High-Performance Computing to Enable Next-Generation Low-Temperature Waste Heat Recovery

The Oak Ridge National Laboratory (ORNL), in collaboration with Eaton Corporation, has performed computational research and development to design an innovative, direct-contact heat exchanger (DCHE) that is optimized for a low-temperature organic Rankine cycle. A computational fluid dynamics (CFD) model of DCHE was developed in STAR-CCM+ which was later calibrated and validated against the experimental data from literature. The validated CFD model was used to develop an industry-relevant liquid-liquid direct-contact heat exchanger system with water and pentane working fluids. This work heavily relied on high-performance computing (HPC) resources to investigate multiple designs and to identify a baseline design. The innovative design consists of two chambers connected by a converging-diverging nozzle. Phase change for pentane, from liquid to vapor, occurs in the first chamber, whereas the second chamber serves as a separator. Outlets in the second chamber are staggered to prevent entrainment of the liquid water by the gaseous pentane. CFD results confirm that the design behaves as expected and the addition of baffles enhances mixing and heat transfer for higher flow rates while preventing entrainment of gaseous pentane by the liquid water.

Stochastic Optimization for Long Term Capital Structures, Systems, and Components Refurbishment and Replacement

As commercial nuclear power plants (NPPs) pursue extended plant operations in the form of Second License Renewals (SLRs), opportunities exist for these plants to provide capital investments to ensure long-term, safe, and economic performance. Several utilities have already announced their intention to pursue extended operations for one or more of their NPPs via SLR2. The goal of this research is to develop a risk-informed approach to evaluate and prioritize plant capital investments made in preparation for, and during the period of, extended plant operations to support decisions in NPP operations. In order to prioritize project selection via a risk-informed approach we developed a single decision-making tool that integrates safety/reliability, cost, and stochastic optimization models to provide users with data analysis capabilities to more cost effectively manage plant assets. Both stochastic analysis methods — such as Monte Carlo-based sampling strategies — and multi-stage stochastic optimization strategies are employed to provide priority lists to decision-makers in support of risk-informed decisions. We applied the proposed method to a trial application of projected replacement/refurbishment expenditures for plant capital assets (i.e., structures, systems, and components [SSCs]). The objective is to optimize the SSC replacement/refurbishment schedule in terms of economic constraints, data uncertainties, and SSC reliability data, as well to generate a priority list for maximizing returns on investment.

Theoretical Analysis of Solar Thermal Desalination Performance Limitation

Solar thermal desalination systems utilize concentrated or non-concentrated sunlight to produce heat to drive a phase change separation process and produce freshwater. It could be an effective solution for increasingly scarce freshwater resources and energy shortages across the globe. In order to explore the performance limits and operating parameters that affect specific water production (SWP), this paper proposes a thermodynamic model of the ideal solar-driven thermal desalination process. The model compares two different heating configurations of solar collector system and considers surface temperature of solar collector, concentration ratio, recovery ratio and inlet saline water salinity to find maximum specific water production. The results show that under reversible condition, a flat plate collector with inlet saline water salinity of 35 g/kg will experience an increase in SWP from 29.9 gs−1m−2 to 52.7 gs−1m−2 if the recovery ratio decrease from 70% to 10%. For a system with concentration ratio of 10, when the surface temperature of solar collector is 507K, the maximum specific water production can reach 166.3 gs−1m−2 as the recovery ratio approaches zero. Reduction in incoming fluid salinity can further increase these performance limitations. The work fundamentally demonstrates the thermodynamic process of solar thermal desalination, and proposes a method to evaluate the performance limitation.

Effects of Ambient Air Humidity on Emissions and Efficiency of Large-Bore Lean-Burn Otto Gas Engines in Development and Application

