Hourly Time Series Data Research Articles

Probabilistic CDF-based load forecasting model in a power distribution system

The study searches for a probabilistic forecasting model for electric power demand based on the correlation between weather temperature and electric power demand using hourly time series data spanning a year. The model is based on the temperature and power demand data's cumulative probability distribution functions or CDFs. In probabilistic forecasting, a range of potential values and their associated probabilities are predicted in addition to a single value for a given range. The qualities of CDF modeling allow it to cover these uncertainties. For scenario 1, case 2, and case 3, the mean absolute percentage error (MAPE)-based forecasting accuracy for the hourly load demand is 3.93, 4.64, and 2.56, respectively. The Gaussian distribution-based CDFs' capacity to forecast outside of the present data range is an extra advantage. The world's electric power networks are growing increasingly intricate and unpredictable through the integration of several distributed energy resources (DERs) and cutting-edge technology from renewable energy resources (RERs). Other variables, such as variations in weather patterns, shifts in the status of the economy, or adjustments to energy policies, can also add to uncertainty and not be foreseen. A body of literature demonstrates a shortage of research on LTLF and general-purpose forecasting methods. The proposed approach is flexible and might establish a standard for many different forecasting tasks. The proposed method is intended to be a flexible strategy that may be used for different utilities with varying time horizons.

Short-term solar irradiance forecasting based on a novel Bayesian optimized deep Long Short-Term Memory neural network

Accurate forecasting is indispensable for improving solar renewables integration and minimizing the effects of solar energy's intermittency. Existing research on time series solar forecasting confronts challenges such as determining the accurate hyperparameters and flexibility in considering meteorological parameters. This study proposes a novel deep learning model, namely an optimized stacked Bi-directional Long Short-Term Memory(BiLSTM)/ Long Short-Term Memory(LSTM) model to forecast univariate and multivariate hourly time series data by integrating stacked LSTM layers, drop out architecture, and LSTM based model. The performance of the model is enhanced by Bayesian optimization with the tuning of six relevant hyperparameters. To evaluate the model, standard Global Horizontal Irradiance (GHI) and observed Plane of Array (POA) irradiance with meteorological real-world solar data from Sweihan Photovoltaic Independent Power project in Abu Dhabi, UAE, and NREL solar data for year-round data are forecasted. Furthermore, the performance of the proposed algorithm is also evaluated under weather uncertainty for different climate types. The forecasting accuracy is evaluated based on various performance metrics and it is observed that the proposed model offered the best R2 values, 0.99 for univariate as well as multivariate models using GHI data and 0.97 using POA data. The findings suggest that the proposed model is a reliable technique for solar prediction due to its comparable performance with both GHI and POA in terms of accuracy.

Storage requirements in a 100% renewable electricity system: extreme events and inter-annual variability

In the context of 100% renewable electricity systems, prolonged periods with persistently scarce supply from wind and solar resources have received increasing academic and political attention. This article explores how such scarcity periods relate to energy storage requirements. To this end, we contrast results from a time series analysis with those from a system cost optimization model, based on a German 100% renewable case study using 35 years of hourly time series data. While our time series analysis supports previous findings that periods with persistently scarce supply last no longer than two weeks, we find that the maximum energy deficit occurs over a much longer period of nine weeks. This is because multiple scarce periods can closely follow each other. When considering storage losses and charging limitations, the period defining storage requirements extends over as much as 12 weeks. For this longer period, the cost-optimal storage needs to be large enough to supply 36 TWh of electricity, which is about three times larger than the energy deficit of the scarcest two weeks. Most of this storage is provided via hydrogen storage in salt caverns, of which the capacity is even larger due to electricity reconversion losses (55 TWh). Adding other sources of flexibility, for example with bioenergy, the duration of the period that defines storage requirements lengthens to more than one year. When optimizing system costs based on a single year rather than a multi-year time series, we find substantial inter-annual variation in the overall storage requirements, with the average year needing less than half as much storage as calculated for all 35 years together. We conclude that focusing on short-duration extreme events or single years can lead to an underestimation of storage requirements and costs of a 100% renewable system.

