Surface Roughness In Real Time Research Articles

Real-time surface roughness estimation and automatic regrinding of ground workpieces using a data-driven model and grinding force inputs

Real-time surface roughness estimation and automatic regrinding of ground workpieces using a data-driven model and grinding force inputs

FE-trans-net: Feature enhancement based single branch deep learning model for surface roughness detection

Real-time surface roughness measurement is crucial for modern manufacturing. Machine vision-based roughness measurement algorithms have the advantages of non-contact, high efficiency and adaptability. Light source is important for online machine vision-based roughness measurement. And different materials may need different light sources to extract more features. From the previous study, dual light sources are better than single light source for surface roughness prediction. However, it is inconvenient to assign two light sources for online detection equipment, and the two-branch model structure for dual light sources doubles the size of the model. To fully utilize the advantages of dual light sources, and avoid its disadvantages, a feature enhancement based single branch deep learning model FE-Trans-Net is proposed in this paper for surface roughness detection. Firstly, a dataset of different materials (Cu, Zn, Al, and Fe) shined with different light sources (white, red, green, red laser and green laser) is collected. Two most suitable light sources for different materials are chosen for dual light sources model as offline pre-training. In this model, lightweight cross-attention fusion (CAF-Trans) and graph convolutional channel attention (GCC-Atten) modules are designed to fuse the feature information of different light sources from the spatial and channel perspectives respectively, enabling the single-branch model to acquire the feature learning capability of different light sources. Then a single-branch network layer is remained as the online prediction. With the feature enhancement, it preserves the advantages of dual light source images, and has only half size of the dual light source model. Extensive experiments validate the effectiveness and generalization performance of the CAF-Trans and GCC-Atten modules, and the proposed FE-Trans-Net network achieves the best accuracy-efficiency tradeoff compared to other popular deep neural networks.

Measurement of EDMed surfaces roughness using convolutional neural network

In addition to dimensions, surface roughness measurement is crucial in every manufacturing. In this study, a trustworthy method for characterising the surface roughness of electrical discharge machined surfaces was developed using a convolutional neural network. Since feature extraction is incorporated into the convolution process of the network, this technique eliminates it. Images of EDMed surfaces were taken using a mobile camera. MATLAB software was used to process a signal vector that was created from the intensity of the picture pixels. A database of specimens with recorded surface roughness values was created. When samples with known surface roughness are given, the proposed technique is a strong contender for real-time surface roughness measurement. The generated predicted values are compared with the measured values acquired from a profilometer using a stylus. The applicability and accuracy of five loss functions are considered before they are chosen and examined for the prediction models. The accuracy and performance of this digital model suggest that it has the capability to assess the surface roughness very well.

Surface stability in stress-assisted corrosion: a peridynamic investigation

In this paper, the stability of corrosion front of metallic material under tension is investigated by using a stress-assisted corrosion peridynamic model. The electrolyte diffusion tends to smoothen the corrosion front, while the mechanical loading tends to roughen the corrosion front and leads to higher stress concentration. We introduce a critical stress, below which the roughness of the corrosion front does not blow up, and the corrosion front is stable. Real-time surface roughness evolution is utilized to find the critical stress. Numerical results for both initially convex/concave sinusoidal and random surface contours show that corrosion without stress flattens the corrosion front, and when stress is higher than a certain value, the stress-assisted corrosion makes the corrosion front rougher. The effect of mechanochemical coefficient and applied electrical potential on the critical stress is analyzed numerically. It is found that the critical stress is less affected by the external voltage than by the mechanochemical coefficient. We also apply the model for corrosion in metal with initial holes or insulate impurities, revealing that hard impurities, such as dispersed oxides, can help maintain the stability of corrosion front, thus reducing the risk of stress-corrosion cracking.

Wear model based on real-time surface roughness and its effect on lubrication regimes

Abstract In this study, the running-in tribological behaviors of brass sliding against 100Cr6 under lubricated conditions are investigated by an in-situ tribometer. The wear model based on the relationship between real-time wear rate and real-time surface roughness during the test is established. The evaluations of the wear model indicate that it can predict wear rate effectively under certain working conditions. It is found that the tribological behaviors strongly depend on the real-time roughness. On this basis, the lubrication number Z parameter of the Stribeck curve is modified by considering the impact of the real-time surface roughness. Furthermore, the running-in tribological behaviors and the wear mechanisms under different regimes are discussed through the obtained Stribeck curve.

