On-line Operational Data Research Articles

A novel load allocation strategy based on the adaptive chiller model with data augmentation

Model-based load allocation strategy is an impactful solution to enhance energy efficiency of multiple-chiller system. Its performance is heavily dependent on the accuracy of chiller model. Data-driven model is a pretty-good solution. However, in real multiple-chiller system, the range of operation condition in historical data is commonly narrow, so it is challenging to develop an accurate data-driven model of chiller throughout full range of operation condition. In this paper, data augmentation algorithm is presented to generate the data outside of historical data, which is based on conditional generative adversarial network (CGAN) and elastic weight consolidation algorithm (EWC). Combined historical data and generated data, augmented training dataset is set up and updated by online operation data. Trained by online updated augmented training dataset periodically, adaptive chiller model is set up. Based on adaptive chiller model, a novel load allocation strategy presented for multiple-chiller system. The proposed strategy is verified by field test in multiple-chiller system. The results show that adaptive chiller model, with the aid of data augmentation algorithm, is more accurate. The proposed strategy can achieve 5.03 % energy saving compared with fixed set-point strategy, and the EER of proposed strategy is 6.27 % higher than that of fixed set-point strategy.

X-Stor: A Cloud-Native NoSQL Database Service with Multi-Model Support

In recent years at Tencent, we have observed that the use of multiple NoSQL databases for storing business data with diverse models has led to increased programming and deployment costs, as well as inefficient maintenance and underutilized resources. In this paper, we report X-Stor, a cloud-native NoSQL database system that supports multiple data models by extending different storage engines and efficiently managing them through a unified control plane. This design significantly reduces expenses and enables rapid expansion of new models, while seamlessly supporting their complete functionality through storage engine extensions. By consolidating multi-tenant services and data models on the same physical machines, X-Stor significantly enhances the utilization of cluster resources. Additionally, X-Stor introduces a standardized metric called Request Unit (RU) to measure tenant resource consumption for consumption-based pricing purposes. Leveraging this metric, we design RU-based resource management strategies and achieve efficient multi-tenant resource isolation and system load balancing. Currently, X-Stor manages a storage capacity of over 12PB for online operational data, including more than 100,000 tables with multiple data models. It handles 700 billion requests per day with a peak of 30 million requests per second. We evaluate the performance of X-Stor on popular benchmarks and production workloads. The results show that X-Stor performs well under diverse data models.

Identification of User Requirements and their Influencing Factors Based on Online Reviews and Operational Data

It is crucial for enterprises to clearly identify user needs during the process of formulating product design improvement plans. Therefore, it is essential to comprehensively and accurately identify user needs, explore the reasons behind the emergence of these needs, and incorporate user opinions into the process of product design improvement. A method is proposed to comprehensively and accurately capture user requirements and address the challenge of identifying the underlying causes of user requirements. This method utilizes online comments and operational data to identify user requirements and their influencing factors. First, text sentiment analysis techniques are employed to quantify the importance and performance values of product feature topics. Second, we construct a quadrant model to identify product features requiring improvement, and the original negative comments related to these features are traced. However, the quadrant model alone is insufficient to reflect specific product issues that users are concerned about. Therefore, a functional structure model based on product issues is designed to filter and identify factors that influence user requirements using operational data. Finally, a Bayesian network inference approach is utilized to identify the key influencing factors on user requirements, enabling analysis of the causes behind user requirements and the proposal of product design improvement strategies. The feasibility and effectiveness of the proposed method are validated through experiments conducted on heavy-duty truck data. By analyzing the original negative comments related to the power characteristics, specific user demands regarding the insufficient power of the product were identified, such as “obviously insufficient power when climbing slopes” and other issues. Based on the vehicle power system functional structure model, combined with expert knowledge and operational data, factors related to the state of parts and user behavior that may affect “insufficient vehicle power” were identified. Based on the analysis results, suggestions were made to improve the engine intake air temperature control strategy and to enhance vehicle performance by promoting correct user behavior through informational campaigns.

