Showing papers on "Process variable published in 1991"

Process control using neural network

Abstract: A control system and method for a continuous process in which a trained neural network predicts the value of an indirectly controlled process variable and the values of directly controlled process variables are changed to cause the predicted value to approach a desired value.

Experimental optimization of an anisotropic etching process for random texturization of silicon solar cells

Abstract: A multifactor experimental investigation of silicon surface texturing was conducted using aqueous potassium-hydroxide (KOH) solutions with isopropyl alcohol (IPA) added as a complexing agent. Czochralski, magnetic-Czochralski, and float-zone silicon wafers of different resistivities with both polished and lapped surfaces were included in the experiment. Process variables considered were solution temperature, time in solution, degree of mechanical mixing, KOH concentration, and IPA concentration. Using hemispherical reflectance as the primary gauge of success, process variables that resulted in an effective surface texture with reflectance less than 12% prior to antireflection coating were identified. Of particular interest was a low-temperature (70 degrees C) process with less than 2% concentration of both KOH and IPA and wide process variable tolerances. >

Statistical process control for air separation process

Abstract: Statistical process control system for an air-separation plant using off-line and on-line computer means to define control chart limits, trend thresholds and dynamic models for use with on-line process variable data for determination of statistical process control.

Real-time monitoring and control in plasma etching

Abstract: Process control strategies have been developed for plasma etching of silicon (Si) and silicon dioxide (SiO/sub 2/) in a CF/sub 4//O/sub 2/ plasma. The analysis considered four measured variables, four manipulated variables, and up to seven performance variables. Empirical input-output models were developed by regression analysis. Relative gain array analysis and singular value decomposition were used to select manipulated/process variable control loop pairings and to evaluate potential difficulties in control system performance. Singular value decomposition was also used to determine process/performance variable pairings. Block relative gain analysis of multivariable interactions in the process indicated that partial decoupling was necessary for adequate control, and this was verified by simulation. >

Adaptive controller in a process control system and a method therefor

Abstract: A process control system (1) for controlling a process variable (pv) of a process (10) includes an adaptive controller (22) tuned to the process. The process variable output signal of the process is coupled to a difference device (23) and also coupled to the adaptive controller, thereby permitting the adaptive controller to monitor the process variable signal independently of an error signal (e) outputted from the difference device. Any changes detected in the process variable signal without a corresponding change in the setpoint signal inputted to the difference device initiates an adjustment in the tuning parameters of the adaptive controller thereby providing continual, optimal, control of the process.

Selection of process variables for die casting

Abstract: Selection of process variables is the phase of design for die casting in which the compatibility of part geometry and specifications, die variables, and machine setup are considered. In this research, a system is developed to help users select process variable values and find an acceptable setup point. The research covers topics on the methodology of the analysis, the system structure, and the modelling tools applied in the selection process. The analysis is divided into layers, a window layer and a variables layer. The window layer is used to find an operational window by defining limits on fill time, gate speed and final metal pressure. Heat transfer, solidification, and flow simulation are the essential tools used. The variables layer is used to obtain a set-up point by analyzing the die and machine simultaneously. The system is composed of five basic units; database, working memory, window layer, variables layer, and monitor. The window layer and variables layer are built from modules. Each module analyzes a single phenomena based on the process modelling scheme. Data is transferred between modules to consider interaction. Several new models were developed. These are (1) fill time and pressure, where pressure developed in the die cavity is used as a criterion for diecasting quality; (2) venting and vacuuming, which calculates volume of trapped air and relates it to the process variables; and (3) an optimization module, that maximizes flexibility of the setup point. Some of the models were validated experimentally using a shot end simulating unit. Analysis of diecastings collected from several industrial sources were performed to check the system and to validate its predictions. The results obtained are comparable to actual variable values in most cases.

