On-line Estimation Research Articles

A novel sampled-data asynchronous CatBoost model with knowledge-based guidance for product quality estimation in the industrial processes

For process industry with asynchronous sampling, delayed laboratory analysis of the product quality restricts operation control and optimization. However, it is challenging to ensure the estimation accuracy due to the fact that online estimation model is disturbed by the dynamic change of production conditions. Therefore, a product quality estimation strategy for asynchronous sampling industrial processes under data-driven and knowledge-based guidance is proposed. First, domain knowledge and process mechanism are used to establish the mechanics models and identify the critical variables. Then, the dimensionality of the samples is reduced to extract effective information which is used to partition data samples. Besides, the data sample is balanced by the adaptive synthetic sampling technique under different production conditions. Finally, a CATBoost estimation model is built for product quality online acquisition to improve interpretability and adaptability. The reliability of the proposed estimation approach is verified by a sodium aluminate solution evaporation application case.

A fast, accurate and uncalibrated robotic puncture method

AbstractBackgroundRobotic puncture system (RPS) consists of an optical tracking system (OTS) and a robotic arm gripping the puncture needle. Typically, the RPS requires hand‐eye calibration before the surgery in order to obtain the relative position between the OTS and the robotic arm. However, if there is any displacement or angular deviation in either the robotic arm or the OTS, the calibration results become invalid, necessitating recalibration.MethodsWe propose an uncalibrated robotic puncture method that does not rely on the hand‐eye relationship of the RPS. By constructing angle and position graph jacobian matrices respectively, and employing Square Root Cubature Kalman Filter for online estimation. This enables obtaining control variables for the robot to perform puncture operations.ResultsIn simulation experiments, our method achieves an average error of 1.3495 mm and an average time consumption of 39.331 s.ConclusionsExperimental results indicate that our method possesses high accuracy, low time consumption, and strong robustness.

Filtering-based gradient joint identification algorithms for nonlinear fractional-order models with colored noises

The main focus is to realize the on-line parameter estimation of the nonlinear fractional-order model with colored noises in this paper. Firstly, an auxiliary model gradient descent algorithm is derived to synchronously produce the estimates of the parameters, including the fractional-order. In order to decrease the noise interference, the filtering-based gradient descent estimation framework by constructing a linear filter provides a feasible method. Furthermore, the forgetting factor is introduced to the proposed filtering-based algorithm for the sake of improving the convergence rate. Comparative simulation results demonstrate the proposed effectiveness and high approximation accuracy of the proposed algorithms.

Adaptive Regression Analysis of Heterogeneous Data Streams via Models with Dynamic Effects

Streaming data sequences arise from various areas in the era of big data, and it is challenging to explore efficient online models that adapt to them. To address the potential heterogeneity, we introduce a new online estimation procedure to analyze the constantly incoming streaming datasets. The underlying model structures are assumed to be the generalized linear models with dynamic regression coefficients. Our key idea lies in introducing a vector of unknown parameters to measure the differences between batch-specific regression coefficients from adjacent data blocks. This is followed by the usage of the adaptive lasso penalization methodology to accurately select nonzero components, which indicates the existence of dynamic coefficients. We provide detailed derivations to demonstrate how our proposed method not only fits within the online updating framework in which the old estimator is recursively replaced with a new one based solely on the current individual-level samples and historical summary statistics but also adaptively avoids undesirable estimation biases coming from the potential changes in model parameters of interest. Computational issues are also discussed in detail to facilitate implementation. Its practical performance is demonstrated through both extensive simulations and a real case study. In summary, we contribute to a novel online method that efficiently adapts to streaming data environment, addresses potential heterogeneity, and mitigates estimation biases from changes in coefficients.

LoRa-based outdoor localization and tracking using unsupervised symbolization

This paper proposes a long-range (LoRa)-based outdoor localization and tracking method. Our method presents an unsupervised localization approach that utilizes symbolized LoRa received signal features, such as RSSI, SNR, and path loss, where each symbol represents a system state. To identify the partitioning boundaries between the symbols in time series, we employ maximum entropy partitioning. The D-Markov machine is used to construct nondeterministic finite-state automata for extracting temporal patterns. We incorporate the Chinese restaurant process for online estimation, especially in scenarios with an unbounded number of probable areas around each LoRa gateway. An adaptive trilateration approach is then used to localize the target node from the estimated ranged radii of areas. The point-wise localization data was used for time-series continuous tracking. We collected a dataset using three LoRaWAN gateways, sensor nodes powered by single-use batteries, and a Chirpstack server on a sports oval. We thoroughly evaluated the proposed method from the perspectives of localization accuracy and tracking capability. Our method outperformed state-of-the-art machine learning-driven range-based and fingerprint-based localization techniques.

