Statistical Process Control Techniques Research Articles

Integrated APC-controlled SPC monitoring chart for quality improvement

There are two widely used process control techniques for the reduction of process output variability. The first technique is automatic process control (APC) that adjusts the process using information about its current level or deviation from a desired target. The APC actions for process adjustments are achieved by the minimum mean square estimate (MMSE) controller or proportional, integral and derivative (PID) controller which minimize the output deviations from the quality target. The MMSE controller is optimal in terms of minimizing mean squared residual errors when the model and its parameters are exactly known. Whereas the PID controller is very efficient and also robust against non-stationarity due to the fact that it can continuously adjust the process whenever the data is auto-correlated. The second technique is statistical process control (SPC) which utilizes control charts. The goal of SPC for improving quality is to monitor and detect process variability, so that the special causes of the process shifting are investigated. While SPC has been successfully used in industry for identifying and eliminating the assignable cause of variations, APC techniques are widely employed in the continuous process industry to reduce common cause variations. For an improved performance of the process for the industry practitioner, both the monitoring and the adjustments of process are needed to receive the full benefit of each approach. Recently, integration of APC and SPC successfully resulted in the reduction of process output variability and improved process efficiency. In this paper, we integrate SPC and APC for various types of industries including liquefied natural gas (LNG) processes. By applying both the statistical process control and the advanced process control to a process we can dramatically improve the quality of process output. Both statistical process control and automatic process control techniques have been widely applied in industry to detect causes of variability by monitoring the key variables in the process. We investigate both techniques, their integration and methods for shift detection in the process for the monitoring of a process.

CUSUM charts for monitoring clinical practice quality using primary care prescribing data: a case study of an initiative to encourage generic prescribing of proton pump inhibitors

Expectations on organizations to monitor quality of care are growing. Whilst relevant data are increasingly becoming available it is, in many cases, difficult to distinguish real effects from background variation. Here, the potential usefulness of cumulative sum (CUSUM) charts for monitoring the use of medicines is explored through a case study of an initiative to encourage the prescribing of lowest cost proton pump inhibitors (PPI) in the context of implementation of national guidelines for the management of dyspepsia. This was a longitudinal study involving analysis of routinely collected prescribing data, set in all 12 primary care trusts (PCT) in the North East Strategic Health Authority. In it, comparison (by subtraction) of the time-series of the percentage of generic PPI prescription items for Gateshead with the mean of the other 11 PCTs was used to reduce both variation and bias. This was followed by the construction of a CUSUM chart displaying the effect of the Gateshead initiative. The simple process of comparison was very successful both in removing extraneous trends and reducing background variation, and the CUSUM highly effective for displaying the evidence for the hypothesized step-change in prescribing behaviour consequent on the Gateshead initiative. The effectiveness of the CUSUM here is strongly linked to the success of the preliminary comparison step. CUSUM, a statistical process control technique, has already been tested as a tool for interpreting hospital and general practice mortality rates. Here, its potential for more general applications in quality monitoring is demonstrated using routinely collected prescribing data. Such data contain valuable information about changes in clinical practice and CUSUM charts, when coupled with the idea of removing time trends and extraneous variation by reference to average behaviour, can provide a simple but effective technique for extracting it.

A METHOD OF DETERMINING ACCURACY AND PRECISION FOR DOSIMETER SYSTEMS USING ACCREDITATION DATA

A study of the uncertainty of dosimeter results is required by the national accreditation programs for each dosimeter model for which accreditation is sought. Typically, the methods used to determine uncertainty have included the partial differentiation method described in the U.S. Guide to Uncertainty in Measurements or the use of Monte Carlo techniques and probability distribution functions to generate simulated dose results. Each of these techniques has particular strengths and should be employed when the areas of uncertainty are required to be understood in detail. However, the uncertainty of dosimeter results can also be determined using a Model II One-Way Analysis of Variance technique and accreditation testing data. The strengths of the technique include (1) the method is straightforward and the data are provided under accreditation testing and (2) the method provides additional data for the analysis of long-term uncertainty using Statistical Process Control (SPC) techniques. The use of SPC to compare variances and standard deviations over time is described well in other areas and is not discussed in detail in this paper. The application of Analysis of Variance to historic testing data indicated that the accuracy in a representative dosimetry system (Panasonic® Model UD-802) was 8.2%, 5.1%, and 4.8% and the expanded uncertainties at the 95% confidence level were 10.7%, 14.9%, and 15.2% for the Accident, Protection Level-Shallow, and Protection Level-Deep test categories in the Department of Energy Laboratory Accreditation Program, respectively. The 95% level of confidence ranges were (0.98 to 1.19), (0.90 to 1.20), and (0.90 to 1.20) for the three groupings of test categories, respectively.

