Product Quality Prediction Research Articles

Sodium-chloride reduction by substitution with potassium, calcium and magnesium salts in wheat bread

Sodium chloride has various functions in bread, but its reduction is a topic that should be focused on due to nutritional recommendations. This study addresses effects of sodium reduction on both dough and bread properties. 50% NaCl was systematically substituted by different combinations of KCl, MgCl2 and CaCl2. Stiffness of the yeasted doughs decreased with increasing calcium content, and crumb structure as well as bread volume and sensory perception was also affected by divalent cations, which can be attributed to a weakened protein network. In contrast there was no significant influence of potassium on dough stiffness and extensibility. Crumb cell structure and crust firmness were comparable to the reference and a reduction of sodium chloride seems feasible for levels up to 50%, corresponding to 0.75 g NaCl per 100 g, without negatively affecting sensory properties. The correlations identified between salt level, dough properties and bread characteristics are helpful to predict product quality.

Simulations of Die Casting With Uncertainty Quantification

Abstract Die casting is a type of metal casting in which a liquid metal is solidified in a reusable die. In such a complex process, measuring and controlling the process parameters are difficult. Conventional deterministic simulations are insufficient to completely estimate the effect of stochastic variation in the process parameters on product quality. In this research, a framework to simulate the effect of stochastic variation together with verification, validation, and uncertainty quantification (UQ) is proposed. This framework includes high-speed numerical simulations of solidification, microstructure, and mechanical properties prediction models along with experimental inputs for calibration and validation. Both experimental data and stochastic variation in process parameters with numerical modeling are employed, thus enhancing the utility of traditional numerical simulations used in die casting to have a better prediction of product quality. Although the framework is being developed and applied to die casting, it can be generalized to any manufacturing process or other engineering problems as well.

Just-in-time learning based soft sensor with variable selection and weighting optimized by evolutionary optimization for quality prediction of nonlinear processes

Just-in-time (JIT) learning based soft sensors have been widely used for predicting product quality variables of nonlinear processes. They dynamically build online local models by selecting the samples most relevant to the query data from a historical database whenever an estimate is requested. However, building high-performance JIT soft sensors remains challenging due to difficulties defining similarity measures and the selection of input variables for facilitating efficient relevant sample selection and model building. In this study, we propose a novel soft sensing framework, referred to as JIT learning with variable selection and weighting (JIT-VSW). In this framework, a mixture weighted similarity (MWS) measure is defined by combining multiple weighted Euclidean distance (WED) based similarity measures. The MWS measure enables variable weighting embedded in WED measures to account for the relevance between input and output variables and facilitates the handling of highly complex process characteristics through the mixture-type similarity measure. Meanwhile, a wrapper optimization approach using evolutional algorithms is proposed for input variable selection. Further, the selection of input variables and the determination of MWS parameters, i.e., weights assigned to input variables and mixture coefficients of WED similarity measures, are formulated as a mixed integer optimization problem and solved simultaneously by using the mixed integer genetic algorithms (MIGA). The effectiveness and superiority of JIT-VSW are verified through three real-world applications.

Quality prediction of petroleum products based on KPCARNN

With the application of the Internet of Things in oil production, a large number of equipment parameters and environment parameters in production process can be collected. The data can be applied to production equipment troubleshooting and product quality prediction. Prior gasoline quality testing has strict requirements for equipment and operation, but with a low efficiency. The paper applies RNN in a new situation and proposes a gasoline production quality prediction model fused with dimensionality reduction which is based on improved RNN, to achieve the real-time monitoring for gasoline production quality. Since the collected data is time-series data, the data features are complex, high-dimensional, and contain a large amount of noise and outliers. Therefore, the data needs to be cleaned, and feature engineering is used to achieve feature selection and dimensionality reduction. The model is trained and validated by an actual data set of gasoline production in an offshore petroleum field. The results show that the established prediction model fused with dimensionality reduction technique and based on improved RNN can predict the quality of gasoline products in real time and ensure a high accuracy.

