Multivariate Statistical Control Charts Research Articles

Integration of the multivariate statistical control chart and machine learning to identify faults in the quality characteristics for polylactic acid with glass fiber composites in injection molding

Complex processing parameters need to be adjusted for expected qualities in injection molding processing. Once the process is abnormal, it is essential to spend time and human work on fault diagnosis. In this study, we focus on fault diagnosis of injection molding processing parameters for polylactic acid/glass fiber composites. The injection molding processing parameters include the melt temperature, injection speed, packing pressure, packing time, and cooling time. The qualities include tensile strength, hardness, impact strength, and flexure strength. When processing parameters deviate from the optimal process condition, the multivariate statistical control chart monitors downgraded qualities. The machine is operated at the optimal process conditions to generate normal samples and the corresponding four qualities of data are chosen as the historical data. Hotelling’s T2 is used to calculate the upper control limit (UCL) from the historical data to detect abnormal samples. If the T2 value exceeds the UCL, the corresponding sample is considered abnormal. Then, the residuals of qualities for abnormal samples are obtained by a residual control chart. They are chosen as the feature values for the backpropagation neural network (BPNN) to identify the abnormal processing parameters. The experimental results proved that the BPNN can achieve a 100% recognition rate for single-factor abnormal samples. For the single-/double-factor mixture, the accuracy rate of double-factor classification can reach 97.44%. This proposed study has the advantage of high stability, being non-destructive, high precision, and low cost, and can be widely promoted in injection molding industries.

Integration of Multivariate Statistical Control Chart and Machine Learning to Identify the Abnormal Process Parameters for Polylactide with Glass Fiber Composites in Injection Molding; Part I: The Processing Parameter Optimization for Multiple Qualities of Polylactide/Glass Fiber Composites in Injection Molding.

This paper discusses the mixing of polylactide (PLA) and glass fiber which use injection molding to produce a functional composite material with glass fiber properties. The injection molding process explores the influence of glass fiber ratio, melt temperature, injection speed, packing pressure, packing time and cooling time on the mechanical properties of composite. Using the orthogonal table planning experiment of the Taguchi method, the optimal parameter level combination of a single quality process is obtained through main effect analysis (MEA) and Analysis of variance (ANOVA). Then, the optimal parameter level combination of multiple qualities is obtained through principal component analysis (PCA) and data envelopment analysis (DEA), respectively. It is observed that if all the quality characteristics of tensile strength, hardness, impact strength and bending strength are considered at the same time, the optimal process conditions are glass fiber addition 20 wt %, melt temperature 185 °C, injection speed 80 mm/s, holding pressure 60 MPa, holding time 1 s and cooling time 15 s, and the corresponding mechanical properties are tensile strength 95.04 MPa, hardness 86.52 Shore D, impact strength 4.4408 J/cm2, bending strength 119.89 MPa. This study effectively enhances multiple qualities of PLA/GF composite.

Development of dietary fibre incorporated tuna sausage employing response surface methodology and quality evaluation during chilled storage using multivariate control charts

In the present study, dietary fibre incorporated functional sausage was prepared by optimising the ingredients viz. tuna mince, wheat and oats dietary fibre using response surface methodology. D-optimal mixture response surface experimental design with 13 runs was formulated for the development of dietary fibre incorporated tuna sausage. The optimum combinations of sausage were found to be 62.5% fish mince, 2.5% wheat and 5% oats fibre and 66.8% fish mince and 1.6% of oats and wheat fibre each. The quality evaluation of combination sausage prepared at optimum conditions along with control sample was carried out based on biochemical, physical and sensory parameters during storage in chilled condition at 2°C. Storage stability studies of the optimised sausage samples indicated a shelf life of 14 days under chilled conditions. Multivariate statistical control charts based on T2 and PSE statistics were developed to monitor quality variations during storage.

