M. Mrówczyńska

Elements of an algorithm for optimizing a parameter-structural neural network

Abstract The field of processing information provided by measurement results is one of the most important components of geodetic technologies. The dynamic development of this field improves classic algorithms for numerical calculations in the aspect of analytical solutions that are difficult to achieve. Algorithms based on artificial intelligence in the form of artificial neural networks, including the topology of connections between neurons have become an important instrument connected to the problem of processing and modelling processes. This concept results from the integration of neural networks and parameter optimization methods and makes it possible to avoid the necessity to arbitrarily define the structure of a network. This kind of extension of the training process is exemplified by the algorithm called the Group Method of Data Handling (GMDH), which belongs to the class of evolutionary algorithms. The article presents a GMDH type network, used for modelling deformations of the geometrical axis of a steel chimney during its operation.

Proposition of determination of displacements using the TDRA 6000 laser station

With modern measurement technologies, it is possible to achieve higher precision and solve complex measurement tasks. The tested device (TDRA 6000 laser station) features a direct drive technology ensuring a 3D point accuracy of 0.25 mm. TDRA 6000 combined with proper software is a perfect tool to monitor engineering objects. The article presents the measurement technique and the method of calculation of 3D displacements of control and measurement networks using the TDRA 6000 laser station from Leica. The values presented herein derive from periodic measurements carried out in control points stabilised at a ground the above-ground crossing between buildings located on the premises of the University of Science and Technology campus in Bydgoszcz.

Approximation of the process of changes in deformation of land surface using artificial neural networks

Artificial neural networks are an interesting method for modelling phenomena, including spatial phenomena, which are difficult to describe with known mathematical models. The properties of neural networks enable their practical application for solving such problems as: approximation, interpolation, identification and classification of patterns, compression, prediction, etc. The article presents the use of multilayer feedforward artificial neural networks for describing the process of changes in land surface deformation in the area of the Legnicaof the Fore Sudetic Monocline. Results provided by geodesic monitoring, which consists of land surveying and interpreting data obtained in this way, are undoubtedly significant in terms of identifying the impact of mining on the land surface the results of measurements carried out by precise levelling in the years 19672014 were used to determine changes in land deformation in the Legnicantre. The concept of a flexible reference system was used to assess the stability of points in the measurement and control network stabilized in order to determine vertical displacements. However, the reference system itself was identified on the basis of the critical value of the increment of the square of the norm of corrections to the observations.

Indicating vertical deviation of historical buildings using geodetic methods - case study of brick and wood tower in Nowe Miasteczko

Abstract The article presents the possibilities of applying geodetic methods to determine the vertical deviation of historical buildings. In particular, the results of measurements obtained for a brick and wood Town Hall Tower located in the town of Nowe Miasteczko have been presented. Geodetic measurements of vertical deviation taken before and after carrying out repairs which were aimed at stopping or eliminating the destructive processes of degradation taking place, especially in the wooden part of the tower. During the renovation works, attention was also given to improving the technical condition of the building, which was reflected by the results of the measurements and calculations.

The SVM Method As An Instrument For The Classification Of Vertical Displacements

Abstract The article presents the basic rules for constructing and training neural networks called the Support Vector Machine method as well as possible applications for this kind of network. SVM networks are mainly used for solving tasks of classifying linearly and non-linearly separable data and regression. However, in recent years more applications have been found for them. The networks also solve such problems as the recognition of signals and images as well as speech identification. In this paper, non-linear SVM networks have been used for classifying linearly separable and non-separable data with a view to formulating a model of displacements of points in a measurement-control network. The points of the measurement-control network were placed on a civil engineering object located on expansive soil (linearly separable data) and represented a mining exploitation area (linearly non-separable data). The task of training SVM networks requires the use of quadratic programming in search of an optimum point of the Lagrangian function in relation to the parameters being optimised. In the case of linearly non-separable data, the SVM method makes it possible to find a hyperplane which classifies objects as correctly as possible, and at the same time is located possibly far away from concentrations typical of each class.

EVOLUTIONARY STRATEGY (μ+λ) AS AN INSTRUMENT FOR DETERMINING DEFORMATION PARAMETERS OF STEEL STRUCTURES

Abstract The article presents the use of an evolutionary algorithm for determining the shape of the guy rope sag of a steel smokestack. The author excludes the analysis of the operation of the rope, and discusses only the problem of determining parameters of the function of the adaption of the rope sag curve into empirical data, obtained by the geodetic method. The estimation of parameters of the curve and the characteristics of the accuracy of its adaption into experimental data were carried out by means of an evolutionary algorithm with the use of an evolutionary strategy (μ+λ). The correctness of the strategy presented in the paper, as an instrument for searching for a global minimum of a criterion function, has been presented using as an example the minimisation of a certain two dimensional function and the estimation of parameters of an ordinary and orthogonal regression function. Previous theoretical analyses have also been used for determining parameters of the guy rope sag of a steel smokestack, which is measured periodically. In addition approximate values of the pull forces in the guy ropes have been calculated.