Temperature enhanced early detection of internal short circuits in lithium-ion batteries using an extended Kalman filter

Fuel shortages and environmental concerns primarily drive the recent growth in electrification. Lithium-ion batteries are the preferred power source for electric vehicles and other applications due to their high energy and power densities. Among potential issues that can affect Li-ion batteries, thermal runaway is a significant concern, believed to be primarily caused by internal short circuits. Therefore, early internal short circuit detection has become a critical task for any Li-ion battery-powered engineering system prioritizing safety. This paper presents a novel online internal short circuit detection method based on the state vector augmentation of an extended Kalman filter with: (i) voltage and surface temperature observations, (ii) a hysteresis state, and (iii) a state related to the internal short circuit. The proposed method is assessed numerically, mimicking an electrical vehicle battery working cycle. The framework allows for an online estimation of the internal short circuit state while remaining computationally lean, thus potentially allowing for implementation into commercial BMSs, and it is proven to capture the internal short circuit occurrence within safety limits. Additionally, the advantages of including both voltage and surface temperature observations have been highlighted. Future work envisaged towards field implementation of the technique in BMS is eventually and briefly discussed.

Adaptive neural identification and non-singular control of pure-feedback nonlinear systems

This paper proposes a new constructive identification and adaptive control method for nonlinear pure-feedback systems, which remedies the ’explosion of complexity’ and potential control singularity encountered in the traditional adaptive backstepping controllers. First, to avoid using the backstepping recursive design, alternative state variables and the corresponding coordinate transformation are introduced to reformulate the pure-feedback system into an equivalent canonical model. Then, a high-order sliding mode (HOSM) observer is used to reconstruct the unknown states for this canonical model. To remedy the potential singularity in the control, the unknown system dynamics are lumped to derive an alternative identification structure and one-step control synthesis, where two radial basis function neural networks (RBFNN) are adopted to online estimate these lumped dynamics. In this framework, the online estimation of control gain is not in the denominator of controller, and thus the division by zero in the controllers is avoided. Finally, a new online learning algorithm is constructed to obtain the RBFNNs’ weights, ensuring the convergence to the neighborhood of true values and allowing accurate identification of unknown dynamics. Theoretical analysis elaborates that the convergence of both the tracking error and the estimation error is obtained simultaneously. Simulations and practical experiments on a hydraulic servo test-rig verify the effectiveness and utility of the suggested methods.

Enhancing active disturbance rejection design via deep reinforcement learning and its application to autonomous vehicle

Taking the velocity regulation of autonomous driving as an example, this paper developed an enhancing active disturbance rejection design employing the deep reinforcement learning-deep deterministic policy gradient (DDPG). In this scheme, the active disturbance rejection control (ADRC) is adopted to online estimate and compensate disturbances and uncertainties, and feasible regions of control parameters are obtained through the Lyapunov method. Then DDPG is combined with ADRC to online adaptively tune control parameters in response to the changing environments, where safety, comfort, and energy-saving are considered in the reward design, and mapping relation between the defined action and state is constructed for the maximal reward. Besides, numerical simulations demonstrate the better performance and stronger robustness of the enhancing design when facing uncertainties, sensor noise, and mechanical faults, and comfort and energy consumption have also been improved to some extent compared with the general ADRC and model predictive control (MPC).

Dynamic performance improvement of an isolated wind-diesel system with intelligently controlled supercapacitor energy storage system

This paper aims to develop a supervisory control scheme to enhance the effectiveness and profitability of a small-rating super capacitor energy storage system (SCESS) used in load-frequency-control applications. The proposed approach, which uses one-step-ahead adaptive predictive control (APC), adeptly handles the operational limitations of the SCESS. For the purpose of online estimation of system parameters, the recursive least square (RLS) algorithm has been employed. The system can be characterized as a two input two output system, wherein the real and reactive powers required by the SCESS serve as control signals issued by the controller. The SCESS voltage is subject to constraints to limit energy trade within specific bounds. The proposed control scheme successfully maintains the voltage constraints of the SCESS while significantly reducing frequency and voltage deviations in the presence of two disturbances, viz; load disturbance and wind disturbance. The effectiveness of the proposed scheme is demonstrated through simulation experiments on an isolated hybrid wind diesel power system, which addresses several modeling and design aspects.

