Abstract
Heavily utilised reconfigurable production equipment has uneven deterioration rate caused by frequently changing configurations and complex subsystem interaction. Thus, traditional maintenance approaches that rely on reliability modelling are ineffective when applied on this class of equipment. This paper introduces a throughput-driven condition-based maintenance framework intended to perform predictive maintenance for heavily utilised reconfigurable production equipments. The effectiveness of the proposed method is experimentally validated using actual production records of the semiconductor test handlers. The primary benefit of the proposed framework is its prediction efficacy while accounting for the complex subsystem interaction through Bayesian statistics. It is low cost to implement because it only uses existing resources already available in most manufacturing plants.
Similar content being viewed by others
References
Fouquin M, Guimbard H, Herzog C, Unal D (2010) CEPII World economic overview. CEPII, France. http://www.cepii.fr/anglaisgraph/bdd/chelem/panorama/panorama.pps. Accessed 09 June 2011
Custer W (2009) Business outlook - global electronics industry. Custer Consulting Group. http://www.ttiinc.com/docs/IO/14314/20090322.pdf. Accessed 09 June 2011 http://www.gartner.com/it/page.jsp?id=985912. Accessed 06 May 2011
Texas Instruments (2011) TI to acquire national semiconductor. Texas Instruments. http://www.ti.com/ww/en/acquire/index.shtml. Accessed 08 May 2011
Peng Y, Dong M, Zuo M (2010) Current status of machine prognostics in condition-based maintenance: a review. Int J Adv Manuf Tech 50(1):297–313. doi:10.1007/s00170-009-2482-0
Liao W, Pan E, Xi L (2010) Preventive maintenance scheduling for repairable system with deterioration. J Intell Manuf 21(6):875–884. doi:10.1007/s10845-009-0264-z
Tu PYL, Yam R, Tse P, Sun AO (2001) An integrated maintenance management system for an advanced manufacturing company. Int J Adv Manuf Tech 17(9):692–703. doi:10.1007/s001700170135
Przytula KW, Choi A (2007) Reasoning framework for diagnosis and prognosis. Proceedings of 2007 IEEE Aerospace Conf, 3–10 Mar 2007, Big Sky, MT (USA), art. no. 416164
Sheppard JW, Kaufman MA (2005) Bayesian diagnosis and prognosis using instrument uncertainty. Proceedings of AUTOTESTCON, 26–29 Sep 2005, Orlando, FL (USA), vol 2005, pp 417–423
Kobbacy KAH, Murthy DNP, Budai G, Dekker R, Nicolai RP (2008) Maintenance and production: A review of planning models. In: Complex system maintenance handbook. Springer Series in Reliability Engineering. Springer London, pp 321–344. doi:10.1007/978-1-84800-011-7_13
Aetrium Incorporated (2005) 55V16 component test handler operation and maintenance manual, 1010216 Revision D. (Chapters 6–14)
Brotherton T, Jahns J, Jacobs J, Wroblewski D (2000) Prognosis of faults in gas turbine engines. Proceedings of the IEEE Aerospace Conf, 18–25 Mar. 2000, Big Sky, MT (USA), vol 6, pp 163–171
Garga AK, McClintic KT, Campbell RL, Yang CC, Lebold MS, Hay TA, Byington CS (2001) Hybrid reasoning for prognostic learning in CBM systems. Proceedings of the IEEE Aerospace Conf, 10–17 Mar 2001, Big Sky, MT (USA), vol 6, pp 2957–2969
Choi S (1995) Development of an on-line fuzzy expert system for integrated alarm processing in nuclear power plants. IEEE Trans Nucl Sci 42(4):1406–1418
Frelicot C (1996) A fuzzy-based prognostic adaptive system. RAIRO-APIII-JESA. J Eur Sys Autom 30(2):281–299
Kobbacy KAH, Murthy DNP, Percy DF (2008) Preventive maintenance models for complex systems. In: Complex system maintenance handbook. Springer Series in Reliability Engineering. Springer London, pp 179-207. doi:10.1007/978-1-84800-011-7_8
Luo J, Namburu M, Pattipati K, Liu Q, Kawamoto M, Chigusa S (2003) Model-based prognostic techniques [maintenance applications]. Proceedings of AUTOTESTCON, IEEE Sys Readiness Tech Conf, 22–25 Sep 2003, Anaheim, CA (USA), vol 22–25, pp 330–340
Zhang S, Ganesan R (1997) Multivariable trend analysis using neural networks for intelligent diagnostics of rotating machinery. Trans ASME J Eng Gas Turbine Power 119:378–384
Yam RCM, Tse PW, Li L, Tu P (2001) Intelligent predictive decision support system for condition-based maintenance. Int J Adv Manuf Technol 17:383–391
Sheau-Chyi L (2007) IC handler throughput evaluation for test process optimization. Proceedings of 2007 IEEE Instrum and Measurement Tech Conf, 1–3 May 2007, Warsaw (Poland), pp 2423-2428
SiliconFarEast (2005) Test equipment load boards/interface boards. SiliconFarEast. http://www.siliconfareast.com/loadbrds2.htm. Accessed 02 June 2011