An electric-thermal coupling modeling method for lithium-ion battery using the state of charge normalization calculation method

The improved equivalent circuit model for Lithium-ion battery (LIB) has gained popularity in engineering due to its practical and straightforward structure, encompassing core parameters of interest such as open-circuit voltage, internal resistance, temperature, and state of charge (SOC). However, the nonlinear and coupled nature of electrical and thermal parameters makes it challenging to solve these parameters online in a battery management system. Thus, this paper proposes a battery model to decouple the electric-thermal coupling relationship, comprising analytical sub-models for open-circuit voltage, internal resistance and thermal parameters. The proposed method involves designing sets of off-line experiments under constant-current battery conditions and normalizing SOC calculation by utilizing a coulomb efficiency coefficient. The electrical and thermal parameters are then decoupled using off-line parameter identification to obtain open-circuit voltage, internal resistance, and temperature data under specific conditions. An open circuit voltage modeling method is proposed based on the Nernst equation and fitting method. Then, an internal resistance model of the battery is also modeled based on the Arrhenius equation and fitting method. In addition, according to the principles of heat generation and heat transfer, a thermal model is constructed and its model parameters are identified. Through the verification of internal resistance and terminal voltage under different ambient temperature and current conditions, the results illustrate that the modeling method proposed in this paper is able to accurately estimate the open-circuit voltage, internal resistance, and temperature based on online operational data of the battery.

Online Rotor Systems Condition Monitoring Using Nonlinear Output Frequency Response Functions Under Harmonic Excitations

Operation conditions such as rub-impact and misalignment usually accumulate and can eventually induce severe faults in rotor systems. These abnormal conditions often make rotor systems behave nonlinearly. Consequently, nonlinear signal and systems analyses have been applied to address rotor system condition monitoring and fault diagnosis problems. Nonlinear frequency response functions (NOFRFs) are a concept proposed to extend the well-known concept of frequency response function to the nonlinear case, and have recently been exploited to solve problems with signal analysis based rotor system condition monitoring and fault diagnosis. However, existing NOFRFs based analyses require conducting offline experiments on rotor systems. This cannot satisfy the industrial need of continuous monitoring of rotor system operational conditions, which is important for online fault diagnosis. In order to resolve this problem, the present article proposes a novel approach that performs the rotor system condition monitoring using the NOFRFs determined directly from the rotor system online operational data which are inherently harmonics. This approach involves the identification of a data-driven NARX (nonlinear auto regressive with eXegenous input) model of rotor systems from harmonic excitations and corresponding responses, evaluation of rotor system NOFRFs using the identified NARX model, and the use of evaluated NOFRFs for rotor system operational condition monitoring and fault diagnosis. The approach, for the first time, enables the NOFRFs to be applied to perform online monitoring of rotor systems. Experimental studies have been conducted to verify the effectiveness and potential applications of the proposed approach as well as its advantages over commonly used signal analysis based techniques.

Partial domain adaption based prediction calibration methodology for fault detection and diagnosis of chillers under variable operational condition scenarios

Data-driven fault detection and diagnosis (FDD) for building energy system has drawn much attention from both industry and academy due to their ability to learn complex relationships from big data. Most existing FDD studies only focus on the model diagnosis accuracy where the operational conditions of test dataset are similar with training dataset. However, it is rare for online operational data to be similar with offline training data. To better simulate the online environment in the aspect of operational conditions, a novel online data simulation method, which is named as variable operational conditions test, is proposed in this study. Furthermore, in order to improve the poor diagnosis performance of FDD models under variable operational condition scenarios, an unsupervised partial domain adaption algorithm is employed to calibrate the predictions of the FDD model within a short period of time. A comprehensive data experiment was conduct based on ASHARE RP-1043. The experimental results indicate the prediction calibration method yields decent performance improvement compared with three traditional data-driven-based FDD models, the highest and overall improvements are 55.07% and 14.5%, respectively. The research outcomes provide practical guidelines for the online application of FDD models when the training data cannot cover the whole sample space.