Improved electric-thermal-aging multi-physics domain coupling modeling and identification decoupling of complex kinetic processes based on timescale quantification in lithium-ion batteries

Unraveling the kinetic behavior inside the battery is essential to break through the limitations of mechanistic studies and to optimize the control of the integrated management system. Given this fact that the battery system is multi-domain coupled and highly nonlinear, an improved lumped parameter multi-physical domain coupling model is first developed to capture the electrical, thermal and aging characteristics of the battery in this paper. On this basis, an adaptive multi-timescale decoupled identification and estimation strategy is proposed based on the quantified timescale innovation, which realizes the online monitoring of the battery state and the accurate identification of the model parameters. The specific idea is that the decoupled identification of kinetic parameters inside the cell, the terminal voltage prediction and the real-time monitoring of the internal temperature with the online estimation of the available capacity are distinguished under different timescales. Meanwhile, the response time characteristic of the different kinetics is extracted and analyzed as a distinction between the coupled internal electrochemical processes. In this idea, four functionally different sub-observers are developed independently. Significantly, adaptive time-scale driven methods designed based on the fundamental timescale of the system, the amount of variation of the state of charge, and the amount of transfer charge are used separately for the observer implementation at different timescales. In addition, the coupling of the fast and slow kinetic parameter discriminators is achieved by diffusion voltage, and the internal temperature observer as well as the available capacity observer are coupled to each other based on the estimation results. Experimental results for two long-time operating conditions at 5, 25 and 45 °C show that the proposed strategy has fast convergence and reliable accuracy in monitoring the battery state characteristics. Compared with the traditional fixed timescale algorithm, the proposed multi-physics domain coupling modeling strategy based on independent timescale driven design is more competitive in practical embedded applications.

Positioning and antiswing control of overhead crane systems: A supervisory scheme

Overhead crane, as a common handling mechanical equipment, plays an irreplaceable role in modern industry. To ensure its efficient and reliable running, there is an urgent need to develop advanced control strategies. In this paper, a novel supervisory control scheme is proposed for overhead crane systems. The scheme involves a high-level decision making mechanism, called a supervisor, to evaluate which controller is suitable at the current instant of time among a family of candidate controllers. Detailedly, for all possible payloads, a multi-estimator is first introduced to achieve the appropriate estimations of trolley position, trolley velocity, payload angle and angular velocity. Then, based on the estimator signals, a family of candidate controllers are designed, each of which is workable for a certain payload. Finally, by constructing monitoring signals and online evaluating the values of the signals under different candidate controllers, the controller corresponding to the minimum value of the signals is applied to the system at the current instant of time. Under the proposed supervisory controller, the payload can be driven to the desired position, and meanwhile the swing of the payload is eliminated. Remarkably, uncertainties dynamical compensation and performance online evaluation mechanisms are included in the scheme, which endows the scheme powerful ability to suppress payload mass uncertainties, to guarantee fast setting-time, and to achieve online estimation of the payload masses. The simulation and experimental results show better workability of the proposed supervisory control scheme in application, including strong robustness, fast setting-time, accurate payload mass estimation, and convenient parameter selection.

Net Power Enhancement of PEMFC System Based on Dual Loop Multivariable Coordinated Management

A dual loop multivariable coordinated management (DLMCM) is proposed in this paper dealing with the existed nonlinearity and strong coupling between operating parameters such as stack temperature and oxygen excess ratio (OER), to achieve net power enhancement of proton exchange membrane fuel cell (PEMFC) system. The proposed optimization method includes: 1) outer net power optimization loop and 2) inner data-driven multivariable control loop. In solving the outer net power optimization, the experimental data based net power smooth surface model and the security constraints is proposed for maximum net power regulator design. Regarding inner data-driven multivariable control loop, firstly, pseudo partial derivative linearization (PPDL) is proposed for online modeling and estimation of the multi input and multi output (MIMO) system. Then, PPDL based discrete adaptive integral terminal sliding mode control is proposed to achieve maximum net power trajectory tracking. Experimental tests carried out on 5-kW PEMFC system show that the proposed DLMCM has better tracking ability, and internal disturbance rejection compared with sliding mode control (SMC). Moreover, the comparative study with manual guided control, temperature control and air flow control testifies the effectiveness of proposed DLMCM for net power enhancement with a low level of parasitic power consumption ratio.

