Predictive
Construction
Engineering By and
I.
Hideo Tateshi
Control
System
Work* UCHI
YAMA,
* * Isao
ICHIHARA,
* * Hiromi
** ***
SHIKATA**
KOSEKI***
Introduction In civil engineering constructions, there is an increasing tendency for field measurements to be made during construction to discover and minimize the unknown factors in the design and execution. This philosophy was proposed by Terzaghi and called the Observational Method (Observational Procedure) long ago.') In our case, we started constructing the Mizushima Works in 1962, and since the site was on extremely soft and weak ground, we proceeded with construction while developing an observational construction control system out of necessity. And within the integrated steel works construction field, we were able to apply the system successfully in a variety of cases. Following that, we applied the system to the construction of the Chiba Works' West Plant and more recently to a steel works in Brazil. The fields of application for this system can be roughly classified as follows : (1) Construction of various kinds of yards on unstable and soft ground, by improving the ground without the use of piles.2) (2) Works requiring deep excavation such as scale pits and earthquake-proof foundation for heavy structures.3~ (3) Construction of port and harbor structures including piers, revetments and breakwaters. (4) Construction of general foundations for various plants and others where the use of piles is an undecided factor. Most of the above four cases, though somewhat different, can be analyzed and controlled using a common system flow. There follows an example of a deep excavation coming under category (2). In March 1974, we decided to construct the No. 6 blast furnace, the largest in the world at that time, at Chiba Works on a site with weak and soft cohesive soil having a subsoil layer 30 m thick. Since we adopted as the foundation type featuring an earthquake-proof design a double interlocked steel pipe pile well foundation, 30 m deep excavations had to be made. Consequently, the difficulty in design and work execution was unprecedented. We felt that the response to design and measurement during construction of conventional site measuring methods had reached its limits in view of its inability to forecast quantitatively. In the circumstances, we developed a Real-time Construction Control System (hereinafter *
in Civil
called " RCC " system) capable of quantitatively estimating the behavior of retaining wall structures which change with the progress of excavation work. This system is capable of predicting the behavior based on soil parameters estimated from the field measurement data. Employing this system in the foundation works of the blast furnace enabled the construction to be completed successfully by identifying unrealistic factors in the original design as the work progressed and by promptly feeding back modified data for design and work execution. Another example in which this system was used in foundation works was a case where it became clear, unlike the above example, that the original design was uneconomical. The design was able to be changed during execution of the work, with a resulting reduction of costs and shortening of construction period. In this way, this system has made it possible to control construction works taking into account both safety and economy simultaneously; something previously thought to be extremely difficult. This is because the designed safety margin can be brought closer to the optimum value by correcting the overor the underestimate in the original design from forecasts made by this system. II.
Field Measurement struction Control
and System
Computerized
1.
Conventional Field Measurement
Con-
Process
In the field, measurement is performed mainly on the displacement and stress in structures. On the other hand, the effect of soil strength on the construction could be estimated only indirectly and qualitatively from the displacement and stress of the structure, because of the great difficulty in the measurement of soil strength. Consequently in the case of conventional field measurement control, analysis and evaluation of the data were largely dependent on the engineers' experience and skill. Furthermore, with the recent trend in extremely large constructions, the number of measuring items has become large. Thus, with the vast quantities of data, coupled with restrictions on construction period and others, it had been impossible to analyze the data satisfactorily. In these circumstances, it is impossible to promptly feed back measured data to modify the design. Consequently con-
Received December 7, 1981, © 1982 ISIJ Systems Planning and Data Processing Department, Chiba Works, Kawasaki Steel Corporation, Kawasaki-cho, Chiba 260. Civil Engineering Department, Engineering Division, Kawasaki Steel Corporation, Uchisaiwai-cho, Chiyoda-ku, Tokyo 100.
Technical
Report
(743)
( 744 )
Transactions
struction basis.
control
ISIJ, Vol. 22, 1982
has
been
made
on
,,
a " one-way
2.
Introductionof a Real-time ConstructionControl System As a result of the successful development of the RCC system, it has become possible to review the design based on a quantitative forecast. That is to say, it has become possible to revise the design in a series of cycles, i.e., design -~ construction and measurement --+design. This means that the optimum design satisfying both safety and economy has become achievable. (see Fig. 1) III.
Estimating
assumed to form a continuous wall in both depth and lateral length. (Modulus of elasticity: E, Moment of inertia: I) (2) The struts are all treated as elastic supports to the retaining piles. (Spring constant: EH). (3) The restraining of the top of the retaining sheet pile is represented by a coil spring. (Spring constant : KB) (4) It is assumed that the active earth pressure
Method
1. Structure Model of Retaining Wall As a simple model to explain the b ehavior lowing model was set up. (see Fig. 2) (1) In this model, the retaining sheet
the
piles
fol-
are
Fig.
1.
Construction
Fig . 2.
control
A model
usin g RCC
for
system.
structural
analysis.
