StartLux Industrial: Conveyor Roller Fault Diagnosis

Which of three conveyor drives is abnormal, and what should be measured next? This StartLux Decision simulation turns diagnosis into a sequence of evidence-gathering choices: compare acquired data, request additional tests, select a hypothesis, isolate the drive and measure again. What the video shows - 00:00–around 00:14: The gantry probe visits the three drives to obtain speed and load data. M2 has higher torque and lower speed than its peers. - Around 00:14–00:19: The model selects vibration measurement on M2 and prioritizes it for investigation. Its measured vibration is still below this simulation's normal reference, so high load alone is not presented as a vibration fault. - Around 00:19–00:23: Speed-reduction and drive-off coast-down tests provide additional evidence for a bearing-drag hypothesis. - Around 00:23–00:27: The simulated PLC isolates and brakes M2. A separate verification measurement follows. The final frame reports consistent evidence, the selected roller stopped and isolated, ten accepted choices and seven evidence records. The probe positions itself and measures; it does not strike or repair the bearing. A diagnostic judgment updates the report, while isolation is executed by the equipment controller. Neither requires a new probe movement for every decision. From evidence to action The model receives completed measurements, public reference values and currently allowed actions. The injected fault and signals that have not yet been measured are excluded. The harness handles alignment, sampling gates, operating-condition changes and completion checks, then returns new evidence to the decision loop. In this recording, M2's load indicator is about six times the reference while its measured radial vibration is 0.144 mm RMS. These readings are not presented as a damage probability. Speed-reduction and coast-down tests help distinguish possible causes. A fresh post-isolation window then checks whether the drive actually stopped, instead of treating the old evidence as proof of a completed intervention. Value and scope This example illustrates active diagnosis, test selection and post-action verification. The supplied documentation says vibration is computed from simulated bearing-housing displacement, and faults use simplified damping, imbalance or torque-limit mechanisms. It is not a complete simulation of laser vibrometry or microscopic bearing damage. The demonstrated outcome is localization and isolation, not repair. This entry was prepared from the supplied 1080p silent recording and project documentation. The original interface text is Chinese. No simulation rerun, independent audit of raw inference traces or physical conveyor test was performed. Choice probabilities are not fault probabilities or an industrial safety certification.