No description available.

Thesis (MScEng (Mechanical and Mechatronic Engineering))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: Large air‐cooled heat exchangers and condensers make use of fan arrays to provide cooling. The performance of the fan arrays are however negatively affected by distorted inlet conditions. Given the size of these fans, quantifying exactly what the detrimental effects are is practically impossible. This serves as motivation for developing a scaled multi‐fan testing platform that is able to mimic the behaviour of a full‐scale array but at a more measureable and manageable scale. This investigation was conducted in two parts. A Reynolds number investigation was conducted to determine what effect incorporating Reynolds number effects will have on the scaled fan performance. Computational fluid dynamics (CFD) was used to determine where turbulent transition onset occurs on the full‐scale fan blades and trip wires were then sized and positioned appropriately to recreate the same transition effect. From this initial investigation it was found that the trip wires have an allround negative impact on fan performance but when compensated for by increasing the blade angle, the tripped fan static pressure rise performance is comparable with the non‐tripped fan performance. This suggests that the Reynolds number effects may be ignored for this scaled fan testing investigation. Scaled equivalents of the fans in cooling arrays were tested in a three row multi‐fan array which, using symmetry, represents a typical bay in a full‐scale installation. Various platform heights were simulated and the corresponding decrease in system performance was compared to a model formulated to predict how fan volumetric effectiveness decreases with platform height. The model deviated from the test data at very low platforms heights. A replacement fan for the scaled equivalent fan was tested while installed as the edge fan of a multi‐fan array, the fan worst affected by any inlet disturbance, and was found to perform similarly to the scaled equivalent fan for most platform heights but exceeded the scaled equivalent fan’s performance for the lowest platform height and the most adverse conditions. This proves the effectiveness of the replacement fan as an edge fan. / AFRIKAANSE OPSOMMING: Groot lugverkoelde warmteruilers maak gebruik van waaier matrikse om verkoeling te voorsien. Die verrigting van die waaieropstelling word egter negatief beïnvloed deur versteurde inlaattoestande. Gegewe die grootte van hierdie waaiers is daar geen praktiese metode om die nadelige invloed van die versteurde inlaattoestande te kwantifiseer nie. Hierdie dien as motivering vir die ontwikkeling van ‘n multi‐waaier toetsplatform wat daartoe in staat is om die gedrag van die volskaal opstelling na te boots maar op ‘n baie meer meetbare en hanteerbare skaal. Hierdie ondersoek was uitgevoer in twee dele. ‘n Reynoldsgetal ondersoek was uitgevoer om te bepaal watter impak dit sal hê op die skaalwaaier verrigting indien die Reynolds getaleffekte ingesluit word in die toetswerk. Berekende vloeidinamka (BVD) was gebruik om te bepaal waar turbulente oorgang voorkom op die volskaal waaierlemme en pooitjiedrade was geselekteer en geposisioneer hiervolgens om dieselfde oorgang te herskep. Vanuit hierdie aanvanklike ondersoek was dit gevind dat die pooitjiedrade ‘n algehele afname in verrigting tot gevolg het, maar wanneer dit oorkom word deur die lemhoek op te stel, die gedrag en verrigting van die gepooitjiede waaier soortgelyk is aan die van die nie‐gepooitjiede waaier. Hierdie gedrag stel voor dat die Reynoldsgetalle maar geïgnoreer kan word vir hierdie skaalwaaier toetswerk. Gelykwaardige skaalwaaiers van die wat in bedryf is in volskaal opstellings was getoets in ‘n drie ry multi‐waaier opstelling wat, deur simmetrie, verteenwordigend is van ‘n tipiese straat in ‘n volskaal opstelling. Verskeie platformhoogtes was gesimuleer en die ooreenstemmende afname in stelsel verrigting was vergelyk met ‘n model wat geformuleer is juis om te voorspel hoe die volumetriese effektiwiteit afneem met platformhoogte. Die model wyk af van die toetsdata by baie lae platform hoogtes. ‘n Vervangingswaaier vir die aanvanklike geskalleerde waaier was getoets as ‘n randwaaier, die waaier wat die ergste benadeel word deur versteurde inlaattoetstande, in die multi‐waaier opstelling. Die vervangingswaaier het soortgelyk aan die aanvanklike waaier verrig vir meeste platformhoogtes, maar oortref die aanvanklike waaier se werksverrigting by die laagste platformhoogte en mees ongunstige toestande. Hierdie bewys die vermoëns van die vervangingswaaier as ‘n randwaaier.

