Condition Monitoring Algorithm Research Articles

Design and implementation of an interactive networked condition monitoring strategy for plant-wide production equipment toward Industry 4.0

Equipment condition monitoring holds immense significance in modern industrial factories, facilitating sustained operation with heightened efficiency, safety, and cost-effectiveness. This paper presents an interactive networked condition monitoring strategy to enhance plant-wide production equipment condition monitoring, leveraging networked control and edge computing approaches. This strategy offers a rapid and efficient solution that comprises four key components: drag-and-drop monitoring algorithm design, executable program generation, networked algorithm implementation, and interactive monitoring process. Initially, algorithms for equipment condition monitoring are designed using pre-defined components and then seamlessly compiled into executable programs with a single click. Subsequently, these algorithms are implemented within an edge computing-based monitoring system using the networked control approach, where the generated programs are remotely deployed into edge nodes positioned adjacent to the equipment. Moreover, the monitoring process features visual displays of parameters and results using configurable widgets, enabling interactive adjustments. Real-world application cases from an aluminum manufacturing factory showcase the strategy’s effectiveness in comprehensive condition monitoring of hundreds of equipment across nine workshops, significantly reducing the equipment inspection time to tens of minutes compared to manual inspection. With extensive implementation in various industrial factories, this strategy emerges as a promising solution for equipment condition monitoring in smart factories.

Time–Frequency Approach for Cutting Tool Power Signal Separation in Face Milling Operations

Face milling is among the processes that can produce a high-precision surface finish. Tool condition monitoring and signal processing algorithms are under extensive research to improve production quality and productivity in machining processes. In the present research, the time–frequency analysis technique was applied to the signal obtained from a sensor integrated into the primary AC power circuitry during the milling of steel bars to evaluate its applicability in detecting the current variations associated with the cutting force. The signal acquired from the sensor was processed in the time–frequency domain using wavelet analysis, and the results were compared with the traditional time and frequency analyses. The results showed that the signal variations produced by the cutting force were well localized in the frequency spectra with both approaches. However, the wavelet processing method yielded a poorly defined cutting force signal shape due to the limited resolution inherent in the sub-bands containing the frequencies of interest.

Auxiliary input-enhanced siamese neural network: A robust tool wear prediction framework with improved feature extraction and generalization ability

Tool wear monitoring is essential for automated and resilient manufacturing, as it can prevent catastrophic failures caused by severe wear on cutting edges during machining. The conventional tool wear monitoring approaches depend on features extracted from signals, which require sensitive signals and consistent tool wear. In practical scenarios, however, neither the sensitivity of collected signals to the tool wear status nor the consistency of the actual tool wear evolution is hard to meet the requirement of the tool condition monitoring algorithm, which greatly limit the wide spread of its industrial applications. To overcome this challenge, we propose an Auxiliary Input-enhanced Siamese Neural Network (AISNN) framework by incorporating a Siamese structure into the feature extraction part of a convolutional neural network (CNN), and introducing an auxiliary input to its nonlinear regression part. The Siamese structure, instead of extracting features directly from signals, distinguishes the difference between the features extracted from signals of the examined cut and the first cut, and uses this difference as the indicator of tool wear status. Moreover, the auxiliary input provides an additional feature that has heavy dependence on the tool wear, which enables the model learning the general wear evolution of the examined cutting tool. The effectiveness of the proposed AISNN framework is verified in a set of milling experiments where input signal is insensitive to the flank wear of cutting tool and different tools’ wear evolution exhibits obvious inconsistency. Compared to the traditional CNN, the proposed AISNN significantly improves the accuracy on the verification set from 63% to 95% and on the testing set from 50% to 81%. The results demonstrate that the AISNN framework achieves significant improvement in feature extraction and generalization ability. The proposed AISNN, as a universal framework, can empower most existing deep learning-based tool wear prediction methods, enhancing their robustness in handling insensitive signals and inconsistent wear evolution and thereby promoting more industrial applications.