Bellman R (1961) Adaptive control processes: a guided tour. Princeton University Press, New Jersey
Scott DW (1992) Multivariate density estimation: theory, practice, and visualization. Wiley Series in Probability and Statistics Wiley. doi:citeulike-article-id:2226994
Park SY, Bera AK (2009) Maximum entropy sutoregressive conditional heteroskedasticity model. J Econometrics 150:219–230
Mathworks (2011) Neural network toolbox trainlm algorithm. R2011b documentation. http://www.mathworks.com/help/toolbox/nnet/ref/trainlm.html. Accessed 03 Jan 2012
Mathworks (2011) Neural network toolbox dividerand function. R2011b documentation. http://www.mathworks.com/help/toolbox/nnet/ref/dividerand.html. Accessed 03 Jan 2012
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
1.1 A—Semiconductor production data records field summary
Field | Type | Description |
Index time | Raw | Time between end of test of a unit and start of test of the next |
Test time | Configuration | The test time depending on the test programme used |
Jam time | Raw | The length of time that the equipment spends idle while jammed |
Jam count | Raw | The number of jams that occur during the run |
Lot run time | Raw | The time taken for the lot to complete its run |
Lot size | Configuration | The size of the lot |
Net units | Raw | Number of units that pass the test |
Expected sites | Configuration | Number of test sites configured to be used during the run |
Actual sites | Raw | Average number of test sites actually operating |
MCBJ | Processed | Average number of units tested in between jams |
MSE | Processed | Average percentage of working sites during individual runs |
Yield | Processed | The percentage of units in a test lot that passes the test |
PPH | Processed | The number of units tested within the window of 1 hour |
Handler model | Configuration | The model of the test handler used |
Tester model | Configuration | The model of the attached tester |
Run type | Configuration | The run type of the lot run: Normal run, Retest rejects and Quality control run |
Configuration parameters are the inputs. Raw parameters are outputs from each lot. Processed parameters are that which are calculated from raw values. This type of data is normally automatically logged by the semiconductor test equipment in the Standard Test Data Format (STDF), which is a format used by all major semiconductor test equipment companies such as Teradyne and Verigy. It should be noted that on a case by case basis, the records for a production plant that does not adopt STDF would likely have raw and post-processed data fields.
1.2 B—Artificial neural network setup for CBM
State-of-the-art neural network toolbox in MATLAB was used to build ANN for CBM. 10,000 ANNs were trained with randomly generated network architectures. For each handler, the best network was chosen for benchmarking trials according to the following 2 criteria:
-
1.
Best correlation to corrective action
-
2.
Lead time of more than 5 lots
These criteria are exactly the same as those in Table 3. A lead time of 5 lots advance is chosen as it provides just enough time for maintenance to be scheduled. The randomisation parameters for the network architecture are 1–2 hidden layers, 5–15 nodes per hidden layer and the transfer function used are positive linear/tangent sigmoid combination.
Each of the networks is trained by 500 production records that have known maintenance decisions. It is ensured that a ratio of 1:1 between records with ‘Yes’ and ‘No’ decisions is maintained in the training set to enable sufficient exposure to bad handler during training phase. Six inputs (IT, TT, Jam time, Site count, MCBJ, PPH) and one output (Yes/No maintenance action) are used for the training process. These variables were selected as they were understood to have influence the health of the handler earlier in Sections 3.2–3.3.
The neural networks are all trained with the Levenberg–Marquardt optimization (trainlm), back-propagation supervised-learning training algorithm [24]. By default, trainlm invokes MATLAB’s dividerand function, which segregates training data into 3 sets, with a ratio of 70 to 15 to 15 [25]. These sets are used in succession for training, validation and testing of the neural network.
Rights and permissions
About this article
Cite this article
Guan, C.S., Kuang, Y.C. & Ooi, M.PL. Throughput-driven condition-based maintenance for frequently reconfigured mass production equipment. Int J Adv Manuf Technol 65, 1349–1361 (2013). https://doi.org/10.1007/s00170-012-4261-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00170-012-4261-6