Development of a new mechano-chemical model in boundary lubrication

A newly developed tribochemical model based on thermodynamics of interfaces and kinetics of tribochemical reactions is implemented in a contact mechanics simulation and the results are validated against experimental results. The model considers both mechanical and thermal activation of tribochemical reactions instead of former thermal activation theories. The model considers tribofilm removal and is able to capture the tribofilm behaviour during the experiment. The aim of this work is to implement tribochemistry into deterministic modelling of boundary lubrication and study the effect of tribofilms in reducing friction or wear. A new contact mechanics model considering normal and tangential forces in boundary lubrication is developed for two real rough steel surfaces. The model is developed for real tribological systems and is flexible to different laboratory experiments. Tribochemistry (e.g. tribofilm formation and removal) and also mechanical properties are considered in this model. The amount of wear is calculated using a modified Archard’s wear equation accounting for local tribofilm thickness and its mechanical properties. This model can be used for monitoring the tribofilm growth on rough surfaces and also the real time surface roughness as well as changes in the λ ratio. This model enables the observation of in-situ tribofilm thickness and surface coverage and helps in better understanding the real mechanisms of wear.

Viability of an optoelectronic system for real time roughness measurement

Surface roughness evaluation is very important to characterize surface finish. Roughness parameter can be calculated in either two-dimensional (2D) or three-dimensional (3D) form. 2D profile analysis has been widely used in science and engineering for more than half century. In recent years, there was an increases need for 3D surface analysis. The aim of this paper is to present a new optoelectronic system based on two lasers diode for 3D real-time surface roughness measurement. 3D surface roughness parameter (Sq), of a surface characterized by a stationary Gaussian random process, is estimated using Spectral Speckle Correlation (SSC). In the paper, theory is described and experimental results are given.

MECHANICAL PULPING: Investigation of the surface topographical differences between the Cross Direction and the Machine Direction for newspaper and paperboard

Real time paper surface-web measurement is one of the challenging research fields. The traditional laboratory method has many limitations and is unable to measure the entire tambour during the manufacturing process. It has been necessary to develop an online technique that could measure the surface topography in real time. An optical technique was developed, based on laser triangulation, and is applied to develop a new prototype device, which characterizes high speed paper-web surfaces over a wide scale of spatial wavelengths spectrum and computes the surface roughness in real time. The used multi channel pulsed laser diode, source of illumination onto the paper-web, is of benefit due to its low coherence length and is capable to deliver a powerful burst of light beam over a 1 µs duration, which delivers energy of 100 µJ per pulse. The short exposure time avoids blurriness in the acquired images which could possible due to the high speed and vibrations on the paper-web.The laser beam is shaped into a narrow line-of-light using cylindrical lenses and is projected onto a paper-web surface, which covers a physical length of about 210 mm. The created line-of-light cross section full width at half maximum, FWHM Gaussian distribution, is 2-3 pixels on the image. The line-of-light is projected onto the paper-web perpendicular to the plane of the surface. The low angled, low specular, reduced coherence length, scattered reflected laser line is captured by the 3 CCD sensors, which are synchronized with the laser source. The low specular light ensures to avoid saturation of the imaging sensors if the surface is very smooth, and obliquely captures the z-directional fine feature of the surface.The scattered phenomenon of the reflected light is responsible for the surface irregularity measurements. The basic image processing algorithm is applied in order to remove noise and cropped the images widthwise so that only pixels above a preset threshold gray level can be processed, which enables efficient real time measurement. The image is transformed into a 1D array using the center of gravity, COG. The accuracy and precision of the COG depends on the line-of-light FWHM, which, in turn, is responsible for the accuracy, noise and the resolution of the developed technique. The image subpixel resolution achieved is 0.01 times a pixel and uuncertainty in the raw data is 0.43 µm while it is 0.05 µm in the rms roughness.The signal processing steps combining the B-Spline filter and the filter in the spatial frequency domain were employed in order to separate roughness, waviness, and form and position error in the raw profile. The prototype is designed to measure online surface roughness and to characterize surface in a spatial wavelength spectrum from 0.09 to 30 mm, which is extendable to any required spatial range in order to cover a wide scale surface feature such as micro roughness, macro roughness and waviness. It is proven that exploitation of a simple laser triangulation technique could lead to an improvement in the overall quality and efficiency in the paper and paperboard industries and it can also be of potential interest for the other surface characterization problems.

