Monitoring Process Upsets in Production
Heat Treating and Other
Thermal Processing Applications

This month's NDT Solution should be of interest to people involved in material processing, nondestructive evaluation, and quality control. The authors have demonstrated the applicability of a pulsed eddy current method to noninvasive process monitoring. They present two case studies involving components treated by heat treating and thermal plasma coating processes. The unique advantages of using the electromagnetic properties of the component for process monitoring are discussed. G.P. Singh Associate Technical Editor

 

Figures 1-3
Figures 4-6

Introduction
When metals undergo thermal processing such as in heat treating or plasma coating operations, the basic metallurgical structure of the untreated metal is altered as a result of the processing. For many such operations, the ultimate quality parameter (hardness, case depth, coating thickness, etc.) may be nondestructively monitored using conventional single or multifrequency eddy current methods. In typical applications, numerous items are produced under varying process conditions and subsequently tested both destructively and nondestructively to establish a correlation between the two measurements which assure acceptable quality parameters in the shortest practical time.

---

Pulsed eddy current technology can monitor changes, or upsets, in the heat treating or thermal forming process.

---

Commonly used destructive tests are mechanical hardness measurements including the Rockwell, Brinnel, and Vickers standards. The measurements are localized on one small area of the test item and provide information only on the surface hardness but no information of the condition throughout the specified volume to be treated. Consequently, items that have been treated in a malfunctioning process may show an acceptable mechanical hardness measurement on the surface but may not conform to the desired overall mechanical specifications. It is well known in the industry that the electromagnetic properties of metals will be altered under thermal processing and the change in those properties is detectable in many (but not all) cases using eddy current testing. Since, however, conventional single frequency eddy current methods are also sensitive only to the near surface conditions, calibrating such instruments against the mechanical surface hardness measurements still fails to detect underlying changes associated with process upsets that do not alter the surface hardness but do alter the subsurface condition. For example; the authors have found that in induction heat treating operations (including postheat oil quenching), it is possible to reduce the quench time to far below process specification without significantly altering the mechanical (Rockwell) surface hardness measurement, although the subsurface condition may not be to specification.

Advanced state of the art eddy current methods, including pulsed eddy current, provide for eddy current depth of penetration far in excess of that obtainable with conventional instruments and field excitation with a broad eddy current frequency spectrum from DC to hundreds of kilohertz. This yields a capability to detect small changes in the electromagnetic properties of an item both on and below the surface to depths specified in a variety of heat treating or thermal forming operations.

 

Case I — Automotive Engine Valves
Automotive engine valves are designed for operation by opening and closing under the action of a valve lifter which pushes against the tip of the valve stem. Consequently, the tips must be hardened to reduce wear. In the production operation considered here, the valve tips are processed by induction hardening followed by quenching in a continuous process at rates up to 2000 valves/h. Random samples are pulled from the production line and destructively tested for surface hardness and, in some cases, hardness profile through the specified depth of hardening. The total heat applied to each part in the induction hardening is controlled by the voltage applied to the induction coils. Quenching is done in a continuous flow process through three stages, each with a duration of approximately five seconds, for a total quench time of approximately 15 s. Variations in applied heat and total quench time affect the quality of the end product in terms of surface hardness and hardness profile.

It has been found that significant variations in either the total applied heat or quench time do not yield a significant change in the measured Rockwell C hardness surface. In conjunction with testing performed to establish an online eddy current instrument for 100 percent inspection of all parts continuously on line, tests were also performed to compare the Rockwell C hardness measurement sensitivity to process variables to that of the pulsed eddy current instrument. Initially, 31 samples were prepared with varying applied heat sufficient to yield an unacceptable degradation in surface hardness. The Rockwell C hardness on all samples were measured along with the pulsed eddy current response from an eddy current instrument to establish the correlation of the pulsed eddy current response to Rockwell C hardness measurements. Figure 1 shows the typical valve and the pencil type pulsed eddy
current probe used to measure the pulsed eddy current response at the valve tips. The valve stem and tip diameter is 0.34 in. (8.4 mm) and the probe sensor diameter is 0.25 in. (8 mm). Figure 2 shows the results of both sets of measurements and the linear regression line fitting the data with a correlation coefficient of 0.82 That correlation is adequate for production monitoring under that limited range of process and hardness variables.

