Search

Understanding energy demands and costs is important for policy makers and the energy sector, especially in the context of residential heating and cooling systems. To estimate the thermal demand of a residential house, a grey-box modelling method with a resistance–capacitance (RC) analogy was implemented. The architectural properties used to parameterize the grey-box model were derived from a house used for research purposes in Vaughan, Ontario, Canada (TRCA-House A). The house model accounts for solar irradiance on exterior building surfaces, thermal conductivity through all surfaces, solar heat gains through windows, and thermal gains from ventilation. Two parallel short- and long-term calibrations were performed such that model outputs reflected the real-world operation of the house as best as possible. To define the unknown model parameters (such as the conductivity of building materials and some constant parameters), a hybrid optimization scheme including a genetic algorithm (GA) and the Quasi-Newton algorithm was introduced and implemented using Bayesian approximation and Markov Chain Monte Carlo (MCMC) methods. The temperature outputs from the model were compared to the data retrieved from TRCA-House A. The final iteration of the model had an RMSE for interior zone temperature estimation of 0.22 °C when compared to the retrieved interior zone temperature data from TRCA-House A. Furthermore, the annual heating and cooling energy consumption values are within 1.50% and 0.08% of target values, respectively. According to these preliminary results, the introduced model and optimization techniques could be adjusted for different types of housing, as well as for smart control applications on both a short- and long-term basis. [ABSTRACT FROM AUTHOR]

Laser cladding multi-channel forming is a complex, nonlinear process characterized by multi-parameter coupling. To determine the technology parameters, an optimization method of multivariate black-box prediction model (MVBM) matching non-dominated sequential genetic algorithm (NSGA-II) was constructed. Multi-channel LC orthogonal experiments were designed. The MVBM was established with process technology parameters as inputs and characteristic parameters reflecting the quality (surface hardness, dilution rate) as responses. The set of optimal solutions for technology parameters by the NSGA-II. The experimental results showed that, under the optimum parameters, the average hardness of the optimized increased by 6.1%, and the dilution rate was reduced by 49.8%. The dilution rate was maintained within the optimal range. The study's findings can provide theoretical support for optimizing LC multi-objective technology parameters and improving coating quality control. [ABSTRACT FROM AUTHOR]

Abstract Slug liquid holdup (SLH) is a critical requirement for accurate pressure drop prediction during multiphase pipe flows and by extension optimal gas lift design and production optimization in wellbores. Existing empirical correlations provide inaccurate predictions because they were developed with regression analysis and data measured for limited ranges of flow conditions. Existing SLH machine learning models provide accurate predictions but are published without any equations making their use by other researchers difficult. The only existing ML model published with actual equations cannot be considered optimum because it was selected by considering artificial neural network (ANN) structures with only one hidden layer. In this study, an ANN-based model for SLH prediction with actual equations is presented. A total of 2699 data points randomly divided into 70%, 15%, and 15% for training, validation, and testing was used in constructing 71 different network structures with 1, 2, and 3 hidden layers respectively. Sensitivity analysis revealed that the optimum network structure has 3 hidden layers with 20, 5, and 15 neurons in the first, second, and third hidden layers, respectively. The optimum network structure was translated into actual equations with the aid of the weights, biases, and activation functions. Trend analysis revealed that this study’s model reproduced the expected effects of inputs on SLH. Test against measured data revealed that this study’s model is in agreement with measured data with coefficient of determinations of 0.9791, 0.9727, 0.9756, and 0.9776 for training, testing, validation, and entire datasets, respectively. Comparative study revealed that this study’s model outperformed existing models with a relative performance factor of 0.002. The present model is presented with visible mathematical equations making its implementation by any user easy and without the need for any ML framework. Unlike existing ANN-based models developed with one hidden layered ANN structures, the present model was developed by considering ANN structures with one, two, and three hidden layers, respectively.

Slug liquid holdup (SLH) is a critical requirement for accurate pressure drop prediction during multiphase pipe flows and by extension optimal gas lift design and production optimization in wellbores. Existing empirical correlations provide inaccurate predictions because they were developed with regression analysis and data measured for limited ranges of flow conditions. Existing SLH machine learning models provide accurate predictions but are published without any equations making their use by other researchers difficult. The only existing ML model published with actual equations cannot be considered optimum because it was selected by considering artificial neural network (ANN) structures with only one hidden layer. In this study, an ANN-based model for SLH prediction with actual equations is presented. A total of 2699 data points randomly divided into 70%, 15%, and 15% for training, validation, and testing was used in constructing 71 different network structures with 1, 2, and 3 hidden layers respectively. Sensitivity analysis revealed that the optimum network structure has 3 hidden layers with 20, 5, and 15 neurons in the first, second, and third hidden layers, respectively. The optimum network structure was translated into actual equations with the aid of the weights, biases, and activation functions. Trend analysis revealed that this study's model reproduced the expected effects of inputs on SLH. Test against measured data revealed that this study's model is in agreement with measured data with coefficient of determinations of 0.9791, 0.9727, 0.9756, and 0.9776 for training, testing, validation, and entire datasets, respectively. Comparative study revealed that this study's model outperformed existing models with a relative performance factor of 0.002. The present model is presented with visible mathematical equations making its implementation by any user easy and without the need for any ML framework. Unlike existing ANN-based models developed with one hidden layered ANN structures, the present model was developed by considering ANN structures with one, two, and three hidden layers, respectively. [ABSTRACT FROM AUTHOR]

