On-line Identification Research Articles

A Predictive Out-of-Step Identification Method Based on Wide-Area Velocity-Acceleration Trajectory

Online identification of the transient instability is essential in power system analysis, as it facilitates the grid operator to decide and coordinate system failure correction control actions. The rapid technological development of phasor measurement units (PMUs) provides an opportunity for online transient out-of-step identification in real-time. This paper analyzes the characteristics of angular velocity-acceleration trajectories in detail, and then proposes a predictive transient out-of-step identification method among coherent groups of generators based on the velocity-acceleration trajectory sampled by PMUs. To prove that the method can distinguish the transient out-of-step event following disturbance reliably, a mathematical proof of the criterion extended to multi-machine system is made, and the influences of the exciter and emergency control are taken into consideration to a certain extent. Finally, the effectiveness and predictability of the proposed method are demonstrated on a 39-bus New England test system and two large-scale power system simulation models of China by numerical studies.

Vertical load estimation in tractors via in-wheel optical sensing

This paper presents a novel approach for estimating the vertical load exerted on the rear axle of agricultural vehicles during fieldwork using tire compression measurements. The system comprises an optical time-of-fight (TOF) sensor integrated into the wheel rim, a microcontroller, and an online identification algorithm. The sensor exhibits minimal measurement error, approximately 1%, which, when combined with effective preprocessing and a piecewise linear regression model fueled by tire pressure data, yields a root mean square error (RMSE) of around 7%, with a maximum drawdown of 3% when considering periodic updates. This technology promises to reduce production costs while enabling optimal tire pressure calibration, thereby ensuring reduced fuel consumption and improved comfort for agricultural vehicle operators.

Multi-stage Dynamic Latent Variable Model and Lightweight PCANet Based Anomaly Identification of Fused Magnesium Furnaces

The fused magnesium furnace (FMF) is a crucial component in the production of magnesium. The inability to timely detect abnormal operation can result in a semi-molten state, posing a serious risk to production safety. The limitations of current monitoring technologies hinder effective anomaly detection due to their failure to extract intrinsic data features and fully utilize multi-source information. This paper introduces a novel multi-source information fusion approach for anomaly identification of FMFs, leveraging dynamic feature extraction. A two-layer multi-stage dynamic latent variable model is established for dynamic feature extraction. Meanwhile, a lightweight principal component analysis network (PCANet) is established for video deep-learning online identification. The lightweight model can improve the recognition efficiency and reduce the computational cost. Subsequently, through decision-level fusion of multi-source information, the final diagnostic outcomes facilitate the classification and identification of FMF operating conditions. Experiments on real industrial data substantiate the efficacy and advancement of the proposed method. The final accuracy reached 98.43%.

An information theory approach for recursive LPV-ARX model identification via LS-SVM

When modeling a dynamical system, linear models are always the first choice due to their simplicity. However, many times the system is too complex so that employing linear models results in poor performances. In this context, Linear Parameter Varying (LPV) models allow to represent non-linear input/output relationships while preserving the simple structure of linear models. The method of Least Squares Support Vector Machines (LS-SVM) is one of the most common approaches to identify a LPV model in an ARX formulation (LPV-ARX). However, due to its computational cost, it is difficult to identify a LPV-ARX model using LS-SVM in online applications, where the model must be updated every time new data are collected. An efficient update algorithm has been presented for online identification of such models, where the idea is to update the model only upon certain data that are considered informative. However, this approach requires the tuning of some hyperparameters and in certain conditions can stop updating the model even when data are informative. This paper proposes an information-based algorithm to overcome these drawbacks. Evaluation on simulated and experimental data show the effectiveness of the proposed approach on both identification and computational sides.

PID control of data-driven patient response with fixed ratio co-administration of drugs for depth of hypnosis

In this work we make use of knowledge from clinical practice to deliver a trusted environment for feedback control of co-administration of two drugs to induce depth of hypnosis during general anesthesia. The clinical data we have provides valuable insight into the ratio used to determine the Propofol and Remifentanil infusion rates by means of specifying their corresponding effect site concentrations. A feedback closed loop with a fixed parameter PID controller tuned on population dynamics is used. To account the strong variations in gain of the system (patient) we propose an online identification of the dose-effect response from clinical data available during clinical protocol. Theoretical analysis is provided to justify the approach and simulations with real data from patient are given to support the theoretical insight. Limitations of this approach are discussed as to justify why ratio based co-administration is used in practice solely during the induction phase of general anesthesia.

A fluorescence approach for an online measurement technique of atmospheric microplastics.

