Research on Short-Term Prediction Method of Liquefied Gas Concentration based on Mixed Intelligence

In order to improve the accuracy of ultra short-term power prediction of the photovoltaic power generation system, a short-term photovoltaic power prediction method based on an adaptive k-means and Gru machine learning model is proposed. This method first introduces the construction process of the model and then builds a short-term photovoltaic power generation prediction model based on an adaptive k-means and Gru machine learning models. Then, the network structure and key parameters are determined through experiments, and the initial training set of the prediction model is selected according to the short-term photovoltaic power generation characteristics. And the adaptive k-means is used to cluster the initial training set and the photovoltaic power on the forecast day. The Gru model is trained on the initial training set data of each category, and the generated power is predicted in combination with the trained Gru model. Finally, considering three typical weather types, the proposed method is used for simulation analysis and compared with the other three traditional photovoltaic power generation single prediction models. The comparison results show that the proposed short-term photovoltaic power generation prediction method based on an adaptive k-means and Gru network has better effect, better robustness, and less error.

Unplanned events present significant challenges for operations and management in metro systems. Short-term ridership prediction can help agencies to better design contingency strategies under unplanned events. Though many short-term prediction methods have been proposed in the literature, most studies focused on typical situations or planned events. The study develops methods for the short-term metro ridership prediction under unplanned events. It explores event impact representation mechanisms and deals with the imbalanced data training problem in building the prediction model under unplanned events. Typical machine learning and deep learning methods are developed for exploration. A large-scale automatic fare collection (AFC) dataset and event record data for a heavily used metro system are used for empirical studies. The analysis found that the same type of unplanned event shares a similar and consistent demand change pattern (with respect to the demand under typical situations) at the station level. The synthetic minority oversampling technique (SMOTE) can enrich the ridership observations under unplanned events and generate a balanced dataset for model training. Given the occurrence of unplanned events, the results show that a combination of demand change ratio and the SMOTE oversampling technique enables the prediction models to learn the impact of unplanned events and improve the prediction accuracy under unplanned events. However, the oversampling methods (i.e., SMOTE and replication) slightly deteriorate the prediction accuracy for ridership under normal conditions. The findings provide insights into mechanisms for disruption impact representation and oversampling imbalanced data in model training, and guide the development of models for short-term prediction under unplanned events.

Guangzhou’s energy consumption accounts for 24.9% of the energy consumption of Guangdong-Hong Kong-Macao Greater Bay Area (GBA). The improvement of Guangzhou’s energy management system has a significant impact and exemplary role on the energy management of GBA. Therefore, Guangzhou needs annual, quarterly and monthly energy consumption prediction for the requirements of energy consumption control and establishment of refined energy management system. Meanwhile, due to the lack of historical data, it also needs a short-term prediction method of energy consumption with short cycle and small sample. The paper provided variety of city level prediction methods and all districts energy consumption prediction method to predict the energy consumption of Guangzhou in 2018.Then this paper verify the prediction methods, and apply them to the prediction of energy consumption in 2019 and 2020. The results show that in the case of short-term prediction and small data samples, we can use a variety of factors to predict the energy consumption by refining the data collection area. And the result of all districts energy consumption prediction method is closer to the actual value than that of other methods. However, there are some difficulties in data collection of districts. Thus, in order to establish an accurate short-term prediction method, it is also necessary to strengthen the combination of city and district level energy monitoring and management systems, and expands the channels for collection and summary of district level energy data. Finally, in the future, the prediction method can be applied to GBA after the establishment of unified energy information statistics system of each city in GBA.

The short-term prediction of single well production can provide direct data support for timely guiding the optimization and adjustment of oil well production parameters and studying and judging oil well production conditions. In view of the coupling effect of complex factors on the daily output of a single well, a short-term prediction method based on a multi-agent hybrid model is proposed, and a short-term prediction process of single well output is constructed. First, CEEMDAN method is used to decompose and reconstruct the original data set, and the sliding window method is used to compose the data set with the obtained components. Features of components by decomposition are described as feature vectors based on values of fuzzy entropy and autocorrelation coefficient, through which those components are divided into two groups using cluster algorithm for prediction with two sub models. Optimized online sequential extreme learning machine and the deep learning model based on encoder-decoder structure using self-attention are developed as sub models to predict the grouped data, and the final predicted production comes from the sum of prediction values by sub models. The validity of this method for short-term production prediction of single well daily oil production is verified. The statistical value of data deviation and statistical test methods are introduced as the basis for comparative evaluation, and comparative models are used as the reference model to evaluate the prediction effect of the above multi-agent hybrid model. Results indicated that the proposed hybrid model has performed better with MAE value of 0.0935, 0.0694 and 0.0593 in three cases, respectively. By comparison, the short-term prediction method of single well production based on multi-agent hybrid model has considerably improved the statistical value of prediction deviation of selected oil well data in different periods. Through statistical test, the multi-agent hybrid model is superior to the comparative models. Therefore, the short-term prediction method of single well production based on a multi-agent hybrid model can effectively optimize oilfield production parameters and study and judge oil well production conditions.

