An enhanced bilayer long short-term memory method for energy consumption estimation of electric buses with real-time passenger load

Transit bus passenger loading changes significantly over the course of a workday. Therefore, time-varying vehicle mass as a result of passenger load becomes an important factor in instantaneous energy consumption. Battery-powered electric transit buses have restricted range and longer “fueling” time compared with conventional diesel-powered buses; thus, it is critical to know how much energy they require. Our previous work has shown that instantaneous transit bus mass can be obtained by measuring the pressure in the vehicle’s airbag suspension system. This paper leverages this novel technique to determine the impact of time-varying mass on energy consumption. Sixty-five days of velocity and mass data were collected from in-use transit buses operating on routes in the Twin Cities, MN metropolitan area. The simulation tool Future Automotive Systems Technology Simulator was modified to allow both velocity and mass as time-dependent inputs. This tool was then used to model an electrified and conventional bus on the same routes and determine the energy use of each bus. Results showed that the kinetic intensity varied from 0.27 to 4.69 mi−1 and passenger loading ranged from 2 to 21 passengers. Simulation results showed that energy consumption for both buses increased with increasing vehicle mass. The simulation also indicated that passenger loading has a greater impact on energy consumption for conventional buses than for electric buses owing to the electric bus’s ability to recapture energy. This work shows that measuring and analyzing real-time passenger loading is advantageous for determining the energy used by electric and conventional diesel buses.

Integrating an electronic farebox with a location system can provide location-stamped records of passenger boardings, a valuable source of information on passenger travel patterns. However, this information is of small value unless the pattern of passenger alightings can also be determined, since most relevant measures of interest—passenger loads, passenger miles, and origin–destination (O-D) patterns—require a knowledge or at least estimate of passenger alightings by stop. The assumption of symmetry—that the pattern of passenger alightings in one direction mirrors the daily boardings pattern in the opposite direction—is explored. Estimation methods using this assumption are tested at the trip, route, and system levels using a full-day’s set of on–off counts on five Los Angeles area routes. Tests at the route level indicate that although perfect symmetry does not exist, patterns are substantially similar on many routes. Based on the Los Angeles data, it can be found that systemwide estimates of passenger miles made using this method satisfy U.S. National Transit Database precision requirements; however, this finding should be confirmed using data from other cities. Proposed and tested, with a small amount of success, is a method for estimating trip-level O-D patterns using location-stamped farebox data based on the symmetry principle and a gravity model. Location-stamped farebox data can also be used to estimate passenger loads in real time to support control measures such as conditional priority at traffic signals without requiring automatic passenger counters.

A smart grid is a new type of power system based on modern information technology, which utilises advanced communication, computing and control technologies and employs advanced sensors, measurement, communication and control devices that can monitor the status and operation of various devices in the power system in real-time and optimise the dispatch of the power system through intelligent algorithms to achieve efficient operation of the power system. However, due to its complexity and uncertainty, how to effectively perform real-time prediction is an important challenge. This paper proposes a smart grid real-time prediction model based on the attention mechanism of convolutional neural network (CNN) combined with bi-directional long and short-term memory BiLSTM.The model has stronger spatiotemporal feature extraction capability, more accurate prediction capability and better adaptability than ARMA and decision trees. The traditional prediction models ARMA and decision tree can often only use simple statistical methods for prediction, which cannot meet the requirements of high accuracy and efficiency of real-time load prediction, so the CNN-BiLSTM model based on Bayesian optimisation has the following advantages and is more suitable for smart grid real-time load prediction compared with ARMA and decision tree. CNN is a hierarchical neural network structure containing several layers such as a convolutional layer, pooling layer and fully connected layer. The convolutional layer is mainly used for extracting features from data such as images, the pooling layer is used for the dimensionality reduction of features, and the fully connected layer is used for classification and recognition. The core of CNN is the convolutional operation, a locally weighted summation operation on the input data that can effectively extract features from the data. In the convolution operation, different features can be extracted by setting different convolution kernels to achieve feature extraction and classification of data. BiLSTM can capture semantic dependencies in both directions. The BiLSTM structure consists of two LSTM layers that process the input sequence in the forward and backward directions to combine the information in both directions to obtain more comprehensive contextual information. BiLSTM can access both the front and back inputs at each time step to obtain more accurate prediction results. It effectively prevents gradient explosion and gradient disappearance while better capturing longer-distance dependencies. The CNN-BiLSTM extracts features of the data and then optimises them by Bayes. By collecting real-time data from the power system, including power, load, weather and other factors, our model uses the features of CNN-BiLSTM to deeply learn real-time load data from smart grids and extract key features to achieve future load prediction. Meanwhile, the Bayesian optimisation algorithm based on the model can optimise the model’s hyperparameters, thus improving the model’s prediction performance. The model can achieve accurate prediction of a real-time power system load, provide an important reference for the dispatch and operation of the power system, and help optimise the operation efficiency and energy utilisation efficiency of the power system.