Microplastic particles in the atmosphere are regularly detected in urban areas as well as in very remote locations. Yet the sources, chemical transformation, transport, and abundance of airborne microplastics still remain largely unexplained. Therefore, their impact on health, weather and climate related processes lacks comprehensive understanding. Single particle detection presents a substantial challenge due to its time-consuming process and is conducted solely offline. To get more information about the distribution, fluxes and sources of microplastics in the atmosphere, a reliable and fast online measurement technique is of utmost importance. Here we demonstrate the use of the autofluorescence of microplastic particles for their online detection with a high sensitivity towards different widely used polymers. We deploy online, single particle fluorescence spectroscopy with a Wideband Integrated Bioaerosol Sensor WIBS 5/NEO (Droplet Measurement Technologies, USA), which enables single particle fluorescence measurements at two excitation wavelengths (280 nm and 370 nm) and in two emission windows (310-400 nm and 420-650 nm). We investigated shredded (<100 μm) everyday plastic products (drinking bottles and yogurt cups) and pure powders of polyethylene terephthalate (PET), polyethylene and polypropylene. For the broad range of typical plastic products analyzed, we detected fluorescence on a single particle level using the WIBS. The online detection can identify particles smaller than 2 μm. In the case of microplastic particles from a PET bottle, 1.2 μm sized particles can be detected with 95% efficiency. Comparison with biological aerosols reveals that microplastics can be distinguished from two abundant pollen species and investigation of the complete fluorescence excitation emission maps of all samples shows that online identification of microplastics might be possible with fluorescence techniques if multiple channels are available.

An Adaptive PID Control System for the Attitude and Altitude Control of a Quadcopter

Abstract In adaptive model-based control systems, determining the appropriate controller gain is a complex and time-consuming task due to noise and external disturbances. Changes in the controller parameters were assumed to be dependent on the quadcopter mass, which was the process variable. A nonlinear model of the plant was used to identify the mass, employing the weighted recursive least squares (WRLS) method for online identification. The identification and control processes involved filtration using differential filters, which provided appropriate derivatives of signals. Proportional integral derivative (PID) controller tuning was performed using the Gauss–Newton optimisation procedure on the plant. Differential filters played a crucial role in all the developed control systems by significantly reducing measurement noise. The results showed that the performance of classical PID controllers can be improved by using differential filters and gain scheduling. The control and identification algorithms were implemented in an National Instruments (NI) myRIO-1900 controller. The nonlinear model of the plant was built based on Newton’s equations.

Recursive Delta-Operator-Based Subspace Identification with Fixed Data Size

This paper proposes a recursive delta-operator-based subspace identification method with fixed data size. The majority of existing subspace identification methods are constrained to discrete-time systems because of the disparity in Hankel matrices. Additionally, due to the storage cost, LQ-factorization and singular value decomposition in identification methods are best suited for batch processing rather than online identification. The continuous-time systems are transformed into state space models based on the delta-operator to address these issues. These models approach the original systems when the sampling interval approaches zero. The size of the data matrices is fixed to reduce the computing load. By fading the impact of past data on future data, the amount of data storage can be decreased. The effectiveness of the proposed method is illustrated by the continuous stirred tank reactor system.

Smartphone-integrated paper-based sensing platform for the visualization and quantitative detection of pymetrozine

Pymetrozine (PYM) is an effective pyridine insecticide for controlling aphids, while its residues pose a serious threat to human health. Herein, a europium complex (Eu-DBPA, DBPA represents deprotonated 2,5-dibromoterephthalic acid ligand) probe was prepared for the detection of PYM via fluorescence quenching. The detection process has the advantages of short response time (2 min), wide linear range (0–4 and 4–45 mg/kg) and low detection limit (2.2 μg/kg). Furthermore, a portable detection platform was designed by integrating Eu-DBPA-based paper strip with smartphone and applied for the visual detection of PYM in real cucumber, tomato, cabbage and apple samples, obtaining satisfactory recovery (99.00 %–107.00 %) and low standard deviation (RSD < 3.4 %). In addition, a logic gate device was designed to simplify the detection process. The smartphone-integrated paper-based probe detection platform provides a new strategy for intelligent and online identification of hazards in environmental and biological samples.