Application of DEKF algorithm for SOC estimation of lithium batteries

A technique based on a double-extended Kalman filter (DEKF) is created to estimate the SOC of a lithium-ion battery pack in real time. The Kalman filter technique and the Thevenin battery model are used to develop a state-space expression for the battery model. Improved accuracy and a more faithful representation of the battery’s internal state are achieved using the least squares approach and the DEKF algorithm to determine the model’s parameters. Experiments are developed for testing batteries that adhere to the presented idea of the online estimate of charge status using the DEKF algorithm. The results of the experiments demonstrate the algorithm’s excellent accuracy and environmental flexibility in calculating SOC online in various settings. The greatest error is less than 4.5 percent. The DEKF method is shown to successfully address the issues of erroneous initial estimates and cumulative errors and to have strong convergence and resilience.

Physics-Based Model Predictive Control for Power Capability Estimation of Lithium-Ion Batteries

The power capability of a lithium-ion battery signifies its capacity to continuously supply or absorb energy within a given time period. For an electrified vehicle, knowing this information is critical to determining control strategies such as acceleration, power split, and regenerative braking. Unfortunately, such an indicator cannot be directly measured and is usually challenging to be inferred for today's high-energy type of batteries with thicker electrodes. In this work, we propose a novel physics-based battery power capability estimation method to prevent the battery from moving into harmful situations during its operation for its health and safety. The method incorporates a high-fidelity electrochemical-thermal battery model, with which not only the external limitations on current, voltage, and power, but also the internal constraints on lithium plating and thermal runaway, can be readily taken into account. The online estimation of maximum power is accomplished by formulating and solving a constrained nonlinear optimization problem. Due to the relatively high system order, high model nonlinearity, and long prediction horizon, a scheme based on multistep nonlinear model predictive control is found to be computationally affordable and accurate.

Adaptive Output Regulation of a Class of 1-D Hyperbolic PDEs With Unknown Boundary Scaled Parameters

This paper considers the problem of adaptive output regulation of a class of 1-D anti-collocated hyperbolic PDEs. The model is subject to unknown scaled parameters in both the boundary condition and boundary measurement. An adaptive boundary observer, providing online estimates of the system state and parameters, is designed. Particularly, to realize that the pure boundary state at left side tracks a reference signal, a challenging problem with the estimate of the true value of unknown scaled parameter appearing in measurement (at the left side) is tackled in this paper. It is shown that the proposed observer is exponentially convergent and furthermore the exponential output regulation and stabilization are achieved.

Confidence region for distributed stochastic optimization problem via stochastic gradient tracking method

Since stochastic approximation (SA) based algorithms are easy to implement and need less memory, they are very popular in distributed stochastic optimization problems. Many works have focused on the consistency of the objective values and the iterates returned by the SA based algorithms. It is of fundamental interest to know how to quantify the uncertainty associated with SA solutions via the confidence regions of a prescribed level of significance for the true solution. In this paper, we discuss the framework of constructing the asymptotic confidence regions of the optimal solution to distributed stochastic optimization problem with a focus on the distributed stochastic gradient tracking method. To attain this goal, we first present the asymptotic normality of Polyak–Ruppert averaged distributed stochastic gradient tracking method. We then estimate the corresponding covariance matrix through online estimators. Finally, we provide a practical procedure to build the asymptotic confidence regions for the optimal solution. Numerical tests are also conducted to show the efficiency of the proposed methods.

Bayesian Vehicle Load Estimation, Vehicle Position Tracking, and Structural Identification for Bridges with Strain Measurement

Vehicle load estimation and health monitoring of bridges are of great importance for the health monitoring of bridge structure under vehicle loads. Traditional methods for the estimation of vehicle load require the positions of the vehicles. The vehicle position tracking is generally conducted in offline manner and requires the installation of additional sensors. To resolve these problems, we developed a Bayesian probabilistic approach for the online estimation of vehicle loads, vehicle positions, and structural parameters for bridges. The crux is to model the vehicle load vector as a modulated filtered Gaussian white noise due to the fact that the vehicle-bridge interaction forces are in essence the responses of the vehicle-bridge coupled system under the excitation of the road roughness described by Gaussian random field and the constant vehicle weights. Furthermore, the vehicle speed vector is introduced to track the unknown positions of vehicles. There are three appealing features in this approach. First, it allows the simultaneous estimation of vehicle loads, vehicle positions, and structural parameters in an online manner. Second, this method allows for time-varying vehicle speed tracking. Third, the proposed method is applicable to the case with multiple vehicles. Examples for the case where single/multiple vehicles pass across bridges with uniform speeds/variable speeds are presented to demonstrate the feasibility of the proposed method for vehicle load estimation, vehicle position tracking, and bridge structural identification using only strain measurements.