A short-term power prediction method for wind farm cluster based on the fusion of multi-source spatiotemporal feature information

In order to improve the accuracy of urban rail transit passenger flow prediction, a machine learning based method for short-term urban rail transit passenger flow prediction, which efficiently fuses the date-related characteristics, and the autocorrelation of time series changes of passenger flow is proposed. By analyzing the historical passenger flow of urban rail transit stations, the correlation of passenger flow time series and the date-related characteristics of multiple scales composed of date stage, week and work-and-rest day are obtained. With time nodes such as legal holidays, summer vacation and winter vacation, the date within a year can be divided into several time stages, and the stage coefficient is introduced to quantify it to get the stage characteristic. The time series of station historical passenger flow is combined with multiscale date-related characteristics to establish the back propagation neural network prediction model, which realizes the short-term prediction of the inbound passenger flow of urban rail transit stations. The proposed method is validated and compared with the long-short term memory neural network (LSTMNN) model. The experimental result shows that the proposed method is more stable and accurate than the LSTMNN in general, and it has good applicability for different stations.

With the widespread adoption of distributed energy systems, accurately predicting photovoltaic (PV) power generation is essential for optimizing residential energy scheduling. However, existing methods face challenges in balancing prediction accuracy and cost-effectiveness. This study proposes a low-cost short-term PV power prediction method and applies it to residential energy systems (RES) to improve their scheduling performance. The prediction model uses a 3D convolutional neural network (3DCNN) to extract features from sequences of sky images, which are then integrated with historical power data through a DLinear-based multi-step prediction model. The proposed prediction model is further embedded into a multi-agent reinforcement learning (MARL) framework to assess the impact of short-term prediction on renewable energy scheduling. A case study is conducted on RES located in Changping District, Beijing, China. A physical model is developed, incorporating the proposed prediction model into the scheduling strategy to enable precise energy flow control and dynamic supply-demand balancing. Validation results demonstrate that the sequence model can achieve 60-minute-ahead prediction. In system-level simulations, when short-term prediction information is included as part of the state input for MARL, the long-term results show that the proposed approach reduces the operational cost of the RES and optimizes energy flow.

A short-term wind power prediction method is proposed in this paper with experimental results obtained from a wind farm located in Northeast China. In order to improve the accuracy of the prediction method using a traditional back-propagation (BP) neural network algorithm, the improved grey wolf optimization (IGWO) algorithm has been adopted to optimize its parameters. The performance of the proposed method has been evaluated by experiments. First, the features of the wind farm are described to show the fundamental information of the experiments. A single turbine with rated power of 1500 kW and power generation coefficient of 2.74 in the wind farm was introduced to show the technical details of the turbines. Original wind power data of the whole farm were preprocessed by using the quartile method to remove the abnormal data points. Then, the retained wind power data were predicted and analysed by using the proposed IGWO–BP algorithm. Analysis of the results proves the practicability and efficiency of the prediction model. Results show that the average accuracy of prediction is ~11% greater than the traditional BP method. In this way, the proposed wind power prediction method can be adopted to improve the accuracy of prediction and to ensure the effective utilization of wind energy. A short-term wind power prediction method is designed and tested with experimental results obtained from a wind farm located in Northeast China. In order to improve the accuracy of the prediction method, the improved grey wolf optimization algorithm has been adopted to optimize its parameters.

With the development of social economy, intelligent transportation system (ITS) has been booming. It is mainly to realize the omni-directional, real-time, accurate and efficient guidance and control of transportation in a large range. Traffic flow prediction, especially short-term traffic flow prediction, is the basis of urban traffic control and guidance, so traffic flow prediction system is one of the important subsystems in its. Short-term traffic flow prediction technology belongs to the important research field of intelligent traffic control and vehicle guidance. It can help cities to conduct intelligent traffic guidance and enable users to choose the optimal path. Prediction of road traffic flow state and auxiliary guidance are very important basic theories. Based on the theory of artificial intelligence, this paper analyzes the short-term traffic flow prediction and auxiliary induction methods, points out the limitations of the existing prediction methods, and proposes a short-term traffic flow intelligent prediction method combined with artificial intelligence technology.

