Digital FIR filter design for diagnosing problems in gears and bearings using Xilinx's system generator

Abstract

This paper discusses a common problem in industry to detect machine failures, especially in gears and bearings through vibrational analysis. Vibrational analysis is a powerful method for analyzing any machine in the industry, but this method uses expensive signal conditioning/processing devices and significant expertise for the analysis of signals in the frequency domain. In this context, this study proposes a new system of failure detection in machine can be done with digital Finite Impulse Response (FIR) filters. The systems of data Adquisition are compose with the test machine, accelerometer (model-353B17) and PCI-4474 Data Acquisition Card. The data from test machine are evaluated in both optimal and failure conditions. To evaluate the FIR filters system, a system was designed with the Xilinx System Generator. The structure design had four filters for evaluating gears and bearings separately. In failure condition the filters 1, 2, 3 detected failures in their amplitudes, but for bearings just the filter 4 detected a failure. To validate the system it was implemented in a Field Programmable Gate Array (FPGA) Spartan 3E where the design optimizes hardware resources and has the same responses than Xilinx System Generator. The response of the filters was evaluated with the ISO 10816-3 standard.