Process Control Limits Research Articles

Quality management of head and neck patient treatments using statistical process control techniques.

The treatment, planning, simulation, and setup of radiotherapy patients contain many processes subject to errors involving both staff and equipment. Cone-beam-CT (CBCT) provides a final check of patient positioning and corrections based on this can be made prior to treatment delivery. Statistical Process Control (SPC) techniques are used in various industries for quality management and error mitigation. The utility of SPC techniques to monitor process and equipment changes in our Head and Neck patient treatments was assessed by application to CBCT results from a quality-focused longitudinal study. Individuals and moving range (XmR) as well as exponentially-weighted moving average (EWMA) techniques were explored. The SPC techniques were sensitive to process changes and trends over the 12 years of data collected. A reduction in the random component of patient setup errors needing correction was observed. Systematic components of error remained more stable. An uptick in both datasets was observed correlating with the COVID-19 pandemic. Process control limits for use in prospective process monitoring were established. Challenges that arose from using SPC techniques in a retrospective study are outlined.

Machine Learning-Enabled NIR Spectroscopy. Part 3: Hyperparameter by Design (HyD) Based ANN-MLP Optimization, Model Generalizability, and Model Transferability

Data variations, library changes, and poorly tuned hyperparameters can cause failures in data-driven modelling. In such scenarios, model drift, a gradual shift in model performance, can lead to inaccurate predictions. Monitoring and mitigating drift are vital to maintain model effectiveness. USFDA and ICH regulate pharmaceutical variation with scientific risk-based approaches. In this study, the hyperparameter optimization for the Artificial Neural Network Multilayer Perceptron (ANN-MLP) was investigated using open-source data. The design of experiments (DoE) approach in combination with target drift prediction and statistical process control (SPC) was employed to achieve this objective. First, pre-screening and optimization DoEs were conducted on lab-scale data, serving as internal validation data, to identify the design space and control space. The regression performance metrics were carefully monitored to ensure the right set of hyperparameters was selected, optimizing the modelling time and storage requirements. Before extending the analysis to external validation data, a drift analysis on the target variable was performed. This aimed to determine if the external data fell within the studied range or required retraining of the model. Although a drift was observed, the external data remained well within the range of the internal validation data. Subsequently, trend analysis and process monitoring for the mean absolute error of the active content were conducted. The combined use of DoE, drift analysis, and SPC enabled trend analysis, ensuring that both current and external validation data met acceptance criteria. Out-of-specification and process control limits were determined, providing valuable insights into the model’s performance and overall reliability. This comprehensive approach allowed for robust hyperparameter optimization and effective management of model lifecycle, crucial in achieving accurate and dependable predictions in various real-world applications.Graphical

Comparability study between downstream methodologies using physicochemical characteristics of hyperimmune serum for human use

This study compared two downstream methodologies of antivenom production and statistically evaluated the results of quality control between batches of Antiloxoscelic Serum (approved or not) of these methods. The methodologies present significant downstream differences related to the quality control results. The document analysis and statistical evaluation of the results were obtained using a control chart. All batches showed physical-chemical results within the specification limits, but not all downstream B results were within the process control limits. In all quantitative results analyzed (total protein, non-protein nitrogen, phenol, total solids, pH, chloride), it was possible to verify that the downstream B samples did not meet the process control criteria established by the control chart. The conclusion is that method B is outside the control of the stipulated production process.

Astigmatic electron beam propagation

An astigmatic electron beam with an ellipticity ratio of approximately 2:1 was examined using the Enhanced Modified Faraday Cup (EMFC) diagnostic to measure its properties and show how the power density distribution varies near the beam waist. Results show that the beam has two power density peaks, one on either side of the beam crossover point, resulting from different focal distances of the major and minor axes of the elliptical beam shape. Quantification of the beam properties was used to establish a depth-of-field where the beam diameter and peak power densities are relatively constant. For the 1 kW beam studied here, the depth-of-field measured approximately ±5 mm from the beam crossover point, illustrating how electron beam diagnostics can be used to identify process control limits for repeatable and reliable welds under non-ideal electron beam power density distribution conditions.

