Tunnel construction using a TBM involves a highly complex process chain. Such processes generate large amounts of data that can be used for monitoring, reporting and analysis. Major TBM manufacturers have developed software systems to support tunnel contractors and their site teams in both data management and analysis. These programs are mostly webbased and have many advantages. Data acquisition cannot prevent breakdowns from occurring but can facilitate investigations to quickly determine the root cause of a breakdown and implement corrective actions. This paper analyses these data acquisition tools and presents case studies, primarily involving EPBMs, to illustrate how the formation of critical interpretations can be made from user-defined charts and diagrams to diagnose issues and optimise TBM operational parameters.

INTRODUCTION

The storage and visualisation of measured values acquired by sensors and recorders is a crucial element of TBM tunnelling. All of the work being performed by the machine is documented in terms of the recorded data to allow the complete or partial tracing of the tunnel construction in real-time or after completion.

This information could help engineers and operators examine a very large and complex set of data related to TBM operation that cannot be ascertained in the field by the TBM engineers or work crew, particularly when visualised in a graphic format. The examples of measured data and sample graphs presented in this paper are mainly taken from EPBMs, but the logic behind the interpretation of these examples can also be applied to data from hard rock or slurry TBMs.

DATA ACQUISITION AND VISUALISATION SYSTEM

The purpose of a TBM data and acquisition system can be summarised as the "acquisition, processing, storage, display and evaluation of all data connected to the tunnelling machine." A TBM data acquisition system continuously records and visualises all measured data in a predefined cycle. Logging, however, occurs only at specific times. The average time period between logs can be individually selected for each measuring point but is set to 10s for most parameters.

The operating phases of the tunnelling machine are generally classified according to three periods: advance, ring building and standstill. These three phases form a unit called a ring. The data for each ring are usually stored in separate consecutively numbered files.

An immediate correlation to the respective construction phase can be made based on the ring number, file date and fi le time. The measured data acquisition program automatically opens after each system restart and loads all required program components into its memory. It then acquires, stores and visualises the currently-available, measured data.

INTERPRETATION OF TBM OPERATIONAL GRAPHS IN CASE STUDIES

Some of the most common graphs representing the general status of TBM operations are ram extension, rate of advance (ROA), thrust force, cutter head torque, EPB/slurry pressures, weight/volume of excavated material, and grout volume. Of course, illustrating too many parameters on one chart makes interpretation more difficult, so there must always be a compromise between amount of information given and the clarity of the graphics.

EPB PRESSURE GRAPHS
Case study one

In successful EPB operation, face pressures should be maintained at all times and monitored with the data acquisition system. Pressure of material in the chamber could be assessed by information available from EPB cells. The TBM operator closely monitors excavated material and adjusts the type and amount of water, bentonite, polymers, and foam to ensure that the material is properly forming a plug to resist piezometric and ground pressures. Graphs from pressure cell data are among the most used graphs in EPB tunnelling and demonstrate the difference between correct (Figure 1) and incorrect (Figure 2) operation. Excavations similar to Figure 1 result in safe and steady progress while performances similar to Figure 2 are usually linked with significant loss of ground and surface settlement.

Case study two

The EPB pressures for the top, middle and bottom sensors used in this case study are presented in Figure 3, which shows that the bottom sensors record higher pressures due to the higher density and greater compaction of the excavated material. The top sensors record the least pressure and fluctuation because they have less direct contact with the soil and mud in the chamber.

The proper estimation of material contact and density in an excavation chamber is important in multiple stages of a project. It is common practice for the operator to perform and complete an excavation with full level of material in chamber when using EPBMs. However, TBMs must sometimes be operated in semi-open, in which only a portion of the face is balanced by excavated material. These operating conditions are generally determined by engineers based on the ground conditions and stoppage time. Compared to the semi-open mode, full material contact in a chamber requires more thrust and torque from the TBM and increases the equipment wear and the cost of replacing excavation tools on the cutterhead. Working in a semi-open mode could alleviate these issues but is not advisable if there is a high risk of ground collapse and overexcavation.

Other scenarios, such as preparation for cutter head maintenance or leaving the TBM unused for long periods of time, could also influence decisions regarding the level of material that should be present in a chamber during tunnelling operations.

