Search

Remaining useful life (RUL) prediction of aircraft engine (AE) is of great importance to improve its reliability and availability, and reduce its maintenance costs. This article proposes a novel deep bidirectional recurrent neural networks (DBRNNs) ensemble method for the RUL prediction of the AEs. In this method, several kinds of DBRNNs with different neuron structures are built to extract hidden features from sensory data. A new customized loss function is designed to evaluate the performance of the DBRNNs, and a series of the RUL values is obtained. Then, these RUL values are reencapsulated into a predicted RUL domain. By updating the weights of elements in the domain, multiple regression decision tree (RDT) models are trained iteratively. These models integrate the predicted results of different DBRNNs to realize the final RUL prognostics with high accuracy. The proposed method is validated by using C-MAPSS datasets from NASA. The experimental results show that the proposed method has achieved more superior performance compared with other existing methods.

Deep-learning-based methods have been successfully applied to fault diagnosis of rotating machinery. However, the domain mismatch among different operating conditions significantly deteriorates diagnostic performance of these methods in industrial applications. To solve this problem, a new fault diagnosis model based on capsule neural network (Cap-net) is constructed, and a novel online domain adaptation learning method based on deep reinforcement learning (DRL) is proposed in this article to improve the adaptivity of the fault diagnosis model. In this method, the Cap-net is first introduced into the DRL as an agent to extract representative features and diagnoses fault. Moreover, the online domain adaptation learning of the agent is conducted based on the Q-learning of the DRL so as to adapt to different operating domains that have never been experienced. Specifically, an online feature dictionary combined with cosine similarity is designed to coarsely label the online data collected from different operating domain, while a reward mechanism is defined to evaluate the obtained label. Subsequently, the online data, the corresponding label, and the reward are used to optimize the agent to obtain the desired diagnostic model. Two experiment studies are implemented to verify the effectiveness of the proposed method. The experimental results show that the proposed method has more excellent diagnostic performance and adaptivity than the existing popular methods.

Remaining useful life (RUL) prediction is a key solution to improve the reliability, availability and maintainability of engineering systems. Long short-term memory (LSTM) and convolution neural networks (CNN) are the current hotspots in the field of RUL prediction. However, the LSTM-based prognostic approach has a slow loop step to process large-scale time-series data since the dependence of the data processing process at each time on the output of the previous time limits parallelism, and the CNN-based prognostic approach is not fit for time-series data although it can process the data in parallel. In this paper, a new auto-encoder quasi-recurrent neural networks (AEQRNN) based prognostic approach is proposed for RUL prediction of the engineering systems. The AEQRNN contains convolution components that can process input data in parallel, and pooling components which has two LSTM-like gate structures to process time-series data. In addition, the AEQRNN can automatically extract hidden features from monitoring signals without manual feature design. The effectiveness of the proposed prognostic approach is validated by three prognostic benchmarking datasets, including a turbofan engine dataset, a rolling bearing dataset, and a machining tool dataset. Experimental results demonstrate that this approach has both superior prognostic performance and training speed in comparison with other kinds of recurrent neural network-based approaches and various state-of-the-art approaches in the recent literature.

Intelligent manufacturing (IM) talent education is becoming considerably relevant owing to the increasing versatility and multi-level compound requirements of academic and applied personnel. Serving as the core promoter of IM talent education, the actions and orientation of higher education are essential for IM society. However, there is no coherent framework in current engineering teaching programs, which can determine the key advancements of IM to facilitate theoretical learning and professional skills. Hence, this study presents a systematic reference training system in a highly student-centered interpretation for IM formation by establishing (1) innovation education platform, to bridge theoretical learning and hands-on practice and (2) enhanced curriculum development, which advocates broad training outcomes from multi-perspectives. First, the IM system architecture and the required attributes of IM talents are analyzed. Second, an innovation platform is constructed by incorporating interdisciplinary IM frontiers. Third, combining multimedia and interactive instruction, the IM training course clusters are reformed by covering overall cognition, module practice, and comprehensive application. In particular, a typical IM production line is depicted to increase students’ interests in participation, which can simultaneously develop disciplinary knowledge and practical skills. From the perspectives of students and instructors, evaluations and assessments demonstrate that this presented reference training system is beneficial for supporting modern IM talent education.

In order to reduce environment impacts of manufacturing processes and fill in research gaps that most previous separated optimization of process planning and scheduling ignored influences of process planning on scheduling, a multi-objective integrated optimization model of process planning and scheduling for reducing carbon emissions in manufacturing processes is proposed. The model aims at minimizing makespan and carbon emissions in manufacturing processes by integrated optimizing machining methods for all machining features of workpieces, machine allocations of processes, process routes and machining sequence of workpieces. Because there are many parameters in the proposed model needed to be optimized and they are interactional, a four segment encoding method is designed and a Non-dominated Sorting Genetic Algorithm II is adopted to solve the proposed model. A case study including three workpieces with twenty-three machining features to be processed by turning, milling, drilling, boring and grinding is used to verify the proposed integrated model and algorithm. Results show that the proposed integrated optimization method can further reduce carbon emissions and makespan in manufacturing processes compared with conventional separated optimization of process planning and scheduling. The proposed integrated optimization method is validated.

Remaining useful life (RUL) prognosis is of great significance to improve the reliability, availability, and maintenance cost of an industrial equipment. Traditional machine learning method is not fit for dealing with time series signals and has low generalization and stability in prognostic. In this article, a novel ensemble long short-term memory neural network (ELSTMNN) model for RUL prediction is proposed to enhance the RUL prognosis accuracy and improve the adaptive and generalization abilities under different prognostic scenarios. The ELSTMNN contains a series of long short-term memory neural networks (LSTMNNs), each of which is trained on a unique set of historical data. A novel ensemble method is first proposed using Bayesian inference algorithm to integrate multiple predictions of the LSTMNNs for the optimal RUL estimation. The effectiveness of the ELSTMNN-based RUL prognosis method is validated using two characteristically different turbofan engine data sets. The experimental results show a competitive performance of the ELSTMNN in comparison with other prognostic methods.