A cloud-based energy management strategy for hybrid electric city bus considering real-time passenger load prediction

The article summarizes modern approaches to estimating energy consumption by households. In particular, international experience in determining the amount of final energy consumption by households by purpose using data from special household surveys and modeling methods. Possibilities of adaptation and modification of the specified approaches to an estimation of households’ energy consumption are defined. The article aims to highlight the results of practical application in Ukraine of the developed methodology of modeling and appropriate tools for estimation of the energy consumption by households. The study used methods of statistical indicators estimation based on the special household sample survey microdata, developed statistical models, in particular on dummy variables, graphical and tabular methods of data analysis, etc. The peculiarities of the use of different types of energy by different categories of households are assessed: it depends on the geographical location, area of residence, type of housing, and socio-economic characteristics of households. The method of estimation of energy consumption by Ukrainian households is offered to take into account features of the accessible information base, sources of power supply, and financial possibilities of households.

The current state-of-the-art technologies available at the disposal of the aerospace industry lacks the ability to measure the aerodynamic forces and moments acting on an aircraft in real-time during it's flight. Since the entire flight of an aircraft is based on the balance and controlled manipulation of these forces and moments, the appropriate real-time estimation for these parameters is of utmost interest. The work presented herein addresses the issues associated with the real-time aerodynamic load estimation problem through the use of a distributed Flush Air Data System (FADS) sensor network and the development of appropriate estimation methods. This work showcases a method to design the sensor network to capture the critical aerodynamic information in the aircraft pressure signature. It also elaborates upon a neural-network based estimation method to extract the aerodynamic load information from the pressure information captured by the sensor network. This research also focuses on the use of the real-time aerodynamic load estimations on building new aircraft applications for aircraft safety and control. This work shows that the incipient stall conditions can be detected using the real-time aerodynamic load information. The idea and implementation of a prognostic control is also presented in this work. It is shown here that the prognostic control based on the real-time estimates of aerodynamic forces and moments can anticipate the change in aircraft states and therefore employ appropriate control action before a traditional controller.

Estimation of the hydrodynamic loads based on strip theory using the Morrison equation provides an inexpensive method for load estimation for the offshore industry. The advantage of this approach is that it requires only the undisturbed wave kinematics along with inertia and viscous force coefficients. Over the recent years, the development in numerical wave tank simulations makes it possible to simulate nonlinear 3-h sea states, with computational times in the order of real time. This presents the possibility to calculate loads using wave spectrum input in numerical simulations with reasonable computational time and effort. In the current paper, the open-source fully nonlinear potential flow model REEF3D::FNPF is employed for calculating the nonlinear wave kinematics. Here, the Laplace equation for the velocity potential is solved on a σ-coordinate mesh with the nonlinear free surface boundary conditions to close the system. A technique to calculate the total acceleration on the σ-coordinate grid is introduced which makes it possible to apply strip theory in a moving grid framework. With the combination of strip theory and 3-h wave simulations, a unique possibility to estimate the hydrodynamic loads in real time for all discrete positions in space within the domain of the numerical wave tank is presented in this paper. The numerical results for inline forces on an offshore wind mono-pile substructure are compared with measurements, and the new approach shows good agreement.