Maneuvering motion prediction is central to the control and operation of ships, and the application of machine learning algorithms in this field is increasingly prevalent. However, challenges such as extensive training time, complex parameter tuning processes, and heavy reliance on mathematical models pose substantial obstacles to their application. To address these challenges, this paper proposes an LSTM-based modeling algorithm. First, a maneuvering motion model based on a real USV model was constructed, and typical operating conditions were simulated to obtain data. The Ornstein–Uhlenbeck process and the Hidden Markov Model were applied to the simulation data to generate noise and random data loss, respectively, thereby constructing a sample set that reflects real experiment characteristics. The sample data were then pre-processed for training, employing the MaxAbsScaler strategy for data normalization, Kalman filtering and RRF for data smoothing and noise reduction, and Lagrange interpolation for data resampling to enhance the robustness of the training data. Subsequently, based on the USV maneuvering motion model, an LSTM-based black-box motion prediction model was established. An in-depth comparative analysis and discussion of the model's network structure and parameters were conducted, followed by the training of the ship maneuvering motion model using the optimized LSTM model. Generalization tests were then performed on a generalization set under Zigzag and turning conditions to validate the accuracy and generalization performance of the prediction model. [ABSTRACT FROM AUTHOR]

The growing need for in-depth analysis of predictive models leads to a series of new methods for explaining their local and global properties. Which of these methods is the best? It turns out that this is an ill-posed question. One cannot sufficiently explain a black-box machine learning model using a single method that gives only one perspective. Isolated explanations are prone to misunderstanding, leading to wrong or simplistic reasoning. This problem is known as the Rashomon effect and refers to diverse, even contradictory, interpretations of the same phenomenon. Surprisingly, most methods developed for explainable and responsible machine learning focus on a single-aspect of the model behavior. In contrast, we showcase the problem of explainability as an interactive and sequential analysis of a model. This paper proposes how different Explanatory Model Analysis (EMA) methods complement each other and discusses why it is essential to juxtapose them. The introduced process of Interactive EMA (IEMA) derives from the algorithmic side of explainable machine learning and aims to embrace ideas developed in cognitive sciences. We formalize the grammar of IEMA to describe human-model interaction. It is implemented in a widely used human-centered open-source software framework that adopts interactivity, customizability and automation as its main traits. We conduct a user study to evaluate the usefulness of IEMA, which indicates that an interactive sequential analysis of a model may increase the accuracy and confidence of human decision making. [ABSTRACT FROM AUTHOR]

This study examines the relationship between Leaf Area Index and soil moisture in the Beheshtabad watershed through field sampling, MODIS imagery, and black-box modeling based on various factors. For this purpose, climate data such as rainfall, evaporation, transpiration, number of rainy days, and temperature were collected from 2003 to 2015 in this watershed. Additionally, to determine the physical characteristics of the area and prepare maps of soil moisture, data on soil texture, land use, topography, geology, Digital Elevation Model, and drainage network were gathered. During the field visits in 2016 and 2017, data on soil moisture, Leaf Area Index, and vegetation characteristics were collected for the land use in the area. The findings indicate that vegetation cover requires time to respond to changes in soil moisture, with a developmental delay of approximately four months observed in the study area (coefficient of determination = 0.69). Land use, slope, and soil texture separation factors have differing impacts on the relationship between Leaf Area Index and soil moisture, which is nonlinear. This study highlights the importance of understanding the reciprocal effects between environmental factors and vegetation cover for water and soil resource management. [ABSTRACT FROM AUTHOR]

Let Ω be a finite set of finitary operation symbols and let be a nontrivial variety of Ω -algebras. Assume that for some set Γ ⊆ Ω of group operation symbols, all Ω -algebras in are groups under the operations associated with the symbols in  Γ. In other words, is assumed to be a nontrivial variety of expanded groups. In particular, can be a nontrivial variety of groups or rings. Our main result is that there are no post-quantum weakly pseudo-free families in , even in the worst-case setting and/or the black-box model. In this paper, we restrict ourselves to families (H d | d ∈ D) of computational and black-box Ω -algebras (where D ⊆ { 0 , 1 } * ) such that for every d ∈ D , each element of H d is represented by a unique bit string of length polynomial in the length of d. In our main result, we use straight-line programs to represent nontrivial relations between elements of Ω -algebras. Note that under certain conditions, this result depends on the classification of finite simple groups. Also, we define and study some types of post-quantum weak pseudo-freeness for families of computational and black-box Ω -algebras. [ABSTRACT FROM AUTHOR]

Due to the nature of black-box machine learning (ML) models used in the spatial modelling field of environmental and natural hazards, the interpretation of predictive model outputs is necessary. For this purpose, we applied four interpretation techniques consisting of interaction plot, permutation feature importance (PFI) measure, shapley additive explanation (SHAP) decision plot, and accumulated local effects (ALE) plot to explain and interpret the output of an ML model applied to map land subsidence (LS) in the Nazdasht plain, Hormozgan province, southern Iran. We applied a stepwise regression (SR) algorithm and five ML models (Cforest (as a conditional random forest), generalized linear model (GLM), multivariate linear regression (MLR), partial least squares (PLS) and extreme gradient boosting (XGBoost)) to select important features and to map the LS hazard, respectively. Thereafter, several interpretation techniques were used to explain the spatial ML hazard model output. Our findings revealed that a GLM model was the most accurate approach to map LS in our study area, and that 24.3% of the total study area had a very high susceptibility to the LS hazard. According to the interpretation techniques, land use, elevation, groundwater level and vegetation were the most important variables controlling the LS hazard and also the most important variables contributing to the model’s output. Overall, human activities, especially the diversion of the route of one of the main tributaries feeding the plain and the recharging of groundwater five decades ago, intensified the current LS occurrence. Therefore, management activities such as water spreading projects upstream of the plain can be useful to mitigate LS occurrence in the plain.

