·Table of Contents
·Methods and Instrumentation
Eddy Current Application for Non Destructive Test in Light Metal Industry
Head Laboratory of Metrologys
Mechanical Eng. Dept., Industrial Management & O.R. Section
National Technical University of Athens 15780 Zografos GREECE
Tel: +301 7723585, Fax: +301 7723571,
This paper presents the experience of an on going project, co-financed by the Greek General Secretariat of Research and Technology, aiming the installation of a non destructive quality control instrument and a capability study in a traditional production line of light metal industry (collapsible tubes). A specialised instrument for the non-destructive measurement of the thickness of the internal layer of the varnishing colour has been installed in the quality control laboratory (Fischer - Helmut Fischer GMBH+Co). The system measures the coating thickness of aluminium tubes, using Eddy current technique. The installation of the NDT method gave the company the possibility to establish control chart limits without material losses The QI Analyst tool has been used for the development of control charts.
Keywords:Eddy Currents, Aluminium tubes, Control chart limits
1.1 Initial conditions and problems
The markets' requirements for sophisticated high quality products in small amounts and quick response to the demand necessitate production systems with high flexibility and acceptable quality . Quality, meaning meeting customers' requirements in the widest possible sense, is a vital competitive strategy. Two aspects of quality are important in the environment of the company under consideration. First design quality, a measure of how well the product specifications meet the needs and expectations of intended customers. Quality levels are assessed in an "as designed" context to ensure that design targets and specifications truly reflect the business's requirements and the customers' needs. The second aspect of quality relates to the product in its "as produced" state. This aspect, conformance quality, describes how consistently the product as manufactured, achieves the design requirements. Conformance quality measures the freedom from deficiencies or non-conformities and the consistency with which the product is manufactured .
Inspection and testing is an explicit requirement for quality systems, concerned with establishing whether the inputs into the work process, the outputs and work at various stages in between, meet defined requirements thus assessing conformance quality. Arguably an efficient quality system is mainly based on preventive actions and less on the after the event inspection.
Even though an effective system should not depend on inspection and testing, the supplier has to inspect, test and identify product as required by the quality plan or the documented procedures, establish product conformance to specified requirements by use of process monitoring and control methods. In addition the supplier has to hold product until the required inspection and tests have been completed and identify non-conforming products.
One valuable tool for inspection and testing is the statistical process control (SPC). A primary goal of SPC is to detect non random behaviour in a process as soon as possible after it starts. This way the process can be adjusted to correct the problem and bring it back to control before heavy material losses occur. A first important step is to define a standard for the process based on its past performance data in order to compute control chart limits. A heavy restriction to the definition of the limits is that data points for which assignable causes can be found must be removed. The proper interpretation of control charts is critical to their use and several out-of-control criteria have been suggested.
In order to detect the assignable causes, strict precautions that will establish unchanged production conditions during the measurements must be taken. In addition the sequence of the measurements must be the same with the sequence the products go under the process which determines the characteristic measured. As a result, several sets of measurements are often necessary. Destructive methods lead often to prohibiting costs. Additional costs and material losses are associated with the samples that must be periodically measured in order to monitor the process.
In addition, when the main purpose of the process monitoring involves discovering and implementing measures that will reduce sources of variation and maintain minimal variation around targets a never ending improvement effort begins. This effort cannot be conducted without economic regard and destructive methods appear as a major impediment.
1.2 Objectives and methodology
The implementation of quality systems i.e. an established and proven set of procedures which are revised to take into account of a changing environment, increases conformity to customers' requirements and therefore quality .
The company under consideration developed a quality system, certified according to ISO 9002, including statistical process control (SPC) procedures for critical points in the quality plan as the thickness of the internal layer of the varnishing colour. First step of SPC is to compute control chart limits based on past performance data. A restriction is that data must be collected under identical production conditions. The destructive method, implemented so far consisted in the cut of the tube and measurement with micrometer. This method resulted in considerable costs thus a non destructive technique was necessary.
A specialised instrument for the non-destructive measurement of the thickness of the internal layer of the varnishing colour has been installed in the quality control laboratory. Data collected from the measuring system are eventually used to create the control charts and assess the processes' capability. A series of tests (Shewhart rules) are performed in order to verify that the measurments used come indeed from the same production lot and are considered in the correct sequence. When one or more of the out-of-control rules are violated a refinement of the process description unveils non random behavior and a new set of measurements is performed. When all the necessary particularities of the production process have been taken into account the control charts are established.
This procedure leads to an accurate model of the process, which can be furthermore used to point the main variation factors and begin corrective and preventive actions to reduce variation.
The paper is organised as follows: In Section 2, the case study is presented. Section 3 includes a brief introduction to the Eddy current technique and a description of the installed system. In Section 4 the sampling plans are presented along with the results of the statistical treatment of the data. Successive measurement sets led to the control charts, developed as the analysis proceeds in greater depth. Section 5 includes the conclusions, then opportunities are identified for further research in this area.
2. The case study
3. Installation of a non-destructive quality control system
3.1 Principle of operation of the eddy-current method
The principle of non-destructive material testing is such that the material to be tested is put into a suitable physical state of energy where energy flows in the material. Defects and irregularities in the material cause anomalies in the flow of energy, and these can also be established externally without penetrating destructively into the material.
When an AC current flows in a coil in close proximity to a conducting surface the magnetic field of the coil will induce circulating (eddy) currents in that surface. The magnitude and phase of the eddy currents will affect the loading on the coil and thus its impedance.
