps deploy these models, allowing you to…
- Easily connect models to online, live data.
- Connect models to one-another, feeding the inputs of one model with the outputs of another.
- Avoid server load and violating license limits by hosing multiple models in a single application,
- Answer complex questions using optimisation algorithms and goal-seeking,
- Avoid server load by only updating values when changes occur,
- Distribute your output data
The Modular Output System allows you to push your data to your choice of systems - from log files to full OPC-UA servers.
ModelHost also includes a framework for building math-based models. It allows much more complex equations than what is available using the ARDI Calculation driver - you can include custom logic, perform experiments on your models and perform tasks that are difficult or impossible to perform in historians or PLCs, such as goal seeking.