Design of Experiment for factors affecting Contact Resistance in Metal to Carbon Relays

Size: px

Start display at page:

Download "Design of Experiment for factors affecting Contact Resistance in Metal to Carbon Relays"

Kellie Scott
6 years ago
Views:

Design of Experiment for factors affecting Resistance in Metal to Carbon Relays Hemant Kagra Abstract The reliability of Metal to Carbon Relays used in railway signaling for operating various

from the Failure Mode Effect Analysis, the major reason of failure of these relays was determined as High Resistance between the silver / Silver Impregnated s.

1 Design of Experiment for factors affecting Resistance in Metal to Carbon Relays Hemant Kagra Abstract The reliability of Metal to Carbon Relays used in railway signaling for operating various signaling functions viz. signals / points / track circuits etc. is very poor.from the Failure Mode Effect Analysis, the major reason of failure of these relays was determined as High Resistance between the silver / Silver Impregnated s. The aim of this paper is to determine the relay design parameters that will prohibit development of HCR, through Design of Experiment. The control factors, as determined by Cause and Effect analysis, were silver and. The experiment was conducted as per the L9 Orthogonal Array design with total 18 relays. Relays were operated continuously and numbers of operations till first failure were recorded. The results, analyzed through ANOVA, indicate that of 6R and of 30 gms. were the best settings for minimizing High Resistance. The Silver (Metal) and Silver Impregnated Graphite (Carbon) contacts [4] are fixed on springs as shown in figure 1, while the fixing arrangement of these contacts in the metal-to-carbon relays is as shown in figure 2. Figure 1. Silver / Silver Impregnated Graphite s on springs Index Terms -Metal to Carbon relays, Silver Impregnated Graphite (SIG), High contact resistance, FMEA, DOE I.METAL TO CARBON RELAYS Metal to Carbon relays are used in railway signaling systems for configuring various signaling circuits. These relays have silver (metal) and silver impregnated graphite (carbon) contacts [1]. The make and break sequences of these contacts simulates various selections in the signaling circuits, consequently, governing the change in aspect of railway signals from red to yellow to green, as per the requirement of the train operation, and as commanded by the train controller.when the contacts of the various relays are available, as per the pre-determined logical sequence, then the end relay gets picked up [2]. The current then flows from the source to the equipment i. e. to either a signal or a point or a track circuit etc. Even if one of the contacts of any relay is not available, the path of the current shall not be fulfilled and the circuit shall not get completed. In that case, the function viz. signal / point / track circuit shall not operate [3]. Figure 2. Fixing arrangement of contacts in metal to carbon relays The reliability of metal-to-carbon relays is very poor.an analysis of signaling equipment failures during last four years ( to ) on Indian Railways reveals that the failures of metal to carbon relays amount to approx. 7% of the total signaling equipment failures. This failure rate is quite substantial and totally unacceptable. Manuscript received Dec Hemant Kagra, RDSO, Indian Railways, Mumbai, India, As per Failure Modes Effect Analysis, failure modes are the errors or defects in a process, design, or item, potential or actual. A high RPN indicates that the failure shall cause maximum damage / problems to the customer and drastically reduce the product reliability [5]. Accordingly, FMEA of metal to carbon relays was done and the potential failure modes of these relays and their causes were examined.as per the analysis, the highest RPN (576) was found that of High 140