Experimental verification of a data-driven algorithm for drive-by bridge condition monitoring

As the world’s transport infrastructure ages, the importance of bridge condition monitoring is becoming increasingly acknowledged. Large-scale deployment of existing inspection and monitoring techniques is infeasible due to cost and logistical challenges. The concept of using sensors located within vehicles for low cost ‘drive-by’ monitoring has become the focus of much attention in recent years. This paper presents a new data-driven approach for drive-by bridge monitoring. Machine learning techniques are leveraged to allow the influence of vehicle speed to be considered and the Operating Deflection Shape Ratio (ODSR) is presented as an alternative damage-sensitive feature to the commonly used frequency spectrum. Extensive laboratory experiments demonstrate that the method is capable of detecting midspan cracking and seized bearings. A statistical classification approach is adopted to classify damage indicators as either ‘damaged’ or ‘healthy’. Classification accuracy is seen to vary between 65-96% and is similar whether using the frequency spectrum or ODSR. Based on the results of the laboratory testing, it is expected that this approach could be implemented on a large scale to act as an early warning tool for infrastructure owners to identify bridges presenting signs of distress or deterioration.

On the performance comparison of diagnostic techniques in machine monitoring

Predictive maintenance can save a lot of efforts in modern industry and condition monitoring is attracting a lot of attention accordingly. New algorithms for fault detection appear frequently in the technical literature, however an objective, quantitative and widely accepted approach to performance comparison is still lacking. In this paper, we propose a new method leading to a fair and reproducible performance assessment. The proposed solution is based on vibrational analysis and consists of searching and detecting the theoretical cyclic frequencies that appear as a specific “signature” of a fault. Each algorithm for condition monitoring relies on a metric, then the main idea is to quantitatively characterize the peaks of the metric emerging from the machine noise. We think that the wide adoption of the proposed approach could significantly foster the research in the fields of condition monitoring and predictive maintenance.

Application of Machine Learning Algorithms for Tool Condition Monitoring in Milling Chipboard Process.

In this article, we present a novel approach to tool condition monitoring in the chipboard milling process using machine learning algorithms. The presented study aims to address the challenges of detecting tool wear and predicting tool failure in real time, which can significantly improve the efficiency and productivity of the manufacturing process. A combination of feature engineering and machine learning techniques was applied in order to analyze 11 signals generated during the milling process. The presented approach achieved high accuracy in detecting tool wear and predicting tool failure, outperforming traditional methods. The final findings demonstrate the potential of machine learning algorithms in improving tool condition monitoring in the manufacturing industry. This study contributes to the growing body of research on the application of artificial intelligence in industrial processes. In conclusion, the presented research highlights the importance of adopting innovative approaches to address the challenges of tool condition monitoring in the manufacturing industry. The final results provide valuable insights for practitioners and researchers in the field of industrial automation and machine learning.

Gaussian assumptions-free interpretable linear discriminant analysis for locating informative frequency bands for machine condition monitoring

Interpretable learning models become an emerging topic in the domain of machine condition monitoring to connect signal processing algorithms with statistical learning and machine learning. Compared with traditional signal processing algorithms that show strong ability in data analysis, signal-processing-related interpretable learning models could generate interpretable learnable weights/parameters as advanced physically interpretable fault features for both machine condition monitoring and fault diagnosis. It is well-known that linear discriminant analysis (LDA) is one of the most popular and interpretable algorithms for machine condition monitoring. However, this popular algorithm needs Gaussian assumptions in their derivations and parameter estimations. In this paper, Gaussian assumptions-free interpretable LDA is proposed as an interpretable learning model to physically locate informative frequency bands and fault characteristic frequencies for machine condition monitoring. Firstly, statistical decision theory is introduced to connect the nature of regression with that of classification, which poses a foundation for the Gaussian assumptions-free interpretable LDA for machine condition monitoring. Secondly, two propositions are given to mathematically show that LDA can be realized by an equivalent linear regression analysis, which provides a perspective for the Gaussian assumptions-free interpretable LDA for simultaneous machine condition monitoring and fault diagnosis. Finally, linear regression analysis with a sparse Lp-norm regularization term is introduced to realize the Gaussian assumptions-free interpretable LDA for physically locating informative frequency bands and fault characteristic frequencies for machine condition monitoring. Two case studies are provided as illustrative examples to experimentally demonstrate that the Gaussian assumptions-free interpretable LDA is capable of indicating informative frequency bands and fault characteristic frequencies.

