Turbogenerator System Research Articles

Bearing Faults Diagnosis of a Turbogenerator System Using Machine Learning Training Functions

Bearing Faults Diagnosis of a Turbogenerator System Using Machine Learning Training Functions

Modeling, Analysis, and Identification of Parallel and Angular Misalignments in a Coupled Rotor-Bearing-Active Magnetic Bearing System

Abstract The standard techniques used to detect misalignment in rotor systems are loopy orbits, multiple harmonics with predominant 2X component, and high axial vibration. This paper develops a new approach for the identification of misalignment in coupled rotor systems modeled using two-node Timoshenko beam finite elements. The coupling connecting the turbine and generator rotor systems is modeled by a stiffness matrix, which has both static and additive components. While the magnitude of static stiffness component is fixed during operation, the time-varying additive stiffness component displays a multiharmonic behavior and exists only in the presence of misalignment. To numerically simulate the multiharmonic nature coupling force/moment as observed in experiments, a pulse wave is used as the steering function in the mathematical model of the additive coupling stiffness (ACS). The representative turbogenerator (TG) system has two rotor systems, each having two disks and supported on two flexible bearings—connected by coupling. An active magnetic bearing (AMB) is used as an auxiliary bearing on each rotor for the purposes of vibration suppression and fault identification. The formulation of mathematical model is followed by the development of an identification algorithm based on the model developed, which is an inverse problem. Least-squares linear regression technique is used to identify the unbalances, bearing dynamic parameters, AMB constants, and importantly the coupling static and additive stiffness coefficients. The sensitivity of the identification algorithm to signal noise and bias errors in modeling parameters has been tested. The novelty of paper is the representation and identification of misalignment using the ACS matrix coefficients, which are direct indicators of both type and severity of the misalignment.

Performance analysis and availability optimization to improve maintenance schedule for the turbo-generator subsystem of a thermal power plant using particle swarm optimization

The work reported in this paper evaluates the performance of the turbo-generator subsystem of a thermal power plant by employing a particle swarm optimization method. The majorly selected systems considered for the study are turbine governing, turbine lubrication, generator oil system, generator gas system, and generator excitation system. The Markov based availability modeling is adopted for the turbo-generator system. The Markov probabilistic approach based availability simulation model is reported. Through this study, the Markov based results reveal that the turbine lubrication and generator excitation subsystems affect the system availability at most. Therefore, these critical subsystems are on higher priority from a maintenance perspective. Furthermore, system availability is optimized to get the optimized availability parameters using the particle swarm optimization method. The PSO based results revealed that the maximum optimized availability of 98.9394 % is obtained for 30 particle size. The results reveal that with an increase in particle size of more than 30, the system availability remains constant. Further, the optimized availability parameters are used to recommend a suitable maintenance strategy for the plant's turbo-generator system.

Characteristic Parameters Estimation of Active Magnetic Bearings in a Coupled Rotor System

Abstract Present research inspects the performance of rotor–bearing–coupling system in the presence of active magnetic bearings (AMBs). A methodology is suggested to quantify various fault characteristics along with AMB characteristic parameters of a coupled turbine generator system. A simplest possible turbogenerator system is modeled to analyze coupling misalignment. Conventional methodology to estimate dynamic system parameters based on forced response information is not enough for AMB-integrated rotor system because it requires current information along with displacement information. The controlling current of AMB is tuned and controlled with a controller of proportional–integral–derivative (PID) type. A numerical technique (Lagrange's equation) is applied to get equations of motion (EOM). Runge–Kutta technique is used to obtain EOM to acquire the time domain responses. The fast Fourier transformation (FFT) is applied on obtained responses to acquire responses in the frequency domain, and full spectrum technique is applied to propose the methodology. A methodology that depends on the least squares regression approach is proposed to evaluate the multifault parameters of AMB-integrated rotor system. The robustness of the algorithm is checked against various levels of noise and modeling error and observed efficient. An appreciable reduction in misalignment forces and moments is observed by using AMBs.

Study of the electric-booster and turbo-generator system and its influence on a 1.5 L gasoline engine

To effectively recovery the energy in exhausts and improve the transient performance of internal combustion engine. In this paper, a new electric turbo-compounding layout called electric-booster and turbo-generator system (EBTG) was established and studied in a 1.5 L gasoline engine. In the EBTG system, the turbo and compressor operate independently. Its impact on a gasoline engine was analyzed from the perspective of fuel economy, power output, availability distribution and transient performance. Results show that the proposed EBTG system is effective to improve the thermal efficiency of engine at high-speed and wide-open throttle conditions. Simultaneously, it is also valuable to solve the turbo-lag of engine. By applying EBTG system, the maximum reduction of specific fuel consumption was 2.6%. Exergy analysis showed that the EBTG system can reduce the irreversibility in turbine and recovery more exergy from exhausts. There was a 0.8 percentage point exergy efficiency increment at 4400 rpm and full-load operation. The step response time of boost pressure in EBTG engine decreased to about half of it in baseline at 2000 rpm and full-load condition.