Dynamical Characteristics in Time Series Between PM10 and Wind Speed

We study the temporal variation characteristics of PM10 and wind velocity in eight South Korean cities. We employ the detrended cross-correlation analysis (DCCA) method to extract the overall tendency of the hourly variation. We ascertain from three-daily and one-weekly intervals that Busan has the negative largest, while Donghae has the positive largest in the DCCA cross-correlation coefficient between PM10 and wind velocity. As a result of Asian dust events, the cross-correlation is statistically significant for the hourly time series data less than two days. Particularly, we discuss whether a cross-correlation is statistically significant or not from random number surrogation and shuffled time series surrogation

Assessment of Wind Speed and Power Density Using Weibull and Rayleigh Distributions at Turbat, Balochistan, Pakistan

The meager economic situation, indefinite energy crisis, industrial modernization, and hazardous contamination impulse the empiricist to transcend the attention on renewable energy resources in Pakistan. One of the most rapidly growing renewable energy sources is wind energy. The main goal of this research work is to examine the wind characteristics and wind potential at the site of Turbat, Balochistan, Pakistan. For this purpose, the measured hourly time series data was collected from the Pakistan Meteorological Department (CPDC, Karachi) for 21 months (Jan 2012 – Sep 2013). After evaluating the monthly average wind speed (> 4 m/s), the average value of most probable wind speed (3.83 m/s), the average value of wind speed carrying maximum energy (7.732 m/s), and the standard deviation of the data (1.699 – 3.306), the results are used to statistically evaluate the data by Weibull and Rayleigh distributions for the selected site. The monthly average value of wind power and energy densities of the selected site is 140.145 W/m2 and 101.775 kWh/m2, respectively. A comparison was made between the mean power potential of the site and the power potential assessed using the Weibull and Rayleigh distributions. It was revealed that the Weibull distribution depicted the data more accurately. This statement is further enriched by the assessment of the performance of both distributions with the RMSE, χ2, and R2 tests

An LSTM model for power grid loss prediction

With the changes that renewable energy sources bring to the electricity markets in all over the world, prediction of grid losses gets more complex as current methods have limited capability to take local weather conditions into account. This paper suggests a Long Short-Term Memory (LSTM) recurrent neural network model for power grid loss prediction. The model learns long-term relations of hourly time series data from electricity markets, local weather and calendar. We apply the model to predict the total transmission grid losses in Finland. We find that the proposed model outperforms both the reference method currently used in industry and linear regression proposed by previous studies.

Multi-Factorial Comparison for 24 Distinct Transposition Models for Inclined Surface Solar Irradiance Computation in the State of Palestine: A Case Study

Solar energy industries region (MENA). Therefore, this article serves two main objectives. Firstly, it compares comprehensively 24 different transposition models. Several statistical methods as Root Mean Square Error (RMSE), Mean Bias Error (MBE) and Percentage Average Deviation (PAD) and t-statistics (t-stat) are used to quantify the performance of the transposition tilted surface models. Furthermore, the transposition models are compared with real measured hourly timeseries data during 15 months (June 2017 to August 2018) for several Palestinian cities (Gaza and Rafah -in Gaza Strip and Herbon, Jericho, Tulkarm and Nablus -in West Bank) to identify the promising and most accurate model. The analysis was carried out on three bases: annually, monthly and clearness index. The transposition models prove their ability to represent the measured data during the annual and monthly based analysis, but they all failed to achieve complacency in the clearness index (Kt) for the clear sky condition (Kt>0.78).HDKR model is considered the most accurate model for depicting the sky diffuse inclined solar irradiance for majority of Palestinian territories. However, for Nablus the Skartveit-Olseth model has require accurate estimation for global solar irradiation particularly on inclined planes. This improves the accuracy of the sizing procedures and optimizes the performance of the solar energy platforms: as photovoltaic modules and flat-plate solar collectors. Variety of the transposition models are developed and reported to determine solar irradiance incident on an inclined surface. However, there is lack in the literature regarding identifying the most promising transposition model particularly for Middle East and North Africa better overall performance than the other candidate models. It even produces 3.4% better than the HDKR model in predicating solar irradiance for inclined plane. Secondly, the article advises simple, reliable and accurate transposition models for the area of MENA for clear sky conditions. The proposed model was tested for the sites under investigation. The proposed model produces significant better performance than the candidate transposition models, as RMSE and PAD values revealed.

