Engineering Proceedings

In the case of electric vehicles, due to the favorable torque-speed characteristics of the electric motors used to drive the vehicle, the traditional multi-stage gearboxes common in the automotive industry can be omitted, and in many cases it is sufficient to use a single-stage constant ratio gear unit. Different types of gear units are used for vehicles with different performance levels. In the case of low-performance electric vehicles, a typical solution is to drive only one wheel of the vehicle, e.g., electric motors, bicycles, tricycles, electric racing cars, and mopeds. In this case, a typical solution is to use traction drives. Traction drives can be used as the main drive. This article focuses on the creation of a mathematical model for the energetic modeling of electric vehicles built with chain drives among traction drives. The study examines the possibility of supplementing the electric motor drive of the vehicle with a chain drive, assuming that such a supplement can reduce the energy consumption of the vehicle. With the help of the created model, the chain drive can later be integrated into a complex vehicle model at the design level, and thus the vehicle’s drivetrain can be examined from an energetic perspective in the design phase.

This study investigates how combustion-event sequencing affects the performance of a 600 cm³ four-stroke spark-ignition Formula Student engine. A validated 0D/1D model was applied at full load from 3000 to 13,000 rpm to compare the conventional 1-3-4-2 firing order with two uneven 1-3-2-4 sequences. Power, torque, BMEP, peak cylinder pressure, pressure phasing, and delivery ratio were evaluated. The baseline model reproduced dynamometer data with average errors of 3% for power and 4% for torque. The uneven sequences altered cylinder charging and increased cylinder-to-cylinder interaction through the common intake and exhaust systems. Neither uneven strategy improved overall high-speed performance, although limited mid-range benefits were observed. The results also show that a higher delivery ratio does not necessarily yield higher torque when charge distribution, residual-gas effects and pumping behaviour become less favourable. Firing-event spacing is therefore identified as an important gas-dynamic design parameter, although differences comparable with baseline model error should be interpreted cautiously.

Accurate traffic forecasting for telecommunication networks enables proactive operations and capacity planning. We present an event-aware approach for network traffic prediction that explicitly incorporates public holidays and atypical external events. Leveraging large-scale time series data spanning multiple years at five-minute granularity, we engineer calendar, external event, and temporal features to capture traffic pattern variations. We compare multilayer perceptron, long short-term memory (LSTM), and Bidirectional LSTM (BiLSTM) architectures under two feature configurations (V1 and V2), where V2 incorporates richer event-aware features. The training results show that V2 significantly improves recurrent model performance; LSTM-V2 achieves 43.8 and 78.5% mean absolute error reduction for uplink and downlink, compared to LSTM-V1. Prediction visualizations further demonstrate that LSTM-V2 better captures event-driven traffic deviations compared to LSTM-V1. The approach achieves production-ready accuracy and provides practical guidance for integrating event-aware forecasting into operational network management pipelines, enabling telecommunication operators to optimize resource allocation and reduce infrastructure costs.