Automated diagnosis of communicating-automaton networks (CANs) is a complex task, which is typically faced by model-based reasoning, where the behavior of the network is reconstructed based on its observation. This task may take advantage of knowledge-compilation techniques, where a large amount of reasoning is anticipated off-line (when the diagnostic process is not active), by simulating the behavior of the network and by constructing suitable data structures embedding diagnostic information. This (general-purpose) compiled knowledge is exploited on-line (when the diagnostic process becomes active), so as to generate the solution to the problem. Additional reusable (special-purpose) compiled knowledge is generated on-line when solving new problems. A software environment for the diagnosis of CANs has been developed in the C programming language with the support of the PostgreSQL relational database management system, under the Linux operating system. It supports the modeling and preprocessing of CANs as well as the solution of diagnostic problems, including on-line knowledge compilation. The environment has been tested through a variety of experiments. Results are encouraging and provide a valuable feedback for further work. 366S. CERUTTI ET AL. most real-world systems can be viewed as CANs and reasoning about such networks is easier than about continuous systems, from the middle of the 1990s the task of diagnosis of CANs has been receiving an increasing amount of interest from both the artificial intelligence [3][4][5][6][7][8] and the automatic control communities [9][10][11][12][13][14][15][16].Diagnosing a network means computing its candidate diagnoses, each of which is a set of faults that explains the observation collected during the network operation. In the general case, the specific faults of a network cannot be inferred without finding out what has happened to the system [17]. In this way, the system evolutions complying with the observation, called the histories [18], situation histories or narratives [19], paths [20], or trajectories [21], become a product of the diagnostic reasoning.Determining the system evolutions is computationally expensive (see [22] for the difficulties of the diagnoser approach [9,10], or the worst-case computational complexity analysis in [18], or the discussion in [23]). This is why most approaches exploit a trade-off between off-line and on-line computation: some kind of knowledge, implicit in the models of the structure and behavior of the network, is compiled off-line in order to speed up on-line processing.While the approaches by other authors deal with networks wherein the communication between automata is supported by synchronous links, the CANs taken into account here and in previous works by the present authors constitute an adaptation of communicating finite-state machines [24,25], these being networks of finite-state machines that asynchronously exchange messages with each other through FIFO links. In particular, this paper presents a diagnostic environment for a...

No abstract

An enhanced process to Ansaldo Energia Switzerland’s validation approach for vibrational behaviour of gas turbine compressor blading is investigated. The opportunity of relying on less and early-stage measurements is assessed as a driver for cost and time optimisation. Accordingly, an analysis of possible variations of measurement data available during all validation phases is performed in order to confirm observed and expected correlations on a large scale, facilitating measurement chain improvement and simplification. Profile deviation and frequency distribution analysis finds the present coordinate-measuring machine (CMM) data and statically measured eigenfrequencies to be normally distributed. Correlation analysis exposes strong autocorrelations of CMM and static frequency data and distinct linear correlations between them. Consistent correlations with and within rotating machinery cannot be established with the limited data basis at hand as clamping conditions of blades with hammer foot roots introduce an additional degree of variation. CMM data are employed for meta-models translating geometry deviations into static-test-alike frequency information. Average aerofoil thickness deviation is identified as the decisive parameter for eigenfrequency uncertainties founded in geometry deviations. A stage statistics based approach for frequency forecasting in rotating machinery is proposed, allowing the calculation of expected eigenfrequency bands with corresponding confidence levels. Brought up to operational levels it would allow more efficient forecasting of in-engine conditions and constitute a valuable means for exploring compressor blading design space.

The mechanical integrity design of front stages for heavy duty gas turbines is by definition a challenging task. Several design criteria need to be fulfilled, reaching from static stress requirements over sufficient cyclic lifetime to a defined dynamic behavior of the component. This design task becomes even more challenging in case the size of the engine increases. To satisfy the need for designing larger front stage blades, an automated and integrated design tool suite for Mechanical Integrity analysis tasks has been developed. This paper introduces the specific steps of the mechanical assessment in the early concept phase of new designs. The overall process relies on advanced numerical methods, namely Design of Experiments, numerical optimization and probabilistic techniques, which are combined into a single package. The delineated procedure represents an incremental approach, which contrasts to the integrated and partially multi-disciplinary optimization processes which are typically described in literature. The benefits of the incremental method can be identified with regard to computational time, enhancement of interfaces and feedback loops with neighboring disciplines. By applying the process outlined in this paper, a new generation of frontstage blades could be developed with further improved mechanical properties and reduced development efforts.