Display omitted
•Presents a Mxene based virtual sensor array (VSA).•Proposes a convenient strategy for VSA.•We demonstrate correct rates of 90.9%, 90.5%, 90% for the different groups of ...VOCs.•Demonstrates an accuracy of 93.2% for the prediction of ethanol concentrations.
Two-dimensional transition metal carbides/nitrides, known as MXenes, have recently received significant attention for gas sensing applications. However, MXenes have strong adsorption to many types of volatile organic compounds (VOCs), and therefore gas sensors based on MXenes generally have low selectivity and poor performance in mixtures of VOCs due to cross-sensitivity issues. Herein, we developed a Ti3C2Tx-based virtual sensor array (VSA) which allows both highly accurate detection and identification of different VOCs, as well as concentration prediction of the target VOC in variable backgrounds. The VSA’s responses from the broadband impedance spectra create a unique fingerprint of each VOC without a need for changing temperatures. Based on the methodologies of principal component analysis and linear discrimination analysis, we demonstrate highly accurate identifications for different types of VOCs and mixtures using this MXene based VSA. Furthermore, we demonstrate an accuracy of 93.2% for the prediction of ethanol concentrations in the presence of different concentrations of water and methanol. The high level of identification and concentration prediction shows a great potential of MXene based VSA for detection of VOCs of interest in the presence of known and unknown interferences.
The content of free calcium oxide (f-CaO) in clinker significantly determines the quality of the final cement production. However, in practice, the value of f-CaO content in clinker is off-line ...sampled manually with a significant time interval and then analyzed in a laboratory with a large time delay, which could meet the needs for monitoring and control of cement quality. To tackle this problem, this article proposes a data-driven model based on deep dynamic features extracting and transferring methods to build a virtual sensor for f-CaO content prediction. First, in this model, large-scale unlabeled data collected from the process distributed control system (DCS) take a vital effect in extracting nonlinear dynamic features along with the limited labeled data samples. Then, the extracted features are transferred to a powerful regression model, the eXtreme Gradient Boosting (XGBoost), for output f-CaO prediction. Besides, an incremental model updating strategy is proposed for the augment of online data samples, which facilitates the virtual sensor to adapt the process time-variant characteristics. Finally, the proposed virtual sensor is verified by a data set acquired from a real cement production process. Comparing with traditional statistical modeling methods, the prediction accuracy of f-CaO content is significantly improved.
This technical note studies both the stability and tracking recovery problems for a class of continuous-time Markov jump piecewise-affine (PWA) systems against sensor faults. A novel reconfigurable ...control design approach is proposed to recover the mean-square input-to-state stability (ISS) of the closed-loop system and the tracking property of constant reference inputs, the key idea of this approach is to insert a reconfiguration block including a separate virtual sensor between the faulty system and the nominal controller. Then, a novel extended improved mode-region switching paths (extended-IMRSPs) algorithm is developed to overcome the state switching position mismatch between the faulty system and the reconfiguration block due to interleaving different regions. Furthermore, the <inline-formula><tex-math notation="LaTeX">\mathbb {S}</tex-math></inline-formula>-procedure is employed to cope with the affine term, combining with the ellipsoidal outer approximation technique. Finally, the effectiveness and the advantage of both the proposed reconfigurable control strategy and the developed extended-IMRSPs algorithm are demonstrated via an illustrative example.
•An improved decision tree-based fault diagnosis method is proposed for VRFs.•The method combines decision tree with two virtual sensor-based fault indicators.•The method can isolate air-side fouling ...and improper refrigerant charge faults.•The method is evaluated using on-line testing data collected from practical systems.•The method shows more reliable diagnosis results than three tree-based data-driven models.
This paper proposes an improved decision tree (DT)-based fault diagnosis method for practical variable refrigerant flow (VRF) system. The proposed method is a three-stage method combining DT model with virtual sensor-based fault indicators (FIs). First, FIs are developed based on the virtual sensor (VS) theory for VRF faults, i.e., condenser air-side fouling (Fouling), refrigerant undercharge (RU) and overcharge (RO). Second, FIs are employed as additional input variables to induct a DT-based classification model classification and regression tree (CART). Third, the FIs-CART classification model is used to diagnose on-line data. Validation is conducted using two different datasets, the experimental testing dataset and the on-line testing dataset. Results indicate that the method correctly isolates the three faults i.e., Fouling, RU and RO. The improved DT method is also compared with three tree-based data-driven methods including CART, random forest (RF) and generalized boosted regression (GBM). Comparative results reveal that the proposed method has better fault diagnosis performance for both the experimental and the on-line testing datasets.
•Fast demand response strategy for HVAC system without physical flowmeters of AHUs.•Virtual flowmeter modelled to estimate the water flow rate of each AHU.•Self-adjusting water flow supervisor for ...determining water flow rate set-point.•The proposed method extents the application scope of fast demand response.
The power grid is facing the critical issue concerning the power imbalance. To address the issue, demand response programs are increasingly deployed to encourage the end-users to change their load profiles. For buildings, the existing fast demand response strategy has been demonstrated effective in performing quick response to the grid request by reducing the power demand. However, the existing fast demand response methods require a physical flowmeter to be installed in each air handling unit. While in most of the existing commercial buildings, flowmeters are rarely installed in individual air handling units due to the high initial cost. As a result, the existing fast demand response method may not be applicable in these commercial buildings. Thus, this paper presents a virtual sensor based self-adjusting control strategy for fast demand response of building heating, ventilation and air-conditioning system. A virtual flowmeter model is first developed to estimate the water flow rate in each air handling unit. Based on the virtual flowmeter model, a self-adjusting water flow supervisor, in which the online self-adjusting method is integrated to reduce efforts in parameter identification, is then developed to achieve a balanced water flow distribution among different air handling units. The performances of the proposed control strategy have been tested and evaluated on a simulated system. The results show that the virtual flowmeter model has good accuracy for estimating the water flow rate in AHUs. The proposed control strategy can achieve significant and quick power reduction and meanwhile address the hydraulic imbalance problem.