Case study three

Smooth rise (or drop) in EPB graph lines indicates the passage of gaseous or liquid material into (or out of) the excavation chamber, which occurs mostly during the TBM ring build phase. Soft rising curves may be the result of ground water filling the chamber or the injection of ground conditioning material (foam, water or compressed air). Figure 4 shows a smooth drop in cell pressure, suggesting the leakage of air or water through porous ground, a tail shield, purge line or screw conveyor.

THRUST, CUTTERHEAD TORQUE, RPM AND RATE OF ADVANCE

Higher advance rates in TBMs are generally achieved in two ways. A. A higher cutterhead rotation speed, which increases the distance that cutters or rippers travel and thus their work per unit time (mm per minute). In this case, the cutterhead torque will increase. An increase in cutterhead torque can also result from other factors such as poor ground conditioning or high material density in the excavation chamber.

B. Higher forward forces in TBM cylinders to make cutters and rippers excavate more intensively, thereby increasing the cut depth per cutterhead rotation. In this case, both the TBM thrust force and its torque will increase. The TBM thrust can also be increased due to shield friction with the ground or the TBM’s pulling force due to its weight (a factor discussed later in conjunction with contact force).

Case study four

Scenario B is illustrated in Figure 5. The cutterhead rotation speed is set at approximately 2 rpm, so the occasional increase in torque is due to higher thrust forces exerted by the propulsion cylinders at that moment and increases the rate of advance. It should be noted that higher efficiency is usually achieved in soft ground with a lower RPM and higher thrust forces for deeper excavations, whereas cutters break into hard rock by rolling on it. Therefore, better advance rates occur with higher RPMs.

CUTTERHEAD CONTACT AND TBM THRUST FORCES

The graphical representation of the relationship between thrust and contact force is mainly used to identify any opposing forces to the TBM other than the excavation face. In general, the TBM thrust is used to maintain EPB pressure, push the material in the chamber, and pull the gantries and frictional forces of the shield.

The thrust left over from propulsion energy is consumed by the cutterhead in the form of the contact force required to cut through the ground. Because the parameters other than contact force are relatively constant during normal TBM operation, TBM contact and thrust forces are typically synchronised in their fluctuations. Therefore, any mismatch in the graphical patterns between these two forces suggests a status change in other parameters and usually indicates an obstacle during operation (Figure 6).

Case study five

The theoretical graphs in Figures 7 and 9 show a sudden drop in contact force despite a constant increase in the thrust force (Figure 7). These data could indicate collapsed ground around the TBM shield or an entrapped gantry back in the tunnel. Variations between the contact and thrust force that are more gradual (Figure 8) could result from a change in tunnel slope or the accumulation of heavy, dense material in the chamber.

IDENTIFYING OVEREXCAVATION

Most EPBMs today are equipped with weight sensors and laser scanners to estimate the weight and volume of excavated ground (Figure 10). The theoretical weight and volume that a TBM data acquisition system is expected to show is usually calculated manually based on the TBM dimensions, ground properties and advancing distances. These figures are compared with the quantities shown on TBM graphs to check for overexcavation. This information is also useful in the analysis of excessive volume loss and settlement.

Case study six

TBM advance with overexcavation can generally be recognised on TBM data graphs by a higher than-normal grade in the excavation weight or volume line. For example, the theoretical graph in Figure 11 illustrates three sets of data from different advances. Line A, which is the typical advance at a constant rate of excavation, is usually the preferred scenario and ensures the consistency of other parameters, such as ground conditioning and screw conveyor speed during the push. Compared to Line A, Line B has a steeper increase at the beginning and end of excavation period and ends at a higher value. This condition can be interpreted as general overexcavation. Line C represents a normal extraction scenario, except for a very sudden increase over short period of time that could indicate overexcavation with a sudden rush of material through the screw conveyor and out of the chamber. However, other aspects, such as those stated below, must be considered to achieve a realistic understanding of the data.

1. The higher grades in the graph lines based on material weight and volume can be a result of higher advance rates. Thus, the final excavation values must be checked. The screw conveyor rotation speed in those time periods can provide insight as to whether the high extraction rates were intentional or due to ground conditions (Figure 12).

2. The theoretical material weight depends on the advance distance of the TBM and the density of the intact rock, but other factors, such as added water or ground conditioning agents, should also be considered. In regard to ground conditioning material, only the liquid portion will affect the material weight, so the foam expansion ratio (FER) should not be considered in calculations. The FER of the ground conditioning material added to the chamber has no effect full chamber and decides to fill the chamber to its maximum level, less material will be extracted and shown in the data, even though the same amount of material has been excavated from the ground. On the other hand, when a TBM chamber must be emptied during an advance (e.g., the last advance before cutterhead maintenance), the TBM data will show a higher amount of extracted material than average. To eliminate this problem, engineers also look at the rolling average of values for several consecutive rings, which eliminates the effect of chamber space and gives a more realistic picture of the scenario to identify possible overexcavation.