Sensor-Based Recurrence Analysis of Energy Efficiency in Machining Processes

Manufacturing industry incurs a large portion of energy consumption and carbon emission in the economy. Traditionally, smart energy management depends on aggregated measures from billing information, as well as physics-based models, empirical results derived from extensive experiments, which tend to be limited in the ability for real-time monitoring of energy efficiency. There is an urgent need to develop energy monitoring solutions for more transparency about energy use. This paper presents a new sensor-based approach for recurrence analysis of continuous power signals and multi-state modeling of energy efficiency in the machining process. First, we leverage the recurrence plot to characterize the nonlinear variations in power signals and further help delineate different states in the machining process, thereby providing statistics of energy consumption in each state. Second, we compute the composite index of energy efficiency for each workpiece and then develop multivariate statistical control charts for process monitoring of continuous production of workpieces. Third, after an anomaly is detected, we propose the orthogonal decomposition approach to diagnose the root cause of abnormal states in the energy use. The proposed methodology is evaluated and validated on real-world manufacturing of shaft-like parts in a machine shop. Experimental results show that the prediction error of energy efficiency from sensor-based models is within 5% from the ground truth, which show great potentials to implement sensor-based monitoring and analysis of real-time energy efficiency in the manufacturing process.

Multivariate statistical control chart and process capability indices for simultaneous monitoring of project duration and cost

Statistical process control (SPC) techniques have been implemented in a few studies in order to determine action limits on the earned value management (EVM) indices to assist project managers in monitoring project performance. Nevertheless, these indices are correlated, reliance on univariate control charts when more than one variable are involved and they are correlated may lead to unsatisfactory results such as increasing the rate of false alarms. This study presented an approach to overcome this limitation by applying multivariate Hotelling's T2 control chart to take into account possible correlations between EVM indicators and describe the entire capability of the project performance more accurately. Furthermore, in order to quantify how well a project can meet its requirements, some practical multivariate process capability indices (PCIs) are introduced. This approach provides management with more reliable information about the deviations between planned and actual performance both in terms of time and cost simultaneously.

FUNCTION AND DEVELOPMENT OF MULTIVARIATE CONTROL GRAPHS

In the present work, it is intended to highlight in a specific way the utility and advances of multivariate control graphics for the control of quality in modern production systems. The synthesis addresses the development of the most commonly used multivariate control charts as well as identifying their scope and limitations. Once the restrictions to carry out a statistical quality control through multivariate charts have been identified, the study of different techniques that complement the quality analysis in the production systems monitored under said control techniques is addressed. Keywords: Multivariate statistical control charts

Fault Detection and Diagnosis In Hall–Héroult Cells Based on Individual Anode Current Measurements Using Dynamic Kernel PCA

Individual anode current signals in aluminum reduction cells provide localized cell conditions in the vicinity of each anode, which contain more information than the conventionally measured cell voltage and line current. One common use of this measurement is to identify process faults that can cause significant changes in the anode current signals. While this method is simple and direct, it ignores the interactions between anode currents and other important process variables. This paper presents an approach that applies multivariate statistical analysis techniques to individual anode currents and other process operating data, for the detection and diagnosis of local process abnormalities in aluminum reduction cells. Specifically, since the Hall–Heroult process is time-varying with its process variables dynamically and nonlinearly correlated, dynamic kernel principal component analysis with moving windows is used. The cell is discretized into a number of subsystems, with each subsystem representing one anode and cell conditions in its vicinity. The fault associated with each subsystem is identified based on multivariate statistical control charts. The results show that the proposed approach is able to not only effectively pinpoint the problematic areas in the cell, but also assess the effect of the fault on different parts of the cell.

Root Cause Analysis of Quality Defects Using HPLC-MS Fingerprint Knowledgebase for Batch-to-batch Quality Control of Herbal Drugs.