Fault diagnosis is very important to ensure the efficiency and reliability of rotating machinery. Traditional fault diagnosis methods often require manual feature design and extraction, which is an exhausting work in industrial applications. Convolutional neural network (CNN)-based methods have presented a great potential to automatically extract and select representative features for fault diagnosis. In this article, a novel three-step intelligent fault diagnosis method is proposed based on CNN and Bayesian Gaussian mixture (BGM) for rotating machinery. In the fault dataset construction step, multiple binary training datasets are constructed to amplify the difference between one fault mode and the others using a defined fault labeling rule. In the fault feature extraction step, multiple binary classification tasks are, respectively, implemented to learn the representative features for each fault mode through a proposed group of CNN models based on the above-obtained training datasets. In the fault diagnosis step, multiple Gaussian mixture models (GMMs) are adopted to fit the data distributions of the learned fault features, and an BGM model based on GMM parameters and Bayesian network is designed to establish a cause-and-effect relationship between corresponding parameters and fault modes. Two case studies are used to evaluate the effectiveness of the proposed method. The experimental results show that the proposed method has a superior performance over the existing methods.

Accurate health evaluation is crucial to reliable operation of complex degradation systems. Although traditional machine learning methods such as artificial neural network (ANN) and support vector machine (SVM) have been used widely, state assessment schemes based on a single classification model still suffer from low multiclass classification efficiency, high variance, and deviation. To solve these problems, this article proposes a novel health evaluation method based on stacking ensemble learning and generalized multiclass support vector machine (GMSVM) algorithm. The proposed health evaluation framework includes three parts: 1) abnormal value elimination and missing value processing are applied for multiple sensor data; 2) statistical features are extracted from the observed data and the Pearson correlation coefficient is applied for feature selection; and 3) ensemble generalized multiclass support vector machines (EGMSVMs) are utilized to evaluate the health situation of a degradation system. Unlike the binary classifiers and deep-learning-based classifiers, EGMSVMs utilize the stacking-based method to combine several GMSVMs as submodels and random forest as a metamodel, and the metamodel ensembles the results of submodels to reach a satisfied performance. Compared to traditional SVM- and ANN-based algorithms, EGMSVMs, in processing multiclass problems, achieve high efficiency and, meanwhile, low variance and deviation. The proposed method is verified using a hydraulic test rig. The experimental results show the feasibility and applicability of the proposed health evaluation framework.

With the rapid development of computer technology and related softwares for mathematical models, mathematical modelling of scheduling problems is receiving growing attention from researchers. In th...

PurposeEngineering design optimization involving computational simulations is usually a time-consuming, even computationally prohibitive process. To relieve the computational burden, the adaptive metamodel-based design optimization (AMBDO) approaches have been widely used. This paper aims to develop an AMBDO approach, a lower confidence bounding approach based on the coefficient of variation (CV-LCB) approach, to balance the exploration and exploitation objectively for obtaining a global optimum under limited computational budget.Design/methodology/approachIn the proposed CV-LCB approach, the coefficient of variation (CV) of predicted values is introduced to indicate the degree of dispersion of objective function values, while the CV of predicting errors is introduced to represent the accuracy of the established metamodel. Then, a weighted formula, which takes the degree of dispersion and the prediction accuracy into consideration, is defined based on the already-acquired CV information to adaptively update the metamodel during the optimization process.FindingsTen numerical examples with different degrees of complexity and an AIAA aerodynamic design optimization problem are used to demonstrate the effectiveness of the proposed CV-LCB approach. The comparisons between the proposed approach and four existing approaches regarding the computational efficiency and robustness are made. Results illustrate the merits of the proposed CV-LCB approach in computational efficiency and robustness.Practical implicationsThe proposed approach exhibits high efficiency and robustness in engineering design optimization involving computational simulations.Originality/valueCV-LCB approach can balance the exploration and exploitation objectively.

Machine health monitoring is of great importance in industrial informatics field. Recently, deep learning methods applied to machine health monitoring have been proven effective. However, the existing methods face enormous difficulties in extracting heterogeneous features indicating the variation until failure and revealing the inherent high-dimensional features of massive signals, which affect the accuracy and efficiency of machine health monitoring. In this paper, a novel data-driven machine health monitoring method is proposed using adaptive kernel spectral clustering (AKSC) and deep long short-term memory recurrent neural networks (LSTM-RNN). This method include three steps: First, features in the time domain, frequency domain, and time-frequency domain are, respectively, extracted from massive measured signals. And, an Euclidean distance based algorithm is designed to select degradation features. Second, the AKSC algorithm is introduced to adaptively identify machine anomaly behaviors from multiple degradation features. Third, a new deep learning model (LSTM-RNN) is constructed to update and predict the failure time of the machine. The effectiveness of the proposed method is validated using a set of test-to-failure experimental data. The results show that the performance of the proposed method is competitive with other existing methods.

With energy shortage and environmental pollution becoming increasingly severe problems, energy-efficient scheduling is attracting much more attention than before. This paper addresses the flexible job shop scheduling problem (FJSP) with the objective of minimizing total energy consumption. Firstly, the total energy consumption of the workshop is discussed and modelled. Then, based on two different modeling ideas namely idle time variable and idle energy variable, six new mixed integer linear programming (MILP) models with turning Off/On strategy are proposed. The original objective function of the model based on idle time variable is nonlinear and linearization is needed. For the linearization, additional decision variables and constraints are added. The objective function of the model based on idle energy variable is originally linear and concise. Lastly, those six proposed models and the existing one are detailedly compared and evaluated under both the size and computational complexities. The correctness and effectiveness of all MILP models are verified by using CPLEX SLOVER to carry out numerical experiments. The results show that the MILP models based on different modeling ideas vary remarkably in both size and computational complexities, and all the six models proposed in this paper outperform the existing model significantly. The proposed models will help the enterprises rationalize production so as to reduce energy consumption and costs.

