Two computer programs were developed and used to determine the optimum operating parameters of a fedbatch and a continuous two-stage process for fermentation of recombinant bacteria. The study was conducted in three phases: (a) developing two computer programs for simulation and optimization of the above processes, (b) conducting batch culture fermentations to verify the performance of the biokinetic model, and (c) conducting fedbatch and two-stage continuous fermentation experiments to closely examine the simulation and optimization results. The Miao and Kompala (1992) biokinetic model was used for simulation of the bacterial growth and cloned gene expression. The Pattern-Search method, developed by Hooke and Jeeves (1962), was incorporated in the programs to determine the optimum values of the parameters. Extensive studies of the optimization results showed 30-40% higher productivities for the two stage continuous process over the fedbatch process when using the same media in both processes. In addition, increasing the number of stages in the continuous two-stage process resulted in very limited improvement in the productivity of the process (10-12%). The information from the process optimization was then used to design batch, fedbatch nd two stage continuous experiments. Recombinant <i>E. coli </i>(strain BL21DE3) with an inducible gene (sensitive to IPTG, isopropyl-â-D-thiogalactopyranoside) was used throughout the experiments. The experimental results from the fedbatch and two stage continuous processes clearly showed good agreement with the simulation and optimization results $(\cong$15% deviation). The experiments also revealed that the maintenance of plasmid harboring cells over the long-term operation could be an important barrier in achieving the predicted high productivity in the two stage continuous process. Finally, in addition to computer programs for optimization of genetically modified microorganisms, a new computer program with a generic algorithm for optimization of multiple CFSTR fermentation with any kind of biokinetic model was developed. The program was used to optimize multiple CFSTRs with the cybernetic biokinetic model for the first time. Besides finding the optimum residence times for multiple CFSTRs operation, the effect of inaccuracies in different cybernetic model parameters on the overall productivity of the process was investigated. The simulation results illustrated that, a single CFSTR was more sensitive in its operation to inaccuracies in the biokinetic constants as compared to optimized CFSTRs in series (2-8 times more sensitive).

The thesis presents a framework for integrating models, simulation, and experimental data to diagnose incipient failure modes and prognosticate the remaining useful life of critical components, with an application to the main transmission of a helicopter. Although the helicopter example is used to illustrate the methodology presented, by appropriately adapting modules, the architecture can be applied to a variety of similar engineering systems. Models of the kind referenced are commonly referred to in the literature as physical or physics-based models. Such models utilize a mathematical description of some of the natural laws that govern system behaviors. The methodology presented considers separately the aspects of diagnosis and prognosis of engineering systems, but a similar generic framework is proposed for both. The methodology is tested and validated through comparison of results to data from experiments carried out on helicopters in operation and a test cell employing a prototypical helicopter gearbox. Two kinds of experiments have been used. The first one retrieved vibration data from several healthy and faulted aircraft transmissions in operation. The second is a seeded-fault damage-progression test providing gearbox vibration data and ground truth data of increasing crack lengths. For both kinds of experiments, vibration data were collected through a number of accelerometers mounted on the frame of the transmission gearbox. The applied architecture consists of modules with such key elements as the modeling of vibration signatures, extraction of descriptive vibratory features, finite element analysis of a gearbox component, and characterization of fracture progression. Contributions of the thesis include: (1) generic model-based fault diagnosis and failure prognosis methodologies, readily applicable to a dynamic large-scale mechanical system; (2) the characterization of the vibration signals of a class of complex rotary systems through model-based techniques; (3) a reverse engineering approach for fault identification using simulated vibration data; (4) the utilization of models of a faulted planetary gear transmission to classify descriptive system parameters either as fault-sensitive or fault-insensitive; and (5) guidelines for the integration of the model-based diagnosis and prognosis architectures into prognostic algorithms aimed at determining the remaining useful life of failing components.

Thesis (D.Tech.: Mechanical Engineering)-Dept. of Mechanical Engineering, Durban Institute of Technology, 2005 xxi, 265 leaves / The aim of this study is to develop, using neural networks, a model to aid the performance monitoring of operational diesel engines in industrial settings. Feed-forward and modular neural network-based models are created for the prediction of the specific fuel consumption on any normally aspirated direct injection four-stroke diesel engine. The predictive capability of each model is compared to that of a published quadratic method. Since engine performance maps are difficult and time consuming to develop, there is a general scarcity of these maps, thereby limiting the effectiveness of any engine monitoring program that aims to manage the fuel consumption of an operational engine. Current methods applied for engine consumption prediction are either too complex or fail to account for specific engine characteristics that could make engine fuel consumption monitoring simple and general in application. This study addresses these issues by providing a neural network-based predictive model that requires two measured operational parameters: the engine speed and torque, and five known engine parameters. The five parameters are: rated power, rated and minimum specific fuel consumption bore and stroke. The neural networks are trained using the performance maps of eight commercially available diesel engines, with one entire map being held out of sample for assessment of model generalisation performance and application validation. The model inputs are defined using the domain expertise approach to neural network input specification. This approach requires a thorough review of the operational and design parameters affecting engine fuel consumption performance and the development of specific parameters that both scale and normalize engine performance for comparative purposes. Network architecture and learning rate parameters are optimized using a genetic algorithm-based global search method together with a locally adaptive learning algorithm for weight optimization. Network training errors are statistically verified and the neural network test responses are validation tested using both white and black box validation principles. The validation tests are constructed to enable assessment of the confidence that can be associated with the model for its intended purpose. Comparison of the modular network with the feed-forward network indicates that they learn the underlying function differently, with the modular network displaying improved generalisation on the test data set. Both networks demonstrate improved predictive performance over the published quadratic method. The modular network is the only model accepted as verified and validated for application implementation. The significance of this work is that fuel consumption monitoring can be effectively applied to operational diesel engines using a neural network-based model, the consequence of which is improved long term energy efficiency. Further, a methodology is demonstrated for the development and validation testing of modular neural networks for diesel engine performance prediction.