2 Resistance. The other major causes were Sulphation of metallic parts (RPN 128) and Coil Open circuit / short circuit (RPN 120).Some failures were also due to Armature Breakage, Gaps / Cracks in relay base etc. Thus, it is very clear that the failures caused by High contact resistance are the most potential risk. Since the current for various selections of signaling interlocking passes through these contacts, their conductivity is of utmost importance. Due to High contact resistance the current can t pass through these contacts. This disrupts the electrical connection between the contacts, causing the relay to fail, eventually resulting in the failure of signaling system. Thus it is imperative that development of High Resistance in metal to carbon relays should be prevented. II.DESIGN OF EXPERIMENTS Design Of Experiments [6] is the design of an experiment where one or more factors, believed to have an effect on the experimental outcome are identified and manipulated according to a predetermined plan. It is used in evaluating the parameters / properties of physical objects, chemical formulations, structures, components, materials etc. Accordingly, the various steps involved in the DOE for finding the factors mainly responsible for high contact resistance phenomenon in metal to carbon were as under: (i) The Problem Statement The metal to carbon relays fail during operation in field due to development of High Resistance. The failure is observed when the signal lamp fails to light. (ii) The Objective of Experiment Determination of the design parameters that will minimize the High Resistance. (iii) Selection of Quality Characteristics & Measurement System The quality characteristic was the High Resistance causing the first time failure of metal to carbon relays; the failures were measured in terms of the number of operations until the signal lamp fails to light up. (iv) Selection of Factors influencing the High Resistance The Cause and Effect analysis of the various factors, which may be responsible for abnormal increment in the Resistance of the relays, was done, and the Cause Effect Diagram, as obtained through Minitab 16, is as shown in figure 3. Fig 3. Cause & Effect Diagram of HCR As can be seen from the above, the various factors, which may affect the development of High Resistance in relays, would include inadequate profile or radius of silver contact, SIG contact not as per specification, poor quality of solder wire / cleaning agent, Inappropriate Spring cleaning, Improper Spring Tensioning, Unsuitable setting of, Inadequate annealing of magnetic parts, presence of dust inside relay, development of film (oxide / sulphide / hydrocarbon) on contacts, incorrect Resistance / SIG contact testing method, lack of skill or negligence of Technicians etc. (v) Identification of Control & Noise Factors Control Factors are those factors in design or process that can be controlled while Noise Factors are those factors which are either not controllable or are not desired to be controlled on account of the fact that controlling them is either impossible or very expensive and tedious. The control & noise factors in this DOE were determined as follows: Control Factors: Inadequate profile or radius of silver contact; Unsuitable setting of. Noise Factors : SIG contact not as per specification; Film (oxide / sulphide / hydrocarbon) on contacts; Poor quality of solder wire / IPA (cleaning agent); Presence of dust inside relay Accordingly, experiment was designed to study the effect of inadequate profile or radius of silver contact & unsuitable setting of on development of High Resistance in relays. (vi) Selection of Levels & Values for the Factors Each factor can have multiple values, which shall vary over the operating range. Levels are the specific values that factors (inputs) can take. These values are in the operational range of the product. Statistically, these variable values can be either continuous or discrete. Product or process technical expertise is the single most important source for the selection of appropriate values 141

while the of the Silver contact is 19R. Based on the technical expertise and operational experience, the levels and their values for & were set as shown in table 1. (vii) Table 1.

3 for the factor levels. Besides, the product specification / operating limits are also helpful in assigning values for the levels of the factors. The Indian Railway Specification [1] specify the maximum / minimum range of the and. At present, the in the relays is set at 28 gm., which corresponds to armature travel of 56 mm. while the of the Silver contact is 19R. Based on the technical expertise and operational experience, the levels and their values for & were set as shown in table 1. (vii) Table 1. Factors and Levels for DOE of HCR Factors Level 1 Level 2 Level 3 28 gm. 30 gm. 34 gm. 4 R 6 R 19 R Selection of appropriate Orthogonal Array Orthogonality implies that factors can be evaluated independently of one another; the effect of one factor does not bother the estimation of the effect of another factor. The experiments are balanced i.e. all factors get equal opportunity to be tested. Selection of Orthogonal Array predominantly depends on the number of factors & interactions of interest, number of levels for each factor and the desired experimental resolution. The resolution number is the measure of the amount of confounding in a column i.e. whether the main effect is being mixed / overshadowed by some interaction effect. In Orthogonal Array, if one factor is assigned to a particular column and a second factor is assigned to any other column, a specific third column will automatically have the interaction of these factors assigned to that column. Factors were assigned by the help of Factor Assignment table & Interaction Tables. The factor assignment table indicates various columns to which the factor may be assigned while the Interaction table indicates which columns subsequently evaluate the interaction of these factors. In this experiment, the Three Level Orthogonal Array Factor Assignment Table indicates that for L9 OA, Columns 1 & 2 should be used for the main factors. Thus, was assigned to column 1 while was assigned to column 2. The experimental design, with two replicates, as obtained through Minitab 16, was used. The full factorial design provided for total eighteen trial runs with various settings of &. (x) Conducting Experiment as per the trials in Orthogonal Array The Experiment was conducted as per the trial design given above in Para (ix). As explained earlier, in order to enhance the accuracy of trial results, the experiment was replicated two times. Total 18 (9 +9) relays were manufactured with the settings of / as per the trial design. The test setup consisted of relay racks on which the 9 relays were fixed & wired as shown in the figure 4. As there were 2 factors having 3 levels each, from the OA design & selection table, the Orthogonal Array L9 - having total 9 (3 2 ) numbers of trials (different possible test combinations) with a resolution of 4 was selected. (viii) Selection of possible Interaction Effects Many Factors may interact with one another to provide a synergistic effect on the quality characteristics being studied. Interaction is the estimation of effect of one factor over the other factors. This mutual dependency between the factors or interaction may be quite significant and posses high potential to affect the outcome of experiment. In this experiment, since there were only two factors & there was only one interaction effect X. (ix) Assigning Factors to Orthogonal Array & locating Interactions Fig 4. Relay Test setup for DOE The relays were serially interconnected as per the logic of signaling circuit as is generally available in the field and a master PID controller was used to provide and control cyclical feed to the relays. Green and Red LED lamps were used to indicate the final output of the circuit, inter-alia, the appropriate functioning status of relays. These lamps mimicked the Railway Signal in the field, having two aspects Red & Green. The PID controller provided feed to the button relays, which in turn, operated the other relays and the Lamp Checking Relay (ECR) was used to light up the Green / Red LED lamps. When all the relays were picked up 142