The use of large-bore Otto gas engines is currently spreading widely considering the growing share of Power-To-Gas (P2G) solutions using renewable energies. P2G with a Combined Heat and Power (CHP) plant offers a promising way of utilizing chemical energy storage to provide buffering for volatile energy sources such as wind and solar power all over the world. Therefore, ambient conditions like air temperature, humidity and pressure can differ greatly between the location and time of engine operation, influencing its performance. Especially lean-burn Otto processes are sensitive to changes in ambient conditions. Besides, targeted use of humidity variation (e.g. through water injection in the charge air or combustion chamber) can help to reduce NOx emissions at the cost of a slightly lower efficiency in gas engines, being an alternative to selective catalytic reduction (SCR) exhaust gas aftertreatment. The ambient air condition boundaries have to be considered already in the early stages of combustion development, as they can also have a significant effect on generated measurement data in combustion research. To investigate the behavior, a test bench with a natural gas (CNG) powered single-cylinder research engine (piston displacement 4.77 1) at the Institute of Internal Combustion Engines (LVK) of the Technical University of Munich (TUM) was equipped with a sophisticated charge air conditioning system. This includes an air compressor and refrigeration dryer, followed by temperature and pressure control, as well as a controlled injection system for saturated steam and homogenizing containers, enabling the test bench to precisely emulate a widespread area of charge air parameters in terms of pressure, temperature and humidity. With this setup, different engine tests were conducted, monitoring and evaluating the engine’s emission and efficiency behavior regarding charge air humidity. In a first approach, the engine was operated maintaining a steady air-fuel equivalence ratio λ, fuel energy input (Q̇fuel = const.) and center of combustion (MFB 50%) while the relative ambient humidity was varied in steps between 21% and 97% (at 22 °C and 1013.25 hPa). Results show a significant decrease in nitrogen oxides (NOx) emissions (−39.5%) and a slight decrease in indicated efficiency (−1,9%) while hydrocarbon (THC) emissions increased by around 60%. The generated data shows the high significance of considering charge air conditioning already in the development stage at the engine test bench. The comparability of measurement data depends greatly on ambient air humidity. In a second approach, the engine was operated at a constant load and constant NOx emissions, while again varying the charge air humidity. This situation rather reflects an actual engine behavior at a CHP plant, where today often NOx–driven engine control is used, maintaining constant NOx emissions. The decrease in indicated efficiency was comparable to the prior measurements, while the THC emissions showed only a mild increase (5%). From the generated data it is, for instance, possible to derive operational strategies to compensate for changes in ambient conditions while maintaining emission regulations as well as high-efficiency output. Furthermore, the results suggest possibilities, but also challenges of utilizing artificial humidification (e.g. through water injection) considering the effects on THC emissions and efficiency. A possible shift of the knocking limit to earlier centers of combustion with higher humidity is to be investigated. The main goal is the further decrease of NOx emissions, increase of efficiency, while still maintaining hydrocarbon emissions.

Modeling and Control of Subcritical Coal-Fired Power Plant Components for Fault Detection

Due to the increased penetration of renewable power sources into the electric grid, the current number of existing coal-fired power plants shifting from baseload to load-following operations has also increased. This shift creates challenges especially for the power industry as coal-fired power plants were not designed for ramping situations, leading to added stress on major components of these plants. This stress causes the system to degrade over time and eventually develop faults. As boilers are still the primary component that fails and causes forced outages, accurate characterization of faults and fractures of boilers is now becoming increasingly critical to reduce plant downtime and extend the plant life during cycling operations. This work focuses on modeling sections of a subcritical coal-fired power plant and proposes algorithms for fault detection in MATLAB/Simulink. The developed model simulates the process dynamics including steam and feedwater flow regulating valves, drum-boiler, and heat rate on the regulation of pressure, drum level and production of saturated steam. The model also simulates the dynamics of superheaters for increasing the energy content of steam, and a spray section for regulating the temperature of steam upstream of the high-pressure turbine to allow for power output adjustment within a given valve operating range. Furthermore, an extension to a leak detection framework proposed by co-authors in previous work is explored. The new framework includes a modification to the threshold analysis portion of the previous work. The extended framework is then applied to a subcritical coal-fired power plant model for leak detection. In particular, this framework analyzes mismatches or deviations in expected plant dynamics with an identified transfer function model. The mismatch is flagged after it exceeds a threshold. The developed algorithm thus aids in rapid detection of faults to reduce impeded plant performance. The results of this work will support real plant operations by providing an accurate characterization of faults in the operation of coal-fired power plants.

Reactor Network Modeling of Stability and Emissions of Hydrogen and Natural Gas Blends for a Piloted Gas Turbine Combustor

Lean-premixed (LPM) gas turbines have been developed for stationary power generation in efforts to reduce emissions due to strict air quality standards. Lean-premixed operation is beneficial as it reduces combustor temperatures, thus decreasing NOx formation and unburned hydrocarbons. However, tradeoffs occur between system performance and turbine emissions. Efforts to minimize tradeoffs between stability and emissions include the addition of hydrogen to natural gas, a common fuel used in stationary gas turbines. The addition of hydrogen is promising for both increasing combustor stability and further reducing emissions because of its wide flammability limits allowing for lower temperature operation, and lack of carbon molecules. Other efforts to increase gas turbine stability include the usage of a non-lean pilot flame to assist in stabilizing the main flame. By varying fuel composition for both the main and piloted flows of a gas turbine combustor, the effect of hydrogen addition on performance and emissions can be systematically evaluated. In the present work, computational fluid dynamics (CFD) and chemical reactor networks (CRN) are created to evaluate stability (LBO) and emissions of a gas turbine combustor by utilizing fuel and flow rate conditions from former hydrogen and natural gas experimental results. With CFD and CRN analysis, the optimization of parameters between fuel composition and main/pilot flow splits can provide feedback for minimizing pollutants while increasing stability limits. The results from both the gas turbine model and former experimental results can guide future gas turbine operation and design.