Understanding energy demands and costs is important for policy makers and the energy sector, especially in the context of residential heating and cooling systems. To estimate the thermal demand of a residential house, a grey-box modelling method with a resistance–capacitance (RC) analogy was implemented. The architectural properties used to parameterize the grey-box model were derived from a house used for research purposes in Vaughan, Ontario, Canada (TRCA-House A). The house model accounts for solar irradiance on exterior building surfaces, thermal conductivity through all surfaces, solar heat gains through windows, and thermal gains from ventilation. Two parallel short- and long-term calibrations were performed such that model outputs reflected the real-world operation of the house as best as possible. To define the unknown model parameters (such as the conductivity of building materials and some constant parameters), a hybrid optimization scheme including a genetic algorithm (GA) and the Quasi-Newton algorithm was introduced and implemented using Bayesian approximation and Markov Chain Monte Carlo (MCMC) methods. The temperature outputs from the model were compared to the data retrieved from TRCA-House A. The final iteration of the model had an RMSE for interior zone temperature estimation of 0.22 °C when compared to the retrieved interior zone temperature data from TRCA-House A. Furthermore, the annual heating and cooling energy consumption values are within 1.50% and 0.08% of target values, respectively. According to these preliminary results, the introduced model and optimization techniques could be adjusted for different types of housing, as well as for smart control applications on both a short- and long-term basis.

The supply airflow rate is crucial for monitoring, controlling, and detecting faults in rooftop air conditioner units (RTUs). However, the cost and intrusiveness of a supply airflow rate sensor (SARS) make it difficult to deploy in the field. Virtual SARSs have been proposed, but they often require testing or experimentation to train the model, which is not easily scalable. To overcome this limitation, the present study proposed deriving supply airflow using publicly available and scalable original equipment manufacturer (OEM) data of RTU blowers. Two models, the gray-box, and the black-box, were proposed using the OEM data and applied to data from four different manufacturers. Despite limited OEM data, the gray-box model showed an accuracy of ±5%, while the black-box model provided high overall accuracy for the full range of data but yielded low accuracy (up to 27% error) at a lower blower rotation speed. The models were also validated through laboratory testing, with an accuracy of ± 10% for the motor speed range of 50%–100% of the rated speed. Monitoring and controlling the airflow rate in rooftop air conditioner units (RTUs) is essential, but traditional sensors for this purpose are costly and intrusive, making them challenging to use in the real world. To address this issue, researchers have proposed virtual sensors that estimate airflow without physical sensors, but these often require complex training processes that are not easily scalable. In this study, a novel approach is introduced. It leverages readily available data from RTU manufacturers (OEM data) to estimate airflow. Two models, known as the gray-box and the black-box models, are developed using this OEM data and tested on data from four different RTU manufacturers. The gray-box model, despite limited OEM data, achieves impressive accuracy within ±5%. The black-box model performs well overall but struggles with lower blower rotation speeds, resulting in up to a 27% error. To validate the models, laboratory tests were conducted, confirming an accuracy of ±10% for motor speeds ranging from 50% to 100% of the rated speed. This research offers a promising and cost-effective solution for accurately estimating supply airflow rates in RTUs, making it easier to monitor and control these systems efficiently. [ABSTRACT FROM AUTHOR]

Maneuvering motion prediction is central to the control and operation of ships, and the application of machine learning algorithms in this field is increasingly prevalent. However, challenges such as extensive training time, complex parameter tuning processes, and heavy reliance on mathematical models pose substantial obstacles to their application. To address these challenges, this paper proposes an LSTM-based modeling algorithm. First, a maneuvering motion model based on a real USV model was constructed, and typical operating conditions were simulated to obtain data. The Ornstein–Uhlenbeck process and the Hidden Markov Model were applied to the simulation data to generate noise and random data loss, respectively, thereby constructing a sample set that reflects real experiment characteristics. The sample data were then pre-processed for training, employing the MaxAbsScaler strategy for data normalization, Kalman filtering and RRF for data smoothing and noise reduction, and Lagrange interpolation for data resampling to enhance the robustness of the training data. Subsequently, based on the USV maneuvering motion model, an LSTM-based black-box motion prediction model was established. An in-depth comparative analysis and discussion of the model’s network structure and parameters were conducted, followed by the training of the ship maneuvering motion model using the optimized LSTM model. Generalization tests were then performed on a generalization set under Zigzag and turning conditions to validate the accuracy and generalization performance of the prediction model.

