Power optimization of wind turbines with data mining and evolutionary computation

A data-driven approach for maximization of the power produced by wind turbines is presented. The power optimization objective is accomplished by computing optimal control settings of wind turbines using data mining and evolutionary strategy algorithms. Data mining algorithms identify a functional mapping between the power output and controllable and non-controllable variables of a wind turbine. An evolutionary strategy algorithm is applied to determine control settings maximizing the power output of a turbine based on the identified model. Computational studies have demonstrated meaningful opportunities to improve the turbine power output by optimizing blade pitch and yaw angle. It is shown that the pitch angle is an important variable in maximizing energy captured from the wind. Power output can be increased by optimization of the pitch angle. The concepts proposed in this paper are illustrated with industrial wind farm data.     

Virtual models of indoor-air-quality sensors

A data-driven approach for modeling indoor-air-quality (IAQ) sensors used in heating, ventilation, and air conditioning (HVAC) systems is presented. The IAQ sensors considered in the paper measure three basic parameters, temperature, CO₂, and relative humidity. Three models predicting values of IAQ parameters are built with various data mining algorithms. Four data mining algorithms have been tested on the HVAC data set collected at an office-type facility. The computational results produced by models built with different data mining algorithms are discussed. The neural network (NN) with multi-layer perceptron (MLP) algorithms produced the best results for all three IAQ sensors among all algorithms tested. The models built with data mining algorithms can serve as virtual IAQ sensors in buildings and be used for on-line monitoring and calibration of the IAQ sensors. The approach presented in this paper can be applied to HVAC systems in buildings beyond the type considered in this paper.     

Parameter optimization for a PEMFC model with a hybrid genetic algorithm

Many steady‐state models of polymer electrolyte membrane fuel cells (PEMFC) have been developed and published in recent years. However, models which are easy to be solved and feasible for engineering applications are few. Moreover, rarely the methods for parameter optimization of PEMFC stack models were discussed. In this paper, an electrochemical‐based fuel cell model suitable for engineering optimization is presented. Parameters of this PEMFC model are determined and optimized by means of a niche hybrid genetic algorithm (HGA) by using stack output‐voltage, stack demand current, anode pressure and cathode pressure as input–output data. This genetic algorithm is a modified method for global optimization. It provides a new architecture of hybrid algorithms, which organically merges the niche techniques and Nelder–Mead's simplex method into genetic algorithms (GAs). Calculation results of this PEMFC model with optimized parameters agreed with experimental data well and show that this model can be used for the study on the PEMFC steady‐state performance, is broader in applicability than the earlier steady‐state models. HGA is an effective and reliable technique for optimizing the model parameters of PEMFC stack. Copyright © 2005 John Wiley & Sons, Ltd.     

Health monitoring methodology based on exergetic analysis for building mechanical systems

Exergetic analysis is not often performed in the context of retrocommissioning (RCX); this research provides insight into the benefits of incorporating this approach. Data collected from a previously developed RCX test for an air handling unit (AHU) on a college campus are used in an advanced thermodynamic analysis. The operating data is analyzed using the first and second laws and retrofit design solutions are recommended for improved system performance; the second law analysis is particularly helpful because it requires few additional calculations or data collections. The thermodynamic methodology is extended to a building's cooling plant, which uses a vapor compression refrigeration cycle (VCRC) chiller. Existing chiller data collected for the design of automated fault detection and diagnosis methodology is used. As with the AHU analysis, the second law analysis locates irreversibilities that would not be determined from a first law analysis alone. Plant data representing both normal and faulty operation is used to develop a chiller model for assessing performance and health monitoring. Data is analyzed to determine the viability of health monitoring by performing an exergy analysis on existing data. Conclusions are drawn about the usefulness of exergetic analysis for improving system operations of energy intensive building mechanical systems.     