Optimal symbolic entropy: An adaptive feature extraction algorithm for condition monitoring of bearings

Due to their effectiveness in vibration-based fault feature extraction from bearings, entropy-based methods have become a hot research topic. Symbolic dynamic filtering reduces background noise in bearing signals, making it ideal for entropy analysis. However, the partitioning approach selection of symbolic dynamic filtering mainly depends on experience, which may bring over-track and under-track phenomena. The optimal symbolic entropy (OSE) method proposes a solution by using mean spectral kurtosis to evaluate the symbolization performance of partitioning approaches. This method improves the identification of bearing faults through steps such as evaluating frequency and amplitude preservation, selecting the optimal symbolization approach, and using the OSE method with multiscale analysis. Simulative and experimental data analysis demonstrates its superior ability to extract bearing fault characteristics, with better performance and robustness than existing methods.

Wind turbine anomaly detection based on SCADA: A deep autoencoder enhanced by fault instances

An increasing number of deep autoencoder-based algorithms for intelligent condition monitoring and anomaly detection have been reported in recent years to improve wind turbine reliability. However, most existing studies have only focused on the precise modeling of normal data in an unsupervised manner; few studies have utilized the information of fault instances in the learning process, which results in suboptimal detection performance and low robustness. To this end, we first developed a deep autoencoder enhanced by fault instances, that is, a triplet-convolutional deep autoencoder (triplet-Conv DAE), jointly integrating a convolutional autoencoder and deep metric learning. Aided by fault instances, triplet-Conv DAE can not only capture normal operation data patterns but also acquire discriminative deep embedding features. Moreover, to overcome the difficulty of scarce fault instances, we adopted an improved generative adversarial network-based data augmentation method to generate high-quality synthetic fault instances. Finally, we validated the performance of the proposed anomaly detection method using a multitude of performance measures. The experimental results show that our method is superior to three other state-of-the-art methods. In addition, the proposed augmentation method can efficiently improve the performance of the triplet-Conv DAE when fault instances are insufficient.

A Method for Statistical Processing of Magnetic Field Sensor Signals for Non-Invasive Condition Monitoring of Synchronous Generators.

Condition monitoring of synchronous generators through non-invasive methods is widely requested by maintenance teams for not interfering the machine operation. Among the techniques used, external magnetic field monitoring is a recent strategy with great potential for detecting incipient faults. In this context, this paper proposes the application of a simple strategy with low computational cost to process data of external magnetic field time derivative signals for the purposes of condition monitoring and fault detection in synchronous machines. The information of interest is extracted from changes in the magnetic signature of the synchronous generator, obtained from frequency spectra of monitored signals using induction magnetic field sensors. The process forms a set of time series that reflects constructive and operational characteristics of the machine. The Shewhart control chart method is applied for anomaly detection in these time series, allowing the detection of changes in the machine magnetic signature. This method is employed in an algorithm for continuous condition monitoring of synchronous generators, presenting as output a global change indicator for the multivariable problem associated with magnetic signature monitoring. Correlation matrices are used to improve the algorithm response, filtering series with similar variation patterns associated with detected events. The proposed method is validated through tests on an experimental bench that allows the controlled imposition of faults in a synchronous generator. The proposed global change indicator allows the automatic detection of stator and rotor faults with the machine synchronized with the commercial power grid. The proposed methodology is also applied on data obtained from an equipment installed in a 305 MVA synchronous generator of a hydroelectric power plant where the evolution of an incipient fault, i.e., a mechanical vibration fault, has been detected.

