Product Quality Prediction Research Articles

Data-driven real-time control method for process equipment in flow shop towards product quality improvement

Real-time control of complex process equipment is a crucial aspect for product quality assurance in the flow shop. However, there are two bottleneck problems, including the inaccurate quality prediction and the unstable process control, which would lead to poor product quality. To solve these problems, a data-driven real-time product quality prediction and control method for the process equipment is elaborated in this paper. First, to improve the prediction accuracy, a delay-mean rule is designed to eliminate the negative effects of time lag focused on the modeling data. Second, a variable optimization method based on the improved simulated annealing algorithm and a time window is proposed to enhance the stability of control. Finally, a case study regarding the real-time control for a dryer equipment is used to verify the proposed method.

Guideline for Deployment of Machine Learning Models for Predictive Quality in Production

Predicting product quality represents a common area of application of machine learning (ML) in manufacturing. However, manifold challenges occur during the integration of ML models into production processes. Therefore, this paper aims to provide a guideline for the deployment of ML models in production environments. Relevant decisions and steps for deploying models in predictive quality use cases are demonstrated. The results for each component of the proposed guideline - deployment design, productionizing & testing, monitoring, and retraining - have been validated with industry experts including exemplary implementations.

High-Dimensional, Small-Sample Product Quality Prediction Method Based on MIC-Stacking Ensemble Learning

Accurate quality prediction can find and eliminate quality hazards. It is difficult to construct an accurate quality mathematical model for the production of small samples with high dimensionality due to the influence of quality characteristics and the complex mechanism of action. In addition, overfitting scenarios are prone to occur in high-dimensional, small-sample industrial product quality prediction. This paper proposes an ensemble learning and measurement model based on stacking and selects eight algorithms as the base learning model. The maximal information coefficient (MIC) is used to obtain the correlation between the base learning models. Models with low correlation and strong predictive power were chosen to build stacking ensemble models, which effectively avoids overfitting and obtains better predictive performance. To improve the prediction performance as the optimization goal, in the data preprocessing stage, boxplots, ordinary least squares (OLS), and multivariate imputation by chained equations (MICE) are used to detect and replace outliers. The CatBoost algorithm is used to construct combined features. Strong combination features were selected to construct a new feature set. Concrete slump data from the University of California Irvine (UCI) machine learning library were used to conduct comprehensive verification experiments. The experimental results show that, compared with the optimal single model, the minimum correlation stacking ensemble learning model has higher precision and stronger robustness, and a new method is provided to guarantee the accuracy of final product quality prediction.

Feature selection and domain adaptation for cross-machine product quality prediction

Feature selection and domain adaptation for cross-machine product quality prediction

A two-step multivariate statistical learning approach for batch process soft sensing

Statistical machine learning algorithms have been widely used to analyse industrial data for batch process monitoring and control. In this study, we aimed to take a two-step approach to systematically reduce data dimensionality and to design soft-sensors for product quality prediction. The approach first employs partial least squares to screen the entire dataset and identify critical time regions and operational variables, then adopts multiway partial least squares to construct soft-sensors within the reduced space to estimate final product quality. Innovations of this approach include the ease of data visualisation and ability to identify major operational activities within the factory. To highlight efficiency and practical benefits, an industrial personal care product manufacturing process was presented as an example and two soft-sensors were successfully developed for product end viscosity estimation. Furthermore, the accuracy, reliability, and data efficiency of the soft-sensors were thoroughly discussed. This paper, therefore, demonstrates the industrial potential of the proposed approach.

Online quality prediction based on the double cumulative model for multiphase batch processes with transitions

In multiphase batch processes, the final product quality prediction results depend on the common cumulative effects of all the critical phases with transitions on quality rather than a separated one. However, the common cumulative effect are rarely considered at the same time. To address this issue, a double cumulative model is proposed in this paper. The double model includes the internal cumulative model and the external cumulative model. On one hand, the critical local cumulative quality is introduced to isolate the local cumulative effect of different phases and transitions in the internal cumulative model. Especially, for transitions modeling, an average process trajectories method based on the dominant phase identification is developed to deal with the dynamic and uncertainty. On the other hand, the correlation of different cumulative models is extracted by constructing the external cumulative model. In this way, the cumulative effect of all the phases and transitions are considered simultaneously. The proposed method is applied to penicillin fermentation. The simulation results demonstrate the effectiveness and superiority of the proposed method over the competing models.