Deep learning has gained traction in thermal engineering due to its applications to process simulations, the deeper insights it can provide and its abilities to circumvent the shortcomings of classic thermodynamic simulation approaches by capturing complex inter-dependencies. This works sets out to apply probabilistic deep learning to power plant operations using historic plant data. The first study presented, entails the development of a steady-state mixture density network (MDN) capable of predicting effective heat transfer coefficients (HTC) for the various heat exchanger components inside a utility scale boiler. Selected directly controllable input features, including the excess air ratio, steam temperatures, flow rates and pressures are used to predict the HTCs. In the second case study, an encoder-decoder mixturedensity network (MDN) is developed using recurrent neural networks (RNN) for the prediction of utility-scale air-cooled condenser (ACC) backpressure. The effects of ambient conditions and plant operating parameters, such as extraction flow rate, on ACC performance is investigated. In both case studies, hyperparameter searches are done to determine the best performing architectures for these models. Comparisons are drawn between the MDN model versus standard model architecture in both case studies. The HTC predictor model achieved 90% accuracy which equates to an average error of 4.89 W m2K across all heat exchangers. The resultant time-series ACC model achieved an average error of 3.14 kPa, which translate into a model accuracy of 82%.

The performance of parabolic trough based solar power plants over the last 25 years has proven that this technology is an excellent alternative for the commercial power industry. Compared to conventional power plants, parabolic trough solar power plants produce significantly lower levels of carbon dioxide, although additional research is required to bring the cost of concentrator solar plants to a competitive level. The cost reduction is focused on three areas: thermodynamic efficiency improvements by research and development, scaling up of the unit size, and mass production of the equipment. The optimum design, performance simulation and cost analysis of the parabolic trough solar plants are essential for the successful implementation of this technology. A detailed solar power plant simulation and analysis of its components is needed for the design of parabolic trough solar systems which is the subject of this research. Preliminary analysis was carried out by complex models of the solar field components. These components were then integrated into the system whose performance is simulated to emulate real operating conditions. Sensitivity analysis was conducted to get the optimum conditions and minimum levelized cost of electricity (LCOE). A simplified methodology was then developed based on correlations obtained from the detailed component simulations. A comprehensive numerical simulation of a parabolic trough solar power plant was developed, focusing primarily on obtaining a preliminary optimum design through the simplified methodology developed in this research. The proposed methodology is used to obtain optimum parameters and conditions such as: solar field size, operating conditions, parasitic losses, initial investment and LCOE. The methodology is also used to evaluate different scenarios and conditions of operation. The new methodology was implemented for a 50 MWe parabolic trough solar power plant for two cities: Tampa and Daggett. The results obtained for the proposed methodology were compared to another physical model (System Advisor Model, SAM) and a good agreement was achieved, thus showing that this methodology is suitable for any location.

An approach to hybrid cooling technique is proposed using air jets which impinge on a triangular grooved surface with dry grooves and grooves containing water. One major application is for condensers of thermoelectric power plants. The heat and mass transfer analogy were successfully used to evaluate the simultaneous heat and mass transfer. Results showed that hybrid jet impingement produced high heat flux levels at low jet velocities and flow rates. Experimental results were used to characterize the resulting heat transfer under different conditions such as flow open area percentage, array orifices diameter and array to surface stand-off distance. The results have shown that jet impingement is capable of delivering high transfer rates with lower cooling cost rates compared to current industry conventional techniques. Water is efficiently used in hybrid jet impingement because evaporative energy is absorbed directly from the surface instead of cooling air to near wet-bulb temperature. / Master of Science / Array jet impingement cooling experiments were conducted on a triangular grooved surface with the surface at a constant temperature. Results showed that jet impingement can provide high transfer rates with lower rates of cooling cost in comparison to contemporary conventional techniques in the industry. Experiments on the triangular grooved surfaces were performed at dry and wet surface conditions. Under the dry conditions, the objective is to characterize the resulting heat transfer under varying operational conditions such as jet speed, array orifice diameter, array to surface stand-off distance, and flow open area percentage. Results from the triangular surface when dry showed less improvement in heat transfer than the rectangular grooved surface. A hybrid cooling technique approach was proposed and developed by using air jets impinging on a triangular grooved surface with the grooves containing water. The approach is being suggested and experimentally tested for its viability as an alternative to thermoelectric power plant cooling towers. Convection heat and mass transfer coefficients were experimentally measured for different wet coverage of the surface. Results showed that the hybrid jet impingement produced high heat flux levels at low jet velocities and flow rates. The highest heat transfer was consistently found with a 50% coverage of the surface. Hybrid jet impingement showed an improvement up to 500% in heat transfer as compared to jet impingement on a dry grooved surface.