Experimental identification of shaft misalignment in a turbo-generator system

Precise and authentic estimation of the dynamic features of rotating machines and prevention of failure requires accurate experimental characterisation of their critical components, for example, bearings and couplings. These are difficult to model theoretically, and they often suffer from uncertain parameters in their model. Especially, when there is a misalignment in the rotor system, dynamic characterisation of bearings and couplings changes drastically. In the present study, multiple fault parameters (MFPs) of critical components of turbo-generator, that is, bearing and coupling together with residual unbalances (RUs), are evaluated experimentally using model-based methodology. A test rig was developed and used for experimentation in which different levels of misalignment was introduced. After estimating the MFPs, then the accuracy was checked through an impact test on the rotor test rig. The effect of different levels of misalignments on estimated parameters was studied.

Root Cause Analysis of Arcing in Retaining Rings of Turbogenerators

Retaining rings are an important, and the most highly stressed, component of the entire turbogenerator system. Arcing in retaining rings is a very serious problem as it can easily escalate to a full-blown failure. In this paper, we investigate this arcing problem in the retaining rings. We determine the most likely mechanism by which arcing occurs and the type of events that lead to it. Specifically, we try to test two different mechanisms that could have led to the indications of arc. The first one is sparking due to high field effects (or high voltage gradient across the contact junction) and the other is a make-and-break contact arcing owing to the presence of inductance in the system. Experiments performed to measure the contact resistance between the retaining ring and mild iron piece indicate that even very high fault currents cannot produce enough voltage drop to cause sparking. Transient three-dimensional finite-element simulations show that a break in a relatively small contact region, which can be part of a bigger contact area, can generate localized voltage spikes that are high enough to lead to arcing. Interestingly, this can happen at relatively low currents. Furthermore, hardware experiments in the laboratory confirm the possibility of generation of arcs due to a make-and-break contact at low currents.

Quantification of multiple fault parameters in flexible turbo-generator systems with incomplete rundown vibration data

A turbo-generator system of the modern rotating machinery consists of the driver and driven shafts, which are coupled through flexible couplings and mounted on flexible bearings. Dynamic characterisation of vital machine elements of such rotating machinery is a challenging problem for reliable and accurate response predictions. In this paper, a key intention is to estimate the bearing and coupling dynamic parameters along with residual unbalances at predefined planes, and the misalignment forces and moments at the coupling based on the rundown vibration data. To tackle a practical difficulty of limited measurements and a numerical difficulty of the conventional dynamic condensation in the development of identification algorithm, a novel condensation technique has been implemented especially to overcome measurement of transverse rotational DOFs. Numerical examples are also presented to show the effectiveness of the proposed method. The measurement noise has been added in numerically simulated responses that are used in the present algorithm to identify the parameters and it is found to be robust. Modelling errors of few physical parameters are also considered and estimates are found to be very good.

Application of dynamic programming to the optimal management of a hybrid power plant with wind turbines, photovoltaic panels and compressed air energy storage

A model for thermo-economic analysis and optimization of a hybrid power plant consisting of compressed air energy storage (CAES) coupled with a wind farm and a photovoltaic plant is presented. This kind of plant is aiming to overcome some of the major limitations of renewable energy sources, represented by their low power density and intermittent nature, largely depending upon local site and unpredictable weather conditions.In CAES, energy is stored in the form of compressed air in a reservoir during off-peak periods, while it is used on demand during peak periods to generate power with a turbo-generator system. Such plants can offer significant benefits in terms of flexibility in matching a fluctuating power demand, particularly when coupled with renewable sources, characterized by high and often unpredictable variability.A mathematical model, validated in a previous study over the CAES plant in Alabama, US, is coupled with a dynamic programming algorithm to achieve the optimal management of the plant, in order to minimize operational costs while satisfying constraints related to the operation of reservoir, compressors and turbines, also considering their off-design performance. The potential benefits of such plant in terms of energy consumption and CO2 emission are analyzed and discussed, for different configurations and scenarios.