Approach to Condition Monitoring of BLDC Motors with Experimentally Validated Simulation Data

Due to their compact design and low number of wear parts, Brushless Direct Current (BLDC) motors are ideally suited for use in unmanned aerial vehicles (UAVs). In view of the growing areas of application and the increasing complexity of unmanned flight missions, the need for suitable safety mechanisms for the operation of technical components, such as BLDC motors, in unmanned aircraft drive trains is also increasing. The integration of redundant components analogous to manned aviation is often not possible for smaller unmanned aerial vehicles for weight reasons. Therefore, online-capable dynamic diagnosis and prognosis methods for monitoring safety-critical components of unmanned aircraft are subject of ongoing research. One major challenge in the development of data based condition monitoring approaches for safety critical components is the availability of operational data of degraded components. This often leads to an unbalanced database without sufficient information on components’ degradation behavior. In the presented work, this problem is approached by combining bench testing and simulation models. On a test rig, common degradation effects are recreated by targeted manipulation. This allows for a safe and expressive data acquisition of the components’ behavior. In order to reduce the material and time required to build up a sufficient database for condition monitoring with experimental data, the observable effects are replicated in a simulation. This provides the opportunity to create a large database with slight variations in simulation parameters and incorporated noise in the simulation. The BLDC motor manipulation on the test rig includes mechanical, electrical and magnetic manipulation. The effects of the manipulation are analyzed and their representation by parameters in the corresponding simulation is derived. The model is built in MATLAB Simulink and replicates both the electrical and physical behavior of the motor, as well as its commutation behavior. The established simulation data shall be used as a balanced dataset on which condition monitoring algorithms can be trained. This will allow for the comparison of various data based condition monitoring methods in the future. A remaining challenge lies in the time behavior of the analyzed degradation, which has not yet been explored in depth. The proposed approach might also be applied to further unmanned aerial vehicle components, such as servo motors.

Design of modified model of intelligent assembly digital twins based on optical fiber sensor network

Intelligent assembly of large-scale, complex structures using an intelligent manufacturing platform represents the future development direction for industrial manufacturing. During large-scale structural assembly processes, several bottleneck problems occur in the existing auxiliary assembly technology. First, the traditional LiDAR-based assembly technology is often limited by the openness of the manufacturing environment, in which there are blind spots, and continuous online assembly adjustment thus cannot be realized. Second, for assembly of large structures, a single-station LiDAR system cannot achieve complete coverage, which means that a multi-station combination method must be used to acquire the complete three-dimensional data; many more data errors are caused by the transfer between stations than by the measurement accuracy of a single station, which means that the overall system's measurement and adjustment errors are increased greatly. Third, because of the large numbers of structural components contained in a large assembly, the accumulated errors may lead to assembly interference, but the LiDAR-assisted assembly process does not have a feedback perception capability, and thus assembly component loss can easily be caused when assembly interference occurs. Therefore, this paper proposes to combine an optical fiber sensor network with digital twin technology, which will allow the test data from the assembly entity state in the real world to be applied to the “twin” model in the virtual world and thus solve the problems with test openness and data transfer. The problem of station and perception feedback is also addressed and represents the main innovation of this work. The system uses an optical fiber sensor network as a flexible sensing medium to monitor the strain field distribution within a complex area in real time, and then completes real-time parameter adjustment of the virtual assembly based on the distributed data. Complex areas include areas that are laser-unreachable, areas with complex contact surfaces, and areas with large-scale bending deformations. An assembly condition monitoring system is designed based on the optical fiber sensor network, and an assembly condition monitoring algorithm based on multiple physical quantities is proposed. The feasibility of use of the optical fiber sensor network as the real-state parameter acquisition module for the digital twin intelligent assembly system is discussed. The offset of any position in the test area is calculated using the convolutional neural network of a residual module to provide the compensation parameters required for the virtual model of the assembly structure. In the model optimization parameter module, a correction data table is obtained through iterative learning of the algorithm to realize state prediction from the test data. The experiment simulates a large-scale structure assembly process, and performs virtual and real mapping for a variety of situations with different assembly errors to enable correction of the digital twin data stream for the assembly process through the optical fiber sensor network. In the plane strain field calibration experiment, the maximum error among the test values for this system is 0.032 ​mm, and the average error is 0.014 ​mm. The results show that use of visual calibration can correct the test error to within a very small range. This result is equally applicable to gradient curvature surfaces and freeform surfaces. Statistics show that the average measurement accuracy error for regular surfaces is better than 11.2%, and the average measurement accuracy error for irregular surfaces is better than 14.8%. During simulation of large-scale structure assembly experiments, the average position deviation accuracy is 0.043 ​mm, which is in line with the designed accuracy.