This study investigates the Shapley additive explanation (SHAP) of the extreme boosting (XGBoost) model for breast cancer diagnosis. The study employed Wisconsin’s breast cancer dataset, characterized by 30 features extracted from an image of a breast cell. SHAP module generated different explainer values representing the impact of a breast cancer feature on breast cancer diagnosis. The experiment computed SHAP values of 569 samples of the breast cancer dataset. The SHAP explanation indicates perimeter and concave points have the highest impact on breast cancer diagnosis. SHAP explains the XGB model diagnosis outcome showing the features affecting the XGBoost model. The developed XGB model achieves an accuracy of 98.42% [ABSTRACT FROM AUTHOR]

Many applications involving physical systems, such as system control or fault detection, call for a behavioral, black-box, or digital twin of the real system. By observing input-output pairs, a nonlinear system’s black-box twinning model can be built, thus enabling real-time accurate estimation of the system’s health and status. We propose a modeling approach that can be implemented with little hardware resources and predicts system output with acceptable accuracy for a wide range of applications in the IoT and Industry 4.0 application domains, such as cloud and distributed predictive control, maintenance, fault detection, and model drift avoidance. This approach consists of building a compact numerical model, based on the concept of sum-decomposability, with reduced computational complexity and memory requirements, well suited for microcontroller-based IoT applications. The black-box modeling theory, the sizing process, and the learning method are reported. The outputs of two examples of non-linear systems are replicated in real-time using a pioneer experimental setup built around a microcontroller. According to experimental results, online learning and prediction are performed at 1 kS/s with a prediction error comparable to the resolution of the digitalized input-output data. The reduced size of the obtained model calls for real-time sharing and update with cloud and edge-based simulation ecosystems enabling a near real-time digital twinning of field systems.

Considering that current voltage transformer models of electrical multiple units (EMUs) are narrow-band models or transformer models, this paper introduces a wide-band model of EMU voltage transformers based on the vector fitting method, circuit synthesis theory and black-box model theory. The admittances of voltage transformers from 30 kHz to 5 MHz are measured by the vector network analyzer, the branch admittances in the pi-type equivalent circuit are calculated according to the equation of a two-port network equivalent circuit. Based on the vector matching method, the rational function formulas of branch admittances are obtained, and the formulas are converted into the circuit models by circuit synthesis theory. The pi-type equivalent circuit model is constructed in the simulation software, and so is the voltage transformer model in the range of 30 kHz–5 MHz. The frequency sweeping method is used to measure the transmission characteristics from direct current (DC)to 30 kHz. The pi-type model is modified according to transmission characteristics, whereby the wide-band model in DC-5 MHz is obtained. Fast pulse experiments are carried out on the voltage transformer, and the actual injected fast pulse voltage is used as the excitation source in the simulation model. The measurement and simulation results on the secondary side of the voltage transformer show that the wide-band model has a high accuracy. [ABSTRACT FROM AUTHOR]

With the development of artificial intelligence technology, machine learning models are becoming more complex and accurate. However, the explainability of the models is decreasing, and much of the decision process is still unclear and difficult to explain to users. Therefore, we now often use Explainable Artificial Intelligence (XAI) techniques to make models transparent and explainable. For an image, the ability to recognize its content is one of the major contributions of XAI techniques to image recognition. Visual methods for describing classification decisions within an image are usually expressed in terms of salience to indicate the importance of each pixel. In some approaches, explainability is achieved by deforming and integrating white-box models, which limits the use of specific network architectures. Therefore, in contrast to white-box model-based approaches that use weights or other internal network states to estimate pixel saliency, we propose the Optimized Input Sampling Explanation (OISE) technique based on black-box models. OISE uses masks to generate saliency maps that reflect the importance of each pixel to the model predictions, and employs black-box models to empirically infer the importance of each pixel. We evaluate our method using deleted/inserted pixels, and extensive experiments on several basic datasets show that OISE achieves better visual performance and fairness in explaining the decision process compared to the performance of other methods. This approach makes the decision process clearly visible, makes the model transparent and explainable, and serves to explain it to users. [ABSTRACT FROM AUTHOR]

An accurate fuel consumption prediction model is the basis for ship navigation status analysis, energy conservation, and emission reduction. In this study, we develop a black-box model based on machine learning and a white-box model based on mathematical methods to predict ship fuel consumption rates. We also apply the Kwon formula as a data preprocessing cleaning method for the black-box model that can eliminate the data generated during the acceleration and deceleration process. The ship model test data and the regression methods are employed to evaluate the accuracy of the models. Furthermore, we use the predicted correlation between fuel consumption rates and speed under simulated conditions for model performance validation. We also discuss applying the data-cleaning method in the preprocessing of the black-box model. The results demonstrate that this method is feasible and can support the performance of the fuel consumption model in a broad and dense distribution of noise data in data collected from real ships. We improved the error to 4% of the white-box model and the R 2 to 0.9977 and 0.9922 of the XGBoost and RF models, respectively. After applying the Kwon cleaning method, the value of R 2 also can reach 0.9954, which can provide decision support for the operation of shipping companies. [ABSTRACT FROM AUTHOR]