Smartphone Image based Agricultural Product Quality and Harvest Amount Prediction Method

A method for agricultural product quality and harvest amount prediction by using smartphone camera image is proposed. It is desired to predict agricultural product quality and harvest amount as soon as possible after the sowing. In order for that, satellite imagery data, UAV camera based images, ground based camera images are used and tried These methods do cost significantly and these do not work so well due to some reasons, in particular, most of farmers cannot use these properly. The proposed method uses just smartphone camera acquired images. Therefore, it is totally easy to use. If the results of prediction of product quality and harvest amount are not satisfied, then farmers have to add some additional fertilizer at the appropriate time. The experimental results with soy plantations show some possibility of the proposed method.

A new thermodynamic approach for struvite product quality prediction.

Struvite precipitation has drawn much attention in the last decade as a green chemical process for phosphorus removal and recovery. Product purity affects the usefulness, and thus price, of the product when recovered struvite is sold as fertilizer. However, there is currently little research on struvite quality, as well as on models for accurately predicting. This paper presents an alternative approach to the traditional thermodynamic model where the solid with the largest positive saturation index precipitates first, depleting the concentrations of constituent ions before the next solid can precipitate. In the new thermodynamic approach, all solids with a positive saturation index precipitate simultaneously, and deplete the common pool of available ions in tandem. It was validated against experimental data, compared with the traditional thermodynamic models and a previously developed empirical model. The proposed new approach was more accurate than other models, except when both the ammonium nitrogen and magnesium concentrations were very low, a condition not likely to be encountered in industry. Therefore, this model is more suited for predicting the performance of struvite precipitation under varying wastewater conditions.

Opportunities for Process Control and Quality Assurance Using Online NIR Analysis to a Continuous Wet Granulation Tableting Line

This paper investigates the application of online near-infrared measurements as a means to measure blend uniformity in a continuous tableting line. Underlying all the monitoring and control methods is the ability to measure key tablet properties online at a rate suitable for control purposes. The use of NIR to determine any deviations in blend uniformity is demonstrated by interpreting the relevant spectral signature allowing quantitative information to be acquired for process monitoring and quality assurance. In addition to demonstrating the functionality of the NIR probe, the practical issues arising in the application are discussed. The composition of the blend was measured using an NIR probe over a range of concentrations and the results were calculated comparing sub unit dose scale of scrutiny of small populations. This was compared with predicted product quality for whole tablets over the whole production period. This technique has demonstrated how data collected online can be used to successfully predict the quality of the whole production run for the purposes of real-time product quality assurance.

Key quality characteristics selection for imbalanced production data using a two-phase bi-objective feature selection method

The selection of key quality characteristics (KQCs) that are significantly associated with product quality is essential for improving product quality. Production lines generally yield a larger number of regular products than do premium products, which creates an imbalance in production datasets and complicates KQC selection. In this study, a KQC selection method with an excellent ability to predict product quality is proposed based on a two-phase bi-objective feature selection method. The KQC selection model is established as a bi-objective problem of maximizing feature (i.e., quality characteristic) importance and minimizing percentage of selected features, and the geometric mean (G-mean) is selected as the feature importance metric for imbalanced data. To solve this model, a two-phase multi-objective optimization method is proposed; this method yields a set of candidate solutions (KQC sets) using an improved direct multi-search (DMS) strategy and uses the ideal point method (IPM) to select the final KQC sets from the candidate solutions. The experimental results indicate that the proposed method is effective for selecting KQCs for imbalanced production data.

Final quality prediction method for new batch processes based on improved JYKPLS process transfer model

Data-driven methods have been successfully used in modern industrial production. The sufficient data is the basis for implementing these methods. However, it is often impossible to meet the requirement for a new industrial process. In this study, an improved JYKPLS (Joint-Y kernel partial least squares) process transfer model is proposed to solve this issue and perform final product quality prediction for a new batch process. Based on the latent variable transfer technology, the rich information from similar old process data is transferred to accelerate the proceeding of building a new process model. The requirements on the amount of modeling data and prior knowledge of new processes are visibly reduced. Moreover, in order to handle the nonlinear correlation in process data, the kernel function is introduced to make data linear and separable. With actual productions operating, the transfer model is improved gradually by updating it with online data. When the prediction error falls into its confidence interval, the old data with lower similarity will be eliminated to avoid the negative transfer. The prediction results of penicillin concentration verify the effectiveness of the proposed method.