Robust Point M-Scale Estimators for Monitoring the Mean of EWMA Control Chart under the Normal Process

The Exponentially Weighted Moving Average (EWMA) control chart is well known in statistical process control (SPC) for identifying small shifts in process parameters. In general, the EWMA control chart is focused on data set estimates of the process characteristics under study. In this research, for observing the process mean, we modify the control limits of the EWMA control chart by using robust point M-scale estimators under the normal process. We compare five robust point M-scale estimators proposed in SPC literature, namely Sn, Qn, MAD, Tauτ^ and FQn. The traditional control limits are evaluated and distinctions are also made based on true process control limits. The control charts performances under study are estimated via a Monte-Carlo simulation analysis by calculating the interval widths (IWs) and optimal out-of-control average run length (ARL1). From the simulation studies results, it is seen that although all proposed and traditional control limits are closely approximating the true limits of the process but the proposed control limits based on robust FQn, Tauτ^ and MAD estimators are comparatively more stringent for detecting the small shifts in process mean. Moreover the results demonstrates that at optimal size of interval L, the control charts based on MAD, Tauτ^ as well as FQn estimators demonstrate higher sensitivity in comparison to the traditional control chart based on sample standard deviation (S) and proposed control charts based on QnandSn estimators in terms of having smaller ARL1 values. The OC curves of ARL1 indicate similar findings. For the purposes of illustration, a real-life data set is analyzed that also confirms our conclusions from the simulation analysis to a certain degree.

Automated search of process control limits for fault detection in time series data

Manually defined control limits remain a common strategy for quality control in manufacturing due to their ease of deployment on the shop floor compared to more advanced data analysis approaches. Despite their continued importance, there is no systematic method of defining these control limits. However, sub-optimal control limits can lead to undetected faults or cause unnecessary interruption to production. This manuscript presents an algorithm that systematizes this manual process into an efficient search task. We conceptualized the search task as a sequence of sub-problems that are based on the conventional steps taken by process experts when defining control limits. This algorithm can be integrated into an expert tool for shop floor personnel to automate the definition of control limits in annotated time series data. We demonstrate the efficacy of the control limits found by our algorithm by comparing them to those manually defined by process experts in real-world process data from the automotive industry. Furthermore, we show that our algorithm generalizes to traditional time series classification problems and achieves state-of-the-art performance on selected benchmark datasets. Our work is the first effort in automating the otherwise manual definition of control limits for fault detection.

Setting Process Control Limits for Rounded-Off Measurements

Setting Process Control Limits for Rounded-Off Measurements

Application of statistical process control (SPC) in the volume variation of a volatile liquid fuel

Statistical Process Control makes it possible to monitor the quality characteristics of interest, ensuring their maintenance within pre-established limits and indicating when to adopt corrective and improvement actions. The problem addressed in this work The large variation in the net mass of a volatile fuel derived from petroleum (gasoline A) stored in atmospheric tanks at a Fuel Distributor in Manaus-AM. Data on the variation in the volume of gasoline A were studied in the process of transferring between tanks. The data for the study were obtained from the Distributor's Net Cargo Tonnage Certificates (CNCT) and Quality Certificates (QC) and tabulated in an electronic spreadsheet using the Microsoft Excel software from Windows, in which the graphs of the volume variation values, specification limits (LSL and USL) and process control limits (LCL and UCL) were also plotted. All results point to an objective way of demonstrating the high degree of safety required in this process, the studied process control parameters that form the variation of the ambient volume and volume at 20 °C of gasoline A, remained within the limits of control and as for the capacity indexes it was concluded that the process produced high capacity indexes.

Drift Detection Analytics for IoT Sensors

The Industrial Internet of Things (IoT) has an unique opportunity to have a greater impact on the manufacturing sector. Monitoring the health of the expensive equipments in the factory is critical for the business and the opportunities where IoT can be truly transformational. Most often the industries operates on a primitive way of monitoring these equipments on a Statistical Process Control (SPC) Limits. The major flaw in this monitoring system is unable to detect drifts within the static limit and upon triggering of the limits, it is usually too late for the team in the manufacturing floor to take preventive actions before the system goes down. In this paper, we developed a generic model for detecting drifts and identifying potential outliers. The model uses a double linear regression method to identify both aggressive and progressive drift, as well as adjusted boxplot method to detect outliers in both symmetric and skewed distributions. Unlike conventional drift detection approaches, this model has low computational complexity and can be applied to both batch and stream data. This paper will also introduce the infrastructure and architecture on enabling near real-time analytics using the IoT platform and streaming cluster, which reduces the data latency available for analysis to 10 minutes. Enabling real-time monitoring allows the end users to react to the alarms in a timely manner. This system has proven that it is able to provide early detection before the impact was observed as compared to the existing system. Manufacturing operation team could establish a new business process to respond to the early drift alarms by using quality shift left approach.