Reheat optimization of the variable-air-volume box

A data-driven approach for optimizing the reheat process in a variable-air-volume box is presented. Data-mining algorithms derive temporal predictive models from the reheat process data. The bi-objective model formed is solved with a modified particle swarm optimization algorithm. To increase computational efficiency, two levels of non-dominated solutions are introduced while solving the optimization model. A model predictive control strategy is used to generate controls minimizing the reheat output while maintaining the thermal comfort at an acceptable level.     

Modeling and optimization of HVAC energy consumption

A data-driven approach for minimization of the energy to air condition a typical office-type facility is presented. Eight data-mining algorithms are applied to model the nonlinear relationship among energy consumption, control settings (supply air temperature and supply air static pressure), and a set of uncontrollable parameters. The multiple-linear perceptron (MLP) ensemble outperforms other models tested in this research, and therefore it is selected to model a chiller, a pump, a fan, and a reheat device. These four models are integrated into an energy optimization model with two decision variables, the setpoint of the supply air temperature and the static pressure in the air handling unit. The model is solved with a particle swarm optimization algorithm. The optimization results have demonstrated the total energy consumed by the heating, ventilation, and air-conditioning system is reduced by over 7%.     

Anticipatory Control of Wind Turbines With Data-Driven Predictive Models

The concept of anticipatory control applied to wind turbines is presented. Anticipatory control is based on the model predictive control (MPC) approach. Unlike the MPC method, noncontrollable variables (such as wind speed) are directly considered in the dynamic equations presented in the paper to predict response variables, e.g., rotor speed and turbine power output. To determine future states of the power drive with the dynamic equations, a time series model was built for wind speed. The time series model was fused with the dynamic equations to predict the response variables over a certain prediction horizon. Based on these predictions, an optimization model was solved to find the optimal control settings to improve the power output without incurring large rotor speed changes. As both the dynamic equations and time series model were built by data mining algorithms, no gradient information is available. A modified evolutionary strategy algorithm was used to solve a nonlinear constrained optimization problem. The proposed approach has been tested on the data collected from a 1.5 MW wind turbine.      

Energy and economic assessment of desiccant cooling systems coupled with single glazed air and hybrid PV/thermal solar collectors for applications in hot and humid climate

This paper presents a detailed analysis of the energy and economic performance of desiccant cooling systems (DEC) equipped with both single glazed standard air and hybrid photovoltaic/thermal (PV/t) collectors for applications in hot and humid climates. The use of ‘solar cogeneration’ by means of PV/t hybrid collectors enables the simultaneous production of electricity and heat, which can be directly used by desiccant air handling units, thereby making it possible to achieve very energy savings. The present work shows the results of detailed simulations conducted for a set of desiccant cooling systems operating without any heat storage.System performance was investigated through hourly simulations for different systems and load combinations. Three configurations of DEC systems were considered: standard DEC, DEC with an integrated heat pump and DEC with an enthalpy wheel. Two kinds of building occupations were considered: office and lecture room. Moreover, three configurations of solar-assisted air handling units (AHU) equipped with desiccant wheels were considered and compared with standard AHUs, focusing on achievable primary energy savings.The relationship between the solar collector’s area and the specific primary energy consumption for different system configurations and building occupation patterns is described. For both occupation patterns, sensitivity analysis on system performance was performed for different solar collector areas. Also, this work presents an economic assessment of the systems. The cost of conserved energy and the payback time were calculated, with and without public incentives for solar cooling systems. It is worth noting that the use of photovoltaics, and thus the exploitation of related available incentives in many European countries, could positively influence the spread of solar air cooling technologies (SAC). An outcome of this work is that SAC systems equipped with PV/t collectors are shown to have better performance in terms of primary energy saving than conventional systems fed by vapour compression chillers and coupled with PV cells.All SAC systems present good figures for primary energy consumption. The best performances are seen in systems with integrated heat pumps and small solar collector areas. The economics of these SAC systems at current equipment costs and energy prices are acceptable. They become more interesting in the case of public incentives of up to 30% of the investment cost (Simple Payback Time from 5 to 10 years) and doubled energy prices.     

