Short-Term Load Forecasting Based on the Transformer Model

From the perspective of energy providers, accurate short-term load forecasting plays a significant role in the energy generation plan, efficient energy distribution process and electricity price strategy optimisation. However, it is hard to achieve a satisfactory result because the historical data is irregular, non-smooth, non-linear and noisy. To handle these challenges, in this work, we introduce a novel model based on the Transformer network to provide an accurate day-ahead load forecasting service. Our model contains a similar day selection approach involving the LightGBM and k-means algorithms. Compared to the traditional RNN-based model, our proposed model can avoid falling into the local minimum and outperforming the global search. To evaluate the performance of our proposed model, we set up a series of simulation experiments based on the energy consumption data in Australia. The performance of our model has an average MAPE (mean absolute percentage error) of 1.13, where RNN is 4.18, and LSTM is 1.93.

Coordination of similar days and multi-objective posterior process to predict energy of transmission substation

Purpose This paper aims to present a hybrid model to mid-term forecast the load of transmission substations based on the knowledge of expert site and multi-objective posterior framework. The main important challenges in load forecasting are the different behavior of load in specific days. Regular days, holidays and special holidays, days after a holidays and days of load shifting are characterized by abnormal load profiles. The knowledge of these days is verified by expert operators in regional dispatching centers. Design/methodology/approach In this paper, a hybrid model for power prediction of transmission substations based on the combination of similar day selection and multi-objective posterior technique has been proposed. In the first step, the important data for prediction is provided. Posterior method is used in the second step for prediction that it is based on kernel functions. A multi-objective optimization has been formulated with three type of output accuracy measurement function that it is solved by non-dominated sorting genetic technique II (NSGT-II) method. TOPSIS way is used to find the best point of Pareto. Findings The presented method has been tested in four scenarios for three different transmission stations, and the test results have been compared. The presented results indicate that the presentation method has better results and is robust to different load characteristics, which can be used for better forecasting of different stations for better planning of repairs and network operation. Originality/value The main contributions of this paper can be categorized as follows: A hybrid model based on similar days selection and multi-objective framework posterior is presented. Similar day selection is done by expert site that the day type and days with scheduled repair are considered. Hyperparameters of posterior process are found by NSGT-II based on TOPSIS method.

TRANSESOPHAGEAL ECHOCARDIOGRAPHIC ASSESSMENT OF LEFT ATRIAL THROMBUS IN PATIENTS OF RHEUMATIC SEVERE MITRAL STENOSIS WITH NEGATIVE RESULT ON TRANSTHORACIC ECHOCARDIOGRAPHY

Objective: The objective of the study was to assess the frequency of clot in left atrium by transesophageal echocardiographic diagnostic procedure (TEE) in patients with rheumatic severe mitral stenosis (MS) in those patients in whom transthoracic echocardiographic diagnostic procedure (TTE) did not observe any clot in left thrombus. Methodology: The descriptive cross-sectional study was carried out on 369 patients with rheumatic severe MS. Patients having mitral valve are of 1.5 cm2 or less planned for PMV were included in study. Non-cooperative patients were excluded. It was decided to do TEE on a similar day before the intervention to assess LA thrombus. Results: The mean age of the included patients was 45.65 ±11.54 years. There were 122 (33.1%) male patients and 247 (66.9%) female patients. In this study, clot in LA was diagnosed in 29.5% patients having negative TTE. LA clot was found in 33 male patients (30.3%) and 76 female patients (69.7%). Conclusion: There is a high risk of missing LA thrombus on transthoracic echocardiography, therefore we recommend performing TEE in patients with normal TTE study before the procedure.

Research on energy consumption prediction of office buildings based on comprehensive similar day and ensemble learning

Energy consumption prediction can provide reliable data support for energy scheduling and optimization of office buildings. It is difficult for traditional prediction model to achieve stable accuracy and robustness when energy consumption mode is complex and data sources are diverse. Based on such situation, this paper raised an approach containing the method of comprehensive similar day and ensemble learning. Firstly, the historical data was analyzed and calculated to obtain the similarity degree of meteorological features, time factor and precursor. Next, the entropy weight method was used to calculate comprehensive similar day and applied to the model training. Then the improved sine cosine optimization algorithm (SCA) was applied to the optimization and parameter selection of a single model. Finally, an approach of model selection and integration based on dominance was proposed, which was compared with Support Vector Regression (SVR), Back Propagation Neural Network (BPNN), Long Short-Term Memory (LSTM), with a large office building in Xi ‘an taken as an example to analysis showing that compared with the prediction accuracy, root mean square percentage error (RMSPE) in the ensemble learning model after using comprehensive similar day was reduced by about 0.15 compared with the BP model, and was reduced by about 0.05, 0.06 compared with the SVR and LSTM model. Respectively, the mean absolute percentage error (MAPE) was reduced by 12.02%, 6.51% and 5.28%. Compared with several other integration methods, integration model based on dominance reduced absolute error at all times. Accordingly, the proposed approach can effectively solve problems of low accuracy and poor robustness in traditional model and predict the building energy consumption efficaciously.