The Utilization of Six Sigma and Statistical Process Control Techniques in Surgical Quality Improvement

Surgeons have been slow to incorporate industrial reliability techniques. Process control methods were applied to surgeon waiting time between cases, and to length of stay (LOS) after colon surgery. Waiting times between surgeries were evaluated by auditing the operating room records of a single hospital over a 1-month period. The medical records of 628 patients undergoing colon surgery over a 5-year period were reviewed. The average surgeon wait time between cases was 53 min, and the busiest surgeon spent 291/2 hr in 1 month waiting between surgeries. Process control charting demonstrated poor overall control of the room turnover process. Average LOS after colon resection also demonstrated very poor control. Mean LOS was 10 days. Weibull's conditional analysis revealed a conditional LOS of 9.83 days. Serious process management problems were identified in both analyses. These process issues are both expensive and adversely affect the quality of service offered by the institution. Process control mechanisms were suggested or implemented to improve these surgical processes. Industrial reliability and quality management tools can easily and effectively identify process control problems that occur on surgical services.

A Statistical Wavelet-Based Process for Systems Catastrophic Failure Precursor Detection


 
 
 As a consequence of the recent deregulation in the electrical power production industry, new private entrepreneurs with no prior experience in the power plant operation have entered into the power generation business. They hedge the business risks by outsourcing the operation and maintenance activities to third party service providers with whom they share risks/rewards of plant performance. The main maintenance providers are the original equipment manufacturers, who are responsible for the majority of the cost associated with unplanned outages. With the cost-benefit of preventing such unplanned outage as a gas turbine compressor failure hovering around the twenty million dollars mark, techniques for detecting failure precursors to avoid or limit the number of systems catastrophic failure are necessary. In this paper, a methodical process is proposed to detect precursory events that lead to catastrophic systems failure. The wavelet packet transform is used to perform multi-resolution analysis of gas turbines health, condition and vibration sensors data to extract their signal features. Then the probabilistic principal component analysis is utilized to fuse them into a few uncorrelated variables. Next a one-dimensional signal representing the multi-variables data is computed. After that the statistical process control techniques is applied to set the anomaly threshold. Finally, a Bayesian hypothesis testing method is applied for abnormality detection to the monitored signal. As a proof of concept, the proposed process is successfully applied to a gas turbine compressor failure precursor detection.
 
 

Robots with a Social Memory

Robots with a Social Memory

Empirical Procedure for Modeling the Variation Transmission in a Manufacturing Process

ABSTRACT This article describes a method for obtaining a variation transmission model in a manufacturing process. The method uses regression analysis to obtain models that relate quality characteristics to process variables, statistical process control techniques to estimate the variance of the variables, and analysis of variance to estimate the variance components from the observed data. This method can be applied in situations where the process is not always under control and in cases where not all the process variables can be measured directly. The model also includes the measurement errors both of the quality characteristic and of the process variables. Although the method can be applied generally, industrial tests have been conducted to confirm its suitability for a process of manufacturing ceramic tiles. Findings suggest that the method may be appropriate for guiding improvements to the process, although further research would be needed for a more rigorous validation of the procedure.

Statistical estimation of variation transmission model in a manufacturing process

This article describes a method for obtaining a variation transmission model in a multi-stage manufacturing process in situations in which the characteristic that defines the quality of the product is an independent variable which variation is the consequence of the one generated and transmitted through the different process stages. The method uses regression analysis to obtain models that relate the quality characteristic to process variables, statistical process control techniques to estimate the variance of the variables and the analysis of variance to estimate the variance components from the observed data and to verify that data come from a stable process. This method can be applied to processes where the quality characteristics are different at each stage (a usual situation in chemical processes) and in cases where not all the process variables can be measured directly. The model also includes the measurement errors both of the quality characteristic and of the process variables. The new proposed approach has been applied to validate its suitability in a ceramic tiles manufacturing process.

