Incremental Modeling of Rotor Lifetime Consumption During Gas Turbine Operation

In order to enable customers to operate gas turbines (GT) closer to the life limits, a life monitoring tool has been developed which computes rotor life consumption based on operation history of the GT. The tool can account for both creep and fatigue life consumption. The creep module can be applied to extend GTs creep life at high operational gas temperatures, whereas the fatigue module enables more flexible operations meaning more frequent operation at varying load conditions. For creep monitoring a new incremental modeling approach has been developed to calculate rotor radial elongation based on the actual operation history of GTs. In this method the rate of elongation during GT operation is calculated based on a correlation with the total radial elongation and gas temperature. The results of this modeling approach have been validated against the field data of several GTs. A good agreement has been obtained between the simulation and measurement results for different rotor axial positions. For fatigue life monitoring, the temperature based monitoring approach developed in a former publication [1] by the authors has been further adapted to rotating components in particular the rotor for which metal temperature measurement is very challenging. In this method life consumption is calculated based on the gas temperature history and GT power instead of metal temperature. By considering the real-time data, the rotor life consumption can be more precisely calculated and as a result the GT can be operated closer to its life limits. Hence, this method can be used to increase the intervals between inspections and therefore to extend the service life of the GTs.

Hybrid Incremental Model for Imbalance Detection in Ultra-high Precision Rotating Devices

A novel method based on a hybrid incremental modeling approach has been designed and applied to imbalance detection in ultra-high precision rotating machines. The model is obtained by a two-step iterative process that combines an overall model (least-squares fitting) with a local model (fuzzy k-nearest-neighbour) to take advantage of their complementary capacities.  Three normalization strategies of evaluating the effect on accuracy are analyzed.  A comparative study demonstrates that the hybrid incremental model provides better error-based performance indices for detecting imbalance than a nonlinear regression model and an adaptive neural-fuzzy inference system model. The suitability of Mahanolobis normalization for hybrid incremental modeling is also demonstrated in this case study. The proposed strategy for imbalance detection is simple, fast, and non-intrusive, reducing the deterioration in the performance of ultra-high precision rotating machines due to vibrations.

Coping with Complexity When Predicting Surface Roughness in Milling Processes: Hybrid Incremental Model with Optimal Parametrization

The complexity of machining processes relies on the inherent physical mechanisms governing these processes including nonlinear, emergent, and time-variant behavior. The measurement of surface roughness is a critical step done offline by expensive quality control procedures. The surface roughness prediction using an online efficient computational method is a difficult task due to the complexity of machining processes. The paradigm of hybrid incremental modeling makes it possible to address the complexity and nonlinear behavior of machining processes. Parametrization of models is, however, one bottleneck for full deployment of solutions, and the optimal setting of model parameters becomes an essential task. This paper presents a method based on simulated annealing for optimal parameters tuning of the hybrid incremental model. The hybrid incremental modeling plus simulated annealing is applied for predicting the surface roughness in milling processes. Two comparative studies to assess the accuracy and overall quality of the proposed strategy are carried out. The first comparative demonstrates that the proposed strategy is more accurate than theoretical, energy-based, and Taguchi models for predicting surface roughness. The second study also corroborates that hybrid incremental model plus simulated annealing is better than a Bayesian network and a multilayer perceptron for correctly predicting the surface roughness.