Page 316 - Handbook of Materials Failure Analysis
P. 316

1 The Wear Particle Analysis Techniques    313




                  process of distinguishing particles, less expertise is required in identifying different
                  types of wear particles on the ferrogram. A trained technician can conduct particle
                  identification. Furthermore, ferrogram can be heated or chemically treated for fur-
                  ther particle examination.
                     The main disadvantage of analytical ferrography is the poor capability in detect-
                  ing the nonferrous particles. Low efficiency in sample preparation is an additional
                  drawback. Due to these disadvantages, ferrogram lacks its popularity in today’s con-
                  dition monitoring. On the other hand, filtergram has some advantages which can
                  compensate for the disadvantages of ferrogram.
                     Filtergram: Also known as the patch test in the industry, filtergram is actually an
                  old technique. Until the 1990s, filtergram has not been popular in wear particle anal-
                  ysis. Rather than using a magnetic field to deposit particles, filtergram utilizes filter
                  membranes to clog particles, similar to fishing with a net. The filter membrane with
                  the specifications of 1-μm pore size and 47 mm diameter is generally applied in wear
                  particle analysis. Like ferrogram, the particles on filtergram are distinguished under
                  optical microscope by trained technicians. The particles on filtergram can be further
                  studied under SEM/EDS (scanning electronic microscope/energy dispersive spec-
                  trometry). In comparison with ferrogram, filtergram has the following advantages
                  and disadvantages [3]:
                  The advantages in filtergram are:
                  1. Filtergram clogs all the particles which are larger than the membrane pore size.
                     Hence, either ferrous or nonferrous particles can be all deposited. As a
                     quantitative method, filtergram had been applied in ISO 4407:1991 (the
                     microscopic particle count).
                  2. Insamplepreparation,filtergramhas muchhigherefficiency thanthatofferrogram.
                  3. Particle pileup rarely occurs in filtergrams.

                  The disadvantages in filtergram are:
                  1. Particle distinguishing requires the expertise. The examination is also time-
                     consuming.
                  2. For corrosive wear particles, it is difficult to be distinguished. However, such
                     particles are effortlessly spotted via ferrogram, as corrosive particles are piled up
                     at the exit end of the ferrogram.
                  3. As the filter membrane is opaque, bichromatic microscopes cannot be applied in
                     the filtergram.

                  Computer Image Analysis: Due to the time-consuming nature of manual microscopic
                  examination, there is a keen desire to handover the task to computers. In last two
                  decades, some efforts have been put in wear particle image analysis [4–12]. The
                  research on particle images aimed to develop an instrument with artificial intelligence.
                  The emergence of the first commercial instrument of computer image analysis in 1998,
                  named LNF (LaserNet Fines), is a milestone in the history of wear particle analysis.
                     LNF applies laser pulse to count the particles. At the same time, the silhouette
                  images of all particles larger than 20 μm are classified with an artificial neural
   311   312   313   314   315   316   317   318   319   320   321