ارائه روش ترکیبی پیش پردازش داده‌ها در ماشین بردار رگرسیون جهت پیش‌بینی کیفیت گازوییل پالایش شده

نوع مقاله: مقاله پژوهشی

نویسندگان

1 دانشگاه علم و صنعت ایران، دانشکده مهندسی شیمی

2 پژوهشگاه صنعت نفت، پژوهشکده توسعه فرآیند و فناوری تجهیزات

چکیده

از آنجا که دقت داده‌ای اندازه گیری شده فرآیندی در پیش‌بینی کیفیت محصولات بسیار مهم است، در این تحقیق بر روی پیش پردازش داده‌ها تمرکز گردید. برای این منظور حسگر مجازی برای تعیین کیفیت گازوییل خروجی از پایلوت تصفیه هیدروژنی طراحی شد. طراحی حسگر مجازی بر اساس یکی از روش‌های جدید یادگیری ماشین به نام ماشین‌بردار رگرسیون انجام گردید. برای پیش پردازش داده‌ها از تکنیک ترکیبی به صورت پشت سر هم متشکل از آنالیز موجک و کوانتیزاسیون‌برداری به منظور حذف خطاهای تصادفی، متراکم‌سازی داده‌ها و چشم‌پوشی از داده‌هایی که شباهت کمتری به سایر داده‌ها دارند، استفاده گردید. روش‌های متفاوتی از آنالیز موجک برای حذف خطاهای تصادفی به کار برده شد و بهترین روش انتخاب گردید. آزمایشات حذف خطاهای تصادفی با استفاده از آنالیز موجک با تابع پایه هار و دابیچز و با الگوریتم‌های انتخاب آستانهHeursure ،RigrsureMinimaxiو Sqtwolog انجام شد. مقایسه نتایج نشان داد که روش Db4 به همراه روش آستانه‌گیری Rigrsure بهترین نتایج حذف خطا را به دنبال دارد. با استفاده از این روش مقدار عددی AARE و RMSE نسبت به انواع دیگر تابع موجک بهتر است. همچنین، معیار عملکردی AARE برای سنجش دقت پیش‌بینی مدل ماشین بردار رگرسیون استفاده گردید. مقدار AARE برابر 053/0 به دست آمد که نشان‌دهنده دقت بالای مدل در پیش‌بینی غلظت گوگرد خروجی از رآکتور می‌باشد.
 

کلیدواژه‌ها


عنوان مقاله [English]

An Integrated Method of Data Pre-processing in Support Vector Regression for the Quality Prediction of Treated Gas-oil

نویسندگان [English]

  • Saeid Shokri 1
  • Mohammadtaghi Sadeghi 1
  • Mehdi Ahmadi Marvast 2
1 Department of Chemical Engineering, Iran University of Science and Technology (IUST), Tehran, Iran
2 Process & Equipment Technology Development Division, Research Institute of Petroleum Industry (RIPI), Tehran, Iran
چکیده [English]

The accuracy of the measured data is very important for better quality prediction by soft sensors. In order to determine the quality of the treated gas oil, a soft sensor is designed. The soft sensor design is based on a new machine learning technique called support vector regression (SVR). An integrated technique was developed for data preprocessing. In this technique, wavelet analysis and vector quantization were being used sequentially for random error elimination, data compression, and unusual data omitting. Different methods of wavelet analysis were used to remove the random errors and the best method was selected. Random errors were deleted using Harr and Daubechies basis function where Rigrsure, Minimaxi, Heursure, and Sqtwolog were the threshold algorithms. The results showed that the db4 basis function with Rigrsure threshold algorithms provided the best results for error removal. AARE and RMSE for this method was better than the other types of wavelet functions. Additionally, the results of SVR training based on the pilot plant data showed AARE of 0.053. This showed the high accuracy of the SVR model for predicting treated gas oil quality.
 

کلیدواژه‌ها [English]

  • Wavelet Analysis
  • Soft Sensor
  • Support Vector Regression
  • Data Validation
  • Quality Prediction
[1]. Kadlec P., Gabrys B. and Strandt S., “Data-driven soft sensors in the process industry”, Comput. Chem. Eng., Vol. 33, pp. 795, 2009.

[2]. Pan T .H., Wong D. S. H. and Jang S .S., “Development of a novel soft sensor using a local model network with an adaptive subtractive clustering approach”, Ind. Eng. Chem. Res., Vol. 49, pp. 4738–4747, 2010.

[3]. Gonzagaa J. C. B., Meleirob L. A. C., Kianga C. and Filho R. M., “ANN-based soft-sensor for real-time process monitoring and control of an industrial polymerization process”, Comput. Chem. Eng. Vol. 33, pp. 43–49, 2009.