工业厂房空压机冷却散热的回收利用

分析了工厂空压机散热量回收的可行性，通过工程实例介绍了利用回收热量预热空调新风、加热空调热水和工艺及生活热水的方法。     

Improvement of part load efficiency of a combined cycle power plant provisioning ancillary services

According to the type of ancillary service provisioned, operation mode of a power plant may change to part load operation. In this contribution, part load operation is understood as delivering a lower power output than possible at given ambient temperature because of gas turbine power output control. If it is economically justified, a power plant may operate in the part load mode for longer time. Part load performance of a newly built 80 MW combined cycle in Slovakia was studied in order to assess the possibilities for fuel savings. Based on online monitoring data three possibilities were identified: condensate preheating by activation of the currently idle hot water section; change in steam condensing pressure regulation strategy; and the most important gas turbine inlet air preheating. It may seem to be in contradiction with the well proven concept of gas turbine inlet air cooling, which has however been developed for boosting the gas turbine cycles in full load operation. On the contrary, in a combined cycle in the part load operation mode, elevated inlet air temperature does not affect the part load operation of gas turbines but it causes more high pressure steam to be raised in HRSG, which leads to higher steam turbine power output. As a result, less fuel needs to be combusted in gas turbines in order to achieve the requested combined cycle’s power output. By simultaneous application of all three proposals, more than a 2% decrease in the power plant’s natural gas consumption can be achieved with only minor capital expenses needed.     

Turbine inlet cooling with thermal energy storage

Turbine inlet cooling (TIC) is a common technology used to increase combustion turbine power output and efficiency. The use of mechanical or absorption chillers for TIC allows for more air cooling than evaporative methods and also imposes a significant parasitic load to the turbine. Thermal energy storage (TES) can be used to shift this load to off‐peak hours. Use of thermal storage increases the flexibility of turbine power output, which benefits from the application of optimization tools. This paper explores the effect of combining TIC with TES to enhance the performance of a district cooling system that includes a gas turbine for power generation. The work illustrates how the system's performance can be enhanced using optimization. Application of multi‐period optimization to the system that includes TES brings significant operational cost savings when compared with a system without thermal storage. It is also shown how TES provides demand‐side energy management in the district cooling loop and supply‐side management through the use of TIC. In addition to the optimization study, a thorough literature review is included that describes the current body of work on combining TIC with TES. Copyright © 2013 John Wiley & Sons, Ltd.     

Performance improvement of PV/T solar collectors with natural air flow operation

The electrical efficiency of a photovoltaic system drops as its operating temperature rises and PV cooling is necessary. The photovoltaic/thermal (PV/T) system is a relatively recent type of solar collector where a circulating fluid of lower temperature than PV module extracts heat from it, cooling the module to improve its output power while the solar pre-heated fluid provides sensible heat. In the present work, air cooling of a commercial PV module configured as PV/T air solar collector by natural flow is presented, where two low cost modification techniques to enhance heat transfer to air stream in the air channel are studied. The considered methods consist of thin metal sheet suspended at the middle or fins attached to the back wall of the air-channel to improve heat extraction from the module. A numerical model was developed and validated against the experimental data obtained from outdoor test campaigns for both glazed and unglazed PV/T prototype models studied. The validation results show good agreement between predicted values and measured data and thus could be used to study analytically the performance of these PV/T air collectors with respect to several design and operating parameters. The modified systems present better performance than the usual type and will contribute to better performance of integrated PV systems for natural ventilation applications in buildings, both space cooling and heating.     

Spatial optimization of an underhood cooling module – Towards an innovative control approach

The present paper reports a numerical investigation of spatial optimization of heat-exchanger by acting on its positioning in the vehicle’s cooling module. This analysis also elucidates how to act on the different parameters influencing heat-exchanger performance in order to optimize their functioning. A two-dimensional computation code permits optimizing the performance of the cooling module by positioning different heat exchangers, in both the driving and stop phases of the vehicle. The ultimate aim is to apply new control approaches to real vehicles so as to reduce pump and compressor energy consumption and thus fuel consumption. Compared to a reference “in-series” configuration of the cooling module HXs (in which the different HXs are superposed in the airflow direction), an “in-parallel” configuration (in which the different HX surfaces are in a row with respect to the air flow direction) increases the thermal power of the HXs by 4.4% and decreases the pressure losses by 0.9%.     