Apparatus and method for determining gradients of process variables

Abstract: A method and an apparatus for operating directly with the program code of a process for determining the gradients of process variables of interest with respect to process parameters. Each process will contain one or more process variable declarations wherein each process variable declaration defines a process variable. There will also be one or more process function assignments associated with each process variable, wherein each process function assignment assigns a value to the associated process variable. One or more gradient variables will be introduced into the process for each process variable of interest, wherein each gradient variable corresponds to and represents the partial derivative, or gradient, of the process variable of interest with respect to a parameter whose value affects the value of the process variable of interest. One or more gradient function assignments will then be inserted into the process for each occurrence of a process function assignment associated with a process variable of interest. Each gradient function assignment will correspond to one of the gradient variables corresponding to the process variable of interest and will assign a value to the gradient variable.

Method for control of process conditions in a continuous alloy production process

Abstract: A method for controlling process conditions in a continuous skull nozzle process is provided wherein one or more process parameters are controlled to maintain the opera-tion of the process within a process window derived using integral solutions and expressed in terms of dimensionless parameters, the process window defining a range of values within which a steady state solidified skull will be main-tained, and outside of which a steady state solidified skull will not be maintained. A pressure differential between the inside of a crucible holding the melt and outside the cru-cible is one process parameter which is controlled to adjust the melt discharge flow rate to maintain operation within the process window.

Real-time monitoring and control of plasma etching

Abstract: Process control strategies have been developed for reactive ion etching of silicon and silicon dioxide in CFJO2 and CF/H2 plasmas. Four measured variables four manipulated variables and six performance variables were considered for both chemistries. Relative gain array analysis and singular value decomposition were used to select manipulated/process variable control loop pairings for feedback control and to evaluate potential difficulties in control system performance. Block relative gain analysis of multivariable interactions in the process indicated that single loop feedback control would be inadequate for control of both CF/H2 and CFJO2 etching which was subsequently verified by simulation. Multivariable control (partial decoupling) was much more effective in reducing dynamic fluctuations in the process variables. Closed loop simulations have demonstrated a quality of control sufficient for manufacturing purposes. Singular value decomposition was also used to determine which process variable (measured in real-time) correlated best with a given performance variable. Empirical input-output models developed by response surface methodology were used for the above analyses. For the CF/H2 system functions based on semi-theoretical reasoning of process variables were also investigated. For example the function of [F](DC Bias)25/[CF2] was found to be more indicative of process performance than the absolute values of these process variables. Unfortunately this work reveals the need for more real-time measurable process variables in order to achieve sufficient process performance.

An improved self-tuning controller

Abstract: An apparatus and method for automatically adjusting the control parameters of a self-tuning controller used to regulate a process having a measured process variable signal. Using the measured process variable signal, an error signal representing a closed-loop response of the process to an upset condition is generated. Local extrema of the error signal is measured and three successive amplitude values are selected to produce measured decay and overshoot characteristics of the error signal. The three successive amplitude values are selected such that the measured decay characteristic is greater than the overshoot characteristic. Based on the measured decay and overshoot characteristics at least one of the control parameters of the controller is automatically adjusted to improve the difference between one of the measured characteristics and a target characteristic.

Noninvasive sensors for in-situ process monitoring and control in advanced microelectronics manufacturing

Abstract: The combination of noninvasive in-situ monitoring sensors single-wafer processing modules vacuum-integrated cluster tools and computer-integrated manufacturing (CIM) can provide a suitable fabrication environment for flexible and high-yield advanced semiconductor device manufacturing. The use of in-situ sensors for monitoring of equipment process and wafer parameters results in increased equipment/process up-time reduced process and device parameter spread improved cluster tool reliability and functionality and reduced overall device manufacturing cycle time. This paper will present an overview of the main features and impact of noninvasive in-situ monitoring sensors for semiconductor device manufacturing applications. Specific examples will be presented for the use of critical sensors in conjunction with cluster tools for advanced CMOS device processing. A noninvasive temperature sensor will be presented which can monitor true wafer temperature via infrared (5. 35 jtm) pyrometery and laser-assisted real-time spectral wafer emissivity measurements. This sensor design eliminates any. temperature measurement errors caused by the heating lamp radiation and wafer emissivity variations. 1. SENSORS: MOTIVATIONS AND IMPACT Semiconductor chip manufacturing factories usually employ well-established statistical process control (SPC) techniques to minimize the process parameter deviations and to increase the device fabrication yield. The conventional fabrication environments rely on controlling a limited set of critical equipment and process parameters (e. g. process pressure gas flow rates substrate temperature RF power etc. ) however most of the significant wafer process and equipment parameters of interest are not monitored in real© (1991) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Process system identification