Neural network and regression methods for optimizations between two meteorological factors

This paper is concerned with the temporal variation characteristics of two meteorological factors (temperature and humidity) in four metropolitan cities (Seoul, Busan, Daegu, Daejeon) in South Korea. Data are extracted from seven years (2008 to 2014) of hourly time series data in meteorological offices of the Korea Meteorological Administration. Using the detrended cross-correlation analysis (DCCA) method, the DCCA coefficient of temperature is compared to that of humidity from daily time series data during four seasons in the four metropolitan cities. In particular, as window size s increases, the DCCA cross-correlation coefficient approaches 0.034 at a time lag of 14 days in the case of spring in Seoul. We find the weak cross-correlation between temperature and humidity at different time lags of 14, 21 and 28 days in spring in Seoul, and the errors ET in the ANN are relatively larger values than that of any other season in both the ANN and the MRM. Particularly, in the ANN model, there exist relatively a large error value of temperature as the characteristics of the non-stationary, deterministically chaotic and noisy meteorological data extracted for the short-term prediction rather than the long-term prediction.

Multiple regression analysis of performance parameters of a binary cycle geothermal power plant

Regression analysis of a 7.35MWe existing binary geothermal power plant is conducted using actual plant data to assess the plant performance. The thermo physical properties of geothermal fluid and ambient conditions, which are brine (geothermal water) temperature and flow rate, steam and NCGs (non-condensable gases) flow rates and ambient air temperature, directly affect power generation from a geothermal power plant. Generally, amount of power generated is calculated by deterministic formulations of thermodynamics. However, the data would be probabilistic because inputs may be measured by uncalibrated devices or some parameters may be neglected during the calculation. In these cases, the performance of power plant may be estimated by using regression analysis and then changing of plant performance may be monitored over time. All measured parameters on DORA-1 Geothermal Power Plant from 2006 to 2012 and 49,411 hourly time series data are used in this study.A review of the available literature indicates this paper is the first study to focus on the prediction of power generation of a geothermal power plant by using multiple linear regression analysis. In this study, annual multiple linear regression models are developed to estimate the performance of a geothermal power plant. These models are tested by using classical assumptions of linear regressions and positive serial autocorrelation is found in all models. Autocorrelations are eliminated by using Orcutt–Cochran method.Although the performance model trends, from 2006 to 2008, are found to be close, the performance status of the plant is generally variable from year to year. According to annual regression models, since 2009, the plant performance started to decline with 270kWe electricity generation capacity. The total degradation of the plant performance reached 760kWe capacity by 2012.

Dynamical mechanism in meteorological factors using detrended cross-correlation analysis

We simulate and analyze the temporal dependence of fine dust particle with sizes less than 10 microns, denoted as PM10, on influencing factors such as temperature, humidity, and wind velocity in eight South Korean cities. We employ the detrended cross-correlation analysis method to extract the overall tendencies of the hourly variations of those dependences. The relationships between PM10 and the meteorological factors are established by using the cross-correlation coefficients. Particularly, we ascertain from a specific interval 3≤ n ≤168 of the hourly time series data that a city in Korea, Andong has the largest cross-correlation coefficient while another city in Korea, Busan, has the smallest value in the correlation between the dust density PM10 and the wind velocity. A city in Korea, Donghae has the largest negative value of the cross-correlation coefficient between PM10and the humidity. We find that the cross-correlation is statistically significant for the hourly time intervals n = 12, 24, and 48 for meteorological time series data.

