Refine
Document Type
- Article (6)
- Conference Proceeding (4)
- Master's Thesis (1)
Institute
Keywords
- Bubble column humidifier (2)
- Desalination (2)
- Humidification-dehumidification (2)
- Air humidification (1)
- Air-liquid heat transfer (1)
- Air-liquid mass transfer (1)
- Blasensäule (1)
- Bubble Column Humidifier (1)
- Carbon capture (1)
- Clustering (1)
The formation of mixed gas hydrates for pre- or post-combustion capture of carbon dioxide is considered a promising alternative to conventional carbon capture technologies. Yet, to keep up with conventional technologies or even reduce the cost of capture associated with them, a hydrate-based technology must have (1) a short induction time, (2) fast formation kinetics, and (3) moderate process conditions. To date, these requirements can only be met by adding promoters to the system, which comes at its own cost and disadvantages. Here, we show that the requirements can also be met without promoters by forming mixed gas hydrates in a packed bed of ice stabilized by fumed silica. While the high specific surface area of the packed bed warrants short induction times and fast kinetics, low temperatures ensure both moderate formation pressures and a high CO2 selectivity. The favorable properties can be maintained and even improved upon over many capture/regeneration cycles when operated at temperatures lower than 253 K, as this ensures a continuous formation of pores in the ice. We demonstrate the advantages of this route for carbon capture on a bench scale through batch, semi-batch, and continuous experiments. In semi-batch operation at 233 K and 40 bar, the mole fraction of CO2 in a synthetic flue gas is reduced from 15 mol% to 2.5 mol%. At the same thermodynamic conditions, a split fraction of 70% and a specific energy consumption below 3.0 GJ/tCO2 are achieved in continuous operation. The inherent advantages and simplicity of this process, a specific energy consumption comparable with the state of the art even though entirely based on the bench-scale experiment, as well as environmental harmlessness, emphasize the potential of this hydrate-based process to meet the demands of the industry at a minimal cost of capture.
A data-driven regression model for predicting thermal plant performance under load fluctuations
(2024)
The global energy demand is still primarily reliant on fossil-fueled thermal power plants. With the growing share of renewables, these plants must frequently adjust their loads. Maintaining, or ideally increasing operational efficiency under these conditions is crucial. Increasing the efficiency of such systems directly reduces associated greenhouse gas emissions, but it requires sophisticated models and monitoring systems. Data-driven models have proven their value here, as they can be used for monitoring, operational state estimation, and prediction. However, they are also sensitive to (1) the training approach, (2) the selected feature set, (3) and the algorithm used. Using operational data, we comprehensively investigate these model parameters for performance prediction in a thermal plant for process steam generation. Specifically, four regression algorithms are evaluated for the prediction of the highly fluctuating live steam flow with two training approaches and three feature subsets of the raw dataset. Furthermore, manual and automatic clustering methods are used to identify different states of operation regarding the fuel amounts used in the combustion chamber. Our results show that the live steam flow is predicted with excellent accuracy for a testing period of one month (R² =0.994 and NMAE=0.55%) when using a dynamic training approach and a comprehensive feature set comprised of 48 features representing the combustion process. It is also seen that the statically trained model predicts various load changes with strong accuracy and that the accuracy of the dynamically trained model can be approached by incorporating the cluster information into the static model. These models reflect the plant's physical intricacies under varying loads, where deviations from the predicted live steam flow indicate unwanted long-term drifts. They can be directly implemented to help operators detect inefficiencies and optimize plant performance.
The food industry faces significant challenges in managing operational costs due to its high energy intensity and rising energy prices. Industrial food-processing facilities, with substantial thermal capacities and large demands for cooling and heating, offer promising opportunities for demand response (DR) strategies. This study explores the application of deep reinforcement learning (RL) as an innovative, data-driven approach for DR in the food industry. By leveraging the adaptive, self-learning capabilities of RL, energy costs in the investigated plant are effectively decreased. The RL algorithm was compared with the well-established optimization method Mixed Integer Linear Programming (MILP), and both were benchmarked against a reference scenario without DR. The two optimization strategies demonstrate cost savings of 17.57% and 18.65% for RL and MILP, respectively. Although RL is slightly less efficient in cost reduction, it significantly outperforms in computational speed, being approximately 20 times faster. During operation, RL only needs 2ms per optimization compared to 19s for MILP, making it a promising optimization tool for edge computing. Moreover, while MILP’s computation time increases considerably with the number of binary variables, RL efficiently learns dynamic system behavior and scales to more complex systems without significant performance degradation. These results highlight that deep RL, when applied to DR, offers substantial cost savings and computational efficiency, with broad applicability to energy management in various applications.
Vast amounts of oily wastewater are byproducts of the petrochemical and the shipping industry and to this day frequently discharged into water bodies either without or after insufficient treatment. To alleviate the resulting pollution, water treatment processes are in great demand. Bubble column humidifiers (BCHs) as part of humidification–dehumidification systems are predestined for such a task, since they are insensitive to different feed liquids, simple in design and have low maintenance requirements. While humidification in a bubble column has been investigated plentiful for desalination, a systematic investigation of oily wastewater treatment is missing in literature. We filled this gap by analyzing the treatment of an oil–water emulsion experimentally to derive recommendations for future design and operation of BCHs. Our humidity measurements indicate that the air stream is always saturated after humidification for a liquid height of only 10 cm. A residual water mass fraction of 3.5 wt% is measured after a batch run of six hours. Furthermore, continuous measurements show that an increase in oil mass fraction leads to a decrease in system productivity especially for high oil mass fractions. This decrease is caused by the heterogeneity of the liquid temperature profile. A lower liquid height mitigates this heterogeneity, therefore decreasing the heat demand and improving the overall efficiency. The oil content of the produced condensate is below 15 ppm, allowing discharge into various water bodies. The results of our systematic investigation prove suitability and indicate a strong future potential for the use of BCHs in oily wastewater treatment.