Parameter Identification of Permanent Magnet Synchronous Motor with Dynamic Forgetting Factor Based on H∞ Filtering Algorithm

To address system parameter changes during permanent magnet synchronous motor (PMSM) operation, an H∞ filtering algorithm with a dynamic forgetting factor is proposed for online identification of motor resistance and inductance. First, a standard linear discrete PMSM parameter identification model is established; then, the discrete H∞ filtering algorithm is derived using game theory reducing state and measurement noise influence. A cost function is defined, solving extremes values of different terms. A dynamic forgetting factor is introduced to the weighted combination of initial and current measurement noise covariance matrices, eliminating identification issues from different initial values. On this basis, a dynamic forgetting factor is added to weigh the combination of the initial measurement noise covariance matrix and the current measurement noise covariance matrix, which eliminates the influence of the discrimination error caused by the different initial values. Finally, the identification model is built in MATLAB/Simulink for simulation analysis to verify the feasibility of the proposed algorithm. The simulation results show the proposed H∞ filtering algorithm rapidly and accurately identifies resistance and inductance values with significantly improved robustness. The forgetting factor enables quick stable recognition even with poor initial values, enhancing PMSM control performance.

Metaheuristic adaptive control based on polynomial regression and differential evolution for robotic manipulators

Adaptive control of systems consists of the online adjustment of their control parameters to deal with parametric uncertainties and disturbances. Indirect Adaptive Control (IAC) of electromechanical systems and robotic manipulators based on metaheuristic optimization has been shown to be competitive compared to other classical and advanced control schemes. In this approach, the identification of system model parameters is stated as an optimization problem and then solved online by metaheuristics. Subsequently, the identified model is used in a second stage of online metaheuristic optimization to predict the system behavior and tune its control parameters. Therefore, this approach can be computationally expensive and unaffordable for systems with more complex dynamics whose models are just as complex. Due to the above, this work proposes an indirect adaptive control method for robotic manipulators based on differential evolution (DE) optimization, which aims to reduce the computational cost of the online identification stage by using Polynomial Regressors (PR) whose adjustment can be performed through a closed-form solution. The proposed method is tested in simulation for a 2 d.o.f. robotic manipulator and the results are compared against the original indirect adaptive control approach based on metaheuristic optimization and also a controller optimized offline. The post-hoc pair-wise Friedman tests confirm that the proposed approach provides competitive results compared to the original approach, with significant savings in computational resources.

Field trial of concurrent co-cable and co-trench optical fiber online identification based on ensemble learning.

The co-route optical fibers, comprising both co-cable and co-trench fibers, pose a significant potential risk to network service quality assurance by operators. They are incapable of achieving high-precision recognition and visual state management. In this study, we gathered both static and dynamic optical fiber data using a linewidth tunable light source (LTLS) and introduced a multimodal detection architecture that applies ensemble learning to the collected data. This constitutes what we believe to be the first field trial of concurrent recognition of optical fibers found both in co-cables and co-trenches. To identify co-cable fibers, we employed a double-layer cascaded Random Forest (DLC-RF) model based on the static features of fibers. For co-trench fiber, the dynamic characteristics of fiber vibrations are utilized in combination with multiple independent curve similarity contrast learners for classifying tasks. The proposed architecture is capable of automatically detecting the condition of the optical fiber and actively identifying the same routing segment within the network, eliminating the need for human intervention and enabling the visualization of passive optical fiber resources. Finally, after rigorous testing and validation across 11 sites in a typical urban area, including aggregation and backbone scenarios within the operator's live network environments, we have confirmed that the solution's ability to identify co-routes is accurate, exceeding 95%. This provides strong empirical evidence of its effectiveness.

A monitoring method of hull structural bending and torsional moment

The online identification of wave bending and torsional moment can be used for hull structural safety monitoring, that is, identifying the wave bending and torsional moment of the key cross section of the hull structure, comparing it to the threshold of the bending and torsional moment of the cross section, and finally evaluating the safety of the hull structure online. Therefore, the premise of the aforementioned structural safety monitoring is to accurately identify wave bending and torsional moments. For hull structures without large openings, a stable and high-precision method for identifying the wave bending moment is proposed, in which the position where the long baseline strain gauge (LBSG) installed is optimised using C-optimal and D-optimal design methods, and the mathematical model of the proposed method is established using a full-ship finite element simulation. The accuracy of the method for identifying the wave bending moment was verified using several numerical examples and the on-site monitoring of a large semi-submersible vessel. The results of the numerical simulations and onsite monitoring showed that the method for identifying the bending moment has high accuracy. For hull structures with large openings, a joint dual-section monitoring method (JDSMM) is proposed, in which two adjacent cross sections of the hull structure are monitored simultaneously; the vertical and horizontal bending moments, free torsional moment, and bimoment of the two cross sections being monitored are first identified, and then the bending and torsional moments of the middle section are obtained. The accuracy of the JDSMM was verified using several time-domain numerical examples. The results of the numerical simulations indicated that the JDSMM can accurately identify wave bending and torsional moment.