Neural networks approached for modelling river suspended sediment concentration due to tropical storms

Artificial neural networks are one of the advanced technologies employed in hydrology modelling. This paper investigates the potential of two algorithm networks, the feed forward backpropagation (BP) and generalized regression neural network (GRNN) in comparison with the classical regression for modelling the event-based suspended sediment concentration at Jiasian diversion weir in Southern Taiwan. For this study, the hourly time series data comprised of water discharge, turbidity and suspended sediment concentration during the storm events in the year of 2002 are taken into account in the models. The statistical performances comparison showed that both BP and GRNN are superior to the classical regression in the weir sediment modelling. Additionally, the turbidity was found to be a dominant input variable over the water discharge for suspended sediment concentration estimation. Statistically, both neural network models can be successfully applied for the event-based suspended sediment concentration modelling in the weir studied herein when few data are available.

Predictive models for forecasting hourly urban water demand

Summary One of the goals of efficient water supply management is the regular supply of clean water at the pressure required by consumers. In this context, predicting water consumption in urban areas is of key importance for water supply management. This prediction is also relevant in processes for reviewing prices; as well as for operational management of a water network. In this paper, we describe and compare a series of predictive models for forecasting water demand. The models are obtained using time series data from water consumption in an urban area of a city in south-eastern Spain. This includes highly non-linear time series data, which has conditioned the type of models we have included in our study. Namely, we have considered artificial neural networks, projection pursuit regression, multivariate adaptive regression splines, random forests and support vector regression. Apart from these models, we also propose a simple model based on the weighted demand profile resulting from our exploratory analysis of the data. In our comparative study, all predictive models were evaluated using an experimental methodology for hourly time series data that detailed water demand in a hydraulic sector of a water supply network in a city in south-eastern Spain. The accuracy of the obtained results, together with the medium size of the demand area, suggests that this was a suitable environment for making adequate management decisions.

기후변화가 극한강우와 I-D-F 분석에 미치는 영향 평가

최근 수공시설물의 설계규모를 넘어서는 극한 강우사상이 발생하여 홍수방어를 위하여 구축된 수리구조물이 파괴 되는 등 많은 홍수피해가 발생하고 있다. 따라서 극한 강우사상의 시공간적 발생 특성을 파악하고 미래의 기후변화하에서 극한강우사상이 어떻게 변화하고 설계수명기간(Design period)동안 분포 특성이 어떻게 변화할지를 이해하는 것은 매우 중요하다. 이에 본 논문에서는 미래의 기후변화가 극한 강우에 어떠한 영향을 미치는지를 평가하기 위해 기후변화 시나리오를 이용하여 미래의 극한강우의 특성 분석과 I-D-F 분석을 실시하였다. 본 연구에서는 SRES B2 온난화가스 시나리오와 YONU CGCM 를 이용하여 2030s(2031-2050)를 모의하였으며 통계학적 축소기법을 적용하여 우리나라에 위치한 기상청 산하 관측소별로 일 기상자료를 구축하였다. 또한, 이를 과거 관측 자료와 비교하여 Quantile Mapping 방법으로 편이보정을 실시하였고, 구형펄스(Modified Bartlett Lewis Rectangular Pulse, MBLRP) 모형(Onof과 Wheater, 1993; Onof 2000)과 분해기법(adjust method)을 적용하여 일 강우 시계열자료를 시 강우 시계열 자료로 변환하였으며 지속기간별 빈도별 강우량을 산정하여 I-D-F 곡선을 작성하였다. 본 논문에서는 66개 관측소 중에서 서울, 대구, 전주, 광주 지점의 결과만을 수록하였으며 그 결과 거의 모든 지점에서 현재와 비교하였을 때 지속기간이 길어질수록 강우강도가 증가함을 확인할 수 있었다. Recently, extreme precipitation events beyond design capacity of hydraulic system have been occurred and this is the causes of failure of hydraulic structure for flood prevention and of severe flood damage. Therefore it is very important to understand temporal and spatial characteristics of extreme precipitation events as well as expected changes in extreme precipitation events and distributional characteristics during design period under future climate change. In this paper, climate change scenarios were used to assess the impacts of future climate change on extreme precipitation. Furthermore, analysis of future extreme precipitation characteristics and I-D-F analysis were carried out. This study used SRES B2 greenhouse gas scenario and YONU CGCM to simulate climatic conditions from 2031 to 2050 and statistical downscaling method was applied to establish weather data from each of observation sites operated by the Korean Meteorological Administration. Then quantile mapping of bias correction methods was carried out by comparing the simulated data with observations for bias correction. In addition Modified Bartlett Lewis Rectangular Pulse(MBLRP) model (Onof and Wheater, 1993; Onof 2000) and adjust method were applied to transform daily precipitation time series data into hourly time series data. Finally, rainfall intensity, duration, and frequency were calculated to draw I-D-F curve. Although there are 66 observation sites in Korea, we consider here the results from only Seoul, Daegu, Jeonju, and Gwangju sites in this paper. From the results we found that the rainfall intensity will be increased and the bigger intensity will be occurred for longer rainfall duration when we compare the climate conditions of 2030s with present conditions.