LRProb control chart based on logistic regression for monitoring mean shifts of auto-correlated manufacturing processes

With the growth of automation in manufacturing, process quality characteristics are being measured at higher rates and data is more likely to be auto-correlated. Traditional statistical process control (SPC) techniques of control charting are not applicable in many process industries because process parameters are highly auto-correlated. Several attempts such as some time series based control charts have been made in the previous years to extend traditional SPC techniques. However, these extensions pose some serious limitations for monitoring the process mean shifts. These charts require that a suitable model has been identified for the time series of process observations before residuals can be obtained. In this paper, a logistic regression (LR)-based process monitoring model is proposed for enhancing the monitoring of processes. It is capable of providing a comprehensible and quantitative assessment value for the current process state, which is achieved by the event occurrence probability calculation of LR. Based on these probability values over the time series, a novel chart: LRProb chart, is developed for monitoring and visualising process changes. The aim of this research is to analyse the performance of the LRProb chart under the assumption that only a small number of predictable abnormal patterns are available. To such aim, the performance of the LRProb chart is evaluated on two real-world industrial cases and simulated processes. Given the simplicity, visualisation and quantification of the proposed LRProb chart, this approach is proved from the experiments to be a feasible alternative for quality monitoring in the case of auto-correlated process data.

Decreasing Extremes in Patient Waiting Time

We present a conceptual framework for approaching reducing excessive patient wait time in an outpatient setting. We hypothesized that statistical process control techniques can be used to identify extremes in waiting time; root cause analysis can be used to identify specific delay causes; and minimizing the contribution of the root causes will lead to an improvement in system performance. We conducted a prospective study of waiting times in a private outpatient clinic providing high-risk obstetrical care. The baseline period consisted of 55 clinic sessions, and the intervention period consisted of 101 clinic sessions. Mean waiting time was prolonged during 9 (16.4%) baseline clinic sessions. The root cause analysis determined that appointment schedule, physician tardiness, and patient complexity contributed to clinic delays. After making changes to minimize root causes, there was a significant reduction in prolonged waiting times (16.4% vs 4.9%, Yates chi(2) = 4.37, P = .037); a significant decrease in mean waiting time (32.7 +/- 23.6 minutes vs 29.3 +/- 21.2 minutes, t = 3.42, P < .001); and a significant improvement in the waiting time distribution (Kruskal-Wallis test of homogeneity, P = .003). Our methodology was successful in identifying and reducing factors associated with prolonged wait times. However, although system operation was improved, as defined by a decrease in the occurrence of excessive clinic delays, effecting a large and sustained decrease in patient waiting times was challenging.

An effective corner detection method using subpixel edge detector and Gaussian filter

PurposeThe purpose of this paper is to develop an effective and reliable corner detection algorithm so as to extract all the desired corners successfully. In addition, the influences of edge detection method as well as smoothing technique on the overall performance of corner detection techniques are investigated.Design/methodology/approachIn this paper, an effective corner detection algorithm based on subpixel edge detector and Gaussian filter is presented. First, a subpixel accuracy edge detector is used rather than a pixel accuracy edge detector to detect edges. Second, B‐splines approximation technique is used to eliminate the staircase effect of a digital curve. Third, curvature curve derived from the edges is smoothed by a Gaussian filter. Finally, statistical process control technique is applied to detect vertices.FindingsThe results show that spatial‐moment outperforms chain code as an edge detector. Furthermore, the Gaussian filter should be used to smooth curvature curve instead of smoothing the profile of an object, because the former provides greater impact on the corner detection results.Originality/valueIn addition to object recognition, motion tracking and obstacle avoidance, the proposed method also has many important engineering and manufacturing applications such as dimensional measuring, reverse engineering, and machine vision‐based computer numerical control (CNC) machining of polygonal sheet metal parts.

Quality control methods at a hospital

Healthcare providers need the information contained in patient records to provide high-quality services. To be effective, patient record assembly must be completed in a timely manner. This study aims to analyse the medical records assembly process for a hospital in Southeastern United States having difficulty meeting standard completion times established by the Joint Commission on the Accreditation of Healthcare Organization. Several quality improvement tools were used to evaluate and improve the assembly process. As a result of the study, a new procedure was implemented. Consequently, the hospital reduced the time required to assemble medical records, thereby improving efficiency and effectiveness. There are hopes to further improve the process. The study provides guidance on how statistical process control techniques can be applied to improve hospital services. The techniques employed can be used to analyze and improve any process. However, results are limited to improving medical record assembly processes at one particular hospital. Past studies considered the application of various statistical process control techniques for improving healthcare quality. The study extends research by employing process improvement efforts to understand and develop medical record assembly in a regional hospital via process flow diagramming and control charts.