To improve the motor efficiency and expand the actual external characteristic region of electric vehicle permanent magnet synchronous motor (PMSM) drive systems, the optimal operation of mapping torque to d-q axis current is usually applied. Nevertheless, it is difficult to deal with the complex mechanism factors such as parameter saturation and temperature change for the traditional optimization method based on the basic voltage equation of PMSM. In this paper, a black-box-model-based torque–current optimization method is proposed, which does not rely on any information of the inner mechanism model, and the derivative-free, optimal, improved Nelder–Mead Simplex(NMS) method is used to minimize the copper loss and maximize the electromagnetic torque in the flux-weakening region. Moreover, a synchronous online compensation of the electromagnetic torque and optimal current angle is implemented, in view of the time variation of permanent magnet flux with temperature. Finally, through a comparison experiment with the nominal-parameters-based formula maximum torque per ampere (MTPA) method, the proposed method achieves higher torque accuracy and better efficiency performance in a wide temperature range with regard to a reasonable response speed. [ABSTRACT FROM AUTHOR]

Several studies have addressed the importance of context and users' knowledge and experience in quantifying the usability and effectiveness of the explanations generated by explainable artificial intelligence (XAI) systems. However, to the best of our knowledge, no component-agnostic system that accounts for this need has yet been built. This paper describes an approach called ConvXAI, which can create a dialogical multimodal interface for any black-box explainer by considering the knowledge and experience of the user. First, we formally extend the state-of-the-art conversational explanation framework by introducing clarification dialogue as an additional dialogue type. We then implement our approach as an off-the-shelf Python tool. To evaluate our framework, we performed a user study including 45 participants divided into three groups based on their level of technology use and job function. Experimental results show that (i) different groups perceive explanations differently; (ii) all groups prefer textual explanations over graphical ones; and (iii) ConvXAI provides clarifications that enhance the usefulness of the original explanations. [ABSTRACT FROM AUTHOR]

[Display omitted] • First planning model for rapid formation of aerobic granular sludge (AGS) built. • Accurate prediction of AGS particle size was achieved by gray box modeling. • Six machine learning algorithms were applied to improve predictive model robustness. • The AGS cultivation process was quantified and optimized with intelligent algorithms. • The formation time of model-regulated AGS was shortened from 62 days to 4 days. Aerobic Granular Sludge (AGS) has advantages over Activated sludge (AS) but faces challenges with long granulation periods. In this study, a novel grey-box model is devised to optimize the cultivation of AGS to shorten the formation time. This model is based on an existing white-box model. The modeling process starts with the application of four sensitivity analysis methods to assess the 12 model metrics selected. Subsequently, 12 prediction models were constructed by combining the six Machine learning (ML) algorithms and integrated algorithms, with the best performance selected (R2 = 0.98). Finally, an AGS selection pressure planning model was designed in conjunction with a simulated annealing (SA) algorithm to guide AGS training. The results demonstrate that AGS formation could be achieved within four days under the model's optimal control. Therefore, the establishment of this model provides a new technique for the cultivation of AGS. [ABSTRACT FROM AUTHOR]

Minimum Kolmogorov and Wasserstein distance estimates, θ ^ MKD and θ ^ MWD , respectively, of model parameter, θ (∈ Θ) , are empirically compared, obtained assuming the model is intractable. For the Cauchy and Lognormal models, simulations indicate both estimates have expected values nearly θ , but θ ^ MKD has in all repetitions of the experiments smaller SD than θ ^ MWD , and θ ^ MKD 's relative efficiency with respect to θ ^ MWD improves as the sample size, n, increases. The minimum expected Kolmogorov distance estimate, θ ^ MEKD , has eventually bias and SD both smaller than the corresponding Wasserstein estimate, θ ^ MEWD , and θ ^ MEKD 's relative efficiency improves as n increases. These results hold also for stable models with stability index α =. 5 and α = 1.1. For the Uniform and the Normal models the estimates have similar performance. The disturbing empirical findings for θ ^ MWD are due to the unboudedness and non-robustness of the Wasserstein distance and the heavy tails of the underlying univariate models.Theoretical confirmation is provided for stable models with 1 < α < 2 , which have finite first moment. Similar results are expected to hold for multivariate heavy tail models. Combined with existing results in the literature, the findings do not support the use of Wasserstein distance in statistical inference, especially for intractable and Black Box models with unverifiable heavy tails. [ABSTRACT FROM AUTHOR]

Interpretability of DNN-based object detection has been a rising concern for the research community. The first step towards this goal is a saliency map that visualizes the importance (saliency) of pixels in an image for the object detected by a specific model. Black-box based methods generate a saliency map without the need to look into the internals of a model, thus applicable to all models without the need of adaptation. In addition, they provide more reliable evaluation on the saliency of pixels than white-box methods by means of the absence of these pixels from the image. However, with current black-box methods, the absence of pixels is produced by random image masks. Despite the need of a great number of random masks for sufficient coverage, the quality of the pixel saliency is not assured to be satisfactory. In this work, we propose a more effective black-box framework with hierarchical masking. In this framework, called GSM-HM, pixel saliency is evaluated at multiple levels, with each lower level performing a refinement on the saliency information of the upper level. This hierarchical framework significantly reduces the masking efforts on less valuable pixels, thus it can produce saliency maps with higher qualities. In our experiments, the quality of a generated saliency map is evaluated with four different metrics: deletion, insertion, convergence and RAM (the ratio of average to maximum). Compared with D-RISE, a recent black-block method, GSM-HM generates more accurate saliency maps evaluated by these metrics.

