A fuzzy genetic model for estimating forces in link chains from the measurement of the natural frequencies
Isnardo Cadena Rodríguez. · 2020
Offshore facilities have mooring lines to provide stability, support and holding to the structures.These mooring lines are commonly made up of synthetic fiber ropes, cables and chains.When the load solicitation is high, the mooring lines must be made up of chain.The monitoring of the strength of these chains is vital for the reliability and security of the production of energy.The current methods for supervising the loads on the chains are expensive and have many uncertainties involved.In this context, it is proposed a new methodology for the force estimation in chains through the measurements of their natural frequencies.The present dissertation arises as an improvement of this approach.A fuzzy inference system optimized by a genetic algorithm is introduced to enhance the estimation of the load on the chains.The inputs of the fuzzy models are the natural frequencies of the chains and the output is the estimated force.The Mamdani and Sugeno methodologies were implemented and compared.Triangular and Gaussian membership functions were used to model the inputs and the output.The rules were set according to the relations between the natural frequencies and the force on the chain.To optimize the system, the genetic algorithm can use the results provided by a mathematical model or by a set of measurements as training data.The mathematical model has good agreement with the experimental data.The fuzzy genetic model was simulated and tested providing good accuracy in estimating the force.In addition, the non-singleton fuzzification demonstrated that can be a helpful tool when the entries are noisy.