Toward a framework for risk monitoring of complex engineering systems with online operational data: A deep learning-based solution

A mathematical architecture is developed for system-level condition monitoring. This architecture is built toward performing end-to-end operation risk and condition monitoring. The streaming monitoring data is given to the architecture as the input and system-level and component-level operation health states are computed as the output. This architecture integrates fault trees as the system-level modeling method and Deep Learning (DL) as the components condition monitoring method. A number of different deep learning models are trained using both operation and maintenance data for the components. Then, the fault tree fuses the continuous components’ assessments to provide system-level health insight. The applicability of this architecture is tested by implementing it on a real-world mining stone crusher system. This approach is extendable to dynamic risk assessment of complex engineering systems. However, DL models should be used with caution for safety-critical applications. We show that having DL models with high accuracy is not enough for trusting their predictions. We discuss the calibration of DL-based condition monitoring models and demonstrate how they can improve the trustworthiness and interpretability of DL models in risk and reliability applications.

Requirements for developing high temperature creep life models for ageing pipework systems using power plant condition monitoring and inspection data

A novel creep life assessment methodology is proposed in this paper that uses periodic and routine inspection data in conjunction with online operational data and other computed parameters, such as the ‘rate of change’, and where applicable correlations with data from small specimen tests on ex-service materials. The proposed methodology is critically implemented by the introduction of a life assessment transfer function (LATF). This enables the computation of the residual life of high temperature components and essentially links the plant data routinely collected during outage inspections, via online monitoring or by testing of material samples to the life prediction model. The LATF performs the function of data filtering, pre-conditioning and parameter derivation so that this information can be used effectively in the life prediction model. In this paper high temperature pipework is used as an example of the proposed approach and deployable creep life prediction models are reviewed that could function with these plant datasets and enabled by a suitable LATF. Of critical importance is that the output from the life prediction is used to inform plant operations on any necessary changes in order to mitigate damage accumulation. An example is provided to illustrate the current assessment approach, exploiting real inspection data for estimation of residual life, as well as how the proposed new approach will address the limitations.

Comprehensive understanding of SO3 effects on synergies among air pollution control devices in ultra-low emission power plants burning high-sulfur coal

Abstract The environmental issue of SO3 pollution caused by coal-fired power plants attracts increasing attention. This work focused on the synergies between air pollution control devices in a 660 MW power plant burning high-sulfur coal. Parameters such as gas temperature and flow rate were varied within wide ranges under actual conditions. Both continuously tested SO3 concentration and on-line operational data were combined to establish a correlation between device performance and operational parameter. Results indicated that the gas temperature acted as a bridge of synergies between the electrostatic precipitator (ESP) and wet electrostatic precipitator (WESP). The SO3 removal efficiency across the ESP increased from 60.3% to 91.1% with the gas temperature decreasing from 136.7 to 114.8 °C, leading to the increase in corona current and the decrease in spark-over frequency of WESP. Moreover, additional electric field stages are also effective to enhance the corona current and reduce spark-over frequency. Furthermore, deviations between the design and operational data within 0–100% loading provided a correction to appropriately choose the gas velocity. With the proposed countermeasures, the SO3 emission can achieve a favorable level less than 5 mg/m3. The research findings provide a valuable technical pathway to eliminate the SO3 emission from power plants burning high-sulfur coal.

Virtualization of food supply chains with the internet of things

Internet technologies allow supply chains to use virtualizations dynamically in operational management processes. This will improve support for food companies in dealing with perishable products, unpredictable supply variations and stringent food safety and sustainability requirements. Virtualization enables supply chain actors to monitor, control, plan and optimize business processes remotely and in real-time through the Internet, based on virtual objects instead of observation on-site. This paper analyses the concept of virtual food supply chains from an Internet of Things perspective and proposes an architecture to implement enabling information systems. As a proof of concept, the architecture is applied to a case study of a fish supply chain. These developments are expected to establish a basis for virtual supply chain optimization, simulation and decision support based on on-line operational data. In the Internet of Things food supply chains can become self-adaptive systems in which smart objects operate, decide and learn autonomously.

