Digital Twins

digital twins is a litmus edge (le) feature that lets you create virtual representations of assets (machines, devices, or production lines) and combine static metadata with real time data from your shop floor into a single, structured payload use digital twins to monitor asset performance, feed contextualized data to northbound systems, and optimize operations without relying on those systems to understand how raw data is structured note digital twins requires a scale or growth litmus license how digital twins works digital twins occupies the middle of a five layer stack plcs, sensors, and machines at the bottom layer generate tag and i/o data le collects, normalizes, preprocesses, and filters the data at the edge it also enriches the data with metadata and asset information before passing the structured data to higher layers for processing and analysis digital twins is part of litmus edge and combines that structured data with static metadata to produce a complete digital representation of each physical asset the assembled twin data (including metadata and context) feeds the layers above analytics systems use it for real time kpis, thresholds, historical trending, what if simulations, and ml/ai pattern detection visualization and integration systems use it for dashboards, mes/erp connections, cloud exports, and cross site reuse within litmus edge, digital twins uses a two level hierarchy models and instances a model is a template it defines the structure of a digital twin the static attributes (metadata such as manufacturer, plant, and area) and dynamic attributes (real time data fields such as temperature, current, or pressure) models also hold transformation schemas and topic expression parameters that apply to all instances built from that model a physical asset can have more than one model applied to it for example, the same machine might have a model for energy monitoring, a separate model for maintenance tracking, and another for production output each model produces its own set of instances and topic based data streams an instance is a specific asset it inherits its structure from a model and lets you fill in the actual values the real device name, the specific nats topics that feed live data into each dynamic attribute, and any parameter overrides you can create many instances from a single model, for example, one model for a cnc machine type and one instance per physical machine on the floor when an instance runs, litmus edge assembles a structured json payload combining static metadata and live tag data, then publishes it on a nats topic at the interval you configure models you configure the following in a model static attributes fixed metadata fields such as device id , serial , and model you define them as placeholders in the model and set the real values in each instance dynamic attributes variable data fields such as temperature , vibration , and downtime each dynamic attribute is mapped to a live data topic in an instance transformations rules that convert raw json data into the required format before it enters the payload use json transformations to filter key value pairs or jsonata to apply conditional logic and generate computed outputs parameters named variables that you define in a model and set by instance they allow a single topic expression to resolve to a different topic for each instance this saves time when a model has many dynamic attributes and many instances schema a custom data structure that groups static and dynamic attributes into a named hierarchy for downstream systems that expect a specific json structure instances every instance is associated with a model the model's attribute configuration determines the data the instance collects each instance generates a nats topic and publishes its assembled payload to the internal message broker at the interval you configure when you configure an instance, you can set real values for static attributes map live data topics to dynamic attributes set parameter values for the instance view real time data in the data access tab publish attributes on individual topics publish flattened payloads for tabular databases workflow when you set up a model, follow the data flow because each phase depends on the one previous one for example, a dynamic attribute can only reference an existing transformation schema, and a topic expression can resolve only a previously defined parameter to set up a model configure static attributes for metadata such as manufacturer, plant, and area transform and map data by defining transformation rules, parameters for topic expressions, and dynamic attributes that reference those parameters define the output schema to provide downstream systems with the expected json structure an instance applies that structure to one physical asset it inherits the model's attributes and replaces placeholder values with real ones when you set up an instance, you choose the model to inherit from and interval for publishing to the message broker enter the asset's real static attribute values in place of the model's placeholders map each dynamic attribute to the topic that feeds it, or set the parameter values that resolve its topic expression check the result in the data access tab, which shows the assembled json payload after you set up an instance, start or stop it as needed if you update the model, manually upgrade the instance to apply the changes access digital twins ui to access digital twins log in to litmus edge in the navigation panel, select digital twins the models pane appears next steps create and configure a model docid\ zpadx1lmme0f44v7r6v6q configure transformations docid\ tbkp rlu3ocbdjcnjy4lq configure parameters docid\ tsncjwvos5h4bxmoujicg add an instance docid\ jzoth4t0hcrtrauusdryk use digital twins to contextualize data docid\ vdrp68hiwea7iylgo3ew7