Self-evolution direct thrust control for turbofan engine individuals based on reinforcement learning methods

With the development of high-performance aero engines, the need for faster and more accurate thrust control has become increasingly critical. Remarkable achievements have been made in direct thrust control (DTC) for turbofan engines to meet these demands in recent years. However, most model-based methods only focus on the control law optimization to pursue better thrust control performance and ignore the model difference between the nominal model and the real engine. When there exist individual differences, the offline-designed control laws may suffer from significant online control performance degeneration problems. This paper proposes a novel self-evolution direct thrust control (SDTC) architecture to solve this problem, in which an uncertainty-aware state prediction model (USPM) and a self-evolution thrust controller are included. The USPM is proposed to predict the engine states of the individual engine, in which bootstrapped neural network (NN) ensemble methods are applied to deal with the epistemic uncertainty during the online identification process. To solve the performance degeneration problem, a novel sample evaluation prediction control (SEPC) and online policy evolution (OPE) reinforcement learning (RL) algorithm are proposed. When the nominal control policy is first applied in the onboard controller, the SEPC algorithm will temporarily adjust the output action in order to achieve a stable dynamic performance based on the prediction results of the USPM. The OPE algorithm updates the policy periodically and makes sure the policy is adapted to the new characteristics. Numerical simulation shows that the USPM can achieve high prediction accuracy under epistemic uncertainties, which helps the SEPC achieve satisfactory control performances under individual differences. The OPE algorithm can facilitate policy adaptation to individual characteristics with relatively low computational cost. Compared with other methods, our proposed SDTC architecture takes the lead in both performance and complexity and is instructive for other DTC researches.

Online identification of burn-through and weld deviation in sheet lap MIG welding based on YOLOv5

Online monitoring of the welding process is very important to the development of the modern manufacturing industry. However, monitoring welding defects in real-time and with high accuracy is challenging because welding is inherently a dynamic, non-linear process. In this paper, the forming defects generated during the melt Inert-gas welding of sheet steel as the object of research. We propose a multi-task simultaneous monitoring system for welding defects based on the YOLOv5 model, which can achieve highly accurate simultaneous detection of burn-through and weld deviation defects. Firstly, a passive vision system has been designed that can filter out strong arc light interference. In combination with a self-made camera bracket with flexible adjustment and repeatable positioning, clear images of the molten pool can be acquired. Compared to common methods that require a laser for weld deviation detection, all monitoring tasks can be performed with just one charge coupled device camera, significantly reducing the cost of deploying the system. Secondly, to obtain experimental data closer to the industrial field environment. We simulated different degrees of local and continuous burn-through by changing three parameters: welding current, plate thickness, and lap width. Four different deviations: left, right, and tilt were designed to obtain rich weld deviation image data. Finally, based on the above data, we compared several aspects such as detection speed, recognition accuracy, and loss function. The YOLOv5s model with the smallest model parameters was finally selected as the base model. The results show that our proposed multi-task simultaneous monitoring system achieves an average accuracy of 98.84% for the identification of four molten pool states. Burn-through defect and weld deviation detection with more than 50 frames per second can be achieved in the online detection state. This paper can provide some guidance for the online monitoring of the welding manufacturing process.

Mass estimation of a simple hydraulic crane using discrete extended Kalman filter and inverse dynamics for online identification

Automatization of hydraulic machinery requires accurate information of the current dynamic state of the machinery but also information of the underlying dynamic model characterized by a set of parameters. Some of the parameters can be considered static and well defined, such as machinery dimensions, whereas a part of the parameter set is time varying and needs to be identified based on observations. Particularly, difficult parameters to estimate are the ones, from which no prior knowledge is available. Consequently, the parameter corrections cannot be assumed to be small, which is commonly required for the existing parameter estimation algorithms. This study creates an online capable identification algorithm for estimation of a load mass operated by a hydraulic crane. In the case of load mass estimation, the unknown parameter can be practically any positive value, which implies the parameter corrections to be large. In this study, the estimation problem is divided in two parts: First, the dynamical states of the system are estimated based on the system kinematic relationships and dynamics of the hydraulic circuit. Secondly, the unknown load mass is estimated based on the known hydraulic forces and kinematics using the inverse dynamics of the mechanical structure. The proposed algorithm is tested with both artificially created measurements and with an experimental setup. The results show that both the kinematics of the structure and hydraulic pressures can be accurately estimated using the proposed method. Moreover, the method can be used to further estimate the payload mass. A drawback related to inverse dynamics is that it produces biased estimates in static equilibrium because of the discontinuous nature of static friction force. However, this drawback can be avoided, in part, by not updating the payload estimate in the low-velocity region. The proposed estimation methodology is capable for online identification, and as such, it can be used to adapt the control laws of automated machinery. Moreover, the methodology can be useful to record and document the amount of payload being handled during a work cycle.

