Search

Data-driven soft sensor technology has been widely used in process monitoring, quality prediction, etc. However, there are dynamic time-varying and concept drift problems in industrial processes, making the accurate modeling of the soft sensor is still a challenging task. To solve the above problems, this paper proposes a concept drift adaptive dynamic partial least squares method. The method maps high-dimensional process data into a low-dimensional latent variable subspace. Dynamic information is introduced by establishing the dynamic regression relationship between process latent variables and quality latent variables. At the same time, transfer learning is used to align second-order statistics between latent variables of data from different distributions to solve the concept drift problem. The experiments on multiple industrial datasets show that the method proposed in this paper can effectively reduce prediction errors and improve the generalization ability of the model. [Display omitted] • Distribution alignment and dynamic system modeling are realized under NIPALS. • The objective function is organized into the form of Rayleigh quotient. • There are few hyperparameters and they have better interpretability. • Experiments show that our method can adapt to the change of data distribution well. [ABSTRACT FROM AUTHOR]

With the increasing proportion of solar grid-connected, the establishment of an accurate photovoltaic (PV) power prediction model is very important for safe operation and efficient dispatching of a power grid. Considering the multi-level periodicity of PV power caused by many factors, such as seasons and weather, a short-term PV power prediction model based on transfer component analysis is designed by introducing the idea of transfer learning. In order to measure the uncertainty of numerical weather prediction (NWP) and power sequence, a novel algorithm considering weather similarity and power trend similarity is proposed. First, the intrinsic trend is measured by extracting permutation entropy, variance, and mean from the historical PV power sequence. Second, weighting of NWP is accomplished based on the Pearson correlation coefficient. PV power data are divided into different clusters by K-medoids clustering. At the same time, the transfer component analysis alleviates the time-varying problem of data distribution caused by multi-level time periodicity and effectively improves the prediction accuracy of the model. Finally, simulation experiments are carried out on the PV power output dataset (PVOD). The results show that the prediction accuracy of the proposed method is better than the traditional methods, and the accuracy and applicability of the proposed method are verified. [ABSTRACT FROM AUTHOR]

The traditional soft sensor models are based on the independent and identical distribution assumption, which are difficult to adapt to changes in data distribution under multiple operating conditions, resulting in model performance deterioration. The domain adaptive transfer learning methods learn knowledge in different domains by means of distribution alignment, which can reduce the impact of data distribution differences, and effectively improve the generalization ability of the model. However, most of the existing models established by domain adaptation methods are static models, which cannot reflect the dynamic characteristics of the system, and have limited prediction accuracy when applied to dynamic system modeling under multiple operating conditions. The dynamic system modeling methods can effectively extract the dynamic characteristics of the data, but they cannot deal with the concept drift problem caused by the change of data distribution. This paper proposes a new dynamic transfer partial least squares method, which maps the high-dimensional process data into the low-dimensional latent variable subspace, establishes the dynamic regression relationship between the latent variables and the labels, and realizes the systematic dynamic modeling, at the same time, the model adds regular terms for distribution alignment and structure preservation, which realizes dynamic alignment of data distribution difference. The effectiveness of the proposed method is validated on three publicly available industrial process datasets. [Display omitted] • This paper proposes a novel dynamic transfer partial least squares algorithm (DTPLS). • DTPLS can solve the concept drift problem under the multiple working conditions. • Distribution alignment regularization is designed based on distribution differences. • Structure preservation regularization is designed based on smoothness assumptions. [ABSTRACT FROM AUTHOR]