Abstract: PROCESS SYSTEM IDENTIFICATION Abstract of Disclosure A tool, and the method of making the tool, for process system identification that is based on the general purpose learning capabilities of neural networks. The tool and method can be used for a wide variety of system identification problems with little or no analytic effort. A neural network is trained using a process model to approximate a function which relates process input and output data to process parameter values. Once trained, the network can be used as a system identification tool. In principle, this approach can be used for linear or nonlinear processes, for open or closed loop identification, and for identifying any or all process parameters.

Two degrees of freedom type control system

Abstract: A system for controlling a process system subject to an external disturbance by adjusting a process variable output by the process system to a given target variable. The system includes a target variable filter unit for generating a computed target variable to perform a compensation control operation, a deviation operating unit for computing a deviation between the process variable and the computed target variable, a main control unit for performing at least proportional and integral control operations on the computed deviation between the process variable and the computed target variable to suppress fluctuation of the process variable produced by the external disturbance, and a target variable control unit for changing an input into the deviation operation unit from the computed target variable to the given target variable only when the deviation between the given target variable and the computed target variable computed by the target variable filter unit is less than a predetermined value.

A new approach to process parameter selection: Integration of quality

Abstract: This paper proposes an approach to select the process parameters incorporating quality. We used Taguchi's method of process design for parameter selection. The approach is illustrated using experimental data. Only one quality parameter, surface finish, is considered in the experiment.

Optimizing the Wave Soldering Process

Abstract: Wave soldering is a very effective technique for joining components to printed circuit boards (PCBs) in large quantities. A printed circuit assembly (PCA) consists of the printed circuit board, connectors, and all components. As in all metallic joining applications, there are three basic steps to the soldering process. The first step prepares the surface for soldering. The next step applies molten solder to the bottom side of the assembly. The final step consists of allowing the joint to solidify by cooling. Defects that result in joint reliability issues can result from all of these stages. The goal of wave soldering process optimization is the reduction of joint defects. Several wave soldering process hardware enhancements are available to reduce certain types of defects. The process parameters should also be optimized to truly capitalize on the machine investment. Some examples of process parameters are solder temperature, solder wave height, conveyor speed, and preheat temperature setpoints. The optimal wave soldering machine and wave soldering process depend closely on the PCA design intended for production. Both wave soldering process parameter identification and process hardware selection must be examined to optimize the total process and reduce solder joint defects.

Investigation of the neutral-solution etch process for refractive SOE antireflective surfaces

Abstract: Antireflection of optically clear glass used in photovoltaic concentrator refractive secondary optical elements (SOEs) was investigated using the neutral-solution etch process. Test coupons and SOEs made from barium zinc glass, which does not solarize under ultraviolet exposure, were successfully etched at the center-point process variable conditions at 87 degrees C and 24 hours. Reflectance of the plano-plano coupons dropped from 7.7% to 0.8%, with a corresponding increase in transmission from 91.7% to 98.5%. The etching process uses nonhydrofluoric relatively nontoxic chemicals in a low-cost process well-suited for use by photovoltaic system manufacturers during production. >

Optimal control of positive optical photoresist development

Abstract: This work considers the modeling and control of positive optical photoresist development. The control effort is applied to this final process step in order to compensate for deviations in previous irreversible steps. A new approach to line-width control is studied. State estimation, combining interferometric development measurements with a process model, is used to calculate the time at which developer breakthrough to the substrate occurs. Parameter identification is used to determine any sample-to-sample changes in the process model. The optimal control policy is shown to be bang-bang with switching at the final time. A model describing development after breakthrough is derived and used to calculate the final developer shutoff time, based upon an on-line identified development process parameter. The policy is tested experimentally on different batches of the same resist type. Results indicate that accurate control is achieved, despite changes in environmental conditions and unexpected transient disturbances.