Spacecraft Segment Damage Identification Method Based on Fiber Optic Strain Difference Field Reconstruction and Norm Calculation

Real-time online identification of spacecraft segment damage is of great significance for realizing spacecraft structural health monitoring and life prediction. In this paper, a damage response characteristic field inversion algorithm based on the differential reconstruction of strain response is proposed to solve the problem of not being able to recognize the small damages of spacecraft structure directly by the strain response alone. Four crack damage location identification methods based on vector norm computation are proposed, which realize online identification and precise location of structural damage events without external excitation by means of spacecraft structural working loads only. A spacecraft segment structural damage monitoring system based on fiber optic grating sensors was constructed, and the average error of damage localization based on the curvature vector 2 norm calculation was 2.58 mm, and the root-mean-square error was 1.98 mm. The results show that the method has superior engineering applicability for on-orbit service environments.

Isolation and Identification of Periodic Disturbances in BLDC Motors

A high-gain linear observer is proposed for online identification and reconstruction of a periodic disturbance. It is assumed that the frequency of the periodic disturbance is known, but its amplitude and phase are unknown. This is not a stringent assumption since such supposition can be easily characterized for disturbances like cogging torque, or bearing defects, both well-known problems in the control of electric machines. The observer design is based on the internal model principle in combination with a high-gain observer. Numerical results are included to validate the estimation of the periodic disturbance even when another disturbance of a different frequency is added. The estimated disturbance can be used to design a controller for reducing the disturbance effect. Alternatively, the estimated disturbance can be used for fault detection like in the case of gears and bearing faults, or simply to characterize the cogging torque of a particular machine.

Enhanced robust multimode process monitoring under dirty data via difference-based decomposition of matrix

Traditional data-driven methods generally suppose the training dataset is not corrupted by outliers. However, outliers are inevitable in the real industrial processes even with a relatively high ratio, which degrades the accuracy of data-based models. For multimode process monitoring, outliers may deteriorate the accuracy of both mode identification and fault detection. However, the existing robust methods can hardly deal with a large percentage of outliers, i.e., dirty data in neither single nor multimode processes. In this paper, a robust multimode process monitoring scheme is developed by alleviating the negative effect of dirty data. A difference-based decomposition of matrix (DDM) algorithm is first proposed to divide the data into a basic subpart and a non-basic subpart. The optimization function of the proposed DDM algorithm is solved by the alternating direction method of multipliers (ADMM). In the off-line procedure, an iterative decomposition strategy is designed based on the proposed DDM approach to identify dirty data and obtain the clean dataset of each mode. In the on-line procedure, an on-line sample identification strategy is developed by the type indicator derived from the DDM algorithm to determine whether the current sample belongs to a mode, the dirty data, or the fault. A numerical example and an industrial-scale multiphase flow facility indicate the proposed method can improve the accuracy of both mode identification and fault detection for multimode processes even under dirty data.

Viability of Glycolysis for the Chemical Recycling of Highly Coloured and Multi-Layered Actual PET Wastes.

The chemical recycling of poly(ethylene terephthalate) -PET- fractions, derived from actual household packaging waste streams, using solvolysis, was investigated. This recycling strategy was applied after a previous on-line automatic identification, by near-infrared spectroscopy -NIR-, and a subsequent selective sorting of the different PET materials that were present in the packaging wastes. Using this technology, it was possible to classify fractions exclusively including PET, virtually avoiding the presence of both other plastics and materials, such as paper, cardboard and wood, that are present in the packaging wastes, as they were efficiently recognised and differentiated. The simple PET fractions, including clear and monolayered materials, were adequate to be recycled by mechanical means meanwhile the complex PET fractions, containing highly coloured and multi-layered materials, were suitable candidates to be recycled by chemical routes. The depolymerisation capacity of the catalytic glycolysis, when applied to those complex PET wastes, was studied by evaluating the effect of the process parameters on the resulting formation and recovery of the monomer bis(2-hydroxyethyl) terephthalate -BHET- and the achieved quality of this reaction product. Comparable and reasonable results, in terms of monomer yield and its characteristics, were obtained independently of the type of complex PET waste that was chemically recycled.