Aiming at the problem of mismatch between real-time data distribution and modeling data distribution caused by the change of working conditions in industrial process, which leads to the performance deterioration of the soft sensor model, a multi-source unsupervised soft sensor method based on joint distribution alignment and mapping structure preservation is proposed. Firstly, the method uses the hypergraph to establish the complex structure of feature and label, and clusters the hypergraph matrix in multiple views to completely construct the class pseudo label; then dynamic distribution alignment is used to adapt marginal distribution and conditional distribution between the data of historical working conditions and the current working conditions, and the hypergraph Laplacian operator is introduced for manifold regularization to prevent the mapping relationship between feature and label from being destroyed; finally, similar working conditions are introduced to further enhance the robustness of the model. The experimental results show that compared with the traditional unsupervised soft sensor methods, the method used in this paper can effectively improve the prediction accuracy of the model. • DASP constructs pseudo label to dynamically distribute adapt the continuous regression data. • The new label can retain the mapping relationship between the original data feature and label. • Several experimental results verify the effectiveness of the method. [ABSTRACT FROM AUTHOR]

In the grinding industry, accurate prediction of the mill load is the key to enhancing plant profitability and reducing mill failures. After the condition changes, the model is usually mismatched, and the true label cannot be obtained in time. To solve this problem, we propose a multi-source unsupervised domain adaptation model based on deep learning and fusion features (MUDA_DF), which is suitable for predicting multiple mill load parameters. Considering that the information contained in the data of a single working condition is insufficient to completely cover the information of the unmodeled working condition, we introduce multiple source domains into the model, which can improve the generalization of the model. Since the special features of the source domain also contain useful information, we use the features after the fusion of special features and common features for regression prediction. By designing multiple loss functions, we achieve the prediction of three mill load parameters in unknown operating conditions: material-to-ball volume ratio (MBVR), pulp density (PD), and charge volume ratio (CVR). Experiments were carried out on data sets collected by a laboratory ball mill. Compared with the traditional mill load soft measurement method, the R 2 of this method can reach above 0.9 and the highest can reach 0.99, which proves the effectiveness of this method. • An unsupervised domain-adaptive mill load prediction method is proposed. • Applying multi-source domains to mill load regression prediction. • Applying deep learning to mill load regression prediction. • Extract common and special features simultaneously. [Display omitted] [ABSTRACT FROM AUTHOR]

The industrially collected process data usually exhibit non-Gaussian and multi-mode characteristics. Due to sensor failures, irregular disturbances, and transmission problems, there are unavoidable outliers that make the data exhibit heavy-tailed characteristics. To this end, a variational auto-encoder regression method based on the mixture Laplacian distribution (MLVAER) is proposed, by introducing a type-II multivariate Laplacian distribution in the latent variable space for robust modeling, and further extending it to the mixture form to accommodate multi-mode processes, the corresponding reparameterization trick is finally proposed for the mixture form of this distribution for neural network gradient descent training. The model based on this distribution assumption has higher degrees of freedom than the model based on the traditional multivariate Laplace distribution assumption when the network structure is the same. Numerical simulation and experiments on two industrial examples demonstrate that the proposed algorithm reduces the root mean square error by over 15% compared to other algorithms. [Display omitted] • Using mixture Laplace distribution for robust and flexible modeling. • Introduced mixture type II Laplace distribution combined with deep network. • Resampling strategy for multivariate Laplace distribution in gradient descent. [ABSTRACT FROM AUTHOR]

In order to deal with the unknown data drift problem caused by the change of working conditions in the process of industrial soft sensor modeling, this paper proposes a semi-supervised soft sensor modeling algorithm based on instance transfer learning, named instance transfer partial least squares. Under the framework of nonlinear partial least squares, the algorithm introduces the instance transfer inner model, and realizes the distribution adaptation of the modeling data and the data to be predicted through sample importance weighting. First, a nonlinear partial least squares model is constructed by using labeled samples in the source domain, which consists of a linear outer model and a nonlinear inner model. The linear outer model maps the original data to the hidden variable space, and the nonlinear inner model realizes the nonlinear modeling of hidden variables through kernel ridge regression. Second, a small number of labeled samples from the target domain are used to update the model weights of the nonlinear inner model, enabling adaptation of the latent variable distribution. The effectiveness of the algorithm is verified on a toy dataset, a simulation dataset and a real pulverized coal power plant dataset. • A new instance transfer soft sensor method is proposed to solve data drift problem. • The method aligns the data distribution of latent variable while nonlinear modeling. • The algorithm uses sample importance weighting to achieve distribution alignment. • The algorithm is suitable for multiple working conditions and few-sample scenarios. [ABSTRACT FROM AUTHOR]