[4]. Chitralekha S. B. and Shah S.L., “Application of support vector regression for developing soft sensors for nonlinear processes”, The Canadian Journal of Chemical Engineering. Vol. 88, pp. 696-709, 2010.

[5]. Park T. C., Kim T. Y., and Yeo Y. K., “Prediction of the melt flow index using partial least squares and support vector regression in high-density polyethylene (HDPE) process”, Korean J. Chem. Eng., 27(6), pp. 1662-1668, 2010.

[6]. Dutta S. and Gupta J. P., “PVT correlations of indian crude using support vector regression”, Energy & Fuels, Vol. 23, pp. 5483–5490, 2009.

[7]. Yan W., Shao H. and Wang X., “Soft sensing modeling based on support vector machine and Bayesian model selection”, Comput. Chem. Eng., Vol. 28, pp. 1489–1498, 2004.

[8]. Desai K., Badhe Y., Tambe S. S. and Kulkarni B. D., “Soft-sensor development for fed-batch bioreactors using support vector regression”, Biochemical Engineering Journal 27, pp. 225–239, 2006.

[9]. Guohai L., Dawei Z., Haixia X. and Congli M., “Model optimization of SVM for a fermentation soft sensor”, Expert Systems with Applications, Vol. 37, pp. 2708–2713, 2010.

[10]. Liu Y., Hu N., Wang H.  and Li P., “Soft chemical analyzer development using adaptive least-squares support vector regression with selective pruning and variable moving window size”, Ind. Eng. Chem. Res., Vol. 48, pp. 5731–574, 2009.

[11]. Hong W. C., “Traffic flow forecasting by seasonal SVR with chaotic simulated annealing algorithm”, Neurocomputing, Vol. 74, 2096-2107, 2011.

[12]. Minqiang P., Dehuai Z. and Gang X., “Temperature prediction of hydrogen producing reactor using SVM regression with PSO”, Journal of computers,Vol 5, No.3, 2010.

[13]. Yin J., “LogP prediction for blocked tripeptides with amino acids descriptors (HMLP) by multiple linear regression and support vector regression”, Procedia Environmental Sciences 8, pp. 173–178, 2011.

[14]. Vapnik V. N. The nature of statistical learning theory, 2nd, ed. New York: Springer; 1999.

[15]. Boser B. E., Guyon I. M. and Vapnik V. N. A training algorithm for optimal margin classifiers, In D. Haussler, editor, 5th Annual ACM Workshop on COLT, 144-152, Pittsburgh, PA, ACM Press, 1992.

[16]. Vapnik V. N., Statistical learning theory, Wiley, New York, 1998.

[17]. Basak D., Pal S. and Patranabis D. C., “Support vector regression”, Neural Inf. Process. Vol. 11, pp. 203–225, 2007.

[18]. Cherkassky V. and Ma Y., “Practical selection of SVM parameters and noise estimation for SVM regression”, Neural Networks, Vol. 17, pp. 113–126, 2004.

[19]. Benqlilou C., Data reconciliation as a framework for chemical processes optimization and control, PHD Thesis,March, 2004.

[20]. Singh M. K., Denising of natural images using the wavelet transform, Master's Thesis. San Jose State University, 2010.

[21]. Huang H. P. and Luo K. Y., “On-line wavelets filtering with application to linear dynamic data reconciliation”, Ind. Eng. Chem. Res. Vol. 46, pp. 8746-8755, 2007.

[22]. Unser M. and Blu T., “Wavelet theory demystified”, IEEE Transaction on Signal Processing, 51(2), pp. 470-483, 2003.

[23]. Shukla P. D., Complex wavelet transforms and their applications, Master of Philosophy Thesis, University of Strathclyde , 2003.

[24]. Phinyomark A., Limsakul C., and Phukpattaranont P., “A comparative study of wavelet denoising for multifunction myoelectric control”, inInternational Conference on Computer and AutomationEngineering, pp. 21–25, 2009.

[25]. Jiang C. F. and Kuo S. L., “A comparative study of wavelet denoising of durface electromyographic signals”, in 29th Annual International Conferenceof the IEEE Engineering in Medicine and BiologySociety, pp. 1868–1871, 2007.

[26]. Somasundaram K. and Vimala S., “Fast encoding algorithm for vector quantization”, International Journal of Engineering Science and Technology, Vol. 2(9), pp. 4876-4879, 2010.

[27]. Yu T., Simoff S. and Jan T., “VQSVM: a case study for incorporating prior domain knowledge in to inductive machine learning”, Neurocomputing Vol. 73, pp. 2614–2623, 2010.

[28]. Chang C. C. and Lin C. J., LIBSVM: a library for support vector machines, Software http://www.csie.ntu.edu.tw/~cjlin/libsvm/, 2001.