Synergising stoichiometric modelling with artificial neural networks to predict antibody glycosylation patterns in Chinese hamster ovary cells

In-process quality control of biotherapeutics, such as monoclonal antibodies, requires computationally efficient process models that use readily measured process variables to compute product quality. Existing kinetic cell culture models can effectively describe the underlying mechanisms but require considerable development and parameterisation effort. Stoichiometric models, on the other hand, provide a generic, parameter-free means for describing metabolic behaviour but do not extend to product quality prediction. We have overcome this limitation by integrating a stoichiometric model of Chinese hamster ovary (CHO) cell metabolism with an artificial neural network that uses the fluxes of nucleotide sugar donor synthesis to compute the profile of Fc N-glycosylation, a critical quality attribute of antibody therapeutics. We demonstrate that this hybrid framework accurately computes glycan distribution profiles using a set of easy-to-obtain experimental data, thus providing a platform for process control applications.

Product Quality Prediction for Wire Electrical Discharge Machining with Markov Transition Fields and Convolutional Long Short-Term Memory Neural Networks

This paper proposes a wire electrical discharge machining (WEDM) product quality prediction method, called MTF-CLSTM, to integrate the Markov transition field (MTF) and the convolutional long short-term memory (CLSTM) neural network. The proposed MTF-CLSTM method can accurately predict WEDM workpiece surface roughness right after manufacturing by collecting and analyzing static machining parameters and dynamic manufacturing conditions. The highly accurate prediction is due to the following two reasons. First, MTF can transform data into images to extract data temporal information and state transition probability information. Second, the CLSTM neural network can extract image spacial features and temporal relationship of data that are separated far apart. In short, MTF-CLSTM predicts WEDM workpiece surface roughness with the MTF model and the CLSTM neural network using static machining parameters and dynamic manufacturing conditions. MTF-CLSTM is compared with 10 related research studies in many aspects. There is only one existing method that is like MTF-CLSTM to predict WEDM workpiece surface roughness by using static machining parameters and dynamic manufacturing conditions. Experiments are conducted to evaluate MTF-CLSTM performance to show that MTF-CLSTM significantly outperforms the existing method in terms of the prediction mean absolute percentage error.

Big data analytics in Industry 4.0 ecosystems

Big data analytics in Industry 4.0 ecosystems

Quality prediction for aluminum diecasting process based on shallow neural network and data feature selection technique

This paper aims to develop a prediction algorithm for better product quality of the diecasting process, one of the most essential aspects of quality management in traditional manufacturing industries. Shallow neural network (SNN) architecture has been applied as a regressive classifier engine in order to predict product quality of the manufacturing process to renovate the conventional technology adopted by the mainstream fundamental manufacturing industries. The results obtained have been contrasted with classical multilayer perceptron to establish the efficacy of the proposed strategy. In addition to the predictive classifier, various aspects of the research on the prediction algorithm and optimal input selection for successful yield have also been investigated for the aforementioned process using Fuzzy C Means clustering approach. As a case study for validation, the proposed algorithm applies for data set from an aluminum die-casting process. The results reflect that the proposed analytic model can prove to be an effective tool for the prediction and optimization of the product quality of various fundamental manufacturing processes.