Soft sensor technology has found widespread application in the real-time detection of challenging variables like product quality and key process parameters. However, changes in working conditions can lead to alterations in data distribution, resulting in model inaccuracies. Furthermore, data from new working conditions arrive sequentially and online. Due to technical difficulties, time-consuming measurements, and high costs, there is often a lack of sufficient new data for modeling in the initial phase. Moreover, industrial processes often exhibit dynamic, nonlinear, variable coupling and other factors, posing significant challenges for model establishment. To address these challenges, this study proposed an online transfer kernel recursive algorithm for soft sensor modeling under changing working conditions. The algorithm consists of an offline phase and an online phase. In the offline phase, the algorithm considers the dynamic features of the process and variable coupling, resulting in the establishment of a dynamic inner model. The original data is then projected onto a latent variable space. In the online phase, the proposed method incorporates the concept of online transfer learning and takes into account the nonlinearity of the process. This allows for the establishment of a kernel online recursive method in the latent variable space, integrating parameter transfer and sample transfer. Experimental results conducted on multiple industrial datasets validate the effectiveness of the proposed approach. [Display omitted] • Online transfer learning enables real-time updates with single samples. • A forgetting mechanism gradually weakens old knowledge as new data arrives. • KDiPLS with unified goals for inner model, outer model, and deflation process. [ABSTRACT FROM AUTHOR]

In this paper, a local cloud model similarity measurement (CMSM) is proposed as a novel method to measure the similarity of time series. Time series similarity measurement is an indispensable part for improving the efficiency and accuracy of prediction. The randomness and uncertainty of series data are critical problems in the processing of similarity measurement. CMSM obtains the internal information of time series from the general perspective and local trend using the cloud model, which reduces the uncertainty of measurement. The neighbor set is selected from time series by CMSM and used to construct a prediction model based on least squares support vector machine. The proposed technique reduces the potential for overfitting and uncertainty and improves model prediction quality and generalization. Experiments were performed with four datasets selected from Time Series Data Library. The experimental results show the feasibility and effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Realistic modeling traffic flow dynamics in urban traffic network remains a big challenge at present due to the complex nonlinear characteristics of traffic flow. In this paper, a novel and model-free iterative learning control (ILC) strategy for balancing traffic flow in urban traffic network is proposed. To tackle the randomly varying trial lengths in the iteration domain of traffic system, an iterative-average operator is introduced in the proposed ILC law for tracking tasks with non-uniform trial lengths, which thus mitigates the requirement on classic ILC that all trial lengths must be identical. The learning convergence condition of the ILC strategy in iteration-average and expectation is derived through rigorous analysis. The performance and the effectiveness of the ILC strategy are analyzed by simulations on a test road network. The results show that the proposed ILC strategy can homogeneously balance the accumulation in the network and improve the network mobility. [ABSTRACT FROM AUTHOR]

This paper presents a practical study on how to improve the ℋ ∞ performance and meet the input–output constraints of the two-degrees-of-freedom (DOF) flexible-joint manipulator system (FJMS) with parameter uncertainties and external disturbances. For this reason, a robust constrained moving-horizon ℋ ∞ controller is designed to improve the system ℋ ∞ performance while still satisfying the input–output constraints of the uncertain system. First, the uncertain controlled system model of the two-DOF FJMS is established via the Lagrange equation method, Spong's assumption, and the linear fractional transformation (LFT) technique. Then, the control requirements and input–output constraints of the uncertain system are transformed into the linear matrix inequality (LMI) via the theory of ℋ ∞ control and the full-block multiplier technique. Next, the LMI optimization problem refreshed by the current state is addressed at each sample moment with the idea of the moving-horizon control of the model predictive control (MPC), and the calculated gain is implemented to the nonlinear closed-loop system under the state feedback structure. The validity and feasibility of the designed control scheme is finally verified via the results of simulation experiments. [ABSTRACT FROM AUTHOR]

