Percorrer por autor "Ramos, Daniel"
A mostrar 1 - 9 de 9
Resultados por página
Opções de ordenação
- A Contextual Reinforcement Learning Approach for Electricity Consumption Forecasting in BuildingsPublication . Ramos, Daniel; Faria, Pedro; Gomes, Luis; Vale, ZitaThe energy management of buildings plays a vital role in the energy sector. With that in mind, and targeting an accurate forecast of electricity consumption, in the present paper is aimed to provide decision on the best prediction algorithm for each context. It may also increase energy usage related with renewables. In this way, the identification of different contexts is an advantage that may improve prediction accuracy. This paper proposes an innovative approach where a decision tree is used to identify different contexts in energy patterns. One week of five-minutes data sampling is used to test the proposed methodology. Each context is evaluated with a decision criterion based on reinforcement learning to find the best suitable forecasting algorithm. Two forecasting models are approached in this paper, based on K-Nearest Neighbor and Artificial Neural Networks, to illustrate the application of the proposed methodology. The reinforcement learning criterion consists of using the Multiarmed Bandit algorithm. The obtained results validate the adequacy of the proposed methodology in two case-studies: building; and industry.
- Demonstration of an Energy Consumption Forecasting System for Energy Management in BuildingsPublication . Jozi, Aria; Ramos, Daniel; Gomes, Luis; Faria, Pedro; Pinto, Tiago; Vale, ZitaDue to the increment of the energy consumption and dependency of the nowadays lifestyle to the electrical appliances, the essential role of an energy management system in the buildings is realized more than ever. With this motivation, predicting energy consumption is very relevant to support the energy management in buildings. In this paper, the use of an energy management system supported by forecasting models applied to energy consumption prediction is demonstrated. The real-time automatic forecasting system is running separately but integrated with the existing SCADA system. Nine different forecasting approaches to obtain the most reliable estimated energy consumption of the building during the following hours are implemented.
- Electricity Consumption Forecast in an Industry Facility to Support Production Planning Update in Short TimePublication . Ramos, Daniel; Faria, Pedro; Vale, Zita; Correia, ReginaThe global environmental concerns raise the need to decrease energy, namely electricity consumption. Energy consumption can be reduced by improving energy efficiency and by improving the optimization of energy management in each context. These opportunities are very relevant in buildings and industry facilities. In order to improve the optimized energy management, adequate forecasting tools are needed regarding the load consumption patterns in each building. In the present paper, two forecasting technics, namely neural networks, and support vector machine, are used to predict the consumption of an industry facility for each 5 minutes. The proposed model finds the best method in order to be used in a later stage regarding the updated of production planning. The size of historic data is also discussed. The case study includes one-week test data and more than one-year train data
- Industrial Facility Electricity Consumption Forecast Using Artificial Neural Networks and Incremental LearningPublication . Ramos, Daniel; Faria, Pedro; Vale, Zita; Mourinho, João; Correia, ReginaSociety’s concerns with electricity consumption have motivated researchers to improve on the way that energy consumption management is done. The reduction of energy consumption and the optimization of energy management are, therefore, two major aspects to be considered. Additionally, load forecast provides relevant information with the support of historical data allowing an enhanced energy management, allowing energy costs reduction. In this paper, the proposed consumption forecast methodology uses an Artificial Neural Network (ANN) and incremental learning to increase the forecast accuracy. The ANN is retrained daily, providing an updated forecasting model. The case study uses 16 months of data, split in 5-min periods, from a real industrial facility. The advantages of using the proposed method are illustrated with the numerical results
- Integrated Robotic and Network Simulation MethodPublication . Ramos, Daniel; Almeida, Luis; Moreno, UbirajaraThe increasing use of mobile cooperative robots in a variety of applications also implies an increasing research effort on cooperative strategies solutions, typically involving communications and control. For such research, simulation is a powerful tool to quickly test algorithms, allowing to do more exhaustive tests before implementation in a real application. However, the transition from an initial simulation environment to a real application may imply substantial rework if early implementation results do not match the ones obtained by simulation, meaning the simulation was not accurate enough. One way to improve accuracy is to incorporate network and control strategies in the same simulation and to use a systematic procedure to assess how different techniques perform. In this paper, we propose a set of procedures called Integrated Robotic and Network Simulation Method (IRoNS Method), which guide developers in building a simulation study for cooperative robots and communication networks applications. We exemplify the use of the improved methodology in a case-study of cooperative control comparison with and without message losses. This case is simulated with the OMNET++/INET framework, using a group of robots in a rendezvous task with topology control. The methodology led to more realistic simulations while improving the results presentation and analysis.