Array jet impingement cooling experiments were performed on flat and grooved surfaces with the surface at a constant temperature. For the flat surface, power and temperature measurements were performed to obtain convection coefficients under a wide range of operating conditions such as jet speed, orifice to surface stand-of distance, and open area percentage. Cooling performance (CP) was calculated as the ratio between heat transfer and fan power. An empirical model was developed to predict jet impingement heat transfer taking into account the entrainment effects. Experimental results showed that jet impingement can provide high transfer rates with lower rates of cooling cost in comparison to contemporary conventional techniques in the industry. CP values over 279 were measured which are significantly higher than the standard values of 70 to 95 in current technology. The model enhanced prediction accuracy by taking into account the entrainment effects; an effect that is rarely considered in the literature. Experiments on the grooved surfaces were performed at dry and wet surface conditions. Under dry conditions, results showed 10%~55% improvement in heat transfer when compared to the flat surface. Improvement percentage tends to be higher at wider gaps between the array of orifices and the grooved surface. An improvement of 30%~40% was observed when increasing Re either by increasing orifice diameter or jet speed. Similar improvement was observed at higher flow open area percentages. No significant improvement in heat transfer resulted from decreasing the size of the grooves from 3.56mm to 2.54mm. Similarly, no noticeable change in heat transfer resulted from changing the relative position of the jets striking the surface at the top of the grooves to the bottom of the grooves. Deeper grooves with twice the depth gave statistically similar average heat transfer coefficients as shallower grooves. Under wet conditions, a hybrid cooling technique approach was proposed by using air jets impinging on a grooved surface with the grooves containing water. The approached is proposed and evaluated experimentally for its feasibility as an alternative for cooling towers of thermoelectric power plants. Convection heat and mass transfer coefficients were measured experimentally using the heat mass transfer analogy. Results showed that hybrid jet impingement provided high magnitudes of heat flux at low jet speeds and flow rates. High coefficients of performance CP > 3000, and heat fluxes > 8,000W/m2 were observed. Hybrid jet impingement showed 500% improvement as compared to jet impingement on a dry flat surface. CP values of hybrid jet impingement is 600% to 1,500% more as compared to performance of air-cooled condensers and wet cooling towers. Water use for hybrid jet impingement cooling is efficient since evaporation energy is absorbed from the surface directly instead of cooling air to near wet-bulb temperature. / PHD / This thesis explored the possibility of using air jets on the outside surface of a device that is used to condense steam. An experiment apparatus was used to imitate the conditions of steam condensation in the lab. A flat metallic surface was heated by placing an electric heater beneath it. The metallic surface was cooled using air jets coming out of orifices situated above the hot metallic surface. A fan, connected to an electric motor, was used to create the air jets. The amount of heat transfer was measured by measuring the electric power the heater consumed. This measured power was compared to the power needed to run the fan. The ratio of heat transfer to fan power is called the coefficient of performance CP. The CP values of more than 200 were obtained when air jets were used meaning that we need one kilowatt of mechanical power to remove 200 kilowatts of heat. This CP value is 300% more than the current technology used in the industry where CP ranges from 70 to 90. This means that we can build very efficient steam condensers for power plants. This type of condensers that uses air jets allows the power plant to be efficient and to be able to increase the amount of power generated without extra cost. Further enhancement of the CP can be achieved by making the hot surface grooved instead of flat with the grooves containing water. Air jets, coming out of orifices situated above the grooved surface, were used for cooling. The CP values of more than 3,000 were obtained when air jets were used with wet grooved surface. This CP values is 1,500% more than the current technology used in the industry. This type of condensers that uses air jets on wet grooves allows the power plant to be efficient and to be able to tremendously increase the amount of power generated without extra power and water costs.