• Experiments are conducted to investigate the frequency spectrum characteristics. • Multiple channel mechanical signals are interpreted for mill load parameter (MLP). • A measurement method for the contribution rate of the mechanical signals is proposed. • Mechanical channels must be selected to construct an effective MLP prediction model. Online monitoring load parameters inside the ball mill is the key to improving the production quality and quantity of the mineral grinding process. In this paper, the experimental analysis of wet mill load parameter (MLPs) based on multiple channel mechanical signals is presented. A series of experiments is conducted to investigate the mechanical frequency spectrum characteristics in terms of different grinding conditions, such as only-ball, -mineral, or –water load change. Based on power spectra density (PowSD), multiple channel mechanical signals are interpreted for different MLPs, i.e., mineral-to-ball volume ratio (MBVR), pulp density (PD), and charge volume ratio (CVR), in detail. Experimental results show that the PowSDs of these mechanical signals are positively correlated with CVR and negatively correlated with MBVR and PD. Further, the generation mechanism of these mechanical signals is qualitatively analyzed, and a new measurement method for the contribution rate of multiple channel mechanical signals, i.e., combination estimation index, is proposed. The results show the different contribution rates of these signals to various MLPs under varied grinding conditions. Appropriate mechanical channels for different MLPs must be selected to construct an effective MLP forecasting model. [ABSTRACT FROM AUTHOR]

In this research, a partial least squares (PLS)-based RF with hybrid feature subspace selection is proposed for regression problems. For the problem that average voting strategy of basic RF may decrease method accuracy, PLS is adopted to automatically assign a voting weight to each tree and aggregate the outputs of all trees. To improve feature subspace selection, stratified sampling and embedded feature selection are integrated. First, the variable importance (VI) of each input feature is obtained through embedded feature selection and the features are categorized into two disjointed sets according to VI. During the construction of the trees, stratified sampling is used for feature subspace selection. The effectiveness of PLS aggregation and hybrid feature selection is respectively validated on six regression datasets. The superiority of the proposed RF is demonstrated on historical operation datasets of two power plants through a comparison with five other models. • Partial least square is used to weight each tree of random forest (RF). • An embedded feature selection with stratified sampling is also applied to improve RF. • The variable importance evaluated by RF is used to stratify the features. • The improved RF is applied on UCI and operation datasets of coal-fired boilers. • The results indicate the effectiveness of proposed model for regression problems. [ABSTRACT FROM AUTHOR]

Multimodal data have recently become nearly ubiquitous in the real world. Exploring the multimodal fusion is beneficial to improve the performance of the system. However, it is difficult to ensure that data collected from different sources are full pairing. In this paper, we will focus on the weakly paired case of multimodal data, i.e., each modality is partitioned into multiple groups, only paired information on groups is known instead of full pairing between data samples. A new framework of weakly paired multimodal fusion based on multilayer extreme learning machine (ML-ELM) is proposed in this paper, which will find complex nonlinear transformations of each modality of data such that the resulting representations are highly correlated. In this framework, unsupervised hierarchical ELM performs feature extraction for all modalities separately. Then, the higher-level representations from all modalities perform joint dimension reduction by weakly paired maximum covariance analysis. We evaluate our framework on three challenging cross-modal datasets, and the results have proved the effectiveness of proposed method. [ABSTRACT FROM AUTHOR]