Health degradation monitoring of rolling element bearing (REB) is of great significance to ensure safety and availability of mechanical equipment. This paper presents a new online health degradation monitoring method of REB based on growing self-organizing mapping (GSOM) and clustered support vector machine (CSVM). In the proposed method, multidimensional health degradation features of the REB are extracted to reflect health degradation process, including time-domain statistical features, frequency spectrum features, intrinsic mode function energy features, wavelet packet frequency band energy features. Multiple GSOMs are utilized to adaptively fuse each kind of the extracted features for the health indices. CSVM is constructed to achieve accurate health status identification of the REB. A health degradation experimental case of the REB is analyzed to demonstrate the effectiveness of the proposed method. The results show that the proposed method has obvious superior performance compared to other existing methods.

Health degradation monitoring of machining tools is of great importance in industrial application field. In this paper, a novel multisensory data-driven health degradation monitoring system schema for the machining tools is proposed by using a generalized multiclass support vector machine (GenSVM). In this schema, multidimensional feature extraction is implemented in the time domain, frequency domain, and time-frequency domain based on the time domain statistical analysis, power spectrum analysis, and complete ensemble empirical mode decomposition with adaptive noise, respectively. On the basis, effective features that are strongly associated with the degradation process are picked out using a Pearson correlation coefficient. Meanwhile, a new and flexible GenSVM model is constructed to identify the health state of machining tools, which integrates a simplex encoding and iterative majorization optimization algorithm. A practical application case study is implemented to verify the effectiveness of the proposed method. The results show the superior performance of the proposed method compared with other standard methods.

Time-to-failure (TTF) prognostic plays a crucial role in predicting remaining lifetime of electrical machines for improving machinery health management. This paper presents a novel three-step degradation data-driven TTF prognostics approach for rolling element bearings (REBs) in electrical machines. In the degradation feature extraction step, multiple degradation features, including statistical features, intrinsic energy features, and fault frequency features, are extracted to detect the degradation phenomenon of REBs using complete ensemble empirical mode decomposition with adaptive noise and Hilbert–Huang transform methods. In degradation feature reduction step, the degradation features, which are monotonic, robust, and correlative to the fault evolution of the REBs, are selected and fused into a principal component Mahalanobis distance health index using dynamic principal component analysis and Mahalanobis distance methods. In TTF prediction step, the degradation process and local TTF of the REBs are observed by an exponential regression-based local degradation model, and the global TTF is predicted by an empirical Bayesian algorithm with a continuous update. A practical case study involving run-to-failure experiments of REBs on PRONOSTIA platform is provided to validate the effectiveness of the proposed approach and to show a more accurate prediction of TTF than the existing major approaches.

Bayesian optimization is a metamodel-based global optimization approach that can balance between exploration and exploitation. It has been widely used to solve single-objective optimization problems. In engineering design, making trade-offs between multiple conflicting objectives is common. In this work, a multi-objective Bayesian optimization approach is proposed to obtain the Pareto solutions. A novel acquisition function is proposed to determine the next sample point, which helps improve the diversity and convergence of the Pareto solutions. The proposed approach is compared with some state-of-the-art metamodel-based multi-objective optimization approaches with four numerical examples and one engineering case. The results show that the proposed approach can obtain satisfactory Pareto solutions with significantly reduced computational cost.

This paper investigates an energy-conscious hybrid flow shop scheduling problem with unrelated parallel machines (HFSP-UPM) with the energy-saving strategy of turning off and on. We first analyse t...

Remaining useful life (RUL) prediction of machining tools is a typical multi-sensor information fusion problem. It involves the use of the monitoring information acquired from different types of sensors installed on computer numerical control machine to realize the RUL prediction of the machining tools in cutting process. Owing to the nonlinear and stochastic nature between the extracted features and tool wear level, the promptness and precision of online RUL prediction of machining tools are still difficult to be obtained. In this paper, a multi-sensor information fusion system for online RUL prediction of machining tools is proposed. The system includes sensor signal preprocessing based on ensemble empirical mode decomposition method, statistics feature extraction based on time domain and frequency domain analysis, optimum feature selection based on Pearson correlation coefficient, monotonicity and autocorrelation, feature fusion based on adaptive network based fuzzy inference system and RUL prediction model based on polynomial curve fitting method. We report a practical application of this multi-sensor information system and estimate its prediction performance. The proposed system may be applied to the industrial field. Meanwhile, the comparison between the proposed method and other standard methods is carried out using several statistical indices.

Population-based algorithms, which require a large number of fitness evaluations, can become computationally intractable when applied in engineering design optimization problems involving computational expensive simulations. To address this challenge, this paper proposes an on-line variable-fidelity metamodel assisted Multi-Objective Genetic Algorithm (OLVFM-MOGA) approach. In OLVFM-MOGA, the variable-fidelity metamodel (VFM) is constructed to replace the expensive simulation models to ease the computational burden. Besides, a novel model updating strategy is developed to address the issues of 1) which sample points should be sent for simulation analysis to improve the optimization quality, and 2) whether the low-fidelity (LF) model or the high-fidelity (HF) model should be selected to run for a selected sample point. Six numerical examples and an engineering case with different degrees of complexity are used to demonstrate the applicability and efficiency of the proposed approach. Results illustrate that the proposed OLVFM-MOGA is able to obtain comparable convergence and diversity of the Pareto frontier as to that obtained by MOGA with HF model, while at the same time significantly reducing the computational cost.