Weibull representative compressed wind speed data for energy and performance calculations of wind energy systems

The present article presents the development of compressed wind speed data to be used in wind energy and performance calculations of stand alone or hybrid wind energy systems. Compressed data attempts to represent the important statistics of an entire month of days with d selected days, where d is less than the actual number of days in the month. In the present article, compressed wind speed data are generated based on the Weibull wind speed distribution model. Two different sets of wind speed data are generated, three- and four-day month, each month being represented by 72 and 96 h of wind speed, respectively. The three- and four-day month wind speed data are then used to calculate the monthly energy yields, which, in turn, are compared to those calculated from the measured hourly time series data. A total of eight years long of measured hourly time series data from five different locations throughout the world are used to validate the method presented. The overall errors in estimation of the wind energy yield using the Weibull representative compressed wind speed data are 3.67% and 3.21% for the three- and four-day months, respectively.

Prediction of maximum daily ozone level using combined neural network and statistical characteristics

Analysis and forecasting of air quality parameters are important topics of atmospheric and environmental research today due to the health impact caused by air pollution. As one of major pollutants, ozone, especially ground level ozone, is responsible for various adverse effects on both human being and foliage. Therefore, prediction of ambient ozone levels in certain environment, especially the ground ozone level in densely urban areas, is of great importance to urban air quality and city image. To date, though several ozone prediction models have been established, there is still a need for more accurate models to develop effective warning strategies. The development of such models is difficult because the meteorological variables and the photochemical reactions involved in ozone formation are very complex. The present work aims to develop an improved neural network model, which combines the adaptive radial basis function (ARBF) network with statistical characteristics of ozone in selected specific areas, and is used to predict the daily maximum ozone concentration level. The improved method is trained and testified by hourly time series data collected at three air pollutant-monitoring stations in Hong Kong during 1999 and 2000. The simulation results demonstrate the effectiveness and the reliability of the proposed method.

The chi square periodogram: Its utility for analysis of circadian rhythms

It is proposed that chi-square statistic be employed in constructing periodograms for the analysis of hourly time series data obtained in studies of circadian rhythmicity. We show that even for relatively short (10 day) time series, the integral-valued chi-square periodogram can distinguish circadian-periodic from random series at a level of significance of about 0·01. In addition, we describe the effects of serial correlation and examine the resolving power of the method for two periodic components in the circadian range. We suggest how the method can be most profitably employed in the analysis of event-recorder data for detection of rhythmicity in the range 14 to 34 h., and for the estimation of period to ±0·2 h.