Estimation of the hydrodynamic loads based on strip theory with the Morrison equation provides a fast and inexpensive method for load estimation for the offshore industry. The advantage of this approach is that it requires only the undisturbed wave kinematics along with inertia and viscous force coefficients. Over the recent years, the development in numerical wave tank simulations makes it possible to simulate nonlinear three-hour sea states, with computational times in the order of real time. This provides an opportunity to calculate loads using wave spectrum input in numerical simulations at reasonable computational time and effort. In the current paper, the open-source fully nonlinear potential flow model REEF3D::FNPF is employed for the wave propagation calculations. Here, the Laplace equation for the velocity potential is solved on a sigma-coordinate mesh with the nonlinear free surface boundary conditions to close the system. A technique to calculate the total acceleration on the sigma-coordinate grid is introduced which makes it possible to apply strip theory in a moving grid framework. With the combination of strip theory and three-hour wave simulations, a unique possibility to estimate the hydrodynamic loads in real time for all discrete positions in space within the domain of the numerical wave tank is presented in this paper. The numerical results for inline forces on an offshore wind mono-pile substructure are compared with measurements, and the new approach shows good agreement.

Energy consumption of buildings is one of the major sources of greenhouse gas emissions, frequently accounting for around 40 % of the total energy consumption in urban areas. Reasonable sustainability assessment requires effective and efficient methods for energy consumption estimation of buildings. With an aim to improve the accuracy in building energy consumption estimation, this paper proposes a life cycle building energy consumption estimation method based on the work breakdown structure. For each basic work package, the relevant construction business information can support a more accurate estimation of energy consumption and thus the integration of all basic work packages can enhance the accuracy in energy consumption estimation.

Revisiting the building energy consumption in China: Insights from a large-scale national survey

Four indirect estimations of energy expenditure were examined, (i) O(2) debt, (ii) O(2) deficit, (iii) blood lactate concentration, and (iv) excess CO(2) production during and after 6 exercise durations (2, 4, 10, 15, 30, and 75 s) performed at 3 different intensities (50%, 100%, and 200% of VO(2) max). Analysis of variance (ANOVA) was used to determine if significant differences existed among these 4 estimations of anaerobic energy expenditure and among 4 estimations of total energy expenditure (that included exercise O(2) uptake and excess post-exercise oxygen consumption, or EPOC, measurements). The data indicate that estimations of anaerobic energy expenditure often differed for brief (2, 4, and 10 s) bouts of exercise, but this did not extend to total energy expenditure. At the higher exercise intensities with the longest durations O(2) deficit, blood lactate concentration, and excess CO(2) estimates of anaerobic and total energy expenditure revealed high variability; however, they were not statistically different. Moreover, they all differed significantly from the O(2) debt interpretation (p < 0.05). It is concluded that as the contribution of rapid substrate-level ATP turnover with lactate production becomes larger, the greatest error in quantifying total energy expenditure is suggested to occur not with the method of estimation, but with the omission of a reasonable estimate of anaerobic energy expenditure.

Electric city buses have become a key solution for reducing carbon emissions and improving air quality in urban areas, with over 670,000 electric buses in operation globally as of 2023 over 95% of which are deployed in China. Public transport systems account for around 25% of total urban energy consumption, and improving the energy efficiency of electric buses can reduce fuel costs by up to 30%. However, traditional methods for estimating energy consumption often depend on historical averages, static models, and driverspecific behavior, making them time-consuming, error-prone, and poorly suited for adapting to dynamic traffic and route conditions. To overcome these limitations, this study presents a machine learning-based framework to predict the energy economy (measured in kWh/km) of electric city buses. The proposed approach leverages real-world operational data, including speed profiles, passenger loads, elevation variations, and ambient temperatures, collected via a smart fleet monitoring system across various urban routes and thousands of recorded trips. The data undergoes preprocessing steps such as imputation, normalization, outlier removal, and feature engineering to enhance model performance. While benchmark models like Linear Regression and XGBoost Regressor are used for comparison, they exhibit shortcomings in modeling complex nonlinear relationships. In contrast, the proposed CatBoost Regressor model demonstrates superior performance due to its efficient handling of categorical data, robustness in high-dimensional feature spaces, and lower dependency on manual tuning. The model is trained using an 80-20 train-test split and delivers high predictive accuracy. Test results confirm that the CatBoost model achieves lower Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), making it a valuable tool for energy consumption forecasting and intelligent scheduling in electric bus operations