It is important to work out precise and easy-to-use mathematical models to predict the indoor temperature in buildings for human residence. Such models can support model-based/predictive controls to efficiently maintain the temperature at a comfortable level.Accordingly, as main contribution, the paper proposes a new, easy-to-use, black-box type, multiple linear regression (MLR) based model for the indoor temperature of mobile (office) containers. The model, having low computational demand, could be easily generalized for different types of residence places (in the future).A discretized, physically-based model version of the classical, widely used heat transfer theory (based on energy balance) is recalled for comparison with the MLR-based model.Both models have the same (exogenous) inputs: global solar irradiance, environment temperature and wind speed.Both models are validated based on measured data. The MLR-based model is more precise, its modelling error is 7.1%, which means that it can be used well for general engineering aims.Moreover, the model is detailed in time (it gives an output value per half minute), so it could be properly used for real-time prediction and control purposes.Another contribution of the paper is that the MLR-based model is used to estimate the application potential of solar collectors, installed on the top of the container, for space heating. Based on the results, two solar collectors could extend the time with comfortable indoor temperature by more than 5 h within a three-day period in spring (in Hungary).Finally, conclusions and possible topics for future researches are provided.

Buildings use up to 40% of the global primary energy and 30% of global greenhouse gas emissions, which may significantly impact climate change. Heating, ventilation, and air-conditioning (HVAC) systems are among the most significant contributors to global primary energy consumption and carbon gas emissions. Furthermore, HVAC energy demand is expected to rise in the future. Therefore, advancements in HVAC systems' performance and design would be critical for mitigating worldwide energy and environmental concerns. To make such advancements, energy modeling and model predictive control (MPC) play an imperative role in designing and operating HVAC systems effectively. Building energy simulations and analysis techniques effectively implement HVAC control schemes in the building system design and operation phases, and thus provide quantitative insights into the behaviors of the HVAC energy flow for architects and engineers. Extensive research and advanced HVAC modeling/control techniques have emerged to provide better solutions in response to the issues. This study reviews building energy modeling techniques and state-of-the-art updates of MPC in HVAC applications based on the most recent research articles (e.g., from MDPI's and Elsevier's databases). For the review process, the investigation of relevant keywords and context-based collected data is first carried out to overview their frequency and distribution comprehensively. Then, this review study narrows the topic selection and search scopes to focus on relevant research papers and extract relevant information and outcomes. Finally, a systematic review approach is adopted based on the collected review and research papers to overview the advancements in building system modeling and MPC technologies. This study reveals that advanced building energy modeling is crucial in implementing the MPC-based control and operation design to reduce building energy consumption and cost. This paper presents the details of major modeling techniques, including white-box, grey-box, and black-box modeling approaches. This paper also provides future insights into the advanced HVAC control and operation design for researchers in relevant research and practical fields. [ABSTRACT FROM AUTHOR]

High voltage direct current (HVDC) systems have been increasingly employed to provide a connection amongst various energy sources, especially renewable energies. In order to switch the load current in these systems, the passive resonance breaker (PRB) is employed as a transfer switch. The interruption capability of the PRBs is highly dependent on the dynamic behavior of the arc in this system. Accordingly, this article presents a black-box high-degree of freedom arc-model based on Schwarz and the genetic algorithm as a heuristic optimization to follow the characteristics of the static and quasi-static arc regarding the oscillation frequency of the interruption current. The model has been verified by the published experiments. Subsequently, the PRB operation has been quantified based on the state-space approach along with an accurate dynamic arc model to follow in a wide frequency range of the arc current. The results are indicated that the amplification coefficient alone is insufficient to determine the interruption capability of these protection tools. Therefore, this article attempts to quantitatively determine the interruption capability curve by introducing criteria, such as ΔtPZ (peak to zero time-interval), diarc/dtCZ- and duarc/dtCZ+ to obtain the origins of the interruption failure probabilities and dimensioning of the PRBs with respect to the interruption limits for gas blast CBs. [ABSTRACT FROM AUTHOR]

With the development of artificial intelligence technology, machine learning models are becoming more complex and accurate. However, the explainability of the models is decreasing, and much of the decision process is still unclear and difficult to explain to users. Therefore, we now often use Explainable Artificial Intelligence (XAI) techniques to make models transparent and explainable. For an image, the ability to recognize its content is one of the major contributions of XAI techniques to image recognition. Visual methods for describing classification decisions within an image are usually expressed in terms of salience to indicate the importance of each pixel. In some approaches, explainability is achieved by deforming and integrating white-box models, which limits the use of specific network architectures. Therefore, in contrast to white-box model-based approaches that use weights or other internal network states to estimate pixel saliency, we propose the Optimized Input Sampling Explanation (OISE) technique based on black-box models. OISE uses masks to generate saliency maps that reflect the importance of each pixel to the model predictions, and employs black-box models to empirically infer the importance of each pixel. We evaluate our method using deleted/inserted pixels, and extensive experiments on several basic datasets show that OISE achieves better visual performance and fairness in explaining the decision process compared to the performance of other methods. This approach makes the decision process clearly visible, makes the model transparent and explainable, and serves to explain it to users.