Adaptive process control system for controlling fluid temperature especially

Abstract: PURPOSE: To perform an operation at a high speed and to control a process variable within an accurate limit range by measuring a variable before and after a control point inside a process and performing the control of the variable and the adaptation of a control function corresponding to the change of the variable. CONSTITUTION: Chemicals are made to flow through a temperature control point 12 inside an applicator part 14 for applying microwave energy from a microwave generator 16. The temperature of flow-in fluid is measured by temperature sensors 26 and 28 before and after the control point 12. The temperature of the control point 12 is controlled by controlling the microwave generator 16 corresponding to the temperature change of the temperature sensor 26 and the control function is adaptively controlled corresponding to the temperature change of the temperature sensor 28 when final named change indicates process change. In such a manner, temperature control 15 performed at a high speed.

Pultrusion process characterization

Abstract: Pultrusion is a process through which high-modulus, lightweight composite structural members such as beams, truss components, stiffeners, etc., are manufactured. The pultrusion process, though a well-developed processing art, lacks a fundamental scientific understanding. The objective here was to determine, both experimentally and analytically, the process parameters most important in characterizing and optimizing the pultrusion of uniaxial fibers. The effects of process parameter interactions were experimentally examined as a function of the pultruded product properties. A numerical description based on these experimental results was developed. An analytical model of the pultrusion process was also developed. The objective of the modeling effort was the formulation of a two-dimensional heat transfer model and development of solutions for the governing differential equations using the finite element method.

Method and device for control of automatic device

Abstract: PURPOSE:To facilitate the production of a program by specifying an object needed in an optional process with the process variable defined as an argument and controlling the process and the object with the relationship secured between them. CONSTITUTION:An FAC (flexible assembling center) system 10 comprises a robot 12, a parts supply system 14, a control unit 16, and an input/output device 18. The Nos are assigned to plural processes. Thus these processes are specified by the value of each process variable, and the correspondence between the processes and at most one object that is used in an optional process is stored as a data array where the process variable defined as an argument. Then the object needed in an optional process is specified with the process variable defined as an argument. Thus both the process and the object are controlled with the relationship secured between them by specifying an object needed in an optional process with the process variable defined as an argument. As a result, a control program is easily produced and maintained.

Temperature control systems and method of operation thereof

Abstract: A temperature control system incorporates a temperature sensor characterized by thermal inertia which causes the sensed temperature to lag the actual temperature during the transition from one operating temperature for the controlled environment to a different operating temperature. This control system controls temperature as a function of setpoint temperature and the value of a temperature control variable representing the sensed temperature. The value of the temperature control variable is set equal to the temperature sensed by the sensor during steady state operating conditions. In response to a change in setpoint temperature, the system adjusts the value of the temperature control variable to compensate for the thermal lag of the sensor.

Effect of Angular Errors in Part Registration for PC Board Assembly

Abstract: Automated printed circuit (pc) board assembly is a significant area today, affected by various critical factors. The effects of different assembly process parameter values and their variations (tolerances or errors), on the successful assembly of a part, needs to be properly understood to design an effective and efficient assembly system. Part registration before final assembly is an important phase in the assembly process. This paper analyzes the combined effect of angular and linear errors in the system on successful part registration.

Prediction of Temperature and Thermal Stress in Plasma Sprayed Coatings

Abstract: Plasma spraying is one of the prominent technologies for wear, corrosion and high temperature resistant coatings. The coating quality is very important to increase the performance of the components as well as to protect the outer surface of the component from external environment. The coating quality characteristics depend on many plasma process parameters. Among these parameters, distribution of temperatures and stress in substrate/coating having considerable effect on properties of plasma sprayed coating. Hence, it is necessary to study the mechanisms of stress development to predict and control the temperature and stresses. 1-D model is developed using COMSOL Multiphysics® software to predict the thermal history and stress distribution of substrate/coating during plasma spraying. The mathematical model includes plasma and particle heat flux as well as temperature dependent thermo-physical properties to predict the temperature distribution of substrate/coating by considering semi-infinite body approximation. The effect of coating thickness and stand-off distance are taken into account for simulation. From this study, it is expected that temperature and stress distribution of substrate/coating as well as the effect of coating thickness and stand-off distance on coating temperature can be predicted. Finally, the mathematical model will help to understand the mechanism of stress development within coating as well as to control the plasma process parameter to obtain good and required coating properties for suitable application in economical and effective manner.