Data Driven Modeling for Monitoring and Control of Industrial Fed-Batch Cultivations

A systematic methodology for development of a set of discrete-time sequence models for batch control based on historical and online operating data is presented and investigated experimentally. The modeling is based on the two independent characteristic time dimensions of batch processing, being time within the batch and the batch number. The model set is parsimoniously parametrized as a set of local, interdependent models which are estimated from data for as few as half a dozen batches. On the basis of state space models transformed from the acquired input–output model set, the asymptotic convergence of iterative learning control is combined with the closed-loop performance of model predictive control to form an optimal controller aiming to ensure reliable and reproducible operation of the batch process. This learning model predictive controller may also be used for optimizing control through optimization of the bioreactor operations model. The modeling and preliminary control performance is demonstrated on an industrial fed-batch protein cultivation production process. The presented methods lend themselves directly for application as Process Analytical Technologies.

A Method of Data Integration Check Based on Online Operation Data of Large-Scale Power System

An important part of power grid voltage stability online intelligent warning system was to research into the current Energy Management System real-time power grid operation parameters and power networks offline data splicing integration, form a set of online dynamic simulation trend data, and implement preventive voltage stability calculation. A splicing model of power grid online data integration and practice was established, and a grid online data splicing method was proposed, which had been verified by actual data. The method of practical data online and offline networks data integration can significantly improve the accuracy of the simulation data.

The Role of the Offline/Online Data Mapping Table Based on Online Operation Data of Large-Scale Power System

An important part of the Xinyang power grid voltage stability online intelligent warning system was to research into the current system of EMS real-time power grid operation parameters and power networks offline data splicing integration, forming a set of online PSASP (Power System Analysis Software Package) trend data, implementing preventive voltage stability calculation. By splicing model of Xinyang grid online data integration and practice, summed up the grid online data splicing method and the model, this article mainly introduced the offline and online data mapping change reason, function is established. It had been verified by actual data achieved by the method can significantly improve the accuracy of the data. It was a kind of practical data online and offline networks data integration method.

Online short-term reliability evaluation using a fast sorting technique

To consider time varying system operating conditions and operating reserve units, the online short-term reliability of a power system is investigated. Real-time system reliability based on the online operation data provided by the SCADA/EMS/WAMS is assessed using time-dependent state probabilities of components. Considering both speed and accuracy requirements, a fast sorting technique (FST) is proposed to quickly select the required number of system states in descending order probability. The number of computations and comparisons used in the evaluation is minimum. The relative accuracy of a given reliability index is defined as the stopping rule for the evaluation. As only a small number of system states are required to achieve the high accuracy of the short-term reliability indices when using the FST, the evaluation can be performed in an online environment. The proposed evaluation technique is illustrated by the application to the IEEE-RTS and China Southern Power Grid.

Development and implementation of neural network observers to estimate the state vector of a synchronous generator from on-line operating data

This paper presents a novel technique for developing and implementing artificial neural network (ANN) observers for estimating unmeasurable rotor body currents of a synchronous generator from time-domain online disturbance data. Data for training the observers are generated through off-line simulations of a 7.5 kVA machine model whose parameters are varied in accordance with previously determined online parameter estimates of the generator under consideration. Studies show that observer robustness towards noise can be improved by enhancing the size of the observer input vector. In order to increase observer robustness towards variations in the field-resistance, simulated variations representative of changes in field-resistance were introduced in the training sets. After training, the observers are tested with experimentally obtained online measurements to provide estimates of unmeasurable rotor body currents. The estimated rotor body currents are then used along with experimental measurements to estimate synchronous generator parameters.