Integrating support vector machine and genetic algorithm to implement dynamic wafer quality prediction system

The semiconductor and thin-film transistor liquid crystal display (TFT-LCD) industries are currently two of the most important high-tech industries in Taiwan and occupy over 40% of the global market share. Moreover, these two industries need huge investments in manufacture production equipments (PE) and own a large production scale of the global market. Therefore, how to increase the processing quality of PE to raise the production capacity has become an important issue. The statistical process control technique is usually adopted to monitor the important process parameters in the current Fabs. Furthermore, a routine check for machine or a predictive maintenance policy is generally applied to enhance the stability of process and improve yields. However, manufacturing system cannot be obtained online quality measurements during the manufacturing process. When the abnormal conditions occur, it will cause a large number of scrapped substrates and the costs will be seriously raised. In this research, a virtual metrology (VM) system is proposed to overcome those mentioned problems. It not only fulfills real-time quality measurement of each wafer, but also detects the performance degradation of the corresponding machines from the information of manufacturing processes. This paper makes four critical contributions: (1) The more principal component analysis (PCA) we used, the higher the accuracy we obtained by the kernel function approach; (2) the more support vector data description (SVDD) we used, the higher the accuracy we obtain in novelty detection module; (3) our empirical results show that genetic algorithm (GA) and incremental learning methods increase the training/learning of support vector machine (SVM) model; and (4) by developing a wafer quality prediction model, the SVM approach obtains better prediction accuracy than the radial basis function neural network (RBFN) and back-propagation neural network (BPNN) approaches. Therefore, this paper proposes that the artificial intelligent (AI) approach could be a more suitable methodology than traditional statistics for predicting the potential scrapped substrates risk of semiconductor manufacturing companies.

Statistical management of an instrument setpoint drift based on plant specific as-found/as-left data

Abstract Instrument setpoints are affected by many sources of uncertainty. ANSI/ISA-S67.04.01-2000 ( ISA, 2000a ) and ISA-RP67.04.02-2000 ( ISA, 2000b ) provide a basis for establishing safety-related instrument setpoints under several uncertainty terms including a random drift. However, the current methodologies and the recommended practices for a setpoint drift analysis need to be modified in order to control plant specific setpoint drift of an instrumentation device, especially when there are many ties in the measurements. In this paper, we propose an extended procedure supplemented with non-parametric and graphic tools for managing the setpoint drift, based on the plant-specific as-found/as-left data. The proposed procedure has three major advantages. First, supplementing non-parametric statistical methods can handle non-standard measurement data even when they have many ties. Second, adopting statistical graphic tools facilitates identifying data characteristics. Third, applying statistical process control (SPC) techniques may provide plant staff with an intuitive way of managing the instrumentation setpoint drift for a long term.

A comparison of robust alternatives to Hotelling’s T 2 control chart

Control charts are one of the widest used techniques in statistical process control. In Phase I, historical observations are analysed in order to construct a control chart. Because of the existence of multiple outliers that are undetected by control charts such as Hotelling’s T 2 due to the masking effect, robust alternatives to Hotelling’s T 2 have been developed based on minimum volume ellipsoid (MVE) estimators, minimum covariance determinant (MCD) estimators, reweighted MCD estimators or trimmed estimators. In this paper, we use a simulation study to analyse the performance of each alternative in various situations and offer guidance for the correct use of each estimator.

Practical guidelines for applying statistical process control to blood component production

Legislation, guidelines and recommendations for blood components related to statistical process control (SPC) and the selection of a quality monitoring (QM) sampling regimen are subject to misinterpretation and lack practical guidance on implementation. The aim of this article is: to review and interpret applicable European legislation and guidelines and to develop an SPC strategy that meets these requirements; and to provide practical guidance on the selection and application of appropriate techniques and the interpretation of resultant blood component QM data. A methodology is presented which utilizes: an algorithm to select an appropriate quality-monitoring strategy for the blood component parameter under consideration; a range of straightforward, validated SPC techniques for variable data and an assessment of process capability (Cpk) and blood component parameter 'criticality' to determine the sampling regimen. The methodology was applied to routine National Health Service Blood and Transplant (NHSBT) blood component data for 2005-2006. Cpk values ranged from 0.22 to >3 and their predicted non-conformance rates were close to those observed (23 to <0.001%). Required sample size ranged from 0.01 to 10%. Chosen techniques identified significant deviation from 'as validated' performance within an appropriate time-scale. Thus the methodology was straightforward to apply and prompted the choice of a clinically and operationally appropriate sampling regimen and analysis for each blood component parameter. This evidence-based, targeted use of SPC for blood component monitoring provides an essential focus on processes with a low capability in achieving their specifications.