4. Added water should be considered in theoretical calculations. Occasionally, depending on ground conditions, most of the injected water is absorbed by the ground, and sometimes added water only replaces the water in saturated soils. 5. The calibration of weight scales and laser scanners must be part of a contractor’s regular maintenance program. Some weight sensors are very sensitive to misalignment and curves in the TBM conveyor belt, while laser scanners could have inaccurate readings depending on their position and air/light interference, such as dust. Using two belt scales and observing their averages can also aid in identifying errors and obtaining more realistic results.

GROUTING SYSTEM

Two-component grouting (A+B) systems through the tail shield have been one of the most problematic areas in TBM tunnelling. Proper grouting is important to prevent ground movement and surface settlement due to volume loss at the tail void. Grouting also stabilises segmental lining in the ground and improves a tunnel’s watertightness. Information available in TBM data acquisition systems can show early signs of system malfunctions and indicate which components require attention or which control settings need to be modified.

TBM data loggers typically record flow, pressure and volume parameters for each grout line. To check the quality of grouting behind segments and ensure that the correct dosage of accelerator (B) is mixed with part (A), TBM data for injected volumes should be checked against the theoretical volume of voids behind the segments. Gauge cutter wear should be considered in theoretical calculations, particularly for larger TBMs. Understanding the bore and cut diameters in hard and soft ground types can also lead to more accurate calculations. If the grouting volumes are lower than their theoretical values, other data must be checked to identify and solve the problem. The simultaneous spike in grout pressure and halt of grout flow in in Figure 13 is commonly an indication of blockage in the line. If grout volumes cannot be achieved when all lines are in operation, then the pre-sets and cutoff levels should be checked. Generally, grout pressures must overcome hydrostatic pressures by 1-2 bar behind segments.

Case study seven

If grouting volumes are higher than their theoretical values, assessments must be performed to identify any channeling of grout to the surrounding environment or leakage through the tail shield. In some instances, high pressure grout finds its way to the excavation chamber, mixes with the excavated material and exits through the screw (Figure 14).

PROPULSION CYLINDERS AND RING BUILD

Information and graphs derived from TBM data on propulsion cylinders can be used to analyse several aspects of their operation, including ring build and steering. These data can also explain damage to segments that occurs after installation. TBM data acquisition systems generally display the pressures and extension of ram groups using the sensors on a representative ram from that group.

Figure 15 shows information collected on 19 rams in six groups (A-F). The location of each group’s representative ram is shown in black.

Case study eight

The graph in Figure 16 shows the pressures applied to a group of segments during ring erection. The lines representing each group show a sudden jump from zero, indicating that the rams have been extracted to hold each segment after its erection. The lines also indicate common slow pressure loss due to micro-movement of the TBM and ring compression in the tunnel. However, an excessive loss of pressure in any group could loosen the adjacent segment and cause vertical (step) and horizontal (gap) misalignments.

On the other hand, excessive pressures on cylinders can cause damage, particularly around the circumferential joints of the segments in front of the cylinders. Ring build reports from the guidance system (as shown in Figure 17) must be studied in conjunction with ram pressure graphs to confirm the location of segments in relation to the propulsion cylinders and explain the damage incurred.

CONCLUSION

The graphical representation and measured values of TBM data can assist contractors by providing information that helps TBM crews increase the reliability and productivity of TBM operations. Such an advantage would ultimately lead to fewer breakdowns and lower tunnelling expenses.

Data acquisition and visualisation alone does not benefit the contractor unless the data is accurately interpreted. The utilisation and correct interpretation of data acquisition systems’ outputs could greatly enhance the control of the excavation and operation of various TBM systems.

As TBMs grow in size and complexity, advances in data monitoring and presentation to optimise TBM parameters will likely continue as well. The correct interpretation of these data is essential for the effective utilisation of these tools and to ensure efficient and productive tunnelling operations.

The key to success in EPB tunnelling is proper engineering combined with experienced operators. Data acquisition cannot prevent breakdowns from happening but allows the rapid identification of the root cause of a breakdown and the timely implementation of corrective actions