Development and Scale-Up of Diversion Strategy for Twin Screw Granulation in Continuous Manufacturing

Successful implementation of Continuous Manufacturing technology requires real time product quality monitoring that can result into rejection strategies for material manufactured outside process control limits. In a twin screw granulation process, parameters like water content, powder feed rate, and granulator screw speed can influence granule quality. Deviations in any of these parameters from the set-point may affect granule quality. Having a sound diversion strategy in place can help divert these implicated granules to waste. Residence time distribution experiments were conducted on a 16-mm Thermo Fisher twin screw granulator (TSG) for a range of process parameters, and the data was modelled to predict the needed diversion time as a function of process parameters. Scale-up from the 16-mm to 24-mm granulator was evaluated and data was found to scale based on mass per unit volume of granulator (channel fill), thus enabling 16-mm data to scale to 24-mm. The diversion strategy proposed is based on utilizing a wash out curve derived from residence time distribution to quantify the maximum concentration of implicated material that could be present in the next downstream unit operation(s) (e.g. a fluid bed dryer) and ensuring it is less than a suitable threshold to prevent product quality impact.

IMPROVE CONTROL AMIDST DIE SHRINKAGE AND 3D PACKAGE COMPLICATION

Abstract Die sizes continue to shrink and packaging technologies continue to evolve, but the common thread for all of them is the need for increased precision and tighter process control limits to achieve final package yield. Nearly all packaging technologies require connections in the third dimension, above or below the die, thus adding, quite literally, a new dimension to inspection and metrology requirements. Increased focus on reliability for automotive, health care and even mobile electronics is driving the need for improved process control solutions. The combination of higher packaging complexity and the need for improved reliability are driving changes to the requirements around inspection and metrology. Vertical integration continues to grow at a pervasive rate and the need for improved process control in the third dimension is growing rapidly in order to ensure reliability. Vertical integration is designed into nearly all packaging forms, including TSV, RDL, WLP, Fan-in, Fan-out, with a focus on continued increase in the number of I/Os, the pitch of features (RDL and bump) and the overall package size increasing. This integration drives the need for 3D metrology of feature height and coplanarity. In addition, the need to augment raw 3D metrology with defect inspection and 2D metrology data enables a comprehensive view (insight) into the packaging process. Achieve total bump process control with the combination of data from: (1) 2D defect detection – voids and shorts, foreign material, misprocessing; (2) 2D metrology – bump diameter, bump position, bump presence; (3) 3D inspection – bump too tall, bump too short, statistical process control (SPC); (4) Auto classifications – data must make sense and be easy to interpret. By combining high speed 2D, 3D metrology with defect inspection and advanced analytics, the quality of process control data can be exponentially improved to enable quick time-to-results for both process development and HVM control. This paper describes the inspection and metrology challenges of bumps in advanced packaging and the next generation high-throughput bump inspection methodology for wafers with extremely high bump counts as well as the data analysis

Statistical control of forming process

A qualitative product must be obtained through a stable and repeatable process. In order to control a defining characteristic of the quality of a specific product, various specific methodologies have been developed in time. One of the most known methods of this kind is the statistical control of the process which provides users with a powerful and useful range of tools designed to ensure process stability and improve capacity by reducing variability. These tools have the following great advantages: they are simple to use; are effective and can be applied to all processes. This paper presents a software application designed to conduct a forming process by the statistical control method. The program analyzes a database of measured values of the traceability characteristics and identifies whether these values have a statistically normal distribution. In the affirmative case, the main statistical indicators of the considered sample are calculated: maximum and minimum values, amplitude, average value, median, dispersion, process control limits and process capability index. These statistics are displayed both numerically and graphically. By using the software, it is possible to track the process and, if exists the tendency to overcome the control limits, to change the process parameters that caused the change in the quality characteristics.