Purpose This paper aims to propose hybrid global optimization based on multiple metamodels for improving the efficiency of global optimization. Design/methodology/approach The method has fully utilized the information provided by different metamodels in the optimization process. It not only imparts the expected improvement criterion of kriging into other metamodels but also intelligently selects appropriate metamodeling techniques to guide the search direction, thus making the search process very efficient. Besides, the corresponding local search strategies are also put forward to further improve the optimizing efficiency. Findings To validate the method, it is tested by several numerical benchmark problems and applied in two engineering design optimization problems. Moreover, an overall comparison between the proposed method and several other typical global optimization methods has been made. Results show that the global optimization efficiency of the proposed method is higher than that of the other methods for most situations. Originality/value The proposed method sufficiently utilizes multiple metamodels in the optimizing process. Thus, good optimizing results are obtained, showing great applicability in engineering design optimization problems which involve costly simulations.

Computational simulation models with different fidelities are usually available in the design of engineering products for obtaining the quantity of interest (QOI). To integrate and fully exploit variable fidelity information, a space mapping based variable-fidelity metamodeling (VFM) approach is developed in this work. Firstly, a Gaussian process (GP) model is constructed for the low-fidelity (LF) model. Secondly, a variable-fidelity metamodel is constructed by taking the predicted information from this GP model as a prior-knowledge of the QOI and directly mapped into the outputs space of the high-fidelity (HF) model. This space mapping process is performed by constructing another GP model. A mathematic example is first adopted for illustrating how the proposed approach works under different sample sizes and sample noises. Then, the proposed approach is applied to two real-life cases, modeling of the maximum stress for the structure of a Small Waterplane Area Twin Hull (SWATH) catamaran and predicting weld geometry in fiber laser keyhole welding, to illustrate its ability in support of complex engineering design.

A mixed-model two-sided assembly line is a manufacturing system designed for the production of large-sized products. In order to describe the actual condition, this paper presents a novel multiobjective programming model for balancing a mixed-model two-sided assembly line subject to multiple constraints, in which, additional constraints including zoning, synchronous, and positional constraints are considered besides the traditional constraints, e.g., the precedence constraint. Two objectives are simultaneously to be optimized, one is to minimize the combination of the weighted line efficiency and the weighted smoothness index, and the other is to minimize the weighted total relevant costs per unit of a product. A novel multiobjective hybrid imperialist competitive algorithm (MOHICA) is proposed to solve this problem. In the presented MOHICA, the sigma method is employed to quantify every individual, a novel merging method is introduced to reserve better individuals into the evolutionary population, and late acceptance hill-climbing (LAHC) algorithm is presented as a local search algorithm to achieve accurate balance between intensification and diversification. The experimental results on the selected benchmark instances and a practical case show that the proposed multiobjective algorithm outperforms nondominated sorting genetic algorithm (NSGA)-II, multiobjective improved teaching-learning-based optimization, and NSGA-III existing in the literature.

Purpose Metamodels are widely used to replace simulation models in engineering design optimization to reduce the computational cost. The purpose of this paper is to develop a novel sequential sampling strategy (weighted accumulative error sampling, WAES) to obtain accurate metamodels and apply it to improve the quality of global optimization. Design/methodology/approach A sequential single objective formulation is constructed to adaptively select new sample points. In this formulation, the optimization objective is to select a sample point with the maximum weighted accumulative predicted error obtained by analyzing data from previous iterations, and a space-filling criterion is introduced and treated as a constraint to avoid generating clustered sample points. Based on the proposed sequential sampling strategy, a two-step global optimization approach is developed. Findings The proposed WAES approach and the global optimization approach are tested in several cases. A comparison has been made between the proposed approach and other existing approaches. Results illustrate that WAES approach performs the best in improving metamodel accuracy and the two-step global optimization approach has a great ability to avoid local optimum. Originality/value The proposed WAES approach overcomes the shortcomings of some existing approaches. Besides, the two-step global optimization approach can be used for improving the optimization results.

Multi-fidelity (MF) metamodeling approaches have attracted significant attention recently for data regression because they can make a trade-off between high accuracy and low computational expense by integrating the information from high-fidelity (HF) and low-fidelity (LF) models. To facilitate the usage of the MF metamodeling approaches, there are still challenging issues on the sample size ratio between HF and LF models and the locations of samples since these two components have profound effects on the prediction accuracy of the MF metamodels. In this study, a sequential multi-fidelity (SMF) metamodeling approach is proposed to address the issues of 1) where to allocate the LF and HF sample points, and 2) how to obtain an optimal combination of the high and low-fidelity sample sizes for a given computational budget and a high-to-low simulation cost ratio. Firstly, sequential objective formulations, with the objective to reduce the estimation of prediction error of MF metamodel, are constructed to update the LF and HF sampling data. Secondly, a decision criterion is proposed to determine whether one HF experiment or several LF experiments with the equivalent computational cost should be selected to update the MF metamodel. The proposed criterion is developed according to which selection will have a greater potential value to improve the prediction accuracy of the MF metamodel. To demonstrate the effectiveness and merits of the proposed SMF metamodeling approach, two numerical examples and a practical aerospace application example are used. Results show that the proposed approach can generate more accurate MF metamodels by providing the optimal high-to-low sample size ratio and sample locations.

The radial basis function has been widely used in constructing metamodels as response surfaces. Yet, it often faces the challenge of accuracy if a sequential sampling strategy is used to insert samples sequentially and refine the models, especially under the constraint of computational resources. In this paper, a sensitive region pursuing based active learning radial basis function (SRP-ALRBF) metamodeling approach is proposed to sequentially exploit the already-acquired knowledge in the modeling process for obtaining a desirable estimation of the relationship between the input design variables and the output response. In this method, the leave-one-out (LOO) errors of each sample point are taken to identify the boundaries of sensitive regions. According to the obtained LOO information, the original design space is divided into some subspaces by adopting the self-organization maps (SOMs). The boundary of the most sensitive region, where the output response is multi-modal or non-smooth with abrupt changes, is determined by the topological graph generated by cluster analysis in SOMs. In the most sensitive region, infill sample point searching is performed based on an optimization formulation. Ten numerical examples are used to compare the proposed SRP-ALRBF with four existing active learning RBF metamodeling approaches. Results show the advantage of the proposed SRP-ALRBF approach in both prediction accuracy and robustness. It is also applied to three engineering cases to illustrate its ability to support complex engineering design.