Combining a Fuel Cell and Ultracapacitor Bank to Power a Vertical Take-Off and Landing Unmanned Aerial System

This research investigated the combination of a fuel cell and ultracapacitors to create a hybrid powertrain for a vertical take-off unmanned aerial system (UAS). This replaced the more common battery-only powertrain or the hybrid fuel cell-battery powertrain. A secondary power source, such as a battery or ultracapacitors, is required to assist a fuel cell with immediate load requests because fuel cells are unable to supply instantaneous power. The fuel cell-ultracapacitor was tested using a power profile that was experimentally determined using a battery-powered vertical take-off UAS during take-off, hover, and landing. This tabletop experiment is meant to lead to a more refined solution that can be easily scaled to fit into a smaller future vertical take-off UAS. Two separate ultracapacitor banks were made to be put in parallel with the fuel cell. The first was a series of 14, 650 Farad ultracapacitors and the second was a series of 14, 350 Farad ultracapacitors. Both fuel cell-ultracapacitor powertrains were able to meet the power requirements while also supplying power to the fuel cell itself, without an external power supply. Future work opportunities include scaling for implementation into a UAS platform and coding the power management software to optimally manage the proposed hybrid powertrain.

Prediction of Wind Speed, Potential Wind Power, and the Associated Uncertainties for Offshore Wind Farm Using Deep Learning

Floating offshore wind turbines hold great potential for future solutions to the growing demand for renewable energy production. Thereafter, the prediction of the offshore wind power generation became critical in locating and designing wind farms and turbines. The purpose of this research is to improve the prediction of the offshore wind power generation by the prediction of local wind speed using a Deep Learning technique. In this paper, the future local wind speed is predicted based on the historical weather data collected from National Oceanic and Atmospheric Administration. Then, the prediction of the wind power generation is performed using the traditional methods using the future wind speed data predicted using Deep Learning. The network layers are designed using both Long Short-Term Memory (LSTM) and Bi-directional LSTM (BLSTM), known to be effective on capturing long-term time-dependency. The selected networks are fine-tuned, trained using a part of the weather data, and tested using the other part of the data. To evaluate the performance of the networks, a parameter study has been performed to find the relationships among: length of the training data, prediction accuracy, and length of the future prediction that is reliable given desired prediction accuracy and the training size.

Online Adaptive Control Tuning in a Gas Turbine Hybrid System

Proportional integral and derivative (PID) controllers are the most popular technique used in the power plant industry for process automation. However, the performance of these controllers may be affected due to variations in the power plant operating conditions, such as between startup, shutdown, and baseload/part-load operation. To maintain the desired performance over the full range of operations, PID controllers are always retuned in most power plants. During this retuning process, the operator takes control of the manipulated variable to perform a standard procedure based on a bump test. This procedure is generally performed to characterize the relationship between the manipulated variable and the process variable at each operating condition. After the bump test, the operator generally applies basic guidelines to assign new parameters to the PID controller. In this paper, the Model Reference Adaptive Controller (MRAC) control technique was implemented to update the PID controller parameters online without performing the bump test procedure. This approach allows updating the controller response on-the-fly while the power plant is running and without using the standard procedure based on a bump test. The MRAC was developed and demonstrated in the gas turbine hybrid cycle at the National Energy Technology Laboratory (NETL) to retune a critically damped mass flow PID controller into an over-damped response. Results showed stable performance during mass flow setpoint steps and also a stable update of the controller parameters.

Wind Turbine Blade Coating Fatigue Induced by Raindrop Impact

With the increase of wind energy production demand, the need to manufacture larger wind turbine blades is on the rise. Because of the high tip speed of the large blade, the blade could be impacted by high-speed objects such as raindrops. This research focuses on developing a computational model for analyzing wind turbine blade coating fatigue induced by raindrop impact. A stochastic rain texture model is used to simulate a realistic rain event determined by a rain intensity and a rain duration. A smoothed particle hydrodynamic approach is implemented to calculate the impact stress considering a single raindrop. A stress interpolation method is proposed to accurately and efficiently estimate the impact of stress under a random rain event. Besides, a crack growth law is used to explain the process of coating shedding. Through a method for calculating crack growth length based on stress, this paper analyzes crack growth life as a function of the rain intensity and the rain duration. This function, together with the statistics of rainfall history, provides a new approach for estimating the expected fatigue life of the blade coating.