The batch-to-batch quality consistency of herbal drugs has always been an important issue. To propose a methodology for batch-to-batch quality control based on HPLC-MS fingerprints and process knowledgebase. The extraction process of Compound E-jiao Oral Liquid was taken as a case study. After establishing the HPLC-MS fingerprint analysis method, the fingerprints of the extract solutions produced under normal and abnormal operation conditions were obtained. Multivariate statistical models were built for fault detection and a discriminant analysis model was built using the probabilistic discriminant partial-least-squares method for fault diagnosis. Based on multivariate statistical analysis, process knowledge was acquired and the cause-effect relationship between process deviations and quality defects was revealed. The quality defects were detected successfully by multivariate statistical control charts and the type of process deviations were diagnosed correctly by discriminant analysis. This work has demonstrated the benefits of combining HPLC-MS fingerprints, process knowledge and multivariate analysis for the quality control of herbal drugs.

Heterogeneous recurrence monitoring and control of nonlinear stochastic processes

Recurrence is one of the most common phenomena in natural and engineering systems. Process monitoring of dynamic transitions in nonlinear and nonstationary systems is more concerned with aperiodic recurrences and recurrence variations. However, little has been done to investigate the heterogeneous recurrence variations and link with the objectives of process monitoring and anomaly detection. Notably, nonlinear recurrence methodologies are based on homogeneous recurrences, which treat all recurrence states in the same way as black dots, and non-recurrence is white in recurrence plots. Heterogeneous recurrences are more concerned about the variations of recurrence states in terms of state properties (e.g., values and relative locations) and the evolving dynamics (e.g., sequential state transitions). This paper presents a novel approach of heterogeneous recurrence analysis that utilizes a new fractal representation to delineate heterogeneous recurrence states in multiple scales, including the recurrences of both single states and multi-state sequences. Further, we developed a new set of heterogeneous recurrence quantifiers that are extracted from fractal representation in the transformed space. To that end, we integrated multivariate statistical control charts with heterogeneous recurrence analysis to simultaneously monitor two or more related quantifiers. Experimental results on nonlinear stochastic processes show that the proposed approach not only captures heterogeneous recurrence patterns in the fractal representation but also effectively monitors the changes in the dynamics of a complex system.

The Performance of Robust Multivariate Statistical Control Charts based on Different Cutoff-points with Sustained Shift in Mean

A separate univariate control chart for each characteristic is often used to detect changes in the inherent variability of a process because of ease of computation. Nevertheless, the shortcoming of using separate individual charts would not have detected out of control condition when the characteristics of interests are correlated. The problems get more complicated in the existence of sustained shift in the process mean. In this study, we proposed a robust multivariate control chart which is less sensitive to the sustained shift in mean process. However, the theoretical cutoff-points of the proposed charts are intractable. As an alternative, we proposed two different procedures of empirical cutoff-points. The performance of these two cutoff-points in detecting a step shift in the mean vector is investigated extensively by real example and monte carlo simulation.

Multivariate control charts based on net analyte signal (NAS) for characterization of the polymorphic composition of Piroxicam using near infrared spectroscopy

Near infrared spectroscopic and multivariate statistical control charts based on the net analyte signal (NAS) were applied to the polymorphic characterization of Piroxicam samples. Three different polymorphic forms (I, II and III) were studied, using X-ray powder diffraction (XRPD) and scanning electron microscopy as reference techniques. Samples containing form I were considered inside the quality specifications and forms II and III were impurities. Three control charts were developed: the NAS chart that corresponds to the analyte of interest (polymorphic form I), the interference chart that corresponds to the contribution of other compounds in the sample and the residual chart that corresponds to nonsystematic variation. From the limits estimated for each chart using samples inside the quality specifications, it was possible to identify samples that did not present polymorphic form I. The use of multivariate control charts provides a rapid evaluation of purity and the polymorphic composition of pharmaceutical formulations based on Piroxicam.

Modeling method of multivariate statistical control chart for small-batch manufacturing process quality

Based on the study of multivariate statistical process quality control chart and small-batch manufacturing process, a kind of synthetical modeling method was put forward to deal with multivariate statistical process for small-batch manufacturing. Practical application and emulational instance show that this modeling method can make the best use of the obtained data, build control model dynamically, consequently resolve the problem of lacking modeling data in small-batch manufacturing process.