Along with the battery technology advancements and government policy support, the penetration level of electric buses (EBs) in the urban public transportation system has been increasing in recent years. Considering the potential influence of the increasing EB charging demand on power systems, estimating the real-time energy consumption of EBs has become a principal issue. In this work, a data-driven approach for EB energy consumption estimation is proposed. In particular, a detailed physical model of EB is constructed to model its energy consumption considering the randomness in EB operation, including speed, acceleration, and passenger count. In order to improve the estimation accuracy, the conventional Kalman filter (KF) is modified involving EB mass estimation considering stochastic real-time passenger count, motion data dimension deduction based on EB operation route. To estimate the EB acceleration accurately and reduce the noise caused by the unimportant features, we extended the feature discarding algorithm of decision trees to the regression trees. In the case study, an Android application is developed to collect the EB motion data so that any general Android smartphone can be used for data collection. The performance of the proposed approach is evaluated based on real-world EB operation data collected from St. Albert Transit, AB, Canada. According to the results, our APP can track the real-time EB trace, and the proposed modified KF can filter most of the noises caused by the GPS data collection process and stochastic passenger count. Also, with the extended random forest algorithm, the unimportant features can be discarded and the real-time EB acceleration is estimated efficiently with a small sum of square error (SSE). Compared with the existing approaches, the proposed approach achieves a more accurate real-time energy consumption estimation of EBs, which in turn, provides a better characterization of power system loading and voltage variation.

Problem statement: China’s energy consumption is increasing with a high-speed in recent years. Especially since building energy consumption caught pubic eyes and became a crucial problem of society, it forced the public to make estimation in order to reduce energy consumption efficiently. However, it is very difficult to analyze a non-residential building accurately due to China’s statistical collection system and the lack of national surveys. Approach: This study introduced a methodology of estimating various energy consumption factors by building types, energy end-use (electric power, space heating, space cooling and hot water) in each province. The unit energy consumption factors were determined based on sample cities’ data and modification by using software analysis. Take 2006 year for example, the estimation method was introduced. Results: The non-residential building energy consumption in China in 2006 year was estimated by the method above mentioned. Through the result of analysis, we found out that energy consumption of space heating, space cooling and hot water were greatly affected by space Heating Degree Day (HDD), space Cooling Degree Day (CDD) and regional consumer spending per person (Op-c). Conclusion: A series of formulas were obtained. So by using the formulas we can not only estimate the energy consumption now, but also the energy consumption in future. However, this is the first step of our research. It might be hoped that the further surveys and research on energy consumption of China can be done to promote our research result.

Several methods to estimate energy expenditure (EE) using body-worn sensors exist; however, quantifications of the differences in estimation error are missing. In this paper, we compare three prevalent EE estimation methods and five body locations to provide a basis for selecting among methods, sensors number, and positioning. We considered 1) counts-based estimation methods, 2) activity-specific estimation methods using METs lookup, and 3) activity-specific estimation methods using accelerometer features. The latter two estimation methods utilize subsequent activity classification and EE estimation steps. Furthermore, we analyzed accelerometer sensors number and on-body positioning to derive optimal EE estimation results during various daily activities. To evaluate our approach, we implemented a study with 15 participants that wore five accelerometer sensors while performing a wide range of sedentary, household, lifestyle, and gym activities at different intensities. Indirect calorimetry was used in parallel to obtain EE reference data. Results show that activity-specific estimation methods using accelerometer features can outperform counts-based methods by 88% and activity-specific methods using METs lookup for active clusters by 23%. No differences were found between activity-specific methods using METs lookup and using accelerometer features for sedentary clusters. For activity-specific estimation methods using accelerometer features, differences in EE estimation error between the best combinations of each number of sensors (1 to 5), analyzed with repeated measures ANOVA, were not significant. Thus, we conclude that choosing the best performing single sensor does not reduce EE estimation accuracy compared to a five sensors system and can reliably be used. However, EE estimation errors can increase up to 80% if a nonoptimal sensor location is chosen.