(Received 2 August 2013; accepted 6 January 2015)
Published Online: 2015
| ||Format||Pages||Price|| |
|PDF (1.7M)||10||$25||  ADD TO CART|
Cite this document
Determination of moisture content of soil is absolutely crucial and must be performed frequently to assure the quality of construction work. In a view to suppress the limitations and inaccuracies of the existing methods, a new concept is successfully implemented. Artificial neural networks (ANNs) and state-of-the-art electronic circuitry are embedded to realize an automatic measurement system based on an open-source computing platform. Moreover, to facilitate user interaction with the system, a sophisticated graphical user interface (GUI) is created. This system, from both the hardware and software perspective, brings new ideas in, not only the moisture content determination test, but in geotechnical engineering laboratory and field instrumentation in terms of accuracy, automation, and machine intelligence. It will be obvious from this work that open-source hardware and software embedded in test equipment can automate the whole test procedure for almost all types of tests in geotechnical engineering using appropriate sensor, and data-acquisition and processing routines. This paper presents the technical know-how of the system setup, hardware, and software development, workflow, and system validation for the determination of moisture content of subgrade soil in two distinct test modes—supervised and unsupervised. This work will demonstrate that, not only the test procedure can be automated, but time optimization is also possible, which results in at least 50 % test time reduction.
Shetu, N. S.
Member, ASTM, West Conshohocken, PA
Masum, M. A.
Member, IEEE, Piscataway, NJ
Stock #: GTJ20130128