Influence of Large Turbo-Generator Stator Ventilation Ducts Structural Changes on Stator Temperature

For the complex status of fluid flow in stator radial ventilation ducts of large turbo-generator, the temperature distribution of stator is dramatically affected by the flow status of cooling medium in stator ventilating ducts. In this paper, a new ventilating ducts structure in stator is investigated. According to fixing a wind deflector on the stator teeth adjacent to the ventilation ducts, the fluid flow status of cooling air is changed flowing in stator ventilation ducts. For this reason, the effect of heat transfer in stator is changed. Taking an air-cooled turbo-generator as an example, considering the characteristics of fluid flow and heat transfer in turbo-generator ventilation system, the three-dimensional fluid flow and heat transfer coupling model is established. Using finite volume method, three-dimensional fluid field and temperature field control equations are coupling solved. Based on this, the velocity distribution in ventilating ducts is obtained. Besides that, the velocity distribution is studied with the cooling air flows into radial ventilation ducts at different incident angles. The influences of wind deflector and incident angles on the fluid velocity and temperature distribution are analyzed. Based on that, some useful conclusions are obtained.

Application of HHT for Online Detection of Inter-Area Short Circuits of Rotor Windings of Turbo-Generators Based on the Thermodynamics Modeling Method

This paper focuses on monitoring and predicting the short circuit faults of the rotor windings of large turbo-generator systems. For the purpose of increasing efficiency and decreasing maintenance cost, a method that combines the HHT (Hilbert Huang Transform) with a wavelet has been studied. This method is based on analyzing a classical Albright detecting coil. Due to the Empirical Mode Decomposition (EMD) and the Intrinsic Mode Functions (IMF) of the HHT the exact location of a short circuit of rotor windings may be given. However, a part of the useful information is eliminated by the unreasonable decomposing scale of the wavelet. Based on the thermodynamics modeling method, this study was illustrated with a 50㎿ turbo-generator system that is installed in Northern China. The analysis results, which have very good agreement with those of a previous study, show that the method of combining the HHT with a wavelet is an effective way to analyze and predict the short circuit faults of the rotor windings of large generators, such as supercritical turbo-generator systems and wind turbo-generator systems. This work can offer a useful reference for analyzing smart grids by improving the power quality of a distribution network that is supplied by a turbo-generator system.

Study of differential control method for solving chaotic solutions of nonlinear dynamic system

In this paper, we focus on the need to solve chaotic solutions of high-dimensional nonlinear dynamic systems of which the analytic solution is difficult to obtain. For this purpose, a Differential Control Method (DCM) is proposed based on the Mechanized Mathematics-Wu Elimination Method (WEM). By sampling, the computer time of the differential operator of the nonlinear differential equation can be substituted by the differential quotient of solving the variable time of the sample. The main emphasis of DCM is placed on substituting the differential quotient of a small neighborhood of the sample time of the computer system for the differential operator of the equations studied. The approximate analytical chaotic solutions of the nonlinear differential equations governing the high-dimensional dynamic system can be obtained by the method proposed. In order to increase the computational efficiency of the method proposed, a thermodynamics modeling method is used to decouple the variable and reduce the dimension of the system studied. The validity of the method proposed for obtaining approximate analytical chaotic solutions of the nonlinear differential equations is illustrated by the example of a turbo-generator system. This work is applied to solving a type of nonlinear system of which the dynamic behaviors can be described by nonlinear differential equations.

Demonstration of a Palm-Sized 30 W Air-to-Power Turbine Generator

A compact, high power density turbo-generator system was conceived, designed, and experimentally tested. The air-to-power (A2P) device with a nominal design point of 50 W electric power output operates on high pressure air such as from a plant pneumatic system or from a portable bottle of pressurized air. A concept design study was first carried out to explore the design space for a range of output power at cost efficiency levels specified in collaboration with industry. The cost efficiency is defined as the cost of electrical power over the cost of pressurized air. The key challenge in the design is the relatively low power demand of 50 W while operating at high supply pressures of nominally 5–6 bars. To meet the cost efficiency goal under these conditions, a high-speed turbine and generator (∼450,000 rpm) are required with small blade span (∼200 μm), minimizing the mass flow while achieving the highest possible turbine performance. Since turbines with such small turbomachinery blading are not commercially available, a silicon-based micro electro mechanical systems (MEMS) turbine was designed using 2D and 3D computational fluid dynamics (CFD) computations. To reduce the development time, existing and previously demonstrated custom-made generator and ceramic ball bearing technology were used, resulting in a compact A2P proof-of-concept demonstration. The cylindrical device of 35 mm diameter resembles a tube fitting with a standard M24 adapter. Without load, a top turbine speed of 475,000 rpm was demonstrated, exceeding the design specification. Using load resistors, the proof-of-concept A2P device achieved 30 W of electrical power at 360,000 rpm and a turbine efficiency of 47%, meeting the cost efficiency goal. Higher speeds under load could not be achieved due to thrust load limitations of the off-shelf ball bearings. The demonstrated performance is in good agreement with the projected CFD based predictions. To the authors’ knowledge, this is the first successful demonstration of a self-contained, 50 W class turbo-generator of hybrid architecture where a MEMS turbine disk is joined with a precision machined titanium shaft and aluminum housing.