非均匀热负荷条件下的平板冲击冷却优化

冲击冷却是涡轮冷却中常见的方式,其优化设计涉及多种几何参数,是典型的高维问题。在冲击冷却结构的设计过程中,需要根据涡轮的热负荷情况适应性地设计冷却结构,以提高综合冷却效率和表面温度的均匀性。实验或数值模拟耗时长且成本高,而代理模型可以快速预测结果,配合计算机自动寻优算法可显著提高设计效率和效果。为了降低优化设计的成本、提高优化过程的效率,以平板冲击冷却为研究对象,同时考虑非均匀热负荷的影响,通过数值模拟构建数据集,建立了基于迭代算子神经网络的代理模型,并使用遗传算法对斑状非均匀热载荷条件下孔位置排布进行了优化。优化结果显示：对于优化潜力较低的结构,优化策略保持了靶板平均温度水平不变;对于优化潜力较高的结构,可以降低靶板平均温度约2.6 K;所研究各结构的表面温度标准差普遍降低70%以上。     

Control-orientated thermal model for proton-exchange membrane fuel cell systems

A lumped parameter dynamic model is developed for predicting the stack temperature, temperatures of the exit reactant gases and coolant water outlet in a proton-exchange membrane fuel cell (PEMFC) system. A dynamic model for a water pump is also developed and can be used along with the thermal model to control the stack temperature. The thermal and water pump models are integrated with the air flow compressor and PEMFC stack current–voltage models developed by Pukrushpan et al. to study the fuel cell system under open and closed-loop conditions. The results obtained for the aforementioned variables from open-loop simulation studies are found to be similar to the experimental values reported in the literature. Closed-loop simulations using the model are carried out to study the effect of stack temperature on settling times of other variables such as stack voltage, air flow rate, oxygen excess ratio and net power of the stack. Further, interaction studies are performed for selecting appropriate input–output pairs for control purpose. Finally, the developed thermal model can assist the designer in choosing the required number of cooling plates to minimize the difference between the cooling water outlet temperature and stack temperature.     

Thermal management of solar photovoltaics modules for enhanced power generation

Industry and government interest in solar energy has increased in recent years in the Middle East. However, despite high levels of solar irradiance in the Arabian Gulf, harsh climatic conditions adversely affect the electrical performance of solar photovoltaics (PV). The objective of this study is to compare the annual performance characteristics of solar PV modules that utilize either sun-tracking or water cooling to increase electrical power generation relative to that of stationary, passively cooled modules in the Middle East climatic conditions. This is achieved using an electro-thermal model developed and validated against experimental data acquired in this study. The model is used to predict the annual electrical power output of a 140 W PV module in Abu Dhabi (24.43°N, 54.45°E) under four operating conditions: (i) stationary geographical south facing orientation with passive air cooling, (ii) sun-tracked orientation with passive air cooling, (iii) stationary geographical south facing orientation with water cooling at ambient air temperature, and (iv) stationary geographical south facing orientation with water refrigerated at either 10 °C or 20 °C below ambient air temperature. For water cooled modules, annual electrical power output increases by 22% for water at ambient air temperature, and by 28% and 31% for water refrigerated at 10 °C and 20 °C below ambient air temperature, respectively. 80% of the annual output enhancement obtained using water cooling occurs between the months of May and October. Finally, whereas the annual yield enhancement obtained with water cooling at ambient air temperature from May to October is of 18% relative to stationary passive cooling conditions, sun-tracking over the complete year produces an enhancement of only 15% relative to stationary passive cooling conditions.     