The performance of machine tools depends on their working accuracy. The paper presents a new approach for in-process monitoring the geometric and kinematic accuracy of a machine tools. This approach ...utilizes a virtual sensor for capturing displacements at the tool center point (TCP). The virtual sensor is set up by using a data-driven approach based on extended transmissibility functions. The input data for monitoring are the NC internal data. The transmissibility functions are established via standard experimental procedures such as double ball bar (DBB) tests tests and estimation of frequency response functions (FRF). The data from the virtual sensor can be used for process monitoring and troubleshooting when launching a new process.
•FDD incorporates virtual sensors and fault impact model with low cost sensors.•Coil fouling fault impact is evaluated in order to set thresholds for diagnosis.•Field test was performed to simulate ...refrigerant charge faults and condenser fouling.•Virtual sensors identify and isolate a specific fault from other faults.•Fault impact models determine the severity of a fault needs for service.
The primary goal of this research is to evaluate, implement, and demonstrate a fault detection and diagnostic based on a number of virtual sensors and a fault impact model in the field. The primary bottlenecks to diagnostic implementation in the field are the high initial costs of additional sensors. The other difficulty in applying existing approach is in handling multiple faults that occur simultaneously because the state variables can depend on more than one fault along with the operating conditions. However, the diagnostic approaches based on virtual sensors can identify and isolate specific faults using a number of low-cost physical sensors. As the first of step, an analysis of data from a number of air conditioners was conducted to understand the impacts of condenser fouling faults on performance in order to set thresholds for diagnostics. The field test for the air conditioners was performed to simulate refrigerant charge faults and condenser fouling. The existing charging method would have difficulty in identifying the proper charge amount under condenser fouling conditions. However, the virtual sensor provides an accurate refrigerant charge estimates within 10% of real measurements regardless of the different operating temperatures and condenser fouling faults. In addition, the implementation and demonstration of the automated fault detection and diagnostic has been developed and connected to data obtained from the air conditioner monitored in the field.
This article introduces a novel virtual sensor for predictive maintenance called mini-term. A mini-term can be defined as the time it takes for a part of the machine to do its job. Being a technical ...sub-cycle time, its function has been linked to production. However, when a machine or component gets deteriorated, the mini-term also suffers deterioration, allowing it to be a multifunctional indicator for the prediction of machine failures as well as measurement of production. Currently, in Industry 4.0, one of the handicaps is Big Data and Data Analysis. However, in the case of predictive maintenance, the need to install sensors in the machines means that when the proposed scientific solutions reach the industry, they cannot be carried out massively due to the high cost this entails. The advantage introduced by the mini-term is that it can be implemented in an easy and simple way in pre-installed systems since you only need to program a timer in the PLC or PC that controls the line/machine in the production line, allowing, according to the authors’ knowledge, to build industrial Big Data on predictive maintenance for the first time, which is called Miniterm 4.0. This article shows evidence of the important improvements generated by the use of Miniterm 4.0 in a factory. At the end of the paper we show the evolution of TAV (Technical availability), Mean Time To Repair (MTTR), EM (Number of Work order (Emergency Orders/line Stop)) and OM (Labour hours in EM) showing a very important improvement as the number of mini-terms was increased and the Miniterm 4.0 system became more reliable. In particular, TAV is increased by 15%, OM is reduced in 5000 orders, MTTR is reduced in 2 h and there are produced 3000 orders less than when mini-terms did not exist. At the end of the article we discuss the benefits and limitations of the mini-terms and we show the conclusions and future works.
This article presents an industrial implementation of a virtual sensor in the process of fault detection of an induction motor. An ensemble-learning soft-sensor is developed to detect broken rotor ...bar that is essential to prevent irreparable damage. Most of the existing diagnostic methods assume that the data distribution is static and that all data is available during the training, while in real applications, the data become available as data streams. The proposed method is inspired by the ensemble learning algorithm, which is combined with a new drift detection mechanism. The advantages of the proposed approach are three-fold. First, a fair comparison with other algorithms show the effectiveness of the soft sensor scheme. Second, the presented concept change detection algorithm is capable of detecting a new class in the data stream as well as data distribution change, and last but not least, the efficacy of the proposed algorithm is demonstrated using benchmark concept drift data streams.
Digital twins are able to bridge the physical and the virtual world, which is especially useful in industrial environments. One small, but rather essential, kind of digital twin is the so called ...virtual sensor model. This type of model is used to enhance or replace a physical sensor in industrial settings to reduce costs or enable information retrieval from inaccessible locations within a machine or process. The virtual sensor model is usually trained once, whereat only a predefined amount of information without any adaptation possibilities on the clients side is provided. As manufacturers want to provide the possibility of custom adaptations for their clients, the virtual sensor models require to incorporate adaptive features, which also have to handle incomplete data recordings, e.g. uncensored data. Traditional offline machine learning approaches are often insufficient for such adaptive requirements, therefore the usage of online learning approaches is gaining increased attention, to avoid high computational, storage and temporal costs. This paper covers the continued training of such adaptive virtual sensor models, focusing on the handling and integration of censored online data. Different approaches to tackle the problems of catastrophic forgetting in online learning and correction of censored data are presented as well as the handling of censored data in online learning environments. The experiments sections compares various scenarios with and without censored data using an industrial dataset and demonstrates the positive influence of different online learning approaches.