Metamodels have been widely used in engineering design and optimization. Sampling method plays an important role in the constructing of metamodels. This paper proposes an adaptive sampling strategy for Kriging metamodel based on Delaunay triangulation and TOPSIS (KMDT). In the proposed KMDT, Delaunay triangulation is employed to partition the design space according to current sample points. The area of each partitioned triangle is used to indicate the degree of dispersion of sample points, and the prediction error of Kriging metamodel at each triangle’s centroid is used to represent the local error of each triangle region. By calculating the weight of the area and prediction error for each triangle region using the entropy method and TOPSIS, the degree of dispersion of sample points and local errors of metamodel are taken into consideration to make a trade-off between global exploration and local exploitation during the sequential sampling process. As a demonstration, the proposed approach is compared to other three sampling methods using several numerical cases and the modeling of the aerodynamic coefficient for a three-dimensional aircraft. The result reveals that the proposed approach provides more accurate metamodel at the same simulation cost, which is very important in metamodel-based engineering design problems.

The multifidelity metamodeling method provides an efficient way to approximate expensive black-box problems by using samples of different fidelities. Assisted by abundant cheap low-fidelity data, i...

Variable-fidelity (VF) modelling methods have been widely used in complex engineering system design to mitigate the computational burden. Building a VF model generally includes two parts: design of experiments and metamodel construction. In this article, an adaptive sampling method based on improved hierarchical kriging (ASM-IHK) is proposed to refine the improved VF model. First, an improved hierarchical kriging model is developed as the metamodel, in which the low-fidelity model is varied through a polynomial response surface function to capture the characteristics of a high-fidelity model. Secondly, to reduce local approximation errors, an active learning strategy based on a sequential sampling method is introduced to make full use of the already required information on the current sampling points and to guide the sampling process of the high-fidelity model. Finally, two numerical examples and the modelling of the aerodynamic coefficient for an aircraft are provided to demonstrate the approximation capability of the proposed approach, as well as three other metamodelling methods and two sequential sampling methods. The results show that ASM-IHK provides a more accurate metamodel at the same simulation cost, which is very important in metamodel-based engineering design problems.

A surrogate model built based on a limited number of sample points will inevitably have large prediction uncertainty. Applying such imprecise surrogate models in design and optimization may lead to misleading predictions or optimal solutions located in unfeasible regions (Picheny in Improving accuracy and compensating for uncertainty in surrogate modeling. University of Florida, Gainesville, 2009). Therefore, verifying the accuracy of a surrogate model before using it can ensure the reliability of the design.

An ensemble of surrogate models (EM) is a surrogate model composed of a series of surrogate models combined through a weighted sum. An EM can take advantage of each individual surrogate model to effectively increase the robustness of the prediction. The mathematical expression for an EM can be given as follows

Simulation models are treated as black boxes that generate input–output correspondences. Nevertheless, designers need to choose simulation models with appropriate fidelities to obtain qualities of interest (QOIs) at affordable cost levels. Generally, high-fidelity (HF) simulation models can provide more reliable and accurate simulation results than low-fidelity (LF) models. Consider the example of designing an airfoil, which is an aerodynamic component, the available simulation models for obtaining aerodynamic coefficients may differ in terms of their resolutions (e.g. coarse meshes versus refined meshes in finite element models), levels of abstraction (e.g. two-dimensional models versus three-dimensional models) or mathematical descriptions (e.g. the Euler non-cohesive equations versus the Navier–Stokes viscous Newton equations). Relying entirely on HF models to obtain QOIs for constructing a surrogate model is always time-consuming and may even be computationally prohibitive. On the other hand, LF models are considerably less computationally demanding. However, the QOIs obtained from LF models may result in inaccurate surrogate models or even distorted ones.

Accurate and timely bearing fault diagnosis is crucial to decrease the probability of unexpected failures of rotating machinery and improve the efficiency of its scheduled maintenance. Since convolutional neural networks (CNN) have poor feature extraction capability for sensor data with 1D format, CNN combined with signal processing algorithm is often adopted for fault diagnosis. This increases manual conversion work and expertise dependence while reducing the feasibility and robustness of the corresponding fault diagnosis method. In this paper, a novel sensor data-driven fault diagnosis method is proposed by fusing S-transform (ST) algorithm and CNN, namely ST-CNN. First of all, a ST layer is designed based on S-transform algorithm. In the ST layer, sensor data is automatically converted into 2D time-frequency matrix without manual conversion work. Then, a new ST-CNN model is constructed, and the time-frequency coefficient matrixes are inputted into the constructed ST-CNN model. After the training process of the ST-CNN model is completed, the classification layer such as softmax performs the fault diagnosis. Finally, the diagnosis performance of the proposed method is evaluated by using two public available datasets of bearings. The experimental results show that the proposed method performs the higher and more robust diagnosis performance than other existing methods.

Remaining useful life (RUL) prediction is very important for improving the availability of a system and reducing its life cycle cost. This paper proposes a deep long short-term memory (DLSTM) network-based RUL prediction method using multiple sensor time series signals. The DLSTM model fuses multi-sensor monitoring signals for accurate RUL prediction, which is able to discover the hidden long-term dependencies among sensor time series signals through deep learning structure. By grid search strategy, the network structure and parameters of the DLSTM are efficiently tuned using an adaptive moment estimation algorithm so as to realize an accurate and robust prediction. Two various turbofan engine datasets are adopted to verify the performance of the DLSTM model. The experimental results demonstrate that the DLSTM model has a competitive performance in comparison with state-of-the-arts reported in literatures and other neural network models.