Power plant operators increasingly rely on predictive models to diagnose and monitor their systems. Data-driven prediction models are generally simple and can have high precision, making them superior to physics-based or knowledge-based models, especially for complex systems like thermal power plants. However, the accuracy of data-driven predictions depends on (1) the quality of the dataset, (2) a suitable selection of sensor signals, and (3) an appropriate selection of the training period. In some instances, redundancies and irrelevant sensors may even reduce the prediction quality.
We investigate ideal configurations for predicting the live steam production of a solid fuel-burning thermal power plant in the pulp and paper industry for different modes of operation. To this end, we benchmark four machine learning algorithms on two feature sets and two training sets to predict steam production. Our results indicate that with the best possible configuration, a coefficient of determination of R^2 = 0.95 and a mean absolute error of MAE=1.2 t/h with an average steam production of 35.1 t/h is reached. On average, using a dynamic dataset for training lowers MAE by 32% compared to a static dataset for training. A feature set based on expert knowledge lowers MAE by an additional 32 %, compared to a simple feature set representing the fuel inputs. We can conclude that based on the static training set and the basic feature set, machine learning algorithms can identify long-term changes. When using a dynamic dataset the performance parameters of thermal power plants are predicted with high accuracy and allow for detecting short-term problems.
Bubble columns are recently used for the humidification of air in water treatment systems and fuel cells. They are well applicable due to their excellent heat and mass transfer and their low technical complexity. To design and operate such devices with high efficiency, the humidification process and the impact of the operating parameters need to be understood to a sufficient degree. To extend this knowledge, we use a refined and novel method to determine the volumetric air–liquid heat and mass transfer coefficients and the humidifier efficiency for various parametric settings. The volumetric transfer coefficients increase with both of the superficial air velocity and the liquid temperature. It is further shown that the decrease of vapor pressure with an increase of the salinity results in a corresponding decrease in the outlet humidity ratio. In contrast to previous studies, liquid heights smaller than 0.1 m are investigated and significant changes in the humidifier efficiency are seen in this range. We present the expected humidifier efficiency with respect to the superficial air velocity and the liquid height in an efficiency chart, such that optimal operating conditions can be determined. Based on this efficiency chart, recommendations for industrial applications as well as future scientific challenges are drawn.
Bubble column humidifiers (BCHs) are frequently used for the humidification of air in various water treatment applications. A potential but not yet profoundly investigated application of such devices is the treatment of oily wastewater. To evaluate this application, the accumulation of an oil-water emulsion using a BCH is experimentally analyzed. The amount of evaporating water vapor can be evaluated by measuring the humidity ratio of the outlet air. However, humidity measurements are difficult in close to saturated conditions, as the formation of liquid droplets on the sensor impacts the measurement accuracy. We use a heating section after the humidifier, such that no liquid droplets are formed on the sensor. This enables us a more accurate humidity measurement. Two batch measurement runs are conducted with (1) tap water and (2) an oil-water emulsion as the respective liquid phase. The humidity measurement in high humidity conditions is highly accurate with an error margin of below 3 % and can be used to predict the oil concentration of the remaining liquid during operation. The measured humidity ratio corresponds with the removed amount of water vapor for both tap water and the accumulation of an oil-water emulsion. Our measurements show that the residual water content
in the oil-water emulsion is below 4 %.
The humidification-dehumidification process (HDH) for desalination is a promising technology to address water scarcity issues in rural regions. However, a low humidifier efficiency is a weakness of the process. Bubble column humidifiers (BCH) are promising for HDH, as they provide enhanced heat and mass transfer and have low maintenance requirements. Previous studies of HDH-systems with BCHs draw different conclusions regarding the impact of superficial air velocity and liquid height on the humidification. Furthermore, the impact of flow characteristics has never been investigated systematically at all. In this study, an optimized BCH test setup that allows for optical analysis of the humidifier is used and evaluated. Our test setup is validated, since the influence of water temperature on the humidification, which is exponential, is reproduced. Measurements with seawater show that the normalised system productivity is increased by about 56 % with an increase in superficial air velocity from 0.5 to 5 cm/s. Furthermore, the system productivity is increased by around 29 % with an increase in liquid height from 60 to 378 mm. While the impact of superficial air velocity can be traced back to temperature changes at the humidifier and dehumidifier outlets, the impact of liquid height is shown to be caused by a smaller heat loss surface in the humidifier with an increase in liquid height. For the impact of sieve plate orifice diameter, a clear influence on the humidification is not apparent, this parameter needs to be investigated further. Finally, our new test setup allows for analysing the humidification of air (1) in a systematic way, (2) in relevant measurement ranges and (3) in comparison with optical analyses of the flow characteristics.