In-process surface roughness prediction using displacement signals from spindle motion

A method to predict surface roughness in real time was proposed and its effectiveness was proved through experiment in this paper. To implement the proposed method in machining process, a sensor system to measure relative displacement caused by the cutting operation was developed. In this research, roughness of machined surface was assumed to be generated by the relative motion between tool and workpiece and the geometric factors of a tool. The relative motion caused by the machining process could be measured in process using a cylindrical capacitive displacement sensor (CCDS). The CCDS was installed at the quill of a spindle and the sensing was not disturbed by the cutting. The workpiece was NAK80 and TiAlN coated carbide end mills were used in the test. Model to predict surface roughness was developed. A simple linear regression model was developed to predict surface roughness using the measured signals of relative motion. Close relation between machined surface roughness and roughness predicted using the measured signals was verified with similarity of about 95%.

Computer Vision Based Non-contact Surface Roughness Assessment Using Wavelet Transform and Response Surface Methodology

ABSTRACT This paper describes a new non-contact measurement approach in characterizing manufactured surfaces. Computer vision is applied to capture digital images of three types of anisotropic steel specimen surfaces from shaping, grinding, and polishing processes. Multiresolution wavelet decomposition is used to obtain signatures of surface profiles from the digital images. Relationships between these signatures and surface roughness parameters (Ra and Rq) are built by response surface methodology (RSM). The proposed models thus developed are suitable for predicting roughness in terms of the roughness parameters. Experimental results show that the proposed approach successfully correlates wavelet signals to Ra and Rq values. In addition, they also show repeatable gage capabilities. The proposed method is a good candidate for on-line, real-time surface roughness inspection when specimens of known surface roughness are available.

Real-time spectroscopic ellipsometry as an in-situ probe of the growth dynamics of amorphous and epitaxial crystal silicon for photovoltaic applications

AbstractIn this paper we report on our work using in-situ real time spectroscopic ellipsometry (RTSE) to study the dynamics of hot-wire chemical vapor deposition (HWCVD) of hydrogenated amorphous silicon (a-Si:H) and epitaxial crystal silicon (epi-Si) for photovoltaic applications. We utilize RTSE as both an in-situ diagnostic and a postgrowth analysis tool for a-Si:H/crystalline silicon heterojunction (SHJ) solar cells and epi-silicon films grown by HWCVD. RTSE enables precise thickness control of the 3 to 10 nm thick layers used in the SHJ devices, as well as monitoring crystallinity and surface roughness in real time. With the assistance of in-situ RTSE feedback we have achieved a photovoltaic energy conversion efficiency of 17% on an Al-backed p-type float-zone c-Si wafer. Open-circuit voltages above 650 mV indicate excellent passivation of the c-Si surface by the a-Si:H intrinsic layer. We have used RTSE to obtain information on the degree of crystallinity and the electronic and optical properties of films as a function of deposition conditions. RTSE has indirectly indicated the persistence of a hydrogen layer at the interface between the a-Si:H layer and the crystal silicon substrate. Absorption spectra determined by RTSE have provided guidance in device optimization.We are also applying in-situ RTSE to study the dynamics of HWCVD growth of epi-Si. The goal of this work is to develop low-temperature methods for growing 2-10 μmthick layers of c-Si on c-Si seed layers on glass for solar cell applications. This study presents unique challenges for RTSE, as perfect epitaxial growth of c-Si on a c-Si wafer would produce no change at all in the RTSE spectra. We have found that by monitoring the pseudo-dielectric function in real time during growth we gain immediate feedback on the breakdown of epi-Si growth. Post-deposition analysis of the RTSE data provides quantitative information on the percent of c-Si and a-Si versus film thickness. The RTSE analysis has been confirmed by cross sectional TEM. Based on the rapid feedback provided by RTSE we have surpassed the previous HWCVD maxiumum of 200 nm of epi-Si growth, achieving a maximum thickness of 500 nm of epi-Si. TEM analysis has shown that micron-sized areas of these films achieve 1000 nm of epi-Si thickness.