4 appropriately, the Green LED lamp was lit up, thus indicating that the circuit was functioning properly. Even if one relay in the circuit failed, the Green lamp extinguished and the Red lamp started glowing and further functioning of the circuit stopped. The defective relay was replaced to restore the functioning again. The relays were operated on the rated voltage i. e. 24 Volts, in accordance with the metal to carbon relay specification no. BRS 930. The operating current of a relay is in the range of ma. As specified in the relay test procedure Document no. SIG 0002 Version 0 of Indian Railways, the operating speed of relays was set to 10 operations per minute. The number of operations of relays were counted by a Data logger and the instant of failure of individual relay was recorded with date & time stamp. Accordingly, the Mean Time To Failure (mean time, in terms of number of operations, until the first failure of the relay) was computed. The test results were as shown in table 2. Trial No. Table 2. DOE Test Results Response (No. of ops.) (xi) Analysis of Results of experimental trials The Analysis of Variance (ANOVA) Table for the experimental Response, is as shown in table 3: Table 3: ANOVA Table Source DF Seq SS Adj MS F * Error Total Source P * Error Total S = R-Sq = 98.06% R-Sq(adj) = 96.34% The ANOVA analysis, as given above, proves that the and are the two main factors affecting the contact Resistance. III.CONCLUSION DOE is an excellent analysis tool for improving the quality, reliability and safety of a product/process, prevents problem recurrence and reduces product costs. Implementation of the recommended parametric settings for and shall certainly prohibit the development of High Resistance, consequently enhancing the reliability of metal to carbon relays. Improvement in reliability of metal to carbon relays by reducing / eliminating their failure modes shall enhance the performance of Railway Signaling Systems, resulting into better punctuality of trains and additionally, improving the safety performance of railway signaling systems and reducing public grievances. 143

5 ACKNOWLEDGMENT Sincere thanks are due to Dr. Samsul Ekram, Asst. General Manager from M/s. CGL, for his guidance and support. Thanks also to Signaling & Telecommunication department of Indian Railways for providing the reliability data of signaling equipment. This paper would not have been possible without the help and insights from quality & production team of M/s. Crompton Greaves Limited, Dhar, Indore. REFERENCES Hemant Kagra obtained hisbachelor in Engineeringdegree in Electronics and Telecommunication from MANIT, Bhopal, M.P., India, in He joined the Indian Railway Service of Signal Engineers in 1991 and is currently posted as Director / Quality Assurance in Research Design & Standards Organization of Indian Railways at Mumbai, India. He has published and presented over 20 papers in national & international conferences and journalsand is member of many national & international professional organizations viz. IEEE, IRSTE etc.he has research interest in the area of railway signaling especially metal to carbon relays. [1] Metal-to-carbon Relay Specifications: (i) British Rail Specification 930 & 931 A (ii) British Rail Specification 932A, 933A, 934A, 935A, 937A, 938A (iii) Indian Rail Specification RDSO/SPN/84/88 (iv) Indian Rail Specification STS/Relay/AC Lit LED Signal/99/2002 [2] Australian Rail Track Corporation Ltd., Relay Design Standard, Australia,. [3] Indian Railways, Handbook on Signaling Circuits,India. [4] M/s Westinghouse, Relay reference Book,U. K. [5] D. H. Stamatis, FMEA From Theory to Execution, 2 nd ed. ASQ Quality Press, Wisconsin, USA, 2003 [6] Philip J. Ross, Taguchi Techniques for Quality Engineering, Tata McGraw-Hill, India,

Performance Evaluation of Wedm Machining on Incoloy800 by TAGUCHI Method

Performance Evaluation of Wedm Machining on Incoloy800 by TAGUCHI Method Gagan Goyal Scholar Shri Balaji Collegeof Engineering & Technology, Jaipur, Rajasthan, India Ashok Choudhary Asistant Professor