Health prognosis of rolling bearing is of great significance to improve its safety and reliability. This paper presents a novel health prognosis method for the rolling bearing based on convolutional neural network (CNN) and bidirectional long short-term memory network (BiLSTM) model. First, a new nonlinear degradation indicator (DI) is designed which can be utilized as training label. Then, through learning and capturing the mapping relationship between raw vibration signals and DI of the rolling bearing, a CNN model is introduced to estimate the DI value of the rolling bearing. And, BiLSTM models are set up to carry out health prognosis using the estimated DI, including future DI and remaining useful life prediction. An experiment verification is implemented to validate the effectiveness of the proposed method. Results show the excellent ability of future DI prediction, and demonstrate the superiority of the proposed method in the field of remaining useful life prediction compared with other existing deep learning models.

This paper presents a data-driven intelligent fault diagnosis approach for rotating machinery (RM) based on a novel continuous wavelet transform-local binary convolutional neural network (CWT-LBCNN) model. The proposed approach builds an end-to-end diagnosis mechanism, and does not need manual feature extraction. By feeding the inputting vibration signal, features are captured adaptively and fault condition of RM is diagnosed automatically. Different from traditional CNNs, the proposed CWT-LBCNN utilizes a local binary convolution layer to replace a traditional convolution layer, which enables CWT-LBCNN to have faster training speed and less proneness to overfitting. Two experimental studies including bearing fault diagnosis and gearbox compound fault diagnosis show that the proposed CWT-LBCNN has more stable and reliable prediction accuracy than other existing methods.

Multi-fidelity metamodeling provides an efficient way to approximate expensive black-box problems by utilizing the samples of multiple fidelities. While it still faces the challenge of "curse-of-dimensionality" when used in approximating high dimensional problems. On the other hand, the high dimensional model representation (HDMR) method, as an efficient tool to tackle high dimensional problems, can only handle single-fidelity samples in approximation. Therefore, a hybrid metamodel which combines Cut-HDMR with Co-kriging and kriging is proposed to improve the metamodeling efficiency for high dimensional problems. The developed HDMR, termed as MF-HDMR, can efficiently use multi-fidelity samples to approximate black-box problems by using a two stage metamodeling strategy. It can naturally explore and exploit the linearity/nonlinearity and correlations among variables of underlying problems, which are unknown or computationally expensive. Besides, to further improve the efficiency of MF-HDMR, an extended maximin distance sequential sampling method is proposed to add new sample points of different fidelities in the metamodeling process. Moreover, a mathematical function is used to illustrate the modeling theory and procedures of MF-HDMR. In order to validate the proposed method, it is tested by several numerical benchmark problems and successfully applied in the optimal design of a long cylinder pressure vessel. Moreover, an overall comparison between the proposed method and several other metamodeling methods has been made. Results show that the proposed method is very efficient in approximating high dimensional problems by using multi-fidelity samples, thus making it particularly suitable for high dimensional engineering design problems involving computationally expensive simulations.

Using computationally cheap low-fidelity (LF) model and more accurate but expensive high-fidelity (HF) model, variable fidelity (VF) model has been widely used in engineering design to replace the actual computationally expensive experiments or computer simulations. To further extend the application of VF to reliability-based design optimization (RBDO), a new framework based on sequential linear programming (SLP) is proposed in this paper. Combining the advantages of additive scaling method and multiplicative scaling method, a hybrid scaling method based on least squares (LSHS) is developed. In LSHS method, the VF model is introduced to replace the implicit performance function in RBDO by using the HF function values and gradient values at all evaluated points around the current design. With the failure probability and its gradient calculated by Monte Carlo Simulation (MCS) at current design, SLP is adopted to calculate the next design. A novel method which considers the target reliability index and the influence domain at the current design is also developed to determine the step size in every sub-optimization problem. Two numerical examples and the shape optimization problem of a curved beam are analyzed in order to demonstrate the performance of the proposed methodology. The comparison results show that the proposed method is very accurate and efficient.

Hybrid laser–arc welding (LAW) provides an effective way to overcome problems commonly encountered during either laser or arc welding such as brittle phase formation, cracking, and porosity. The process parameters of LAW have significant effects on the bead profile and hence the quality of joint. This paper proposes an optimization methodology by combining non-dominated sorting genetic algorithm (NSGA-II) and ensemble of metamodels (EMs) to address multi-objective process parameter optimization in LAW onto 316L. Firstly, Taguchi experimental design is adopted to generate the experimental samples. Secondly, the relationships between process parameters (i.e., laser power (P), welding current (A), distance between laser and arc (D), and welding speed (V)) and the bead geometries are fitted using EMs. The comparative results show that the EMs can take advantage of the prediction ability of each stand-alone metamodel and thus decrease the risk of adopting inappropriate metamodels. Then, the NSGA-II is used to facilitate design space exploration. Besides, the main effects and contribution rates of process parameters on bead profile are analyzed. Eventually, the verification experiments of the obtained optima are carried out and compared with the un-optimized weld seam for bead geometries, weld appearances, and welding defects. Results illustrate that the proposed hybrid approach exhibits great capability of improving welding quality in LAW.

This paper proposes a new stacked pruning sparse denoising autoencoder (sPSDAE) model for intelligent fault diagnosis of rolling bearings. Different from the traditional autoencoder, the proposed sPSDAE model, including a fully connected autoencoder network, uses the superior features extracted in all the previous layers to participate in the subsequent layers. This means that some new channels are created to connect the front layers and the back layers, which reduces information loss. To improve the training efficiency and precision of the sPSDAE model, a pruning operation is added into the sPSDAE model so as to prohibit non-superior units from participating in all the subsequent layers. Meanwhile, a feature fusion mechanism is introduced to ensure the uniqueness of the feature dimensions. After that, the sparse expression of the sPSDAE model is strengthened, thereby improving the generalization ability. The proposed method is evaluated by using a public bearing dataset and is compared with other popular fault diagnosis models. The results show that the ability of the sPSDAE model to extract features is significantly enhanced and the phenomenon of gradient disappearance is further reduced. The proposed model achieves higher diagnostic accuracy than other popular fault diagnosis models.