Identification of Suitable Biomass Torrefaction Operation Envelops for Auto-Thermal Operation

Auto-thermal operation of biomass torrefaction can help avoid additional heat investment and the associated costs to the system. This work provides a general method for relating the feedstock-specific parameters to the energy balance and pre-diagnosing the potential of auto-thermal for different biomass torrefaction and pyrolysis systems. Both solid and gas thermal properties under various torrefaction conditions and their influences to the torrefaction system energy balances are considered. Key parameters that influence the process auto-thermal operation are analyzed, which include torrefaction reaction heat, torrefaction conditions, drying method, biomass species, and inert N2 flowrate. Equations of torgas and biomass higher heating values (HHVs), as well as the torrefaction reaction heat at different operating conditions are developed. It is found that torgas and biomass HHVs increase with torrefaction temperature and biomass weight loss. Torrefaction reaction heat has a linear relationship with the biomass weight loss, with a positive slope at 250–260°C, and a negative slope at 270–300°C, which indicates that torrefaction shifts from endothermic to exothermic at ∼270°C. Applying advanced drying technology and avoiding the use of N2 can help the system achieve auto-thermal operation at lower torrefaction temperature and residence time, thus leading to a higher process energy efficiency and product yield. This is the first work to relate the micro level element changes of biomass to the macro level process energy balances of the torrefaction system. This work is important in design and operation of the torrefaction system in both pilot and industrial scales to improve process efficiency and predict product quality in a reliable and economic manner.

A Comparative Study on Supervised Machine Learning Algorithms for Copper Recovery Quality Prediction in a Leaching Process.

The copper mining industry is increasingly using artificial intelligence methods to improve copper production processes. Recent studies reveal the use of algorithms, such as Artificial Neural Network, Support Vector Machine, and Random Forest, among others, to develop models for predicting product quality. Other studies compare the predictive models developed with these machine learning algorithms in the mining industry as a whole. However, not many copper mining studies published compare the results of machine learning techniques for copper recovery prediction. This study makes a detailed comparison between three models for predicting copper recovery by leaching, using four datasets resulting from mining operations in Northern Chile. The algorithms used for developing the models were Random Forest, Support Vector Machine, and Artificial Neural Network. To validate these models, four indicators or values of merit were used: accuracy (acc), precision (p), recall (r), and Matthew’s correlation coefficient (mcc). This paper describes the dataset preparation and the refinement of the threshold values used for the predictive variable most influential on the class (the copper recovery). Results show both a precision over 98.50% and also the model with the best behavior between the predicted and the real values. Finally, the obtained models have the following mean values: acc = 0.943, p = 88.47, r = 0.995, and mcc = 0.232. These values are highly competitive when compared with those obtained in similar studies using other approaches in the context.

Final Prediction of Product Quality in Batch Process based on Bidirectional Neural Network Algorithm

Based on the analysis of time series characteristics of production process based on common batch process end point quality prediction a predictive model based on bidirectional gated loop neural network is proposed to predict final product quality for unequal interval batch processes. Based on the requirement of forecasting value in actual production, loss function adapted to batch process is constructed, which makes model meet forecasting requirement under guarantee prediction precision, thus obtaining greater production benefit. Two-way gated loop (GRU) neural networks with different loss functions are compared with multi-directional partial least squares (MPLS) neural networks support vector regression SVR and gated cyclic unit neural networks. Results show that bidirectional gated loop neural networks have stronger applicability and higher accuracy.

The use of organic waste of the agro-industrial complex for the production of fuel using neural network algorithms

The article addresses the problem of processing organic waste from the agro-industrial complex. To solve the problem of processing biomass, it is proposed to use gas-generating units that convert the resulting biomass into biogas. Particular attention is paid to the principles of operation of gas generators, the description of the stages of gasification, and the ongoing chemical processes. The scientific novelty lies in the proposal for the introduction of intelligent technology based on artificial neural networks to predict product quality and control the gasification process at each stage. As a result of the study, it was revealed that the use of information technology contributes to the optimization of the production process.

Production rescheduling through product quality prediction

In production management, efficient scheduling is key towards smooth and balanced production. Scheduling can be well-supported by real-time data acquisition systems, resulting in decisions that rely on actual or predicted status of production environment and jobs in progress. Utilizing advanced monitoring systems, prediction-based rescheduling method is proposed that can react on in-process scrap predictions, performed by machine learning algorithms. Based on predictions, overall production can be rescheduled with higher efficiency, compared to rescheduling after completion of the whole machining process with realization of scrap. Series of numerical experiments are presented to demonstrate potentials in prediction-based rescheduling, with early-stage scrap detection.