A fault diagnosis and quality prediction method of ball valves based on state transition probability matrix in Markov chain

Product quality is very important because many faults have a real impact on users directly. It is significative that Markov chain theory is applied in product quality prediction. A new method of state transition probability matrix in Markov chain is used to provide fault diagnosis and quality prediction in this paper. The aim of this study is to deliver a quality prediction method for mechanical and electrical product by using the data collection and analysis techniques. Firstly, a new approach of fault diagnosis was presented by using the theory of state transition probability matrix. Then, Markov chain theory is used to build the mathematical patterns in the quality prediction with the state transition probability matrix. Finally, we illustrate the veracity rationality and science of the approved method using a numerical example of ball valves.

Quality prediction for multi-grade processes by just-in-time latent variable modeling with integration of common and special features

To cope with the difficulty of on-line quality prediction for multi-grade processes widely operated in process industries, a just-in-time latent variable modeling method is proposed based on extracting the common and special features of multi-grade processes. Considering the complicated nonlinear characteristics of multi-grade processes encountered in engineering applications, a just-in-time learning (JITL) strategy is developed to choose the relevant samples from different grades with respect to the query sample. A novel common feature extraction algorithm is proposed to determine the common directions shared by different grades of processes. After extracting the common features, a partial least-squares modeling algorithm is used to extract the special directions for each grade, respectively. Hence, product quality prediction can be simply conducted by integrating the common and special parts of each grade for model building in terms of a JITL strategy. A numerical case and an industrial polyethylene process are used to demonstrate the effectiveness and advantage of the proposed method.

Integrating Operator Information for Manual Grinding and Characterization of Process Performance Based on Operator Profile

Due to its high versatility and scalability, manual grinding is an important and widely used technology in production for rework, repair, deburring, and finishing of large or unique parts. To make the process more interactive and reliable, manual grinding needs to incorporate “skill-based design,” which models a person-based system and can go significantly beyond the considerations of traditional human factors and ergonomics to encompass both processing parameters (e.g., feed rate, tool path, applied forces, material removal rate (MRR)), and machined surface quality (e.g., surface roughness). This study quantitatively analyzes the characteristics of complex techniques involved in manual operations. A series of experiments have been conducted using subjects of different levels of skill, while analyzing their visual gaze, cutting force, tool path, and workpiece quality. Analysis of variance (ANOVA) and multivariate regression analysis were performed and showed that the unique behavior of the operator affects the process performance measures of specific energy consumption and MRR. In the future, these findings can be used to predict product quality and instruct new practitioners.

Assessment of variance & distribution in data for effective use of statistical methods for product quality prediction

Abstract Data mining in automated production systems provide high potential to increase the Overall Equipment Effectiveness. Nevertheless, data of such machines/plants include specific characteristics regarding the variance and distribution of the dataset. For modelling product quality prediction, these characteristics have to be analysed to interpret the results correctly. Therefore, an approach for the analysis of variance and distribution of datasets is proposed. The evaluation of this approach validates the developed guidelines, which identify the reasons for inconsistent prediction results based on two different datasets of the same production system.

Online Monitoring for Industrial Processes Quality Control Using Time Varying Parameter Model

A novel data-driven soft sensor is designed for online product quality prediction and control performance modification in industrial units. A combined approach of time variable parameter (TVP) model, dynamic auto regressive exogenous variable (DARX) algorithm, nonlinear correlation analysis and criterion-based elimination method is introduced in this work. The soft sensor performance validation is tested by data set of an industrial SRU. The comparative study indicated the result associated with more robust soft sensor and more appropriate performance index values compared to other methods for SRU soft sensor design in diverse achievements. Due to high prediction accuracy, the low complication of the model and also saving of time, this technique can be very noticeable in industrial processes control.