As an example, assume that there is a deep crack in the surface immediately underneath the coil. This will interrupt or reduce the eddy current flow, thus decreasing the loading on the coil and increasing its effective impedance.
This is the basis of eddy current testing, by monitoring the voltage across the coil in such an arrangement we can detect changes in the material of interest.
A number of factors, apart from flaws, will affect the eddy current response from a probe. Successful assessment of flaws or any of these factors relies on holding the others constant, or somehow eliminating their effect on the results. It is this elimination of undesired response that forms the basis of much of the technology of eddy current inspection.
3.2 Non-destructive measurement system
Using this technique, a specialised instrument for the non-destructive measurement of the thickness of the internal layer of the varnishing colour has been installed in the quality control laboratory (Fischer - Helmut Fischer GMBH+Co). The system measures the coating thickness of aluminium tubes, using Eddy current technique in a separate place in a non-continuous way. It is a specialised food / beverage container measuring system to measure the coating thickness on the inside and outside of aluminium containers. The system consists in two parts namely the instrument Fischerscope MMS and the system TM85 with the following features: Coating thickness 0-1000mm, measurement accuracy 1mm, max length of collapsible tube 260mm, Int. diameter of collapsible tube 11-82mm, min thickness of collapsible tube 70mm
4. Measuring process
4.1 First sampling plan
In order to perform the measurements 50 tubes have been taken and numbered in the same order as they come from the buffer after the external lacquering. For each tube 15 measurements of the thickness of the internal layer of the varnishing colour have been effectuated, five measurements in each critical position "edge", "middle", "tap". The thickness of the internal layer of the varnishing colour has been measured according to the instructions and the resulting control charts are presented in Figures 2, 3 and 4.
Fig 2: Control chart - Edge
Fig 3: Control chart - Middle
Fig 4: Control chart - Tap
In the charts several Shewhart rules are violated, mainly because of the existance of points outside the control limits. It is obvious that this control chart corresponds in a production process, which is not under statistical control. The results lead to the suggestion that the measurements do not belong to the same population. In order to have a more accurate picture of the production two critical pieces of the equipment have been examined in detail namely the internal lacquering machine and the external lacquering machine
The detailed model for the internal lacquering machine pointed that the internal varnishing is performed by two separate couples of sprayers each one spraying every second tube as they are coming from the transportation system. Therefore, as far as the internal lacquering is concerned, the hypothesis that control charts have to be developed for two separate production processes must be examined.
The detailed model for the external lacquering machine pointed that the sequence of the tubes is altered by a special transportation device which reverse the sequence of the tubes coming out of the machine. Special care must be taken in order to order the tubes for measurement in the sequence they have been actually internally lacquered.
4.2 Second sampling plan
Purpose of the second measurement set has been:
- The examination of the difference among the two sprayer couples
- The identification of the natural tolerance limits of the process.
50 tubes have been numbered before the internal lacquering. The numbers have been traced in the external surface of the tubes in order to make sure that the sequence will not be altered. For each tube 15 measurements of the thickness of the internal layer of the varnishing colour have been effectuated, five measurements in each position "edge", "middle", "tap".
The sample has been divided in two sub-samples, "odd" and "even", in order to separate the tubes lacquered by each sprayer couple. In a tube selected randomly, repeated measurements in four specific points assured the reliability of the measurement system.
The thickness of the internal layer of the varnishing colour has been measured according to the instructions and the resulting control charts for the positions "Edge" and "Middle" are presented in Figures 5 and 6.
Fig 5: Control chart - Edge
Fig 6: Control chart - Middle
This control chart corresponds in a production process under statistical control as none of the out-of-control rules is violated.
The clients specification concerning the thickness, dictates that the all the means of 5 measurements in each critical position "edge", "middle", "tap" must be grater than 8µm.
A summary of the results is presented in the following Table 1.
Table 1: Summary of the results
In no position of the tube statistical difference among the two sets (odd and even) seem to exist. Both sets belong to the same population..
As a result the 6 sets of 125 measurements can be merged in three sets of 250 measurements and the following Table 2 presents the results.
|Table 2: Merging of the measurement sets|
The results from the second measurement set indicate an important difference in the coating thickness on the inside of aluminium tubes. On the contrary the thickness is stable for each separate position "Tap", "Middle" and "Edge". In the position edge and middle the mean of the thickness is greater than the position tap. The mean thickness is almost double than the specification resulting in important material losses. Natural tolerance limits:
- Edge µ±3s =18,4±7,4
- Middle µ±3s =18,8±6,8
- Tap µ±3s =13,2±3,8
5. Conclusions and further research
The installation of a non-destructive system (Eddy current) for measuring the thickness of the internal lacquering gave the company the opportunity to conduct extended measurements at no material loss. The control charts produced made the SPC possible. Important material losses have been indicated to the management of the factory who decided modification of the production process in order to reduce the losses.
An interesting subject of research is to further extend this approach, by examining the control limits for the modified processes and compare the results in order to select the optimum.
- Jackson P., Ashton D., Implementing quality through BS5750, Kogan Page, London, 1993.
- Leopoulos V.N.I.). Modeling and simulation of a production line for capability study purposes using Petri Nets, Modern Applied Mathematics Techniques in Circuits, Systems and Control,sel. 354-358, World Scientific and Engineering Society Press, 1999.
- Luftig J., Quality with confidence in Manufacturing, SPSS Inc, 1997.
- Luz H., Neumaier P., Non-destructive Testing of Materials using the Eddy-current Method, Institute Dr.Forster Report-No 199.
- Womack J., Jones D., Lean Thinking, Simon and Schuster, New York, 1996.