The production of sapphire single crystals with high quality and low cost is an important goal of the sapphire industry. The quality of seeding has a direct impact on the final growth quality of sapphire crystals when sapphire is prepared using the Kyropoulos method. Keeping the crystal/melt contact stable is essential for lowering the number of seeding defects. In order to reduce the preparation cost of sapphire, an optimization strategy integrating computational fluid dynamics, artificial neural networks, and genetic algorithm is proposed to optimize the growth parameters of crystal seeding. The optimization goals are to stabilize the crystal/melt growth interface and increase seeding effectiveness. Three growth parameters, namely, melt temperature, seed crystal rotation speed and pulling speed, are selected as optimization variables. The samples produced by numerical calculation are combined with the actual production data to train and test the neural networks. The genetic algorithm is coupled with artificial neural networks to solve the multi-objective conflict between improving seeding efficiency and reducing the formation of crystal defects in the seeding process. The influence of the mechanism of different combinations of seeding parameters on seeding quality is discussed, and the optimization of seeding growth process is realized. The results show that the optimized combination of parameters gained in this study can improve the seeding speed and ensure the stability of the solid–liquid interface. This work is helpful to reduce the formation of seeding defects, improve the seeding quality, and shorten the seeding time. [ABSTRACT FROM AUTHOR]

Point cloud registration precision degrades when the scene contains symmetric structures or repeated patches. Also, the common operation of continuous downsampling in existing registration methods causes the loss of detailed information and leads to further registration precision degradation. In this paper, we propose a high precision point cloud registration network (HPNet) to learn discriminative features by considering both multi-scale information and position information. We first propose a multi-scale feature extraction module similar to the feature pyramid that allows the extracted features to contain multi-scale information. Then, an information interaction is performed on the features to learn global contextual information by using Transformer with a hybrid position encoding, which takes into account both absolute and relative positions of points. Finally, the features obtained from the information interaction module are directly used to predict the point correspondences. Comprehensive experiments on 3DMatch, 3DLoMatch, ModelNet, and ModelLoNet datasets demonstrate the state-of-the-art performance of the proposed method and the implementation of HPNet in SLAM shows its effectiveness in real application. [Display omitted] • Point cloud registration loses detail information during continuous downsampling. • Feature fusion of low-level detail information and high-level semantic information. • Hybrid position encoding consider both absolute and relative positions. positions. • Point cloud registration algorithms applied to SLAM. [ABSTRACT FROM AUTHOR]

For high-precision positioning and navigation systems, the positioning precision of receiver is jeopardized by multipath interference. Multipath suppression methods based on data processing have drawn much attention recently. The critical step of data processing-based method is to estimate multipath parameters. However, most multipath suppression methods falling into the category of data processing-based methods are limited to Gaussian noises, which means the performance of these methods may be degraded in non-Gaussian noises which are encountered quite often in reality. Besides, only static multipath case is studied in most existing literature, which is not sufficient for potential applications since the occurrence and the disappearance of multipath are always changeable along the movement of receiver. To address these problems, the maximum correntropy criterion(MCC) and the generalized maximum correntropy criterion(GMCC) are integrated into the traditional adaptive multipath estimation(AME) algorithm, named as MCC-AME and GMCC-AME, to handle the dynamic multipath estimation problem in non-Gaussian noises. Furthermore, MCC-AME and GMCC-AME are further improved by adopting forgetting factor, named as RMCC-AME and RGMCC-AME, to improve estimation accuracy and reduce time consumption in a recursive way. The four proposed algorithms also address the problems that the formerly proposed entropy-based multipath estimation algorithms are sensitive to the initial estimation and that the assumption of fixed number of multipath is required. The performance of the four proposed algorithms are analyzed and compared. The analytical results show that GMCC-AME outperforms MCC-AME regarding convergent speed, estimation accuracy and robustness, and RGMCC-AME performs even better than RMCC-AME in the same regard. [ABSTRACT FROM AUTHOR]