Glyco-Mapper: A Chinese hamster ovary (CHO) genome-specific glycosylation prediction tool

Glyco-Mapper is a novel systems biology product quality prediction tool created using a new framework termed: Discretized Reaction Network Modeling using Fuzzy Parameters (DReaM-zyP). Within Glyco-Mapper, users fix the nutrient feed composition and the glycosylation reaction fluxes to fit the model glycoform to the reference experimental glycoform, enabling cell-line specific glycoform predictions as a result of cell engineering strategies. Glyco-Mapper accurately predicts glycoforms associated with genetic alterations that result in the appearance or disappearance of one or more glycans with an accuracy, sensitivity, and specificity of 96%, 85%, and 97%, respectively, for publications between 1999 and 2014. The modeled glycoforms span a large range of glycoform engineering strategies, including the altered expression of glycosylation, nucleotide sugar transport, and metabolism genes, as well as an altered nutrient feeding strategy. A glycoprotein-producing CHO cell line reference glycoform was modeled and a novel Glyco-Mapper prediction was experimentally confirmed with an accuracy and specificity of 95% and 98%, respectively. Glyco-Mapper is a product quality prediction tool that provides a streamlined way to design host cell line genomes to achieve specific product quality attributes.

Novel common and special feature extraction method for modeling multi-grade processes

In the processing industries, operating conditions often change to meet the requirements of the market and customers. To cope with the difficulty of on-line quality prediction for such multi-grade processes widely operated in process industries, a novel common and special feature extraction method is proposed for modeling multi-grade processes. A common feature extraction algorithm is proposed to determine the common directions shared by different grades of these processes. After extracting the common features, a partial least-squares modelling algorithm is used to extract the special directions for each grade, respectively. Hence, product quality prediction can be simply conducted by integrating the common and special parts of each grade for model building. A numerical case and an industrial polyethylene process are used to demonstrate the effectiveness and advantage of the proposed method.

An Advanced Method for Prediction and Judgement of Product Quality by Using Multivariate Analysis and Machine Learning

An Advanced Method for Prediction and Judgement of Product Quality by Using Multivariate Analysis and Machine Learning

Product quality prediction based on geoinformation technologies

Product quality prediction based on geoinformation technologies

‘Grey Model'-Based Simulation Tool for Predictive Product Quality Control

A modified grey model is represented for predicting the product quality index and generating the grey process control chart. The model considers the first sample datum in the numerical solution of the whitened differential equation. This datum is introduced by the refitting smoothed cumulative values and numerical solution of the whitened equation with the last value as initial point. Values predicted in this manner and those predicted in the regular GM (1,1) model were calculated with a specially elaborated MATLAB program. The numbers obtained were compared with reference data. The modified model demonstrates higher accuracy compared to the regular GM as demonstrated by several actual examples. The proposed model is used to design a special simulation tool. The graphical interface of the tool allows to the user to introduce the sample data, control limits and required mean value. After this, the tool outputs the predicted quality index with its error and generates a process control chart with the predicted product quality index.

Virtual metrology for copper-clad laminate manufacturing

Copper-clad laminate (CCL), the key material for printed circuit board production, is used in various electronic products; thereby, the demand for CCL is on the rise. The process of CCL manufacturing occurs in three phases: treating, lay-up, and pressing, while the process with the largest influence on quality control is the treating. For effective quality control, the treating process requires intermediate inspection for three important quality factors: treated weight, minimum viscosity, and gel time. However, a manual inspection, which present-day manufacturers perform, incurs heavy cost in terms of time and money, rendering it ineffective. This study proposes the application of virtual metrology for CCL manufacturing to predict product quality derived from processing data without a product quality inspection. The actual process data from a CCL manufacturer in Korea was collected for a duration of approximately 5months. Based on these data, the application builds a prediction model for CCL quality by utilizing the process variables affecting the CCL quality as predictor variables. As a result, four regression algorithms and three methods of variable selection were applied to build the prediction models for virtual metrology. Prediction models were obtained with a high accuracy in specific target variables. It was also verified that quality control was influenced by not only the important predictor variables empirically recognized by process engineers in the field but also by several essential variables previously unknown to the engineers; effective quality control will be possible by focusing on these variables particularly and more efficiently instead of overall monitoring.