Fuzzy-Logic-Based Recommendation System for Processing in Condition Monitoring.

The development of a machine’s condition monitoring system is often a challenging task. This process requires the collection of a sufficiently large dataset on signals from machine operation, context information related to the operation conditions, and the diagnosis experience. The two referred problems are today relatively easy to solve. The hardest to describe is the diagnosis experience because it is based on imprecise and non-numerical information. However, it is essential to process acquired data to develop a robust monitoring system. This article presents a framework for a system dedicated to recommending processing algorithms for condition monitoring. It includes a database and fuzzy-logic-based modules composed within the system. Based on the contextual knowledge provided by the user, the procedure suggests processing algorithms. This paper presents the evaluation of the proposed agent on two different parallel gearboxes. The results of the system are processing algorithms with assigned model types. The obtained results show that the algorithms recommended by the system achieve a higher accuracy than those selected arbitrarily. The results obtained allow for an average of 5 to 14.5% higher accuracy.

An Intelligent Machine Condition Monitoring Model for Servo Systems

The installation of industrial servo systems and the determination of control parameters are limited to the skills and knowledge of the commissioner. In addition, commissioned systems are often not re-optimized if environmental influences or loads change. The goal of this research is to create an artificial neural network (ANN) model for servo systems that will keep the servo system's proportional, integral, and derivative (PID) parameters working optimally. For this process, a machine condition monitoring algorithm developed with the ANN technique, which uses the data such as actual current, torque, power, position to be obtained from the servo system on an industrial controller, for the control and rearrangement of the parameters.

Online condition monitoring algorithm for element failure detection and fault location in double wye shunt capacitor banks

Failure in quick detection and location of internal failures in Shunt Capacitor Banks (SCBs) may lead to the negligence of necessary repair processes, consequently laying undesirable effects on the power system operation. This paper puts forward a failure detection and fault location method in SCBs with the double wye connection. The proposed method provides a criterion for relay decision-making based on the neutral voltage and current fundamental phasor component, capable of distinguishing the faulty capacitors and determining the number of capacitor failures in the case of multiple faulty phases. The proposed algorithm can be applied for double wye SCBs considering different protection designs, namely fuseless and internally fused units. The proposed method is designed to monitor and detect consecutive failures based on the existing data from commercial relays. Implementation and performance evaluations are conducted in PSCAD and MATLAB software and also with the help of practical data. Performance evaluations are conducted under different circumstances, namely harmonic distortion and voltage unbalance conditions, multiple internal fault locations, and external faults from the grid. Simulation results show the significance and salient features of the proposed method under different conditions.

Novel Data Compression Algorithm for Transmission Line Condition Monitoring

For the problem of data accumulation caused by massive sensor data in transmission line condition monitoring system, this paper analyzes the type and amount of data in the transmission line sensor network, compares the compression algorithms of wireless sensor network data at home and abroad, and proposes an efficient lossless compression algorithm suitable for sensor data in transmission line linear heterogeneous networks. The algorithm combines the wavelet compression algorithm and the neighborhood index sequence algorithm. It displays a fast operation speed and requires a small amount of calculation. It is suitable for battery powered wireless sensor network nodes. By combining wavelet correlation analysis and neighborhood index sequence coding, the compression algorithm proposed in this paper can achieve a high compression rate, has strong robustness to packet loss, has high compression performance, and can help to reduce network load and the packet loss rate. Simulation results show that the proposed method achieves a high compression rate in the compression of the transmission line parameter dataset, is superior to the existing data compression algorithms, and is suitable for the compression and transmission of transmission line condition monitoring data.