Highlights • The OCS method abandons the template-based method of locking region of interest. • The OCS method establishes the convergence model of spoke pattern center. • The OCS method improves the effect of seeding operation. Abstract The seeding process is vital in the preparation of large-diameter sapphire single crystal. It is the key to detect the seeding point during the seeding process. The OCS method is proposed in the paper to detect the seeding point. The OCS method improves the detection method of spoke pattern center proposed by Churl Min Kim, based on this, and the convergence model of spoke pattern center is fitted, so that the real seeding point is detected. In the experiment, the OCS method is verified by comparing with the traditional manual seeding method (operated by skilled seeding technologists) and the method proposed by Churl Min Kim. The OCS method has the same effect as the traditional artificial seeding method and can reduce the number of attempts in a single seeding experiment. Compared with the method proposed by Churl Min Kim, the OCS method can meet the needs of actual industrial production in terms of the number of successful seeding, the number of seeding attempts and the average seeding time. [ABSTRACT FROM AUTHOR]

Impurity in transparent-bottled liquid is a serious production accident in the field of beverage and medicine industry. However, the existing detection systems are difficult to distinguish impurities with dynamic interference (bubbles and stains) and detect impurities located at the edge of the bottle. In order to solve the problems stated above, a new machine vision system for detecting tiny and dynamic impurities is proposed in this paper. In the system, circularity calculation, longitudinal frame-difference method, orthogonal-axis inspection and K-Nearest Neighbor (KNN) machine learning algorithm are combined together to realize the automatic and real-time detection. Experimental results demonstrate that, after completing machine learning, the weighted error of the proposed system for detecting impurities can be effectively controlled at about 0.9% even in dynamic interference environment, which is great significance to safety production in beverage and medicine industry. [ABSTRACT FROM AUTHOR]

In mining operation, the longitudinal tear detection of conveyor belt is of great significance to safety production. A novel method, named Dual Band Infrared Detection (DBID), is proposed to detect the longitudinal tear of conveyor belts in this paper. The DBID method is based on the collaboration of mid-infrared and long infrared vision. The DBID sensor device captures the lower surface images of conveyor belt. The feature of tear or scratch and tearing precursor can be extracted by analyzation of captured images. The DBID method is verified by experiments and practice, including the capturing of dual band infrared images (mid-infrared and long infrared images), image preprocessing, and tear detection. The DBID method is not only unlimited by visual condition, but also much more accurate and reliable compared with the Integrative Binocular Vision Detection (IBVD) method presented before which meets the requirement of real-time online detection. [ABSTRACT FROM AUTHOR]

Real-time and reliable detection of conveyor belt longitudinal tear is an important task in mining operations. This paper presents a novel method of Integrative Binocular Vision Detection (IBVD) to detect the longitudinal tears of conveyor belts. Based on infrared and visible fusion, the IBVD sensor device collects the fusion images of the belt. After extracting the tear features by projection method, the progress of potential tears can be evaluated and the tears can be identified. The IBVD method is verified by an experiment platform fulfilling the acquisition, pro-processing and analysis of fusion images for tear detection. The fusion image processing time is less than 18 ms, which satisfies the requirement of real-time online monitoring. Compared to the individual measurement technique of either infrared detection or visible light detection, the average accuracy of the IBVD method reaches up to 96%, the IBVD method is more reliable in tear detection. [ABSTRACT FROM AUTHOR]

