Early detection of agglomeration in a polyethylene fluidized bed at high temperature and pressure by vibration signature analysis

Analysis of vibration signatures was used as a reliable non-intrusive technique for detecting the hydrodynamic state of a polyethylene fluidized bed for early detection of agglomeration. Experiments were carried out in a pilot scale gas-solid fluidized bed at 15-25 bar and 100 °C. The bed was operating in the bubbling regime of fluidization. Accelerometers were used to record wall vibration generated by flow of particles and bubbles through the bed at various operating pressures and initial particle sizes. Time and frequency domains and state space methods were applied to analyze the experimental data. These methods were able to detect the agglomeration only when the bed has become defluidized. Analyzing the experimental data by the S-test showed that changes in the particle size due to agglomeration can be detected at least 4 °C before defluidization of the bed. This study confirmed that using vibration signatures and applying S-test to the vibration data is a satisfactory approach for early detection of agglomeration in fluidized beds.