Effective usage of model based data in turbine engine component machining

In modern machining operations use of process data drives future machining standards. Machining process data reach from the geometrical data over validated process parameters (e.g. speeds, feeds, depth of cut) up to simulation data to describe physical process impact on components surface integrity. The effective data use in production depends on verification of its physical relevance and knowledge of process control limits. In this paper examples of different types of machining process modelling and data usage are set into context of turbine disc machining. This concludes in a definition of an appropriate level of process modelling on today’s applications.

Implementing Outcome-based Care in Pediatric Psychiatry: Early Results and Overcoming Barriers.

Background:Poor utilization of standardized outcome measures for monitoring patient care and progress in mental health services is evident. The objective of this work was to implement computerized diagnostic and history assessments for outpatient mental health visits in the ambulatory psychiatric clinic of a large pediatric health system.Methods:A computerized assessment system was created for the iPad to administer and score a series of validated diagnostics before new patient and follow-up appointments with a psychiatry provider. Outcome measures were a percentage of completed assessments, provider satisfaction, and patient satisfaction.Results:Across all outpatient psychiatry clinics, screener completion rate for new patients increased from 0% to 90% within 1 year of implementation and sustained within 3σ process control limits for 2 years. Return visit assessment completion increased from 0% to 80%. The most substantial completion rate increase was related to scheduling assessment time as part of the visit. Assessment of provider and patient satisfaction through surveys before and after the implementation showed 94% of providers felt that visit efficiency had improved, and overall opinion of the system was highly positive. Patients also reported high satisfaction with the assessment process (4.1 on Likert scale 0–5, 5 = very positive).Conclusions:This quality improvement program demonstrates strategies for overcoming barriers to the use of standardized assessments in outpatient psychiatry. We show that a large pediatric mental health system can achieve systematic outcome data collection with minimal disruption to routine clinical care.

An EWMA sign control chart with varying control limits for finite horizon processes

AbstractThe sign SN distribution‐free control charts are efficient and simple online tools to perform quality monitoring when the normality of observations cannot be assumed a priori. The possibility of designing an SN control chart whichever is the distribution of observations allows it to be implemented with success in processes with finite production horizon (FPH). Today, FPH processes are widespread in industry. Recently, Shewhart SN control charts have been investigated in literature for this kind of quality control problem. In this paper, we investigate the implementation of an exponentially weighted moving average (EWMA) SN control chart for FPH processes. The property to have varying control limits at each inspection allows the EWMA SN control chart to be very sensitive to shifts in the location. A nonhomogeneous Markov chain model is used to evaluate the statistical properties of the EWMA SN control chart with varying control limits. An illustrative example shows the implementation of the EWMA SN control chart with observations collected from a quality monitoring procedure in a company bottling soft drinks.

Statistical analysis of the Nb3Sn strand production for the ITER toroidal field coils

The ITER toroidal field (TF) strand procurement initiated the largest Nb3Sn superconducting strand production hitherto. The industrial-scale production started in Japan in 2008 and finished in summer 2015. Six ITER partners (so-called Domestic Agencies, or DAs) are in charge of the procurement and involved eight different strand suppliers all over the world, of which four are using the bronze route (BR) process and four the internal-tin (IT) process. In total more than 500 tons have been produced including excess material covering losses during the conductor manufacturing process, in particular the cabling. The procurement is based on a functional specification where the main strand requirements like critical current, hysteresis losses, Cu ratio and residual resistance ratio are specified but not the strand production process or layout. This paper presents the analysis on the data acquired during the quality control (QC) process that was carried out to ensure the same conductor performance requirements are met by the different strand suppliers regardless of strand design. The strand QC is based on 100% billet testing and on applying statistical process control (SPC) limits. Throughout the production, samples adjacent to the strand pieces tested by the suppliers are cross-checked (‘verified’) by their respective DAs reference labs. The level of verification was lowered from 100% at the beginning of the procurement progressively to approximately 25% during the final phase of production. Based on the complete dataset of the TF strand production, an analysis of the SPC limits of the critical strand parameters is made and the related process capability indices are calculated. In view of the large-scale production and costs, key manufacturing parameters such as billet yield, number of breakages and piece-length distribution are also discussed. The results are compared among all the strand suppliers, focusing on the difference between BR and IT processes. Following the completion of the largest Nb3Sn strand production, our experience gained from monitoring the execution of the QC activities and from auditing the results from the measurements is summarised for future superconducting strand material procurement activities.