Quality Prediction of Discrete Manufacturing Process Based on CGAN&Catboost Hybrid Model

Accurate prediction of product quality can be used to control the online production process of workpieces, an appropriate prediction model can effectively improve the yield of industrial production. In order to obtain missing data, the classic method is to regress them one by one. However, the traditional regression prediction method destroys the correlation of the data and cannot use the quality inspection index efficiently. This paper presents an improved hybrid model. The essence lies in a novel easily pluggable CGAN module that refines the previous step of regression data. compared with the regression method, the CGAN module can learn the joint d istribution of data. The experiment of real data result shows that our method has high accuracy compared with the existing ones

Deep Transfer Learning for Improved Product Quality Prediction: A Case Study of Aluminum Gravity Die Casting

High product quality and low rejection rates are decisive for the competitiveness of manufacturing systems. Hence, manufacturing shifts towards digitalization for predicting and controlling product quality. In order to implement data-driven approaches in manufacturing successfully, sufficient model performances are necessary. Especially metal casting comes with several hurdles for data-driven approaches, like sparse and imbalanced data, a relatively high human interaction and frequent mounting events. This often leads to hardly generalizable and weak models. Within this paper, transfer learning is investigated for the quality prediction of aluminum gravity die casting to try to overcome these hurdles and create robust, accurate and data-efficient models.

A Hybrid Feature Selection Method RFSTL for Manufacturing Quality Prediction Based on a High Dimensional Imbalanced Dataset

Under Industry 4.0, manufacturing quality prediction has been gaining increased interest from researchers and manufacturers. From the analysis of previous studies on quality predictions using machine learning, it became clear that the high dimensionality and imbalance of data are major and common problems affecting the learning performance. This work uses a hybrid method to address this issue, applying a Synthetic Minority Oversampling Technique & TomekLinks balancing approach to create balanced data and using Random Forest as the feature selecting measurement to reduce the dimensionality of data. In addition, a Fine Gaussian Support Vector Machine (Fine Gaussian SVM) based on the representative set of features selected by the hybrid method utilized is employed in this work to predict product quality. The results of experimentation demonstrate that the hybrid method proposed in this work performs well for manufacturing quality prediction and offers a simple, quick and powerful way to address the problem of feature selection encountered by the imbalanced classification.

Predictive analytics in quality assurance for assembly processes: lessons learned from a case study at an industry 4.0 demonstration cell

Quality assurance (QA) is an important task in manufacturing to assess whether products meet their specifications. However, QA might be expensive, time-consuming, or incomplete. This paper presents a solution for predictive analytics in QA based on machine sensor values during production while employing specialized machine-learning models for classification in a controlled environment. Furthermore, we present lessons learned while implementing this model, which helps to reduce complexity in further industrial applications. The paper’s outcome proves that the developed model was able to predict product quality, as well as to identify the correlation between machine-status and faulty product occurrence.

Large scale predictability analysis of process variables from injection molding machines

This paper analyzes the process variable data from injection molding processes to identify the key process variables, which can be predicted by other process variables, which highlights the interdependence among different process variables in various production scenarios. The available data from injection molding machines provide information for the run-time, setup parameters of machines, and measurements of different process variables through sensors. For predictive modeling, we employed different linear regression models with the recursive backward feature selection and SVM regression models using a radial kernel to predict nonlinear process variables. We also applied the linear and SVM regression models for outlier data in the process variables assuming that the upper bound outliers represent the perturbed state of process variables during production. These perturbations are affected by material type, machine type (age and performance), regime changes, or other external effects and subsequently affect the predictability of process variables and production output. Such cases are different compared to the normal or controlled range of process variables. Our analysis shows that the predictability varies for different material types due to the interdependence of the associated process variables. We further highlight that various process variables exhibit nonlinear relationships and cannot be predicted using linear models. We additionally look for the interdependence of process variables used previously by three studies as input features to predict product quality.