A fuzzy-nets-based in-process surface roughness prediction system in turning operations

A Fuzzy-Nets-based in-process surface roughness prediction (FISRP) system was developed to predict surface roughness in turning operations in a real time fashion. The input variables of the FISRP system were machining parameters, such as feed rate, spindle speed, depth of cut, and machining vibration per revolution. An accelerometer was employed to gather real-time vibration signals. Two groups of data were collected for two cutter bits with nose radii of 0.016 and 0.031 inches, respectively. Fuzzy nets theory was implemented to use the experimental data in developing the system for real-time prediction. The fuzzy nets theory is a five-step learning procedure for developing a knowledge base to predict surface roughness in real time. This FISRP system was tested to have an average prediction accuracy of 95.70%.

Materials and Interface Optimization of Heterojunction Silicon (HIT) Solar Cells Using in-situ Real-Time Spectroscopic Ellipsometry

ABSTRACTWe have applied real-time spectroscopic ellipsometry (RTSE) as both an in-situ diagnostic and post-growth analysis tool for hydrogenated amorphous silicon (a-Si:H)/crystalline silicon (c-Si) heterojunction with intrinsic thin-layer (HIT) solar cells grown by hot-wire chemical vapor deposition. RTSE enables precise thickness control of the 5 to 25 nm layers used in these devices, as well as monitoring crystallinity and surface roughness in real time. Utilizing RTSE feedback, but without extensive optimization, we have achieved a photovoltaic energy conversion efficiency of 14.1% on an Al-backed p-type Czochralski c-Si wafer coated with thin i and n layers on the front. Open-circuit voltages above 620 mV indicate effective passivation of the c-Si surface by the a-Si:H intrinsic layer. Lifetime measurements using resonant coupled photoconductive decay indicate that surface recombination velocities can approach 1 cm/s. RTSE and transmission electron microscopy show that the intrinsic a-Si:H i-layers grow as a mixture of amorphous and nano-crystalline silicon.

Effects of alloying elements on cobalt silicide formation

Alloying elements can substantially affect the formation of cobalt silicide. A comprehensive study of phase formation was performed on 23 Co alloys with alloying element concentrations ranging from 1 at.% up to 20 at.%. Using in-situ characterization techniques in which x-ray diffraction (XRD) and elastic-light scattering are monitored simultaneously, we follow the formation of the silicide phases and the associated variation in surface roughness in real time during rapid thermal annealing. For pure Co silicide, we detect the formation of all stable silicide phases (Co2Si, CoSi, and CoSi2) as well as abnormal grain growth in the Co film and thermal degradation of the silicide layer at high temperatures. The effect of the various additives on phase formation was determined. The roughness of the interface was also measured using grazing incidence x-ray reflectivity (GIXR). We show that by selecting an alloy with a specific composition, we can change the phase-formation temperatures and modify the final CoSi2 film texture and roughness.

A new technique to characterize and predict laser cut striations

The quality of a laser-made cut is of the utmost importance in laser processing. Any improvement in this area would be of considerable significance, in that it would lead to an elimination of post-machining operations. Currently the mechanisms governing the laser cutting process are not fully understood, partially due to the fact that laser cutting is a highly complex thermal process. It is the aim of the authors therefore to critically investigate the dynamic phenomena occurring within the cutting front, viz. the formation of striations, and the effect they have on the resulting cutting quality. A new technique for determining the frequency of the striations formed and the depth of the periodic structure has been developed. This is the first real attempt at accurately determining this most important quality index, surface roughness. Auxiliary information such as kerf width can also be ascertained. This leads to a more complete characterization of laser cutting quality. Results have shown that both quality indices correlate well with those actually obtained. The conceptual model developed supports the sideways burning theory for the formation of striations. It is argued that more than one mechanism for stria formation could exist and, as cutting conditions change, a move from one predominant mechanism to another could occur. This technique can be used in conjunction with theoretical models undertaken previously, whereby prediction of expected cut quality prior to machine operation will be possible. This has the ability of reducing set-up times involving parameter tuning, and leads to an optimized starting solution. The feasibility of detecting striation frequency on-line is currently being assessed through different sensing techniques. This will result in direct real-time surface roughness prediction and monitoring. Results will be published shortly.

Electronic system for surface roughness measurements in turning

Abstract An electronic-optical method of determining surface roughness in real time during finish turning is presented. The electronic design makes it possible to store pixel information from a CCD video image device at selective rather than continuous intervals. Selective triggering allows a direct correlation between electronically and mechanically acquired measurements at the same point on a turned surface. The electronics are interfaced to an IBM compatible computer and an optical encoder which is used to measure the speed of the lathe's spindle drive. Extensive experimentation has been undertaken and the results indicate that the electronic configuration is superior to existing mechanical instrumentation.