<div>The exponential growth of Internet of Things (IoT) and cyber-physical systems is resulting in complex environments comprising of various devices interacting with each other and with users. In addition, the rapid advances in Artificial Intelligence are making those devices able to autonomously modify their behaviors through the use of techniques such as reinforcement learning (RL). There is thus the need for an intelligent monitoring system on the network edge with a global view of the environment to autonomously predict optimal device actions. However, it is clear however that ensuring safety and security in such environments is critical. To this effect, we develop a constrained RL framework for IoT environments that determines optimal devices actions with respect to user-defined goals or required functionalities using deep Q learning. We use anomaly based intrusion detection on the network edge to dynamically generate security and safety policies to constrain the RL agent in the framework. We analyze the balance required between ‘safety/security’ and ‘functionality’ in IoT environments by manipulating the exploration of safe and unsafe benefit state spaces in the RL framework. We instantiate the framework for testing on application layer control in smart home environments, and network layer control including network functionalities like rate control and routing, for SDN based environments.</div>

Indiana University-Purdue University Indianapolis (IUPUI) / A gearless hydraulic wind energy transfer system utilizes the hydraulic power transmission principles to integrate the energy of multiple wind turbines in a central power generation location. The gearless wind power transfer technology may replace the current energy harvesting system to reduce the cost of operation and increase the reliability of wind power generation. It also allows for the integration of multiple wind turbines to one central generation unit, unlike the traditional wind power generation with dedicated generator and gearbox. A Hydraulic Transmission (HT) can transmit high power and can operate over a wide range of torque-to-speed ratios, allowing efficient transmission of intermittent wind power. The torque to speed ratios illustrates the relationship between the torque and speed of a motor (or pump) from the moment of start to when full-load torque is reached at the manufacturer recommended rated speed. In this thesis, a gearless hydraulic wind energy harvesting and transfer system is mathematically modeled and verified by experimental results. The mathematical model is therefore required to consider the system dynamics and be used in control system development. Mathematical modeling also provided a method to determine the losses of the system as well as overall efficiency. The energy is harvested by a low speed-high torque wind turbine connected to a high fixed-displacement hydraulic pump, which is connected to hydraulic motors. Through mathematical modeling of the system, an enhanced understanding of the HTS through analysis was gained that lead to a highly efficient hydraulic energy transmission system. It was determined which factors significantly influenced the system operation and its efficiency more. It was also established how the overall system operated in a multiple wind turbine configuration. The quality of transferred power from the wind turbine to the generator is important to maintaining the systems power balance, frequency droop control in grid-connected applications, and to ensure that the maximum output power is obtained. A hydraulic transmission system can transfer large amounts of power and has more flexibility than a mechanical and electrical system. However high-pressure hydraulic systems have shown low efficiency in wind power transfer when interfaced with a single turbine to a ground-level generator. HT’s generally have acceptable efficiency at full load and drop efficiency as the loading changes, typically having a peak around 60%. The efficiency of a HT is dependent on several parameters including volumetric flow rate, rotational speed and torque at the pump shaft, and the pressure difference across the inlet and outlet of the hydraulic pump and motor. It has been demonstrated that using a central generation unit for a group of wind turbines and transferring the power of each turbine through hydraulic system increases the efficiency of the overall system versus one turbine to one central generation unit. The efficiency enhancement depends on the rotational speed of the hydraulic pumps. Therefore, it is proven that the multiple-turbine hydraulic power transfer system reaches higher efficiencies at lower rotational speeds. This suggests that the gearbox can be eliminated from the wind powertrains if multiple turbines are connected to the central generation unit. Computer simulations and experimental results are provided to quantify the efficiency enhancements obtained by adding the second wind turbine hydraulic pump to the system.

Indiana University-Purdue University Indianapolis (IUPUI) / The Pedestrian Automatic Emergency Braking System (PAEB) can utilize on-board sensors to detect pedestrians and take safety related actions. However, PAEB system only benefits the individual vehicle and the pedestrians detected by its PAEB. Additionally, due to the range limitations of PAEB sensors and speed limitations of sensory data processing, PAEB system often cannot detect or do not have sufficient time to respond to a potential crash with pedestrians. For further improving pedestrian safety, we proposed the idea for integrating the complimentary capabilities of V2V and PAEB (V2V-PAEB), which allows the vehicles to share the information of pedestrians detected by PAEB system in the V2V network. So a V2V-PAEB enabled vehicle uses not only its on-board sensors of the PAEB system, but also the received V2V messages from other vehicles to detect potential collisions with pedestrians and make better safety related decisions. In this thesis, we discussed the architecture and the information processing stages of the V2V-PAEB system. In addition, a comprehensive Matlab/Simulink based simulation model of the V2V-PAEB system is also developed in PreScan simulation environment. The simulation result shows that this simulation model works properly and the V2V-PAEB system can improve pedestrian safety significantly.