Burak Barutçu

Comparisons between the various types of neural networks with the data of wide range operational conditions of the Borssele NPP

This paper addresses a trend monitoring in operating nuclear power plant by use of two types of Recurrent Neural Networks (RNN). The interesting feature of the RNN is intrinsic dynamic memory that reflects the current output as well as the previous inputs and outputs are gradually quenched. The first one Elman type of RNN which has a feed-back from hidden layer to the input layer neurons while in the Jordan type, from the outputs of the neural net to the inputs of the neural net. In this paper the theoretical assessment of the both RNNs is given. Both topological structures including Back Propagation (BP) neural network were implemented to the Borssele NPP. Learning achieved from 30% to 100% nominal power at the starting period of the new core 30 September 2001. After learning period the reactor operation is followed by the neural network. Paper will present the reactor system, the real time data collection and the merits of the three types of the neural network applied while in the learning and continuous processing of the changing of the operational conditions.

Real time reactor noise diagnostics for the Borssele (PWR) nuclear power plant

Abstract After the upgrade of Borssele NPP in 1997, core cycle 24, the power plant operated three years more with 91% availability. The authority of the power plant decided to enhance and upgrade the reactor trend monitoring and plant information recording system with higher frequencies than the plant data processing system (PPS) as well as installing a flexible and multiple-purpose reactor noise analysis system which may support the reactor maintenance group with on-line and off-line capabilities for several different signal processing applications. Two measuring and monitoring systems were built in 2001 and fully taken in implementation during the start-up of the new core 28. In this sense, the new system was used in power operation during the 29 th of September 2001. This paper will introduce the measuring system, the operational tasks, and the results obtained so far on the real-time core-barrel motions (CBM) and the two-primary coolant pump vibrations measured through the reactor noise analysis.

On-line neuro-expert monitoring system for borssele nuclear power plant

Abstract A new method for an on-line monitoring system for the nuclear power plants has been developed utilizing the neural networks and the expert system. The integration of them is expected to enhance a substantial potential of the functionality as operators support. The recurrent neural network and the feed-forward neural network with adaptive learning are selected for the plant modeling and anomaly detection because of the high capability of modeling for dynamic behavior. The expert system is used as a decision agent, which works on the information space of both the neural networks and the human operators. The information of other sensory signals is also fed to the expert system, together with the outputs that the neural networks generate from the measured plant signals. The expert system can treat almost all known correlation between plant status patterns and operation modes as a priori set of rules. From the off-line test at Borssele Nuclear Power Plant (PWR 480 MWe) in the Netherlands, it was shown that the neuro-expert system successfully monitored the plant status. The expert system worked satisfactorily in diagnosing the system status by using the outputs of the neural networks and a priori knowledge base from the PWR simulator. The electric power coefficient is simultaneously monitored from the measured reactive and active electric power signals.

Artificial neural networks for dynamic monitoring of simulated-operating parameters of high temperature gas cooled engineering test reactor (HTTR)

Abstract This paper addresses to the problem of utilisation of the artificial neural networks (ANNs) for detecting anomalies as well as physical parameters of a nuclear power plant during power operation in real time. Three different types of neural network algorithms were used namely, feed-forward neural network (back-propagation, BP) and two types of recurrent neural networks (RNN). The data used in this paper were gathered from the simulation of the power operation of the Japans High Temperature Engineering Testing Reactor (HTTR). For the wide range of power operation, 56 signals were generated by the reactor dynamic simulation code for several hours of normal power operation at different power ramps between 30 and 100% nominal power. Paper will compare the outcomes of different neural networks and presents the neural network system and the determination of physical parameters from the simulated operating data.

72hr forecast of wind power in Mani̇sa, Turkey by using the WRF model coupled to WindSim

Wind power forecasting has recently become important to fulfill the increasing demand on energy usage. Two main approaches are applied to the wind power forecasting which can vary from 6 hours to 48 hours. One way is to model the atmosphere dynamically and the other method is to analyze wind speed and direction statistically. Although dynamical models forecast better than statistical models, since the former solves the problem physically, statistical models can be preferable when short term forecasting is needed due to their quick response time. Most of the currently available wind power forecasting systems analyzes the effect of wind field on wind power based on numerical weather prediction models. However, the resolution of these models is not sufficient enough when the scale of the turbines on the wind farms is considered. It is possible to overcome this problem by using computational fluid dynamics (CFD) models, which can provide both linear and nonlinear solutions on the turbine scale in terms of both wind speed and wind power forecasting. In this study, the WRF model is used for 72-hour wind speed and direction forecasting. The initial and boundary conditions of the model are provided by ECMWF operational forecasting data with the resolution of 0.25 degree. The WRF model is downscaled to 1 km resolution over Manisa Soma wind farm and 72-hour forecasts for each day of 2010 are accomplished. WindSim uses wind speed and direction values, which are solved on the nearest grid point of the WRF model to the location of a wind turbine, to simulate high-resolution wind speed values for 72hours. These WRF to WindSim coupled model results are compared to the wind power observations. As a result, we found that daily wind power generation errors per turbine vary between 90kW and 200kW for the seasons of spring, summer, and fall, whereas the error is about 150-350kW for winter. We also compared the errors of 24 hourly model outputs and we found that there is no significant difference among the first, the second, and the third 24 hourly forecasts. We finally applied model output statistics to the WRF to WindSim coupled model results in order to minimize their errors.

Improving WRF GHI Forecasts with Model Output Statistics

Solar energy applications need reliable forecasting of solar irradiance. In this study, we present an assessment of a short-term global horizontal irradiance forecasting system based on Advanced Research Weather Research and Forecasting (WRF-ARW) meteorological model and neural networks as a post-processing method to improve the skill of the system in a highly favorable location for the utilization of solar power in Turkey.