A decission support system for shield tunneling; for For bored tunnel projects in soft soils in the Netherlands.

1997. A decission support system for shield tunneling; for For bored tunnel projects in soft soils in the Netherlands. / by ing. R.P. van der Putten. ABSTRACT, supervisor(s): Hack & M.T.J.H. Smits (Fugro Engineers, Leidschendam, The Netherlands).  thesis availability: Technical University Delft

ABSTRACT

The lack of space available, especially in the western part of the Netherlands, causes that infrastructural works can merely be suitably situated at the surface. Building tunnels with trench or cut-and-cover methods in densely populated areas is very disturbing for daily life or to the environment. Building tunnels with soft soil Tunnel Boring Machines (TBM) fully takes place underground and causes minimum disturbance at surface.

Determining an optimum alignment combined with the most suitable TBM can reduce the costs by minimising difficulties during the drive and can reduce accidents. Efficient use of geological and geotechnical data and knowledge about shield tunnelling combined in an automated system would support the user to optimise the alignment and select the most suitable TBM.

Therefore the goal of this research is to develop a Decision Support System (DSS), which supports an advisor/project-engineer in the first stage of a project (feasibility study) to select an appropriate Tunnel Boring Machine (TBM) and optimum alignment with the geological and geotechnical data available.

There was no Decision Support System that satisfied the program requirements or could easily be modified to fulfil these requirements. Therefore a new Decision Support System is developed. The DSS contains two programs: LYNX, a 3-Dimensional Geographical Information System (3D-GIS), and an expert system. First a geological model must be created with LYNX with the data gained during ground investigation (boreholes, geophysical surveys, etc.). To make a geotechnical grid model it is necessary to split the geological model in small elements (grid cells) and assign, by interpolation, values to these elements for each geotechnical parameter. The expert system, written in LISP, contains all knowledge necessary to optimise the alignment and to select the most appropriate TBM. This knowledge, selection criteria, relations, etc. are found in the literature. The geotechnical grid model created with LYNX provides the necessary input data to the expert system. The expert system creates a grid model that contains the alignment. This grid model can be imported in LYNX for visualisation. The expert system and LYNX run on the UNIX operating system.

The expert system supports 18 soft soil shield tunnelling machines divided into six classes based on their method of face support. These classes are:

1.     Open faced shields: natural and mechanical face support;

2.     Compressed air shields: compressed air is used to prevent the ingress of groundwater;

3.     Slurry support shields: slurry (clay, bentonite, etc.) supports the tunnel face;

4.     Earth pressure balance shields: the (conditioned) earth itself is the supporting medium;

5.     Combined shields: combinations of the above mentioned methods of face support;

6.     Microtunnelling: for driven small diameter tunnels, uses mainly slurry support.

Selection of the most applicable soft soil TBM is based on the following selection criteria:

·       grainsize distribution: D10, D30, D60 and D90 (Dxx is defined as the mesh size through which xx % of the soil particles will fall);

·       permeability of the soil mass;

·       maximum ground water pressure;

·       maximum shield diameter;

If a TBM is applicable then ‘1’ is assigned to that particular grid cell, else ‘0’. If the necessary geotechnical data is not available then ‘--’ is assigned to that grid cell.

For the design of the tunnel alignment, the maximum gradient, the minimum vertical and horizontal radius of curvature, the internal diameter and the necessary amount of cover are required.

With the expert system the user can select from the following four options:

1.     if the alignment is known, select the most suitable TBM;

2.     if the TBM is known, optimise the alignment;

3.     if the TBM and the alignment are unknown, select for each grid cell the applicable TBM’s;

4.     if the TBM and the alignment are known, calculate/select items like required face support pressures, amount of (differential) settlement etc.

The most suitable TBM is the one with the highest applicability (%) along the alignment. If the (optimised) alignment and most suitable TBM are known the program continues with calculating/selecting the following items:

1.     length of the shield;

2.     required face support pressure;

3.     external loads on the shield;

4.     required strength of the push cylinders;

5.     required cutting wheel torque;

6.     calculation of air consumption;

7.     danger of blow-outs;

8.     amount of (differential) settlement;

9.     necessity of stone crushers or crushing disks/bits;

10.  type of lining;

11.  amount of excavated soil.

The results of these calculations and selections are stored in a report together with the alignment and chosen TBM.

It can be concluded that the developed Decision Support System is capable in selecting the most suitable soft soil TBM and optimising the alignment in the first stage of a project (feasibility study). The expert system is designed to predict if a project is technically feasible. Because it is important to predict if a project is economically feasible it is recommended to include economical parameters (like economical drive length, separation costs, etc.). It is also recommended to improve the module ‘TBM only known’ which optimises the alignment, because this consumes too much computing time. It calculates all allowed paths for the alignment by searching four cells ahead and selects then the best path. This requires many calculations and therefore a lot of time.