Syed Hamid

Neural network training data selection using memory reduced cluster analysis for field model development

A system and method for selecting a training data set from a set of multidimensional geophysical input data samples for training a model to predict target data. The input data may be data sets produced by a pulsed neutron logging tool at multiple depth points in a cases well. Target data may be responses of an open hole logging tool. The input data is divided into clusters. Actual target data from the training well is linked to the clusters. The linked clusters are analyzed for variance, etc. and fuzzy inference is used to select a portion of each cluster to include in a training set. The reduced set is used to train a model, such as an artificial neural network. The trained model may then be used to produce synthetic open hole logs in response to inputs of cased hole log data.

Multi-Response Optimization of Laser-based Powder Deposition of Multi-track Single Layer Hastelloy C-276

The objective of the current work is to develop an empirical relationship between the input variables and quality characteristics in a laser-based powder deposition of Hastelloy C-276 on steel. The influence of laser power, scanning speed, powder flow rate, and percentage overlap on uniformity index, dilution, and microhardness is studied. A central composite design at five levels allowed us to obtain a second-order model to predict the optimal layer characteristics with the terms of interaction between the four chosen influential input variables. A fine equiaxed microstructure with uniform layer thickness is achieved for the laser power of 285 W, scanning speed of 11.3 mm/s, powder flow rate of 0.176 g/s, and a percentage overlap of 34.1%. The developed experimentally based statistical model can be used to predict the important prerequisites of the qualified deposit such as uniform layer thickness, minimal dilution, and homogenous metallurgical bond. The results demonstrated the existence of a critical range for the powder flow rate and overlap percentage below which the scanning speed and dilution (%) holds direct relationship.

Variable Input Neural Network Ensembles in Generating Synthetic Well Logs

This paper discusses a hybrid method for construction of neural network ensembles (NNE) in generating synthetic well logs that is often driven by the needs of simulating unobtainable actual logs, reducing the operational cost, reconstruction of missing and/or bad log data, and minimizing the hazards associated with using radioactive sources. In this method, several computer-driven routines are developed to rank the candidate neural network inputs as a function of data partition, network complexity and initialization. Then a network pool is automatically formed having the selected candidate networks characterized with multi-set inputs and different hidden nodes. The ensemble optimization is performed using a multi-objective genetic algorithm by aggregating the ensemble validation error, complexity, and negative correlation into a single quantity of merit. The simulations applied to actual field examples demonstrate that using multi-set-input NNE is more robust than using single-set-input NNE with significantly reduced uncertainty and improved prediction accuracy on the new data for some applications.

Construction of surrogate model ensembles with sparse data

Construction of neural network ensembles (NNE) with sparse data requires comprehensive performance measure, multi-stage validation and usually a large member size. This paper presents a hybrid method which takes a selective optimization approach and is characterized with several novel features. First, candidate ensembles are widely explored using a multi-objective genetic algorithm. Secondly, the best local ensembles registered with each distinct objective weighting are determined based on the multi-stage validation results. Finally, a large global ensemble is formed by combining several local ensembles and virtually evaluated in the voids of possible parameter space. The demonstration of the proposed method is presented in a case study in which sparse data from FEA simulations are used to construct NNE for expandable pipe design, a novel application in oil and gas industry.

An Experimental Study on Slurry Erosion Resistance of Single and Multilayered Deposits of Ni-WC Produced by Laser-Based Powder Deposition Process

Single and multilayered deposits containing different mass fractions of tungsten carbide (WC) in nickel (Ni)-matrix (NT-20, NT-60, NT-80) are deposited on a AISI 4140 steel substrate using a laser-based powder deposition process. The transverse cross section of the coupons reveals that the higher the mass fraction of WC in Ni-matrix leads to a more uniform distribution through Ni-matrix. The slurry erosion resistance of the fabricated coupons is tested at three different impingement angles using an abrasive water jet cutting machine, which is quantified based on the erosion rate. The top layer of a multilayered deposit (i.e., NT-60 in a two-layer NT-60 over NT-20 deposit) exhibits better erosion resistance at all three tested impingement angles when compared to a single-layer (NT-60) deposit. A definite increase in the erosion resistance is noted with an addition of nano-size WC particles. The relationship between the different mass fractions of reinforcement (WC) in the deposited composite material (Ni-WC) and their corresponding matrix (Ni) hardness on the erosion rate is studied. The eroded surface is analyzed in the light of a three-dimensional (3-D) profilometer and a scanning electron microscope (SEM). The results show that a volume fraction of approximately 62% of WC with a Ni-matrix hardness of 540 HV resulting in the gouging out of WC from the Ni-matrix by the action of slurry. It is concluded that the slurry erosion resistance of the AISI 4140 steel can be significantly enhanced by introducing single and multilayered deposits of Ni-WC composite material fabricated by the laser-based powder deposition process.