In the materials science domain, the accurate prediction of the yield strength of metallic compositions has often resulted in extensive experimental endeavors, leading to inefficiencies in both time and resources. Here, we introduce an innovative approach to predict yield strength, which can be applied to a variety of metallic substances ranging from the simplest pure metals to the most intricate alloys under varying temperatures and strain rates. The fusion of grain boundary sliding mechanism and cutting-edge machine-learning algorithm forges an expansive framework, which can help realize the critical factors influencing yield strength. The validity and wide applicability of the proposed framework were rigorously confirmed through experimental evaluations conducted on selected Fe-based alloys, such as Fe 60 Ni 25 Cr 15 , Fe 60 Ni 30 Cr 10 , and Fe 64 Ni 15 Co 8 Mn 8 Cu 5. This breakthrough study significantly streamlines experimental design processes, optimizes resource utilization, and marks a significant leap forward in creating a reliable predictive framework for realizing material properties. [Display omitted] [ABSTRACT FROM AUTHOR]

Surrogate models provide a powerful method for simplifying calculations within complex simulations. While surrogate models are broadly applied within chemical engineering, little research exists investigating the level of surrogacy's impact on a simplified process model. In this work, artificial neural networks (ANN) and Kriging models are used as surrogate models at the process, process unit, and thermodynamic levels for a CO2 amine scrubbing process. The surrogated models are evaluated against an Aspen Plus simulation for accuracy, convergence behavior, computational cost, and ability to extrapolate. The thermodynamic and process unit models can better handle discontinuous, non‐smooth behavior, and convergence issues in the surrogated truth model, but poor conditioning in the final system of equations results in a lower accuracy and convergence rate than the process level surrogate. Beyond model accuracy, availability of diverse data, intended re‐usability, and the desired outputs must be considered when selecting a level of abstraction. [ABSTRACT FROM AUTHOR]

Unsupervised domain adaptation (UDA) is an emerging technique that enables the transfer of domain knowledge learned from a labeled source domain to unlabeled target domains, providing a way of coping with the difficulty of labeling in new domains. The majority of prior work has relied on both source and target domain data for adaptation. However, because of privacy concerns about potential leaks in sensitive information contained in patient data, it is often challenging to share the data and labels in the source domain and trained model parameters in cross-center collaborations. To address this issue, we propose a practical framework for UDA with a black-box segmentation model trained in the source domain only, without relying on source data or a white-box source model in which the network parameters are accessible. In particular, we propose a knowledge distillation scheme to gradually learn target-specific representations. Additionally, we regularize the confidence of the labels in the target domain via unsupervised entropy minimization, leading to performance gain over UDA without entropy minimization. We extensively validated our framework on a few datasets and deep learning backbones, demonstrating the potential for our framework to be applied in challenging yet realistic clinical settings. [ABSTRACT FROM AUTHOR]

Unsupervised domain adaptation (UDA) is an emerging technique that enables the transfer of domain knowledge learned from a labeled source domain to unlabeled target domains, providing a way of coping with the difficulty of labeling in new domains. The majority of prior work has relied on both source and target domain data for adaptation. However, because of privacy concerns about potential leaks in sensitive information contained in patient data, it is often challenging to share the data and labels in the source domain and trained model parameters in cross-center collaborations. To address this issue, we propose a practical framework for UDA with a black-box segmentation model trained in the source domain only, without relying on source data or a white-box source model in which the network parameters are accessible. In particular, we propose a knowledge distillation scheme to gradually learn target-specific representations. Additionally, we regularize the confidence of the labels in the target domain via unsupervised entropy minimization, leading to performance gain over UDA without entropy minimization. We extensively validated our framework on a few datasets and deep learning backbones, demonstrating the potential for our framework to be applied in challenging yet realistic clinical settings.

An accurate fuel consumption prediction model is the basis for ship navigation status analysis, energy conservation, and emission reduction. In this study, we develop a black-box model based on machine learning and a white-box model based on mathematical methods to predict ship fuel consumption rates. We also apply the Kwon formula as a data preprocessing cleaning method for the black-box model that can eliminate the data generated during the acceleration and deceleration process. The ship model test data and the regression methods are employed to evaluate the accuracy of the models. Furthermore, we use the predicted correlation between fuel consumption rates and speed under simulated conditions for model performance validation. We also discuss applying the data-cleaning method in the preprocessing of the black-box model. The results demonstrate that this method is feasible and can support the performance of the fuel consumption model in a broad and dense distribution of noise data in data collected from real ships. We improved the error to 4% of the white-box model and the R2 to 0.9977 and 0.9922 of the XGBoost and RF models, respectively. After applying the Kwon cleaning method, the value of R2 also can reach 0.9954, which can provide decision support for the operation of shipping companies.

To improve the motor efficiency and expand the actual external characteristic region of electric vehicle permanent magnet synchronous motor (PMSM) drive systems, the optimal operation of mapping torque to d-q axis current is usually applied. Nevertheless, it is difficult to deal with the complex mechanism factors such as parameter saturation and temperature change for the traditional optimization method based on the basic voltage equation of PMSM. In this paper, a black-box-model-based torque–current optimization method is proposed, which does not rely on any information of the inner mechanism model, and the derivative-free, optimal, improved Nelder–Mead Simplex(NMS) method is used to minimize the copper loss and maximize the electromagnetic torque in the flux-weakening region. Moreover, a synchronous online compensation of the electromagnetic torque and optimal current angle is implemented, in view of the time variation of permanent magnet flux with temperature. Finally, through a comparison experiment with the nominal-parameters-based formula maximum torque per ampere (MTPA) method, the proposed method achieves higher torque accuracy and better efficiency performance in a wide temperature range with regard to a reasonable response speed.

