Model-based control methods do not produce satisfactory results with the batch process of Czochralski (CZ) silicon monocrystalline complex nonlinearity, large delay, and time-varying dynamics. Therefore, this paper proposes a data-driven model-free adaptive iterative learning method (MFAILC) to achieve precision process. Firstly, improve accuracy model, novel deep model for crystal growth is established by combining stacked autoencoder (SAE) long short-term memory network (LSTM) extract...

The wake in an evacuated tube transportation system under choked flow is a complex, unstable, three-dimensional structure composed of trailing vortices, shear layers, shock waves, and their interactions, which may threaten the stability safety train operations. In this study, aerodynamic performance at 1000 km/h investigated using Improved Delayed Detached Eddy Simulation (IDDES) method with stress transport k-ω model. numerical algorithm mesh strategy are validated through in-pipe...

The accurate real-time prediction of the crystal quality index v/G is an important reference for monitoring growth status and process optimization adjustment semiconductor silicon single crystals. This paper proposes a data-driven indicator soft sensor model based on multi-timescale feature fusion to achieve effective v/G. Firstly, characteristics in Czochralski are analyzed. Secondly, broken down into several natural components using something called complete ensemble empirical mode...

The growth process of Czochralski (Cz) silicon single crystal is a dynamic time-varying system with nonlinearity, strong coupling, large hysteresis, and uncertain model. Traditional model-based control methods are difficult to achieve satisfactory effects, it ensure that the quality meets actual requirements. Therefore, from perspective data-driven modeling control, this paper proposes model predictive method for V/G soft-sensing measuring quality. First, because value measure obtain...

Photoacoustic waves generated at the tip of an optical fiber consist a compressive shock wave followed by tensile diffraction waves. These overlap along axis and form cloud cavitation bubbles. We demonstrate that shaping through micromachining alters number direction emitted clouds. Shock emission patterns from five distinctively shaped tips have been studied experimentally compared to linear propagation model. In particular, multiple generation strong tension away realized using modified...

The Czochralski (CZ) process is the core technology for producing semiconductor silicon monocrystalline (SMC), and it a complex batch process. However, crystal growth rate diameter, which are key quality indicators, difficult to detect directly online, offline calculation lags seriously, easily causes blind control. Therefore, this paper proposes data-driven mechanism-based hybrid model SMC variables prediction in CZ Firstly, JITL-SAE-ELM based on just-in-time learning (JITL) fine-tuning...

The growth process of silicon single crystal (SSC) is a typical batch production process, which has the characteristics size, variety, complex and intensive technology. In order to accurately control diameter thermal field temperature Czochralski (Cz) SSC ensure that actual requirements electronic-grade are met, this paper proposes an iterative learning-based predictive strategy Model Predictive Control - Iterative Learning Extended State Observer (MPC-ILC-ESO). consists two parts. time axis...

Nanosecond-pulsed laser ablation is often accompanied by adverse thermal effects such as oxidation, debris recast and burr formation. To reduce these effects, in this paper, the authors present underwater milling process using RSA-905 fine-grained aluminium target material for first time. The results show that channels up to 200 μm width, 700 depth bottom roughness around 1 µm Ra could be fabricated with reduced effects. By conducting multi- single-factor experiments, empirical models...

A new control strategy combing the Finite Element Method (FEM) with method is proposed to study of crystal growth process in this article. In strategy, diameter loop and temperature are designed respectively meet requirements on distribution. loop, a bio-heuristic model based PID controller diameter. considering undetectable characteristic distribution inside crystal, transformed into key monitoring point constraint gradient, reference adaptive achieve temperature. Simulation results show...

To improve crystal microdefect control in silicon single growth, this paper proposes a new method, which can not only obtain optimal process parameter track via numerical simulation analysis and optimization, but also realize simultaneous of the diameter micro-defect. Meanwhile, CGSim software is used to simulate influence height pulling rate changes on shape solid liquid interface V/G value during growth φ300mm cusp magnetic field. Then, data employed at isometric stage as well heater power...

The authors propose a method for rapid laser isolation of 12 μm-thick aluminum (Al) metallized polyethylene terephthalate (PET) film. patterning system employed cost-effective, nanosecond-pulsed master-oscillator power amplifier fiber laser. A dynamic focusing configuration was adopted as replacement conventional telecentric lens setup beam and delivery. ablation process caused thermal removal Al, the effects processing parameters on ablated channels' width, depth, edge quality, electrical...

In view of the previous Czochralski(Cz) monocrystalline silicon growth control method based on mechanism model, it is difficult to accurately reflect nonlinear and real-time variation characteristics crystal process. This paper proposes a data-driven iterative learning support vector regression, which combines diameter prediction model by regression (SVR) off-line training with then output key variable information in framework algorithm. Simulation results for process analysis show that...

Heat is the driving force of solid-liquid phase transformation in Czochralski silicon monocrystal growth process, and melt surface temperature directly affects quality monocrystal. Due to complex physical changes, multi-field multi-phase coupling, uncertain model time-varying nonlinear characteristics which makes mechanism difficult obtain. This paper presents a free adaptive discrete sliding mode control (MFASMC) method for control. Firstly, thermal system between heater power established...

Aiming at the complex nonlinear dynamic time-varying characteristics for Czochralski (Cz) silicon single crystal growth process and difficulty in modeling controlling diameter by conventional mechanisms, based on idea of data-driven control, this paper proposes an improved model-free sliding mode iterative learning control (MFA-SMILC) method. First, a model is established using extreme machine (ELM) with actual data; Then, compact-format linear (CFDL) data model, discrete algorithm used to...