Degradation state analysis of the IGBT module based on apparent junction temperature

The multi-chip parallel insulated gate bipolar transistor (IGBT) is the core device in large-capacity power electronic equipment, but its operational reliability is of considerable concern to industry. The application of IGBT online degradation state analysis technology can be beneficial to the improvement of system reliability. The failure mechanism of IGBT devices is discussed in this paper, and a technique for analyzing the degradation state of IGBT based on apparent junction temperature is proposed. First, the distortion consistency of the voltage rise time in various failures is discussed, and the junction temperature dependence of the voltage rise time is then demonstrated. Subsequently, an apparent junction temperature model based on the voltage rise time is established (the fitting accuracy is as high as 94.3%). From the high-frequency model in the switching process of the device, an online extraction technology of key parameters (e.g., voltage rise time) is developed. Finally, an experimental platform for IGBT degradation state estimation is established, and the feasibility of IGBT degradation state estimation based on apparent junction temperature is proved, especially the degradation of bonding-wire and the gate-oxide-layer. The experimental results show that the proposed IGBT degradation state estimation technique based on apparent junction temperature is a reliable online estimation method with non-contact, high accuracy, and comprehensiveness.

Automatic trajectory adaptation for the control of quality characteristics in a powder compaction process

The manufacturing process of sintered components requires the compaction of powder in a rigid die and the sintering of the resulting green parts. To produce green parts of the same constant quality over a long period of time, regular quality checks and trajectory adjustments are needed to account for changing operating conditions due to e.g. varying temperature, humidity, stroke rate, or powder quality. Also the first setup of the machine requires manual measurements and iterative adjustments of trajectory key points. To overcome these labour-intensive, manual measurements and adjustments, we propose an algorithm for the automatic adaptation of trajectory key-points that for the first time allows bringing and keeping more than one quality characteristic (mass and dimensions) fully automatically within predefined intervals. To safe costs, we further present two real-time capable, model-based solutions by employing grey-box and black-box models for the online estimation of the workpiece mass and height instead of their direct measurement. The grey-box model combines a filling and analytical compaction model and allows estimating mass and height rather than density as typically investigated in literature. The black-box model based on neural networks represents a complete novelty in the field as no similar data-driven approach could be found that aims at estimating mass and dimensions of green compacts. We finally compare the performance of these models to the baseline of sensor measurements as well as study their effect on the overall control performance. Results indicate that for both sensor-based and model-based solutions it is possible to reach and keep aforementioned quality characteristics within desired intervals. While overall the sensor-based approach performed best, also for the model-based approaches a good estimation performance is observed with the black-box model based on neural networks outperforming the grey-box model. While the black-box approach is fully data-driven and thus allows to widely avoid manual modelling activities, it still requires the presence of expensive sensors for capturing training data. The proposed grey-box approach, on the other hand, requires some manual modelling and model calibration activities, but provides a cost-effective solution to the problem and represents an alternative to more expensive, sensor-based implementations.

State-space modeling for electrochemical performance of Li-ion batteries with physics-informed deep operator networks

Online estimation of unobservable internal states is significant for safe operation of Li-ion batteries, and it constitutes one of the main functions of battery management system (BMS). The next-generation BMS expects model-based state estimation, especially with electrochemical models, which are accurate but often costly for solving. Therefore, it is required to build more easily executable state-space representation of electrochemical models for online state estimation. However, the traditional numerical methods for time discretization are relatively complicated, and the discretized system is not very flexible in modifying predictive time intervals. To address such issues, we introduce the concept of physics-informed operator learning for state-space modeling. Specifically, we propose an architecture, termed the physics-informed multiple-input operator network (PI-MIONet), to reformulate the state-space representation of the extended single particle (eSP) model. In this work, the PI-MIONet takes the Li-ion concentration of the whole electrode particle and current densities at the current time as the input functions, and predicts Li-ion concentration at any spatial-temporal location, which means that the forward predictions can be realized with user-defined step size. In addition, due to the capability of taking discretized functions as inputs, the PI-MIONet can be used for estimating states in the form of long vectors, and it can be conducted very efficiently, which makes it highly suitable for online applications in BMS. We verify the predictive performance of PI-MIONet through several synthetic experiments, and successfully apply it to the estimation of Li-ion concentration across the full particle with unscented Kalman filter algorithms.