PID-Like Fuzzy Logic Control for a Turbogenerator

In this paper, PlO-Like fuzzy logic control has designed to control the excitation voltage of a single-input single-output (SISO) turbogenerator system connected to an infinite busbar to demonstrate the effectiveness and robustness of these design approach. The performance of this controller is evaluated in comparison with a conventional proportional, integral and derivative (PID) controller in simulation. Also, the performance evaluated from the responses to both small and large disturbances, including short circuit on the transmission lines. It is found that, the PlO-like fuzzy logic controller is much less sensitive to changes in the operating points of the system.

Individual channel analysis of the turbogenerator with a power system stabilizer

A complete new insight into why excitation-governor control with power system stabilizers (PSSs) has been so successful for the control of turbogenerators connected to an infinite bus is provided by the small-signal multivariable analysis framework, individual channel analysis and design (ICAD). The multivariable analysis justifies treating the turbogenerator system as a pseudo-single-input single-output (SISO) system where the governor loop is first closed and the exciter loop is treated as a SISO system for the prime purpose of rejecting voltage disturbances. The function of the PSS is identified as that of overcoming an awkward switch-back frequency-domain characteristic of the excitation channel so as to - 1 permit high-performance excitation channel bandwidths up to 10 rad s that otherwise could not be obtained. Thus, in addition to the control requirements of set point regulation of the terminal voltage and shaft speed, the PSS provides for a second control requirement of strong voltage disturbance rejection over the - 1 important frequency range 0-10 rad s . The PSS control option is also assessed against other control options. Several other results concerning stability robustness to system uncertainties in different system configurations follow from the analysis in a transparent and immediate way.

Reinforcement learning control using interconnected learning automata

This paper is concerned with the application of interconnected learning automata to the control of unknown stochastic dynamic systems. A method of using N learning automata and subsets of actions to search for optimum controller parameters according to a given performance index is presented. The methodology has been applied as an on-line control strategy for control of an unknown turbogenerator system in noise environments without persistent excitation signals. Simulation results are included.

Transputer implementation of adaptive control for a turbogenerator system

The paper describes the design and implementation of digital self-tuning automatic voltage regulators for excitation control of a synchronous generator. Practical aspects such as controller robustness and processing power are discussed. Conventional microprocessor technology is unable to achieve the sampling periods required for industrial application. A parallel architecture using INMOS transputers combined with an advanced VME bus-based measurement system provide an open-ended solution to this problem, and enable the adaptive controllers to operate within the required 10 ms sample period. The scheme was initially evaluated on a computer simulation of a three-phase turboalternator. A laboratory model turbogenerator was then controlled to test the performance and reliability of the system under realistic conditions.

Transputer implementation of adaptive control for a turbogenerator system

Transputer implementation of adaptive control for a turbogenerator system

Parallel adaptive control for a turbogenerator system

Parallel adaptive control for a turbogenerator system

Integration of turbo-generator modules in digital transient network analyzer

The behavior of a small power system consisting of two interconnected generators is simulated in real-time by a prototype digital transient network analyzer (TNA). The prototype digital TNA consists of two computational modules and one I/O module. The modules communicate with each other through ribbon cables. Each computational module simulates one turbo-generator, its transformer, its governor, exciter, and power system stabilizer systems. The numerical integration is shared by two TMS320C30 DSPs at a step-size of 100 /spl mu/s in realtime. The I/O module post-processes the state variables and presents selected information for analog display. The paper presents oscillograms from a test program which includes symmetry checks and behavioral checks against well known waveforms of hunting oscillations, synchronization out-of-phase torques, and subsynchronous resonance phenomena. The success of the digital TNA depends on: (a) the theoretical method of decoupled partitioning so that different portions of the power system can be allocated to different DSP-modules, (b) the architecture of the DSP-modules which can communicate the numerical integration results of one module to its contiguous neighbors with minimum delay. >