The development of the building energy management systems (BEMS) enable users to intelligently control Heating, Ventilation, Air-conditioning and Cooling (HVAC) systems based on digital information. In order to reduce the power consumption cost of the HVAC system while ensuring user satisfaction, a novel HVAC control system for building system based on a multi-step predictive deep reinforcement learning (MSP-DRL) algorithm is proposed in this paper. In the proposed method, the outdoor ambient temperature is predicted first by a featured deep learning method named GC-LSTM, where the Long Short-term Memory (LSTM) is enhanced by the generalized correntropy (GC) loss function to deal with the non-Gaussian characteristics of the collected outdoor temperature. In addition, the proposed temperature prediction model is combined with a reinforcement learning algorithm named Deep Deterministic Policy Gradient (DDPG) aiming to flexibly adjust the output power of the HVAC system under the dynamic changing of electricity prices. Finally, comprehensive simulation based on real world data is delivered. Numerical results show that the GC-LSTM algorithm is more accurate than other counterparts prediction algorithms, and the proposed HVAC control system based on the multi-step prediction deep reinforcement learning algorithm is effective and could save over 12% cost compared to other approaches, where the user comfort is maintained simultaneously. • A multi-step predictive deep reinforcement learning algorithm is proposed. • A GC-LSTM algorithm is adopted in multi-step outdoor temperature prediction. • A model-free DDPG algorithm is adopted to manage HVAC system for building system. • The proposed method performs well in the HVAC system. [ABSTRACT FROM AUTHOR]

Conveyor belt longitudinal tear is one of the most serious problems in coal mining. It is very important to detect it in real‐time before the length of the tear is too long. In this study, a method of longitudinal tear detection based on visible light charge coupled device (CCD) and infrared CCD is proposed. A visible light and infrared vision real‐time detection system for conveyor belt longitudinal tear is designed based on this method. In this method, the infrared CCD uses adaptive histogram equalisation coordinating with Hough transform line detection. The visible light CCD, coordinating with a laser line source, uses the Hough transform algorithm and the features from accelerated segment test algorithm. These two kinds of CCDs work together to make detection results reliable. Experimental results show that the proposed method is effective and adaptive, and meets the requirements for reliable, real‐time and online longitudinal tear detection. Compared to several current methods, the proposed method has a better performance on efficiency of detection. [ABSTRACT FROM AUTHOR]

[Display omitted] • A VPVF detection method is proposed to detect seeding timing of sapphire. • Detection relies on circumferential motion of spokes and stable radial flow. • Perturbation velocity can clearly reflect flow state in sapphire processing. • VPVF method can improve the efficiency of seeding operations for sapphire. In sapphire processing by the Kyropoulos method, the detection of seeding timing is the core technology. Existing detection methods are incapable of tracking spoke feature motion on the liquid surface and cannot effectively guide sapphire seeding. Therefore, this paper proposes an innovative visual-perturbation-velocity-fitting method for sapphire seeding timing detection. First, the concept of perturbation velocity, which can reflect the flow state of the melt surface, is presented. Then, the position of spoke feature points is obtained by computer vision and calculated by multi-frame interactive periodic detection. The fitting function model corresponding to the perturbation velocity in different flow states is established. Finally, the seeding timing is determined by the transition of the flow state and the stability of the flow. The experimental results show that the proposed method can improve the seeding success rate from 76.7% to 93.3% and the average successful seeding time from 5.6 h to 4.5 h. [ABSTRACT FROM AUTHOR]

• Monitor longitudinal tear with infrared thermography. • Overcome the interference caused by the change of infrared image threshold. • Select the appropriate detection spectral passband. • Coefficient T can be used to quantitatively describe tear change characteristics. A novel approach based on infrared spectrum analysis for early-warning of longitudinal tearing of the conveyor belt was proposed in the paper. Unlike most existing methods, the proposed method monitors the change process of the infrared radiation field of the longitudinal tearing through the infrared thermal imaging technology, and judges whether there is the risk of the longitudinal tearing of the conveyor belt through the frequency-domain characteristic coefficient T of the infrared radiation field. Experimental results exhibit that the characteristic coefficient T can quantitatively describe the change characteristics of infrared radiation field during the tearing process of conveyor belt. When the conveyor belt is complete penetration, the T value fluctuates violently from 0.6 to 1.6. This characteristic can be used as the precursor information of the tearing process, which broadens the train of thought for identification and early warning of conveyor belt longitudinal tearing. [ABSTRACT FROM AUTHOR]