- Selection of features in reinforcement learning applied to energy consumption forecast in buildings according to different contextsPublication . Ramos, Daniel; Faria, Pedro; Gomes, Luis; Campos, P.; Vale, ZitaThe management of buildings responsible for the energy storage and control can be optimized with the support of forecasting techniques. These are essential on the finding of load consumption patterns being these last involved in decisions that analyze which forecasting technique results in more accurate predictions in each context. This paper considers two forecasting methods known as artificial neural network and k-nearest neighbor involved in the prediction of consumption of a building composed by devices recording consumption and sensors data. The forecasts are performed in five minutes periods with the forecasting technique taken into account as a potential to improve the accuracy of predictions. The decision making considers the Multi-armed Bandit in reinforcement learning context to find the best suitable algorithm in each five minutes period thus improving the predictions accuracy in forecasting. The reinforcement learning has been tested in upper confidence bound and greedy algorithms with several exploration alternatives. In the case-study, three contexts have been analyzed.
- Short Time Electricity Consumption Forecast in an Industry FacilityPublication . Ramos, Daniel; Faria, Pedro; Vale, Zita; Correia, ReginaThe work in this article uses artificial neural networks and support vector machine to forecast electricity consumption in an industrial facility. The main objective is to show that such a problem should be treated with a contextual approach that identifies the most adequate technic in each moment for a single building, contrary to the previous works in the literature that compare the accuracy of each method for the complete data set representing aggregated loads. 72 different algorithms have been implemented and tested. After that, the three most suitable ones are selected in order to support the automated decisions of the best algorithm according to the context. In this way, the implemented methodology finds the best method for the prediction of each 5 min. It can be later used to update the production planning in the industrial facility. It also discussed the size of historical data and the most suitable learning parameters for each method. The case study includes test data for one week and more than one year of training data.
- Use of Sensors and Analyzers Data for Load Forecasting: A Two Stage ApproachPublication . Ramos, Daniel; Teixeira, Brigida; Faria, Pedro; Gomes, Luis; Abrishambaf, Omid; Vale, ZitaThe increase in sensors in buildings and home automation bring potential information to improve buildings' energy management. One promissory field is load forecasting, where the inclusion of other sensors' data in addition to load consumption may improve the forecasting results. However, an adequate selection of sensor parameters to use as input to the load forecasting should be done. In this paper, a methodology is proposed that includes a two-stage approach to improve the use of sensor data for a specific building. As an innovation, in the first stage, the relevant sensor data is selected for each specific building, while in the second stage, the load forecast is updated according to the actual forecast error. When a certain error is reached, the forecasting algorithm (Artificial Neural Network or K-Nearest Neighbors) is trained with the most recent data instead of training the algorithm every time. Data collection is provided by a prototype of agent-based sensors developed by the authors in order to support the proposed methodology. In this case study, data over a period of six months with five-minute time intervals regarding eight types of sensors are used. These data have been adapted from an office building to illustrate the advantages of the proposed methodology.
- Using diverse sensors in load forecasting in an office building to support energy managementPublication . Ramos, Daniel; Teixeira, Brígida; Faria, Pedro; Gomes, Luis; Abrishambaf, Omid; Vale, ZitaThe increasing penetration of renewable energy sources led to the development of several energy management approaches. One of the main topics in this field is related to the load forecast in buildings, which can contribute to more intelligent and sustainable energy consumption. However, it is necessary to build a proper forecast model, capable of detecting an accurate consumption profile. The minimum effort to achieve this is to extract a historic with energy consumptions to use as input. Additional information should be considered in order to achieve improvements in forecasting results. This way, information regarding the day of the week is discussed as a reliable source of information that may enhance the load forecast. In this paper, two forecasting techniques, namely neural networks and support vector machine, are used to predict the energy consumption of a building for all 5 min from a period. The proposed model finds the best forecasting technique and determines if the additional information regarding the day of the week enhances the load forecast. In this case study, a period of two years and a half data with a 5-minute time interval is used. Moreover, several tests are performed for varied inputs to understand if the insights are consistent for these tests. This data has been adapted from an office building to illustrate the advantages of the proposed methodology.