Engineering

intelligentintelligentsystems

fortherealworld.

In medical glove production, surface inspection often depends on manual visual checking, where fatigue and inconsistency can lead to defect escapes and rejected batches. itZorix designed an in-line AI vision system that inspects every glove and automates reject decisions in real time. Industrial cameras in controlled-lighting enclosures cover the full glove surface, while deep-learning inference runs on NVIDIA Jetson edge units at the line and triggers automated rejection in synconized with conveyor timing detecting stripping damage, bead damage, finger damage, holes, scum, oil marks, dirt and other surface defects in real time. Every event is logged to a live dashboard and data historian, where predictive analytics correlate defect trends against upstream process parameters flagging root causes before they become rejected batches.

itZorix is developing an intelligent solar power forecasting solution for an industrial client’s Plantation solar facility. The project combines plant telemetry, numerical weather predictions, satellite data and solar-position parameters to address the challenges created by Sri Lanka’s rapidly changing tropical weather conditions. The solution is designed to deliver reliable intraday and day-ahead forecasts of solar irradiance and power generation. It includes data-quality monitoring, statistical and machine-learning models, and performance evaluation using metrics such as MAE, RMSE, MAPE and bias. Once implemented, the system will support improved energy scheduling, better management of generation variability and more informed operational decision-making. The project is also establishing a scalable forecasting framework for future solar-energy applications across Sri Lanka.

itZorix delivers an end-to-end Real-Time Data Acquisition & Control solution for the Solar PV plant, enabling reliable real-time monitoring, data acquisition, and remote control. The solution provides live visibility into plant performance, including power generation, energy, electrical parameters, set-points, and alarms. It enables Modbus controlled , remote active-power control through curtailment signals, allowing power output limits to be securely issued to the inverters and applied within seconds, with command acknowledgement and status verification.