Reliability of magnetic particle inspection 3

[smurali1]

I have a cold forging supplier whose parts were detected of having cracks. The parts are cold extruded, machined, carburise hardened and supplied. Right now root cause analysis is going on to find out where crack happens - during extrusion or at HT. To run production, we have asked supplier to do 100% Magnetic Particle Inspection of parts. My question is
1. How reliable is MPI?What is the min. width and depth of the crack it can detect?
2. What is the Repeatability & Reproducibility of MPI checking? i.e. will once accepted part be rejected when inspected again by different person/equipment?

Thank you in advance for the clarifications.

 

[swall]

In a production environment, you can expect MPI to find flaws .15" long and .0002" wide as a minimum.But ultimately, MPI is a visual process--the operator has to see the indication.Visual inspection is never 100% reliable.The MPI inspector also has to be properly trained to have an understanding of the type and orientation of flaws the extrusion process can produce.And he has to know how to choose the proper current and direction of magnetization for the configuration of part he is inspecting.

 

[Rich2001]

1. Reliability depends on many factors -

Type of electrical current- AC, DC, Pulse DC, if AC the frequency (skin effect)

Direct methods - prod, head/tail stock etc. or indirect/induced methods yokes, coils

Dry or wet mganitic particle

Visible or fluorescent particles

2. Because MPI requires a lot of human attention, one technician may find something that the other did not.

 

[unclesyd]

In the hands of a trained and capable operator using the proper procedure MT (Magnetic Particle Inspection) is close to 100% effective in locating cracks breaking the surface. It is one of the more sensitive NDT procedures. As to the width and depth we have located cracks in tool steel that are not discernible at 300X. Any surface break/crack will create a magnetic field that is detectable with the proper procedure.

I have to qualify these statements by saying it takes a good testing procedure and a qualified technician to get the maximum benefit from MT testing, in fact any NDT testing. You should have a eminently qualified person, a ASNT level III MT or equivalent, setup the inspection procedure with the proper controls and the testing be accomplished by a qualified technician, an ASNT level II or equivalent.

The procedure will be setup according to what type defect and the location that you are looking for. If the defect is known to occur in the same location on each part this makes for a simple inspection. If the defects have random locations this brings the inspection procedure and technician into play big time.

I have to say again that having a qualified person oversee the inspection process will pay big dividends both for you and the part supplier.

 

[MetalMan14]

If you are worried about the testing company, may I recommend Non-Destructive Testing Group headquartered out of Caledonia, Michigan. I worked for them as an inspector and they are one of the leading NDT companies in the states. They do a lot of work for Tier 1 companies. You could also try liquid penetrate inspection.

 

[yates]

Unclesyd is right, MPI is extremely sensitive providing the correct technique is chosen in function of part shape and size. The repeatability of the inspection results will also depend heavily upon the same technique being used each time. This is why a level III should set up this technique, and preferably a level III with experience of extrusions since a very important part of establishining a technique is knowing what we are looking for and where to look on the part. In the aerospace field, both FPI (fluorescent) and MPI are used for crack detection, depending upon whether raw material is ferro-magnetic or not. Wherever both are possible (magnetic steels) MPI is preferred. It has the advantage of revealing defects just under the surface because of the resulting interruption of the magnetic field, whereas cracks which have been plated over, peened over (low quality sand-blasting, are excessively dirty and difficult to clean, will often be missed by FPI, which relies upon capillarity (i.e. crack must be open to surface).