Energy flexible buildings in combination with demand response will play a key role in the future smart grid. To implement control strategies, which enable demand response, like model predictive control, thermal building models are necessary. Therefore, three lumped capacitance models, are compared with a k-Nearest Neighbor regression model. All models show accurate prediction results, if the operating condition of the building is similar during parameter identification or rather during training and the validation period. Parameter identification of lumped capacitance models is a time-consuming task. Especially for complex lumped capacitance models, the search space for certain parameters has to be reduced to avoid local minima. The investigated k-Nearest Neighbor algorithm has the advantage of easy implementation, very fast training and minimal effort for parameter identification in combination with accurate predictions. But its seasonal dependency is very strong, which can be easily overcome with periodically data update, as it is an instance-based learning algorithm.

Buildings use up to 40% of the global primary energy and 30% of global greenhouse gas emissions, which may significantly impact climate change. Heating, ventilation, and air-conditioning (HVAC) systems are among the most significant contributors to global primary energy consumption and carbon gas emissions. Furthermore, HVAC energy demand is expected to rise in the future. Therefore, advancements in HVAC systems’ performance and design would be critical for mitigating worldwide energy and environmental concerns. To make such advancements, energy modeling and model predictive control (MPC) play an imperative role in designing and operating HVAC systems effectively. Building energy simulations and analysis techniques effectively implement HVAC control schemes in the building system design and operation phases, and thus provide quantitative insights into the behaviors of the HVAC energy flow for architects and engineers. Extensive research and advanced HVAC modeling/control techniques have emerged to provide better solutions in response to the issues. This study reviews building energy modeling techniques and state-of-the-art updates of MPC in HVAC applications based on the most recent research articles (e.g., from MDPI’s and Elsevier’s databases). For the review process, the investigation of relevant keywords and context-based collected data is first carried out to overview their frequency and distribution comprehensively. Then, this review study narrows the topic selection and search scopes to focus on relevant research papers and extract relevant information and outcomes. Finally, a systematic review approach is adopted based on the collected review and research papers to overview the advancements in building system modeling and MPC technologies. This study reveals that advanced building energy modeling is crucial in implementing the MPC-based control and operation design to reduce building energy consumption and cost. This paper presents the details of major modeling techniques, including white-box, grey-box, and black-box modeling approaches. This paper also provides future insights into the advanced HVAC control and operation design for researchers in relevant research and practical fields.

The existing methods of flexibility index are mainly based on mixed‐integer linear or nonlinear programming methods, making it difficult to readily deal with complex mathematical models. In this article, a novel solution strategy is proposed for finding a reliable upper bound of the flexibility index where the process model is implemented in a black box that can be directly executed by a commercial simulator, and also avoiding the need for calculating derivatives. Then, the flexibility index problem is formulated as a sequence of univariate derivative‐free optimization (DFO) models. An external DFO solver based on trust‐region methods can be called to solve this model. Finally, after calculating the critical point of the model parameters, the vertex enumeration method and two gradient approximation methods are proposed to evaluate the impact of process parameters and to evaluate the flexibility index. A reaction model is studied to show the efficiency of the proposed algorithm. [ABSTRACT FROM AUTHOR]

A black-box model for arc-based current interruption, originally developed by the same authors for single-phase short-circuit tests, is here challenged with multipolar circuit breakers in three-phase interruptions. Multipolar devices are built by composition of elementary, single pole components. The three phases adopt three replicas of the same single-phase model, its parameters having been identified in single pole tests. Splitter-less neutral poles adopt a simplified model, deduced by specialization of the original model after suitably dropping splitter plate features. Various breakers and test cases are considered. Comparisons with experimental results in real short-circuit tests show very good agreement, in line with motivating industrial, system-level applications. Indeed, model parameters are obviously single pole design dependent, but they are invariant to different operating conditions and no fine tuning is required to match empirical evidence. [ABSTRACT FROM AUTHOR]

This paper presents the mathematical modeling of metal hydride based hydrogen storage bed. The novelty of this research work is the identification of transfer function for the dynamic discharge process of hydrogen storage bed is carried out using system identification tool box. A Fuzzy PID controller is modeled to regulate the pressure rate of hydrogen storage bed using MATLAB Simulink toolbox. The proposed controller reduces the complexity in controlling the pressure rate of hydrogen storage bed during desorption process for an adequate use in fuel cell and in other applications. To validate the reliability of the hydrogen storage bed model, the obtained results are verified with the results of other literature. The exact system transfer function model chosen is Box-Jenkin model, based on the simulated results of different black box model structures for the parameters such as percentage of fitness, final prediction error and the cost function. Usually, for the safer discharge of hydrogen gas, the classical PID controller is employed to control the pressure by manipulating the heating temperature of the storage bed. In order to have a better control, a Fuzzy-PID controller is proposed in this study. The proposed controller produces improved time domain response and better performance compared to the conventional PID controller. Image 1 • A mathematical model for Metal Hydride based hydrogen storage bed is developed. • Dynamic discharge process of hydrogen is modeled by Box-Jenkin method. • PID and Fuzzy-PID controllers are designed to control the bed pressure rate. • The Fuzzy-PID controller is found to be better in terms of dynamic response. [ABSTRACT FROM AUTHOR]