The integration of process planning and scheduling is important for an efficient utilization of manufacturing resources. However, the focus of existing works is mainly on deterministic constraints of jobs. This article proposes a novel memetic algorithm for the integrated process planning and scheduling problem with processing time uncertainty based on processing time scenarios. First, a mathematical model for the stochastic integrated process planning and scheduling problem based on the network graph is established. Due to the nonlinearity in the model and the complexity of the problem, a memetic algorithm is then suggested for this problem. A novel local search (variable neighborhood search) algorithm is incorporated into the memetic algorithm. Two effective neighborhood structures are employed in the variable neighborhood search algorithm to improve the overall performance of the population. Furthermore, for the uncertainty in processing times, a set of scenarios have been generated to evaluate each individual. Finally, two performance measures—the expected performance measure and the worst-case deviation measure—are introduced and compared. In the experimental studies, the proposed memetic algorithm is tested on typical benchmark instances. Computational results show that the expected makespan measure performs better than the worst-case deviation measure and the proposed method exhibits high performance especially for large-scale instances. In addition, the results obtained by the proposed memetic algorithm are more satisfactory than those obtained by the algorithm that considers deterministic processing times only.

Metamodels are often used in engineering design optimization problems with expensive simulations to save computational cost. But these metamodels often face "curse-of-dimensionality" when used in approximating high dimensional problems. Therefore, a new high dimensional model representation (HDMR) by combining Cut-HDMR with an enhanced RBF based on ensemble model is proposed. The developed HDMR, termed as ERBF-HDMR, sufficiently utilizes advantages of RBF and ensemble model in the modeling process. It can naturally explore and exploit the linearity/nonlinearity and correlations among variables of underlying problems, which are unknown or computationally expensive. Besides, to improve the efficiency of the ERBF-HDMR, an adaptive sampling method is proposed to add new sample points. Moreover, a mathematical function is used to illustrate the modeling principles and procedures of the adaptive ERBF-HDMR. And a comprehensive comparison between the adaptive ERBF-HDMR and other different Cut-HDMRs in literature has been made on eleven numerical examples with a wide scope of dimensionalities to show the prediction ability of different HDMRs. Finally, the proposed HDMR is used in the structural design optimization of the bearings of an all-direction propeller with the aim of reducing vibration.

Computational simulation models with variable fidelity have been widely used in complex systems design. However, running the most accurate simulation models tends to be very time-consuming and can therefore only be used sporadically, while incorporating less accurate, inexpensive models into the design process may result in inaccurate design alternatives. To make a trade-off between high accuracy and low expense, variable fidelity (VF) metamodeling approaches that aim to integrate information from both low-fidelity (LF) and high-fidelity (HF) models have gained increasing popularity. In this paper, an adaptive global VF metamodeling approach named difference adaptive decreasing variable-fidelity metamodeling (DAD-VFM) is proposed, in which the one-shot VF metamodeling process is transformed into an iterative process to utilize the already-acquired information of difference characteristics between the HF and LF models. In DAD-VFM, support vector regression (SVR) is adopted to map the difference between the HF and LF models. Besides, a generalized objective-oriented sampling strategy is introduced to adaptively probe and sample more points in the interesting regions where the differences between the HF and LF models are multi-model, non-smooth and have abrupt changes. Several numerical cases and a long cylinder pressure vessel optimization design problem verify the applicability of the proposed VF metamodeling approach.

Process planning and scheduling are two crucial components in a manufacturing system. The integration of the two functions has an important significance on improving the performance of the manufacturing system. However, integrated process planning and scheduling is an intractable non-deterministic polynomial-time (NP)-hard problem, and the multiple objectives requirement widely exists in real-world production situations. In this paper, a multi-objective mathematical model of integrated process planning and scheduling is set up with three different objectives: the overall finishing time (makespan), the maximum machine workload (MMW), and the total workload of machines (TWM). A multi-objective memetic algorithm (MOMA) is proposed to solve this problem. In MOMA, all the possible schedules are improved by a problem-specific multi-objective local search method, which combines a variable neighborhood search (VNS) procedure and an effective objective-specific intensification search method. Moreover, we adopt the TOPSIS method to select a satisfactory schedule scheme from the optimal Pareto front. The proposed MOMA is tested on typical benchmark instances and the experimental results are compared with those obtained by the well-known NSGA-II. Computational results show that MOMA is a promising and very effective method for the multi-objective IPPS problem.

Mixed-model two-sided assembly lines are widely used in a range of industries for their abilities of increasing the flexibility to meet a high variety of customer demands. Balancing assembly lines is a vital design issue for industries. However, the mixed-model two-sided assembly line balancing (MTALB) problem is NP-hard and difficult to solve in a reasonable computational time. So it is necessary for researchers to find some efficient approaches to address this problem. Honey bee mating optimization (HBMO) algorithm is a population-based algorithm inspired by the mating process in the real colony and has been applied to solve many combinatorial optimization problems successfully. In this paper, a hybrid HBMO algorithm is presented to solve the MTALB problem with the objective of minimizing the number of mated-stations and total number of stations for a given cycle time. Compared with the conventional HBMO algorithm, the proposed algorithm employs the simulated annealing (SA) algorithm with three different neighborhood structures as workers to improve broods, which could achieve a good balance between intensification and diversification during the search. In addition, a new encoding and decoding scheme, including the adjustment of the final mated-station, is devised to fit the MTALB problem. The proposed algorithm is tested on several sets of instances and compared with Mixed Integer Programming (MIP) and SA. The superior results of these instances validate the effectiveness of the proposed algorithm. Display Omitted

Engineering design, especially for complex engineering systems, is usually a time-consuming process involving computation-intensive computer-based simulation and analysis methods. A difference mapping method using least square support vector regression is developed in this work, as a special metamodelling methodology that includes variable-fidelity data, to replace the computationally expensive computer codes. A general difference mapping framework is proposed where a surrogate base is first created, then the approximation is gained by a mapping the difference between the base and the real high-fidelity response surface. The least square support vector regression is adopted to accomplish the mapping. Two different sampling strategies, nested and non-nested design of experiments, are conducted to explore their respective effects on modelling accuracy. Different sample sizes and three approximation performance measures of accuracy are considered.