In order to grow efficiency in Intelligent Transportation System (ITS), a number of dynamic route guidance scheme has been designed to assist driver in determining the optimal route for their journeys. To determine an optimal route, utilizing real-time traffic information is a key factor in improving traffic efficiency. Not only being able to utilize the real-time traffic information, but prediction of the traffic proves as one of the important aspects in improving traffic efficiency. In this paper, we will discuss various methods related to traffic flow prediction in ITS that will eventually lead to a proposed method. We will review some papers related to traffic flow prediction methods. Traffic flow prediction method is divided into two general types: Short-Term Prediction method and Long-Term Prediction method. Short-Term Prediction method relies on real-time information; however, this method can be redundant for a daily recurring traffic condition. On the other end, Long-Term Prediction relies on time series collective data from traffic condition routine, but this method is vulnerable to atypical traffic conditions like accident or road work. In the end, the author will propose a method to create a new sufficient method. This proposed method is a traffic prediction method that combines dynamic real-time information prediction (short-term prediction) and time series analysis prediction (long-term prediction).

In this paper the authors are to be commended for the significant contribution in the area of wind power prediction. The methodology presented allows the power production calculation through a two-step procedure: in the first step, the wind direction and velocity are predicted some minutes ahead, and then, in the second step, the power output is calculated using power curves direction-dependent. The authors comments on a method for short-term wind power prediction with multiple observation points are appreciated.

Model identification of ship turning maneuver and extreme short-term trajectory prediction under the influence of sea currents

As one of the hot topics in the field of new energy, short-term wind power prediction research should pay attention to the impact of meteorological characteristics on wind power while improving the prediction accuracy. Therefore, a short-term wind power prediction method based on the combination of meteorological features and CatBoost is presented. Firstly, morgan-stone algebras and sure independence screening(MS-SIS) method is designed to filter the meteorological features, and the influence of the meteorological features on the wind power is explored. Then, a sort enhancement algorithm is designed to increase the accuracy and calculation efficiency of the method and reduce the prediction risk of a single element. Finally, a prediction method based on CatBoost network is constructed to further realize short-term wind power prediction. The National Renewable Energy Laboratory (NREL) dataset is used for experimental analysis. The results show that the short-term wind power prediction method based on the combination of meteorological features and CatBoost not only improve the prediction accuracy of short-term wind power, but also have higher calculation efficiency.

A combined method for short-term traffic flow prediction based on recurrent neural network

Kushida and Kushida found that FM radio waves from stations at distances over-the-horizon are received before earthquakes. Based on this finding, since the mid-1990’s, the Kushidas have been practicing “Earthquake Precursor Detection Experiment”. The performance of the Kushida method during 2000–2003 has been evaluated by checking their predictions against the actual seismicity. During the period, there were 92 Kushida predictions mentioning the possibility of M ≥ 5.5 event, whereas there were 49 M ≥ 5.5 earthquakes in the Japanese region. If the criteria for successful prediction are set as: the errors in date is less than one day, epicentral position is roughly within specified area, and error in M is less than 0.5, the success rate was 20% and the alarm rate was 12%. If we relax the criteria to: the errors in dates within 10 days, epicenter within additional 100 km of specified area and the magnitude error less than 1.0, the success rate was 40% and the alarm rate was 27%. These rates may look insufficient for a practical prediction method. Considering, however, the fact that no other short-term prediction has ever been made in Japan so far it is a significant achievement. Moreover, it was found that in almost all failed predictions, meaningful signals were detected although the interpretations were incorrect. This indicates that the method is promising provided further investigation is carried out. The same evaluation at the M ≥ 6.0 level showed that the general performance was similar to the M ≥ 5.5 level, except that both success rate and alarm rate were lower at the M ≥ 6.0 level. If this unexpected finding is real, it might be inherent to the methodology using scattering of short-wave length radio waves as suggested by M. Hayakawa and may contain important information in understanding the earthquake physics and LAI-coupling. The results of the present study indicate strongly that the earthquake prediction research using anomalous transmission of VHF FM radio waves should be enhanced in parallel with complementary research in other frequency ranges.