Online pattern recognition: distillation towers' jet flooding detection
Aiassa, Roberto E.
Online pattern recognition: distillation towers' jet flooding detection - 6 p.
Transcripción del resumen del autor. Originally this application was developed (December'00) in order to detect an abnormal behavior on highly cycling variables (temperatures) on the Delayed Coker unit analyzing firs and second derivatives, inflection, maximum, minimums, and their sequence. The very same algorithm has been used for flooding detection on our Cat Cracker's Light Ends deethanizer tower. Due to lack of appropriate standard measurements (delta pressure and temperature), the algorithm was used to detect singularities on available thermocouples readings. Internal vapor and reflux traffic usually begins to "chatter", or get noisy prior to the onset of jet flooding affecting thermocouples located below the flooding zone. In order to recognize the abnormal behavior, the algorithm just identifies when any temperature's second derivative is beyond its limits, triggering an alarm to the console operator some 45-60 minutes before jet flooding's classical symptoms become evident.
Online pattern recognition: distillation towers' jet flooding detection - 6 p.
Transcripción del resumen del autor. Originally this application was developed (December'00) in order to detect an abnormal behavior on highly cycling variables (temperatures) on the Delayed Coker unit analyzing firs and second derivatives, inflection, maximum, minimums, and their sequence. The very same algorithm has been used for flooding detection on our Cat Cracker's Light Ends deethanizer tower. Due to lack of appropriate standard measurements (delta pressure and temperature), the algorithm was used to detect singularities on available thermocouples readings. Internal vapor and reflux traffic usually begins to "chatter", or get noisy prior to the onset of jet flooding affecting thermocouples located below the flooding zone. In order to recognize the abnormal behavior, the algorithm just identifies when any temperature's second derivative is beyond its limits, triggering an alarm to the console operator some 45-60 minutes before jet flooding's classical symptoms become evident.
