Innovative monitoring instrumentations and methods for landslide risk management and mitigation

During last century, the large spread of human activities, urban areas and infrastructures, together with the abandonment of rural or mountains zones, faces an increment of landslide events and potentially dangerous scenarios. Due to their geographical and economical characteristics, some Countries were significantly involved in natural disasters and words like “Risk Management” and “Monitoring” have become very common, even if often not applied. In particular, as described in Chapter 1, Italy is susceptible to this theme, with a worrying increment of catastrophic events, which caused several deaths and economical loss. The scientific community has concentrated many efforts in the definition and calibration of predictive numerical models and forecasting risk analysis, associating these tools with Civil Protection guidelines and emergency procedures. Nevertheless, while numerous data are available in the hydraulic field, geotechnics is characterized by several uncertainties and a general lack of information. In fact the evolution of technologies that regarded industry and everyday life do not cover geotechnical instrumentations, which often require an operator, have a very low sampling period and are inadequate for risk management purposes. Chapter 2 summarizes the traditional instrumentation available on the market, together with the innovative tools that have been developed during recent years. These tools permits the integration of several sensors and the detection of cause-effect relationships, greatly improving the monitoring process, even if characterized by considerable costs and a series of features that do not cover the complete spectrum of monitoring needs. Therefore, there are some critical aspects related to their manufacturing that may influence the final result and applicability. The increasing number of infrastructures has required to realize underground excavations or cross landslide areas. In this field, each design is hypothetical due to the use of natural materials and underruns the risk of facing surprises. A “learn as you go” approach is often indispensable, with field observations and quantitative measurements. This thesis concentrates on the study of a new automatic and innovative technology, developed starting from the knowledge regarding landslide and infrastructure achieved during last decades at University of Parma. Chapter 3 reports the general characteristics of the new tool, together with the evolution of installation method, definition of calibration procedures and hardware development. Software and Calculation principles are widely discussed, due to their innovative features and introduction of Internet of Things (IoT) principles. A particular pioneering aspect regards the application of theory related to airplanes navigation to geotechnical field, together with the definition of statistical algorithms. The latter, in particular, are able to automatically detect instrumental and spike noise, sensors malfunctions and data trend in a self-learning philosophy. The software, after a training period, knows how to elaborate the raw data. The aim of the new tool is to increment the knowledge about evolution of slopes, both for natural conditions and artificial changes. Another key aspect is the real time control of geotechnical structures and underground excavations, during both construction and exercise phases. Chapter 4 concentrates on time of failure prediction methods, starting from those available in literature for rock or soil slopes and open pit mines. In particular, Fukuzono (Inverse Velocity) method has showed its efficacy in numerous case histories, despite its site-specific applicability. In order to overcome this restriction, a new generalized criterion is proposed starting from data of past events. The method is able to evaluate landslide Time of Failure for a wide range of landslide behaviors and characteristics. The result has permitted the creation of an abacus that forecast the landslide trend approaching to failure. The second part of the chapter concentrates on the real time applicability, with the implementation of the method within a self-learning algorithm. This solution requires to automatically detect the point of activation. For this purpose, an activation criteria is proposed. Finally, Chapter 5 reports a series of case histories where the new tool has been tested and applied to geotechnical structures and landslides, showing the comparison with traditional instrumentations and numerical models. It has been possible to study the ability of prediction of the Time of Failure method described in Chapter 4, when used in conjunction with an Early Warning Monitoring System.