SU-D-BRD-07: Evaluation of the Effectiveness of Statistical Process Control Methods to Detect Systematic Errors For Routine Electron Energy Verification

Purpose: To evaluate the ability of statistical process control methods to detect systematic errors when using a two dimensional (2D) detector array for routine electron beam energy verification. Methods: Electron beam energy constancy was measured using an aluminum wedge and a 2D diode array on four linear accelerators. Process control limits were established. Measurements were recorded in control charts and compared with both calculated process control limits and TG-142 recommended specification limits. The data was tested for normality, process capability and process acceptability. Additional measurements were recorded while systematic errors were intentionally introduced. Systematic errors included shifts in the alignment of the wedge, incorrect orientation of the wedge, and incorrect array calibration. Results: Control limits calculated for each beam were smaller than the recommended specification limits. Process capability and process acceptability ratios were greater than one in all cases. All data was normally distributed. Shifts in the alignment of the wedge were most apparent for low energies. The smallest shift (0.5 mm) was detectable using process control limits in some cases, while the largest shift (2 mm) was detectable using specification limits in only one case. The wedge orientation tested did not affect the measurements as this did not affectmore » the thickness of aluminum over the detectors of interest. Array calibration dependence varied with energy and selected array calibration. 6 MeV was the least sensitive to array calibration selection while 16 MeV was the most sensitive. Conclusion: Statistical process control methods demonstrated that the data distribution was normally distributed, the process was capable of meeting specifications, and that the process was centered within the specification limits. Though not all systematic errors were distinguishable from random errors, process control limits increased the ability to detect systematic errors using routine measurement of electron beam energy constancy.« less

Uncertainty in Bogue-calculated phase composition of hydraulic cements

The Bogue calculations are used for cement manufacturing process control and acceptance limits for some cement classifications. While it is commonly understood that the Bogue calculations are estimates with potential intrinsic biases, the magnitudes of such biases are not generally known. The biases stem primarily from the compositional variation in phase solid solution, resulting in deviation of their bulk chemistry from that assumed in the calculations, and from bulk oxide measurement uncertainties. Uncertainties in Bogue estimates from these sources are quantified here through propagation-of-error analysis combining effects from bulk chemical analysis imprecision and estimates of variability in chemical compositions of the four principal clinker phases. Standard deviation (1σ) values of about 9.6% for alite and belite, and 2.2% and 1.4% for aluminate and ferrite are calculated. This significant increase compared to uncertainty due to bulk oxide measurements is attributable to the imprecision of the Bogue constants.

Adaptive Fault Detection with Two Time-Varying Control Limits for Nonlinear and Multimodal Processes

A novel fault detection method is proposed for detection process with nonlinearity and multimodal batches. Calculating the Mahalanobis distance of samples, the data with the similar characteristics are replaced by the mean of them; thus, the number of training data is reduced easily. Moreover, the super ball regions of mean and variance of training data are presented, which not only retains the statistical properties of original training data but also avoids the reduction of data unlimitedly. To accurately identify faults, two control limits are determined during investigating the distributions of distances and angles between training samples to their nearest neighboring samples in the reduced database; thus, the traditionalk-nearest neighbors (only considering distances) fault detection (FD-kNN) method is developed. Another feature of the proposed detection method is that the control limits vary with updating database such that an adaptive fault detection technique is obtained. Finally, numerical examples and case study are given to illustrate the effectiveness and advantages of the proposed method.

Robust Resilience: Metaphor and Meaning in Assessing System Performance Ranges

The current emphasis on Resilience Week and the International Symposium on Resilient Cognitive Systems highlights a growing awareness of the importance of designing and operating engineering systems under a variety of environmental conditions and in response to dynamic events. Although there has been considerable confusion and drift in the use of the term, ‘‘resilience’’ as a concept dates back to dynamic systems study of complex ecological systems in the 1970s. This original definition relates clearly to quantitative metrics that link also to statistical process control techniques describing system performance as affected by external, ‘‘assignable’’ causes. This paper discusses important elements in the consideration of resilience as a quantitative metric to improve consistency and clarity of evaluation in engineering systems. Rather than simply a binary attribute of systems, resilience should be considered in terms of system performance measures as affected by environmental conditions or events, energy flow couplings, and statistical process control limits. Our estimations of system resilience are seriously compromised when process control estimates are extrapolated beyond linear ranges of environmental conditions or when including discontinuous performance/event outliers exceeding appropriate forecasting estimates.