Applications of population data analysis in on-farm dairy trials1,2

With appropriate management controls and statistical designs, on-farm trials are an increasingly valuable research tool. On-farm trials can speed up technology adoption, particularly with those studies requiring large numbers of animals. Useful designs include longitudinal (pen vs. pen) trials, in which pen is the experimental unit, and crossover or switchback designs, in which treatments are imposed on a schedule over 1 or more experimental groups. A paired-herd design has been used, in which herds are the experimental units in a crossover trial. Others have published similar studies, including a multisite crossover design that used 35 dairy farms to compare milk responses with a protein source by using individual cow records to evaluate differences in milk production. Recently, statistical process control (SPC) techniques have been used to evaluate management changes by using repeated measures on the farm. Although a drawback to SPC may be the lack of traditional statistics to test differences (i.e., the lack of a control group), standard run rules are used to demonstrate with statistical certainty that a process or variable has changed, or to characterize a seasonal change. With SPC, the inference is limited to the herd or group of animals being monitored. Meta-analysis techniques are powerful tools used to combine results from many similar trials in which the response of interest is either small (i.e., continuous variables) or of low frequency (i.e., discrete variables). Meta-analysis can be used to segment a database so as to validate and compare trial methods or to investigate publication bias. Additional design concerns for reproduction studies include the need for adequate numbers of observations and planning for the lag time between an experimental treatment and response measurement (e.g., confirmation of pregnancy).

Multivariate statistical monitoring of multiphase two-dimensional dynamic batch processes

To ensure product quality and operation safety, multivariate statistical process control (MSPC) techniques have been applied to batch process monitoring. Batch processes have several important features which should be taken into consideration in statistical monitoring, such as two-dimensional (2D) dynamics along both time and batch axes and multiple operation phases within a batch. In this paper, a multiphase two-dimensional dynamic PCA (2D-DPCA) method is proposed to deal with these characteristics simultaneously. An iterative procedure is designed for phase division, region of support (ROS) determination for each phase and phase 2D-DPCA modeling. The transition regions from phase to phase are also identified and modeled. Then, the phase and transition 2D-DPCA models are utilized in online monitoring. A three-water-tank process with 2D dynamics and a simulated two-phase batch process are used as case studies, to show the effectiveness of the proposed method.

Improving service delivery in a county health department WIC clinic: an application of statistical process control techniques.

Local public health agencies are challenged to continually improve service delivery, yet they frequently operate with constrained resources. Quality improvement methods and techniques such as statistical process control are commonly used in other industries, and they have recently been proposed as a means of improving service delivery and performance in public health settings. We analyzed a quality improvement project undertaken at a local Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) clinic to reduce waiting times and improve client satisfaction with a walk-in nutrition education service. We used statistical process control techniques to evaluate initial process performance, implement an intervention, and assess process improvements. We found that implementation of these techniques significantly reduced waiting time and improved clients' satisfaction with the WIC service.

RIDES: Robust Intrusion Detection System for IP-Based Ubiquitous Sensor Networks

The IP-based Ubiquitous Sensor Network (IP-USN) is an effort to build the “Internet of things”. By utilizing IP for low power networks, we can benefit from existing well established tools and technologies of IP networks. Along with many other unresolved issues, securing IP-USN is of great concern for researchers so that future market satisfaction and demands can be met. Without proper security measures, both reactive and proactive, it is hard to envisage an IP-USN realm. In this paper we present a design of an IDS (Intrusion Detection System) called RIDES (Robust Intrusion DEtection System) for IP-USN. RIDES is a hybrid intrusion detection system, which incorporates both Signature and Anomaly based intrusion detection components. For signature based intrusion detection this paper only discusses the implementation of distributed pattern matching algorithm with the help of signature-code, a dynamically created attack-signature identifier. Other aspects, such as creation of rules are not discussed. On the other hand, for anomaly based detection we propose a scoring classifier based on the SPC (Statistical Process Control) technique called CUSUM charts. We also investigate the settings and their effects on the performance of related parameters for both of the components.