Feedback-driven error-corrected single-sensor analytics for real-time condition monitoring

In this paper, a single-sensor based output-only algorithm is proposed for real-time condition monitoring of mechanical vibrating systems. Four key aspects of real-time condition monitoring and maintenance are presented through Recursive singular spectrum analysis (RSSA): (a) Filtering, (b) Enhancification, (c) Fault detection, and (d) Modal identification. Recent prominence of eigen perturbation (EP) solutions for condition monitoring has led to the development of RSSA as the go-to real-time algorithm for single-sensor diagnosis. As single-sensor econometrics has been long sought as a viable option for cases involving instrumentation redundancies, non-optimal sensor placement, and cost considerations, RSSA provides replication, scalability, and transferability for real-time fault detection studies. With the output vibration signals streaming in real-time, the Hankel covariance matrix is formed which filters out the noise subspace in the grouping stage. Online enhancification becomes particularly useful when the signal statistics are masked by time-varying non-stationary excitation. Application examples involving AM–FM signals and operational noise in structural systems demonstrates the versatility of RSSA towards spatio-temporal fault detection in real-time. The efficacy of the proposed algorithm is further validated by experimental investigations of real-time complete modal identification from partial sensor information. With applications extending to real-time passive control and aligned to current infrastructure monitoring demands worldwide, RSSA demonstrates potential to establish as a benchmark algorithm for online condition monitoring.

A methodology for alignment of measured rail profiles in turnouts as a basis for reliable vehicle/track interaction simulations

For the development of condition monitoring algorithms for turnouts, an extensive dataset consisting of measurement and simulation results is required. The basis for reliable simulation results are precise and easy applicable measurement techniques of the turnout geometry, like 3D handheld scanners. The challenge in using the data from portable 3D scanners for vehicle/track interaction simulation lies in the correct alignment and preparation of the measured surface. In this paper, a methodology for pre-processing of 3D scanner measurement results is presented, based on turnout measurements with the device HandySCAN 700 from Creaform. Point cloud processing methods like Principal Component Analysis and Iterative Closest Point algorithm are used to align the measured surface to a reference frame. Interpolation of the measured points by CAD reference data is applied for repairing the profile sections due to missing surface points. Furthermore, comparisons were made regarding the quality of simulation results for different input sources. It can be shown that the results have a good quality agreement in respect of different profile measurements techniques. Thus, it could be shown, that the data from the portable 3D scanner can be used for multibody dynamic simulation after appropriate preparation with the presented method.

Model-Based Condition-Monitoring and Jamming-Tolerant Control of an Electro-Mechanical Flight Actuator with Differential Ball Screws

The work deals with the development of deterministic model-based condition-monitoring algorithms for an electromechanical flight control actuator with fault-tolerant architecture, in which two permanent magnets synchronous motors are coupled with differential ball screws in speed-summing paradigm, so that the system can operate even after a motor fault, an inverter fault or a mechanical jamming. To demonstrate the potential applicability of the system for safety-critical aerospace applications, the failure transients related to major fault modes have to be characterised and analysed. By focusing the attention to jamming faults, a detailed nonlinear model of the actuator is developed from physical first principles and experimentally validated in both time and frequency domains for normal condition and with different types of jamming. The validated model is then used to design the condition-monitoring algorithms and to characterize the system failure transient, by simulating mechanical blocks in different locations of the transmission. The operability after the fault, obtained via fault-tolerant control strategy and position regulator reconfiguration, is also verified, by highlighting and discussing possible enhancements and criticalities.

An Improved Decision Tree Algorithm for Condition Monitoring on Storage Power Station of Internet Things

Power station is an important basic power generation organization, and its operation status is related to the continuous power generation capacity. At present, a large number of physical network equipment and intelligent equipment are used in pumped storage power station, which makes its data mass growth and its operation state become a difficult problem. Accurate operation monitoring results can provide decision support that power generation planners and government, but also reasonably dispatch corresponding resources. In the past, decision tree algorithm was used in operation condition monitoring, which has the problem of data distortion and affects the accuracy of monitoring results. Based on the above reasons, this paper combines the wavelet function and decision tree algorithm, proposes an improved decision tree algorithm to eliminate redundant data in order, and uses wavelet function to cluster distorted data, so as to improve the accuracy and computational efficiency of the algorithm. Matlab simulation results show that: decision tree algorithm can eliminate 90% of redundant data, reduce the impact of feature data extraction on decision tree. At the same time, the improved accuracy is 98%, the calculation time is less than 25s is better than that, the decision tree algorithm. Therefore, the improved algorithm can optimize the condition monitoring of pumped storage power station.