VAI product tour

01 · THE PROBLEM

The data is there. Decisions arrive late.

DCS · HISTORIAN · LAST 24HNOT ANALYZEDANALYZED · VAINOWFORECASTLIVEARCHIVEWTH15 · FEED204.6t/hPROCESS GAS948Nm³/tBUSTLE GAS377°CSAMPLE VALUESKnowing quality before the lab resultProcess control without trial and errorHigh energy use and emissionsControl-system data that goes unusedNo coherent approach to data analysis

Your control system records thousands of process values every second.

01 / 080:00 left
IN THIS TOUR · 8 CHAPTERS

In this tour

From control-room data to a decision: connectivity, workflows, the three levels of AI, and DRI Expert as an example. About 5 min.

  1. 01 · THE PROBLEM

    The data is there. Decisions arrive late.

    Your control system records thousands of process values every second. Most of them stay in the archive and are never analyzed. The challenges we see again and again in industry. VAI predicts on that same data, so the decision comes before the lab result.

  2. 02 · PLATFORM

    VAI connects plant data to decisions and action

    VAI runs on a server in your plant and connects to the control system, databases and devices. Data is cleaned and synchronized in a workflow you can see and control. AI models run along the same path, on live data. Results reach the control room as predictions, recommended settings and alarms.

  3. 03 · CONNECTIVITY

    Direct connection to plant data

    Direct connection to the DCS and devices over the standard OPC UA and Modbus protocols. It also reads IoT devices over MQTT, the Historian (the control system's archive), databases and files. All sources in one environment, with scripts, workflows, schedules and dashboards. If the operator approves, the result is written back to the DCS.

  4. 04 · BUILD · RUN

    Scripts, workflows, schedules and apps in one place

    A Python scripting environment for model logic and data processing. A drag-and-drop workflow editor: connect blocks into a pipeline. Runs on a schedule, or whenever an event occurs. The app builder generates a user interface from a script's parameters, so results reach the operator.

  5. 05 · HOW VAKAV THINKS

    Three steps to a decision

    Prediction: product quality and process behavior, online, before the lab result. Recommendation: the optimal range for every parameter, on the control-room display, next to the operator. Optimization: modules are connected so the whole process reaches its optimum. Each level builds on the data and models of the one before and sharpens the decision.

  6. 06 · INSIDE YOUR PLANT

    Entirely inside your plant

    VAI is installed on a server in your plant. Data is read from the DCS, the Historian and devices inside the same network. No dependency on the internet or outside services, even on a network fully cut off from it. Results appear on the control-room display, with your organization's roles and access levels.

  7. 07 · EXAMPLE · DRI EXPERT

    Example: DRI Expert, built on VAI

    DRI Expert is built on VAI for the Midrex direct-reduction process and reads live process data. The metallization of the sponge iron is predicted before the lab result. Recommended ranges for feed rate, process gas flow and gas temperature, next to the operator. Actual and predicted metallization (MD) and carbon (C), side by side. All numbers shown are sample values.

  8. 08 · ENGAGEMENT

    From assessment to operation

    Needs assessment: understanding the process, the available data and its gaps, and setting a measurable goal. Design and development: modules for each unit, with data integrated end to end. Installation and testing: deployment on the plant server and testing under real conditions. Training and support: enabling your team and continuously improving the models.