Subsequent tests were performed in which the two process parameters were varied to establish the sensitivity of both measurements to those process variables. Groups of 10 samples were processed under normal quench time (15 s) but with varying total applied heat from 100 percent to 43 percent of normal and with normal applied heat of 100 percent and two reductions in quench time. The reduction in quench times were accomplished by shutting down the first and then the first and second stages of the quench stages. The results are shown below in Figure 3.

For the conditions of normal quench and variable heat, the pulsed eddy current response tracks the Rockwell C hardness measurement consistent with the earlier data (Figure 2). For the case of fixed, normal heat but with variable quench time, the Rockwell C hardness measurements show little or no change with respect to changing quench time. The pulsed eddy current response, however, shows a very significant change with respect to changing quench time. Clearly, using the Rockwell C hardness measurement as the indicator of part quality in terms of having been processed to specification would fail to sort on the basis of that quality parameter. Setting an accept window or gate between +50 to +60 on the pulsed eddy current output as indicated on the graph of Figure 3 would assure rejecting parts on the basis of low hardness due to low heat and inadequate quenching.

 

Case II — Overhead Valve Cam Tappets
The valve operating tappets in overhead cam engines are composites structures consisting of an aluminum bucket that has an outer steel coating applied by thermal plasma spray. The aluminum buckets are forged to shape and then heat treated in a solution aging process prior to the application of the steel coating. The overall finished tappet diameter is approximately 33 mm (1.3 in.). The cylindrical walls of the aluminum bucket and the thickness of the steel coating are typically 0.78 mm (0.03 in.) and 0.05 mm (0.02 in.), respectively. Figure 4 shows typical tap pulsed eddy currents along with a cross sectioned unit illustrating the construction. The probe is a differential type with two cylindrical tappet positioning disks for testing individual units against a known reference part fixed in one position. The tappets are subjected to an electromagnetic pulsed field parallel to their axes. The pulsed eddy current response is sensed by two receiving coils directly under the positioning disks.

Parts are initially heat treated in batches of a few thousand and confirmed to be appropriately hardened by mechanical hardness measurements on random samples selected from the batch. Following the heat treating, the tappets are plasma spray coated with steel, one at a time, in a continuous process. During setup of that process, it is necessary to establish the proper heat settings which will assure that the hardness of the aluminum shell will be not altered to below minimal acceptable values. This setup involves testing one or two parts by sectioning, polishing, and performing microhardness tests at the lower portion of the aluminum shell. This is a time consuming and costly process. A rapid, nondestructive method of confirming the hardness was sought as an alternative. The pulsed eddy current method was tested and shown to be effective for the purpose.

Three batches of samples were prepared with normal, low, and high heat settings in the plasma spray. Six from each group were destructively tested and the Vickers hardness measured as usual. Another six from each group were tested with pulsed eddy current instrument unit programmed to yield a maximum spread in output over the three batches. Although there is correlation between the measured hardness and the pulsed eddy current output, the hardness measurements show to be much less sensitive to the change in process heat than does the pulsed eddy current measurement.

In Figures 5 and 6, the hardness and the pulsed eddy current outputs are plotted for each of the samples in sequence to show the sensitivities to the change in the heat. Figure 5 shows the hardness, Figure 6, the pulsed eddy current output. Clearly, the pulsed eddy current output varies directly with the heat but the hardness measurements are less sensitive to the change from low heat to normal heat. Setting an accept window or gate between 4 and 7 V on the pulsed eddy current output would assure rejecting parts on the basis of either case of too much or too little heat, in other words, in or out of process specification.

 

Conclusions
Mechanical hardness measurements are not sufficient to assure that metal items subjected to thermal processing designed to yield the end quality in terms of specified processing parameters. Those measurements provide data at highly localized points only on the surface of treated parts and, as such, may not indicate substandard quality in terms of the thermal processing specifications. Modern pulsed eddy current technology provides the capability to monitor not only surface hardness but also changes, or upsets, in the heat treating or thermal forming process.

 

* Intex, Inc., 1419 Forest Drive, Ste. 205, Annapolis, MD 21403; (301) 322-5866; e-mail roy@intexinc.net.

† TRW, Inc., Automotive Valve Div., 1455 East 185th St., Cleveland, OH 44110; (216) 692-4809; e-mail dino.marino@trw.cmd.

 

Copyright © 2000 by the American Society for Nondestructive Testing, Inc. All rights reserved.

[ Materials Evaluation ]