A Flexible Framework for Magnetic Measurements

Abstract

The work presented in this Ph.D. thesis covers the specification, design, prototyping, and validation of a new version of a magnetic measurement control, acquisition, and data analysis software package: the Flexible Framework for Magnetic Measurements (FFMM). FFMM constitutes the software part of the new platform for magnetic measurements, including also new high-performance hardware, developed at the European Organization for Nuclear Research (CERN) in cooperation with the Department of Engineering of the University of Sannio. FFMM is conceived as a unified solution to drive all the existing and future park of measurement systems (mainly magnetic but also optical, mechanical, etc.). The effort for the series test of the LHC superconducting magnets highlighted limitations in the measurement control and acquisition programs, mainly associated with the relatively long time needed for a development iteration (the cycle of specification-programming-debugging-validation). Moreover, the software capabilities needed to be extended to manage the challenges of the new hardware, namely fast rotating-coil transducers (Micro Rotating Unit) and high-performance digital integrators (Fast Digital Integrator), and the need to perform more specialized tests on small/medium magnet batches. FFMM was developed to address these issues. Implemented in C++, it is based (i) on Object-Oriented Programming (OOP), and (ii) on an innovative technology, the Aspect-Oriented Programming (AOP). AOP extends the object-oriented paradigm in order to encapsulate features that are transversal to several functional units (crosscutting concerns) by means of new software modules, the aspects. The framework supports the user in producing measurement applications for a wide range of requirements by limited effort and development time. FFMM includes utilities for (i) fault detection, (ii) software synchronization, (iii) automatic generation of user interfaces, and (iv) a Measurement Domain Specific Language (MDSL) to provide the test engineer with an easy and fast way to write measurement scripts containing formal descriptions of the test protocols. The higher sampling rate of the new transducers and acquisition hardware, capable of increasing by three orders of magnitude the bandwidth of harmonic measurements with respect to the previous ones, produced an exponential rise in storage requirements. Thus, a data reduction algorithm was conceived as part of the system in order to decrease the size of measurement results by controlling the quality loss simultaneously. In addition to that, the framework includes an algorithm for field harmonic resolution enhancement allowing to overcome the limitation of the CERN standard analysis procedure, for the tests on magnets in non-stationary conditions. A numerical study was performed to assess the performance of both the techniques for data analysis (reduction and resolution enhancement). The results highlighted the good performance achieved by the proposed approaches, and in particular their suitability for application to the magnetic measurements carried out at CERN. For both algorithms, these conclusions were confirmed on the field through dedicated test campaigns. The tests on the field were performed with different protocols and measuring equipments. The framework proved its effectiveness in developing software for measurements with very different requirements. The results highlighted the resolution improvement (up to a factor 100) attained in the harmonic estimation thanks to the new platform and the capability of producing quickly and with a limited effort the acquisition and control software for new applications. The framework was designed to be flexible, maintainable, reusable, effcient. To assess the fulfillment of these project goals on the release 3.0 of FFMM, the internal quality of its source code was evaluated by means of suitable metrics according to the reference model defined in the standard ISO 9126. The results highlighted a good average quality level, with possibilities of interventions to decrease the complexity of some hot spots and exploit more profitably the concepts of object-oriented programming. A supplementary analysis was carried out on the aspect-oriented component handling the fault detection to state the advantages of using such a software design. The proposed architecture proved to grant a high level of exibility, maintainability, and reusability, without affecting significantly the run-time performance. Finally, an experimental approach to the software flexibility assessment of measurement frameworks was proposed and applied in the context of FFMM. The results highlighted that the framework achieves increasing degrees of flexibility moving from the point of view of the developer to that of the test engineer. The highest flexibility is attained for the changes involving the measurement procedure, namely at the level where it was mainly required.