Comparison of support vector regression and random forest algorithms for estimating the SOFC output voltage by considering hydrogen flow rates

Solid Oxide Fuel Cell (SOFC) are complex systems in which gas-phase mass transport, heat transfer, ionic conduction, chemical reactions and electrical conduction take place simultaneously. Therefore, reliable simulation tools are needed to control and optimize their operation. Machine Learning (ML) can quickly estimate and generalize the relationship between the input values and the output values in a process. ML algorithms are used in various applications such as modelling, simulation, optimization, control, signal processing, pattern recognition, up to systems like power, production and renewable energy systems. Many methods help the successful design of algorithms for SOFC systems. However, a few researchers have studied and compared regression algorithms.This paper includes an in-depth study to compare the two efficient ML algorithms – namely Random Forest (RF) and Support Vector Regression (SVR). These algorithms are used to predict the performance of a SOFC cell. The algorithms were generated by the experimental data which are measured by using the different temperatures and hydrogen flow rates. Additionally, the effect of the amount of pure hydrogen and the total amount of hydrogen in the content of the fuel mixtures, which fed to the anode side of the SOFC, on the experimental voltage were compared. The experimental data set used for developing the model consists of 1272 records regarding the SOFC operated under different operating conditions. 1122 records of the experimental data set are used for training the regression algorithms mentioned above. Accordingly, the algorithms are tested and then, the experimental data are compared to the results generated by the algorithms to validate prediction performances. The model predicts the cell performance (output voltage) in approximately 0.52 s with the mean absolute percentage errors 1.97% for the RF algorithm and as low as 0.92% for the SVR algorithm. In this article, the SVR algorithm is identified as the most promising model. The process parameters effects on the variation of the output voltage of the SOFC can be examined when the developed models are proven reliability and precision after test with unknown data.     

Short-Term Prediction of Wind Farm Power: A Data Mining Approach

This paper examines time series models for predicting the power of a wind farm at different time scales, i.e., 10-min and hour-long intervals. The time series models are built with data mining algorithms. Five different data mining algorithms have been tested on various wind farm datasets. Two of the five algorithms performed particularly well. The support vector machine regression algorithm provides accurate predictions of wind power and wind speed at 10-min intervals up to 1 h into the future, while the multilayer perceptron algorithm is accurate in predicting power over hour-long intervals up to 4 h ahead. Wind speed can be predicted fairly accurately based on its historical values; however, the power cannot be accurately determined given a power curve model and the predicted wind speed. Test computational results of all time series models and data mining algorithms are discussed. The tests were performed on data generated at a wind farm of 100 turbines. Suggestions for future research are provided.      

Thermoeconomic analysis and optimization of residential split-type air conditioners

A simulation-based optimization methodology for designing unitary residential air conditioners with focus on both energy performance enhancement and cost savings is presented. A steady-state system simulation model was put forward for a 2.5-ton nominal cooling capacity split-type air conditioning unit operating with R-410A as the working fluid. The model predictions for cooling capacity, sensible heat ratio (SHR) and coefficient of performance (COP) were compared with experimental data, when it was found that the model is able to predict the experimental trends within a ±6% error band. The model was then used to find out the condenser and evaporator geometries (face area, heat transfer area) that enhance the system COP for a fixed cost. On one hand, it was observed that the COP can be increased by 7% if the cost is held fixed. On the other hand, cost savings of 33% were achieved when the system COP was held constant.     

The London Heat Island and building cooling design

London’s urban heat island increases the mean air temperature which affects the demand for heating and cooling buildings. Measured air temperature data have been used as input to a building energy simulation computer program to assess the heating and cooling load of a typical air-conditioned office building positioned at 24 different locations within the London Heat Island. It is found that the urban cooling load is up to 25% higher than the rural load over the year, and the annual heating load is reduced by 22%. The effect of raised temperature and urban context are assessed separately, and the sensitivity of the net impact to the internal gains in a building is determined. For the estimation of peak cooling demand, we propose hourly temperature corrections based on radial distance from London’s centre to be applied to standard published temperatures for the region. For more detailed investigations over the cooling season a range of models is available. These are reviewed in this paper and we describe preliminary results of an Artificial Neural Network (ANN) model that predicts location specific hourly temperatures for London, taking into account radial distance from central London, hourly air temperature measured at the meteorological station and associated synoptic weather data.