Engineering design is usually a daunting optimization task which often involving time-consuming, even computation-prohibitive process. To reduce the computational expense, metamodels are commonly used to replace the actual expensive simulations or experiments. In this paper, a new and efficient metamodeling method named prior-knowledge input least square support vector regression (PKI-LSSVR) is developed, in which samples from different levels of fidelity are incorporated to gain an accurate approximation with limited times of the high-fidelity (HF) expensive simulations. The low-fidelity (LF) output serves as a prior-knowledge of the real response function, and then is used as the input variables of least square support vector regression (LSSVR). When the corresponding HF response is gained, a function that maps the LF outputs to HF outputs is constructed via LSSVR. The predictive accuracy of LSSVR models is highly dependent on their learning parameters. Therefore, a novel optimization method, cellular particle swarm optimization (CPSO), is exploited to seek the optimal hyper-parameters for PKI-LSSVR in order to improve its generalization capability. To get a better optimization performance, a new neighborhood function is developed for CPSO where the global and local search is efficiently balanced by adaptively varied neighbor radius. Several numerical experiments and one engineering case verify the efficiency of the proposed PKI-LSSVR method. Sample quality merits including sample sizes and noise, and metamodel performance evaluation measures incorporating accuracy, robustness, and efficiency are considered.

Reliability prediction is of great importance to improve the operational safety of machinery and decreasing their maintenance costs. In this paper, a new method combining double-Wiener process model with Monte Carlo algorithm is proposed to solve the problem of degradation modeling and reliability prediction of machinery with multiple degradation characteristics. Unlike other Wiener process-based prediction methods that only the mean is modelled, this method considers the degradation of both mean and variance, making degradation modelling more accurate. Meanwhile, the proposed method estimates reliability level of the machinery based on its entire monitoring information to date through an expectation maximization algorithm and a Monte Carlo algorithm, which does not require historical degradation data of other machines in a population. A numerical degradation case and a bearing degradation case with a set of degradation data are studied to validate the proposed method. The results demonstrate the effectiveness of the proposed method compared with other existing methods.

In this paper, an efficient surrogate-assisted particle swarm optimization algorithm is proposed to further improve the efficiency for optimization of high-dimensional expensive problems, which sometimes involve costly simulation analysis. Unlike several surrogate-assisted metaheuristic algorithms, the proposed algorithm can effectively balance the prediction ability of surrogates and the global search ability of particle swarm optimization in the optimization process. Specifically, the proposed algorithm efficiently uses the optima obtained from the global surrogate built in the entire design space and the local surrogate built in a neighbor region around the personal historical best particle to update the velocities of the particles. It should be stressed that the neighbor region partition strategy used to obtain the optimums of local surrogates is an essential aspect of the proposed algorithm. This strategy helps to obtain the predicted optima of local surrogates in the neighbor regions to guide the search of particle swarm optimization in the optimization process. In addition, the neighbor region partition strategy considers the diversity of personal historical best particles, which enables the proposed algorithm to efficiently search for different types of problems. Moreover, the optimization efficiency of the proposed algorithm can be enhanced by using the surrogate prescreening strategy. In order to validate the proposed algorithm, it is tested on several high-dimensional numerical benchmark problems and comprehensively compared with several optimization algorithms. The results show that the proposed algorithm is very promising for the optimization of high-dimensional expensive problems.

Cutting force is one of the fundamental elements that can provide valuable insight in the investigation of cutter breakage, tool wear, machine tool chatter, and surface finish in face milling. Analyzing the relationship between process factors and cutting force is helpful to set the process parameters of the future cutting operation and further improve production quality and efficiency. Since cutting force is impacted by the inherent uncertainties in the machining process, how to predict the cutting force presents a significant challenge. In the meantime, face milling is a complex process involving multiple experts with different domain knowledge, collaborative evaluation of the cutting force model should be conducted to effectively evaluate the constructed predictive model. Gene Expression Programming (GEP) combines the advantages of the Genetic Algorithm (GA) and Genetic Programming (GP), and has been successfully applied in function mining and formula finding. In this paper, a new approach to predict the face milling cutting force based on GEP is proposed. At the basis of defining a GEP environment for the cutting force prediction, an explicit predictive model has been constructed. To verify the effectiveness of the proposed approach, a case study has been conducted. The comparisons between the proposed approach and some previous works show that the constructed model fits very well with the experimental data and can predict the cutting force with a high accuracy. Moreover, in order to better apply the constructed predictive models in actual face milling process, a collaborative model evaluation method is proposed to provide a distributed environment for geographical distributed experts to evaluate the constructed predictive model collaboratively, and four kinds of collaboration mode are discussed.

The paper considers the dynamic job shop scheduling problem (DJSSP) with job release dates which arises widely in practical production systems. The principle characteristic of DJSSP considered in the paper is that the jobs arrive continuously in time and the attributes of the jobs, such as the release dates, routings and processing times are not known in advance, whereas in the classical job shop scheduling problem (CJSSP), it is assumed that all jobs to be processed are available at the beginning of the scheduling process. Reactive scheduling approach is one of the effective approaches for DJSSP. In the paper, a heuristic is proposed to implement the reactive scheduling of the jobs in the dynamic production environment. The proposed heuristic decomposes the original scheduling problem into a number of sub problems. Each sub problem, in fact, is a dynamic single machine scheduling problem with job release dates. The scheduling technique applied in theproposed heuristic is priority scheduling, which determines the next state of the system based on priority values of certain system elements. The system elements are prioritized with the help of scheduling rules (SRs). An approach based on gene expression programming (GEP) is also proposed in the paper to construct efficient SRs for DJSSP. The rules constructed by GEP are evaluated in the comparison of the rules constructed by GP and several prominent human made rules selected from literatures on extensive problem sets with respect to various measures of performance.