• Environmental and Resource Economics Review
• /
• v.26 no.4
• /
• pp.519-547
• /
• 2017

This paper explores whether it is better to forecast city gas demand in Korea using national level data directly or, alternatively, construct forecasts from regional demand models and then aggregate these regional forecasts. In the regional model, we consider gas demand for Seoul metropolitan and the other local areas. Our forecast evaluation exercise for 2013-2016 shows the regional forecast model generally outperforms the national forecasting model. This result comes from the fact that the dynamic properties of each region's gas demands can be better taken into account in the regional demand model. More specifically, the share of residential gas demand in the Seoul metropolitan area is above 50%, and subsequently this demand is heavily influenced by temperature fluctuations. Conversely, the dominant portion of regional gas demand is due to industrial gas consumption. Moreover, electricity is regarded as a substitute for city gas in the residential sector, and industrial gas competes with certain oil products. Our empirical results show that a regional demand forecast model can be an effective alternative to the demand model based on nation-wide gas consumption and that regional information about gas demand is also useful for analyzing sectoral gas consumption.

• Journal of Satellite, Information and Communications
• /
• v.10 no.4
• /
• pp.113-116
• /
• 2015

The electric utility has the responsibility of reducing the impact of peaks on electricity demand and related costs. Therefore, they have introduced Direct Load Control System (DLCS) to automate the external control of shedding customer load that it controls. The existing DLCS have been operated only depend on On/Off signal from the electric utility. That kind of DLCS operating has been successfully used until now. But since the number of customer load participating in the DLC program are keep increasing, On/Off signal control from the electric utility is no longer meets the needs of many different kind of customers. Therefore, In this paper, the author suggest the design of direct load control system using weather sensors to meet the diversity of different customer needs.

• Proceedings of the KIEE Conference
• /
• summer
• /
• pp.695-697
• /
• 2004

현재 전력시장에서 발생하는 게이밍을 반영하기 위한 수리적 모델로서 가장 보편적으로 사용되는 이론 중의 하나가 쿠르노 모델이다. 쿠르노 모델을 실제전력시장에 적용할 때 가장 어려운 점 중의 하나는 정화한 해당 모델에 사용되는 수요와 시장가격간의 관계를 정식화한 수요반웅함수(혹은 역수요함수)를 구하는 것이 다. 기존 모델의 경우 장기간에 걸친 탐문조사나 데이터를 바탕으로 가격탄력성을 구하는 방식을 취하고 있다. 그러나 수요는 전기설비의 교체 소비자의 기호 등 여러가지 변수로 지속적으로 변할 수 있기 때문에 이러한 고정적인 가격탄력성을 적용하는 것은 문제점이 될 수 있기 때문에 본 논문에서는 이러한 가격탄력성을 일정 거래주기 마다 갱신해줄 수 있는 방법을 제안하고자 한다.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.65 no.3
• /
• pp.404-410
• /
• 2016

This paper proposed a program of an energy storage system(ESS) for peak shaving of high-speed railway substations The peak shaving saves cost of equipment and demand cost of the substation. To reduce the peak load, it is very important to know when the peak load appears. The past data based load profile forecasting method is easy and applicable to customers which have relatively fixed load profiles. And an optimal scheduling method of the ESS is helpful in reducing the electricity tariff and shaving the peak load efficiently. Based on these techniques, MS. NET based peak shaving program is developed. In case study, a specific daily load profile of the local substation was applied and simulated to verify performance of the proposed program.

• Journal of Wind Energy
• /
• v.15 no.3
• /
• pp.21-29
• /
• 2024

Wind power is highly variable due to the intermittent nature of wind. This can lead to power grid instability and decreased efficiency. Therefore, it is necessary to improve wind power prediction performance to minimize the negative impact on the power system. Recently, wind power prediction using machine learning has gained popularity, and ensemble models in machine learning have shown high prediction accuracy. RF, GB, XGB and LGBM are decision tree-based ensemble models and have high predictive performance in wind power, but these models have problems from over-fitting and strong dependence on certain variables. However, the stacking model can improve prediction performance by combining individual models and compensate for the shortcomings of each model. In this study, The MAE of RF, XGB and LGBM is 310.42 kWh, 217.07 kWh and 265.20 kWh, respectively, while the stacking model based on RF, XGB and LGBM is 202.33 kWh. Stacking models can improve prediction performance. Finally, it is expected to contribute to electricity supply and demand planning.

• Journal of Internet Computing and Services
• /
• v.20 no.2
• /
• pp.51-59
• /
• 2019

In this paper, we analyzed the relationship between charging operation system and electricity charges connected with charging infrastructure among data of many demonstration projects focused on electric vehicles recently. At this point in time, due to the rapid increase in demand for the electric charging infrastructure that will take place in the future, we can prepare for an upcoming era in the sense of forecasting the demand value. At the same time, demonstrating and modeling optimized system modeling centering on sites is a prerequisite. The modeling based on the existing small - scale simulation and the design of the operating system are based on the data linkage analysis. In this paper, we implemented a new optimized system modeling and introduced it as a standard format to analyze time - dependent time - divisional data for each vehicle and user in each point and node. In order to verify the efficiency of the optimization based on the data linkage analysis for the actual implemented electric car charging infrastructure and operation system.

• Journal of Korean Society of Water and Wastewater
• /
• v.32 no.5
• /
• pp.421-434
• /
• 2018

Water supply/intake pumps operation use 70~80% of power costs in water treatment plants. In the water treatment plant, seasonal and hourly differential electricity rates are applied, so proper pump scheduling can yield power cost savings. Accordingly, the purpose of this study was to develop an optimal water supply pump scheduling scheme. An optimal operation method of water supply pumps by using genetic algorithm was developed. Also, a method to minimize power cost for water supply pump operation based on pump performance derived from the thermodynamic pump efficiency measurement method was proposed. Water level constraints to provide sufficient disinfection performance in a clearwell and reservoirs were calibrated. In addition, continuous operation time constraints were calibrated to prevent frequent pump switching. As a result of optimization, savings ratios during 7 days in winter and summer were 4.5% and 5.1%, respectively. In this study, the method for optimal water pump operation was developed to secure disinfection performance in the clearwell and to save power cost. It is expected that it will be used as a more advanced optimal water pump operation method through further studies such as water demand forecasting and efficiency according to pump combination.

• Journal of IKEEE
• /
• v.23 no.1
• /
• pp.120-126
• /
• 2019

Recently, accurate prediction of power consumption based on machine learning techniques in Internet of Energy (IoE) has been actively studied using the large amount of electricity data acquired from advanced metering infrastructure (AMI). In this paper, we propose a deep learning model based on Gated Recurrent Unit (GRU) as an artificial intelligence (AI) network that can effectively perform pattern recognition of time series data such as the power consumption, and analyze performance of the prediction based on real household power usage data. In the performance analysis, performance comparison between the proposed GRU-based learning model and the conventional learning model of Long Short Term Memory (LSTM) is described. In the simulation results, mean squared error (MSE), mean absolute error (MAE), forecast skill score, normalized root mean square error (RMSE), and normalized mean bias error (NMBE) are used as performance evaluation indexes, and we confirm that the performance of the prediction of the proposed GRU-based learning model is greatly improved.