The Road to Advanced Process Control: From DDC to Real-Time Optimization and Beyond

Abstract The progression of computer control in the petroleum refining industry from Direct Digital Control (DDC) in the 1960s to Model Predictive Control and Real-Time Optimization (RTO) in the 1990s is briefly reviewed. The authors’ personal experience is used to illustrate the rapid advances in application of this constantly changing technology. The closed-loop optimization of the catalytic reforming process at the Sunoco refinery in Sarnia, Ontario, Canada will be described in sufficient detail to illustrate the complexity, challenges and practical solutions encountered during the implementation of this unique real-time optimizer. The unit operations are described by steady-state, open equation-based, mechanistic models. Global spline collocation is used to solve the differential equations associated with the reactor models and the catalytic deactivation over time. The entire set of non-linear equations is solved simultaneously using a successive quadratic programming method designed for these large-scale optimization problems. Using the models of the plant which are appropriate at specified times during the time-to-shutdown, the optimization determines that trajectory of operating conditions over the catalyst life which maximizes the total profit over that period. At the same time, the optimization accounts for catalyst deactivation over this period, thus ensuring that the catalyst does not deactivate too fast or leave unused activity. The optimization honours all equipment constraints, product qualities, and production requirements. Model parameters are estimated continuously using on-line data, thus ensuring that the model matches the plant at all times. The optimizer sends the required setpoints for the current operation directly to the model predictive controller that controls the whole plant. The optimizer is activated when steady state operation is achieved, about seven or eight times daily. An in-plant experimental program is established in which the effluent from the four reactors is periodically sampled and analyzed. These analyses along with on-line operating data and the process model are used to estimate the reaction parameters and the catalyst activity in each reactor. After the formulation of the process model, the optimization system, and the actual operating experience of the reformer RTO have been discussed, the authors’ vision for the near future in advanced process control and process optimization will be elucidated, concluding with some challenges for both the academic and technology provider communities.

On-line estimations of settling capacity in secondary clarifier

In a biological nitrogen removal system, efficient performance of the secondary clarifier is essential since the system depends on a high concentration of active biological solids being available in the biological reactor. Estimations of the current capacity of the clarifier can be made from observing maximum feasible solids flux to the clarifier which does not cause escape of sludge. The aim of this work is to demonstrate a method to establish maximum feasible solids flux, the limiting solids flux, from on-line operational data. In this method the solids flux into the clarifier is calculated from on-line measurements of effluent flow and of suspended solids in the biological reactor, and a known value of return sludge flow. From relations between solids flux and sludge blanket height, the value of limiting flux is estimated. This method is based on the assumption that the movement of sludge blanket height is directly related to changes in the solids flux. When the limiting solids flux is estimated from on-line measurements a measurement of the capacity of the clarifier is retrieved. The suggested method also gives an early warning to prevent sludge escape from the clarifier.

On-line estimations of settling capacity in secondary clarifier

In a biological nitrogen removal system, efficient performance of the secondary clarifier is essential since the system depends on a high concentration of active biological solids being available in the biological reactor. Estimations of the current capacity of the clarifier can be made from observing maximum feasible solids flux to the clarifier which does not cause escape of sludge. The aim of this work is to demonstrate a method to establish maximum feasible solids flux, the limiting solids flux, from on-line operational data. In this method the solids flux into the clarifier is calculated from on-line measurements of effluent flow and of suspended solids in the biological reactor, and a known value of return sludge flow. From relations between solids flux and sludge blanket height, the value of limiting flux is estimated. This method is based on the assumption that the movement of sludge blanket height is directly related to changes in the solids flux. When the limiting solids flux is estimated from on-line measurements a measurement of the capacity of the clarifier is retrieved. The suggested method also gives an early warning to prevent sludge escape from the clarifier.

Construction of an on-line fuzzy controller for the dynamic activated sludge process

This paper presents and discusses an alternative approach, called fuzzy control theory, which can forecast and control complex, uncertain and/or highly-non-linear systems properly. Building the set of control rules is a very difficult task in fuzzy control and they are generated through on-line operation data in this study, not through experts' knowledge and/or operators' opinions, which is different from conventional applications. The fuzzy control theory can be conveniently used as an aid for automatic control, especially for complex and highly-non-linear systems. The steps using the fuzzy control technique are issued, compared and discussed in this paper and are demonstrated by a case study of a dynamic activated sludge process which is much more complex than other wastewater treatment systems. The effects of forecast and control of the fuzzy controller established from the derived fuzzy model are shown by computer simulation and verified by experiments. The results indicate that the forecast and control utilities of the fuzzy controllers are good and very helpful to the understanding of complex systems.