N N
Transactions
(Coefficient of earth pressure : KA) and the apparent difference earth pressure (Declivity: i) are both exerted on active side of the retaining wall. (5) The reaction force of the ground exerted on the retaining wall on the passive side is represented by the distributed springs (Spring constant: k) within the elastic range. (6) For the ground on the passive side, the passive earth pressure is used as the ground reaction force of the plasticized part generated near the excavated surface. (Coefficient of passive earth pressure : Kp) This passive earth pressure is continuous with the elastic reaction force at the elastic-plastic interface. Basically the overground and underground members can be represented by the basic equations as follows. EI d4u d ax+b .........Overground members (1) x4 4 El dx4 +ku= ax+b...Underground members(2)
ISIJ,
Vol.
22,
1982
(745)
was a large scale foundation constructed on a soft and weak ground, highly rigid foundation was required. Consequently the substructure was made of interlocked steel pipe piles driven in double circular form, and the intermediate part was excavated to a depth of approximately 30 m and part of it was replaced by reinforced concrete. (see Fig. 3 and Photo. 1). For some time after the excavation started, there were no problems with the existing values and forecast values, and the work progressed smoothly. In the 4th excavation stage (-19 m), however, the displacement increased suddenly due to a sudden inflow of water. The prediction results thereafter all exceeded the allowable values set for residual stress after completion of excavation. Since there was a danger of a residual stress being generated which would ex-
where,
u: Displacement of members ax+b : Earth pressure in the active part x : Depth Provided that, for the underground plastic zone, the equation is the same as Eq. (1). Further, when there are a coil spring and struts, a spring force term is added to Eq. (1). This model has fixed constants (E, I, EH, KB) for the structure and variable soil parameters (KA, i, Kp, k1, k2, ... ). 2. Estimation of Soil Parameters and Prediction Calculations If soil parameters can the measured displacement
be estimated and stress
placement and stress of the wall the excavation can be calculated.
by analyzing data, the dis-
body at any stage of Consequently it is
possible to make a calculation to predict for the future excavation stage. Estimation of soil parameters boils down to obtaining the parameters P satisfying the total sum of residual squares of the measurement displacement u; and the model calculation displacement u; as follows :
SE = 9
(u3-u1(P))2 --~
min ...............(3)
where, j : Measuring point (Depth direction) P : Soil parameters (KA, i, Kp, k1, k2, k3 ... ). This means that the vast amount of data measured in the field can be condensed into only a few soil parameters (KA, i, K~, k1, k2, k3 ... ) through this basic model. Iv.
Results of Application
1. Foundation Workfor No. 6 Blast Furnace in Chiba Works (an Example of Safety Control) The foundation of No. 6 blast furnace at Chiba Works was constructed to support the load of the world's largest blast furnace consisting of a vertical force of 30 500 t, a horizontal force of 7 200 t and a bending moment of 227 000 t•m. Since it
Fig.
3.
Blast
furnace
structure
and
foundation.
Tcnhniral
R onnrff
Transactions
(746)
Photo.
A plotted
1.
out
ISIJ, Vol.
Construction
example
on
22,
of blast
No.
6 B.F.
1982
furnace
foundation.
foundation.
ceed the allowance set at the beginning of planning, countermeasures against the inflow of water were taken, and the remaining construction was re-examined based on the information available at that stage. As a result, it was decided to add 1 stage of RC timbering and to fill the inside of the steel pipes with concrete to meet the increasing residual stress. The results forecast at this point are shown in Fig. 4. From these results, we had some prospect of restarting the excavation, because all the stresses generated at each of the excavation stages thereafter were within the control value of 2 850 kg/cm2. At each excavation stage, the maximum deformation of the interlocked steel pipe pile predicted in the ensuing excavation process was compared with the maximum deformation measured, as shown in Fig. 5. It can be seen from this figure that the accuracy of prediction increases with the progress of excavation. As described above, making proper decisions by effectively utilizing the system to overcome situations unexpected at the beginning, enabled the excavation
Transactions
Fig. 5.
work to be completed satisfactorily. 2. Foundation Workfor No. 3 Blooming Scale Pit (an Example of EconomicControl) This foundation required deep excavation work using an interlocked steel pipe pile retaining wall in an underground pit 48.2 m long, 13.2 m wide and 14.6 m deep. By the time of completion of the 2nd stage of excavation (-6.3 m), the excavation of the final surface with the 3rd stage struts installed was able to be predicted. The value predicted then was only about one half the control value. Consequently in consideration of economy of construction, the possibility of a construction without 3rd stage struts was examined. The results shown in Fig. 6 (see p. 748) suggested that the construction could be completed safely at less than the allowable value. From this we decided to continue the excavation without installing the 3rd struts. As described above, this prediction system is a very effective means of construction control, because expedient countermeasures can be taken during construction work against any uneconomical design. V. Conclusion The results of applying this system to the foundation works for No. 6 blast furnace in Chiba Works
Predicted
ISIJ,
Vol.
22,
values and actual
1982
(747)
measurements.
have shown high reliability of prediction and the resulting accuracy of construction control. The whole system has since been applied to construction control for Kawasaki Steel Corp. with a number of satisfactory results. Although we have not shown any concrete examples in this paper, we have obtained excellent results on all projects in four categories mentioned in the introduction. Parts of this system have been applied not only to the construction of steel works but also to other general constructions including port and harbor structures, filling work in road construction and sewage disposal plants in urban districts. The system has been favorably received on account of its successful balancing of safety and economy. REFERENCES 1) 2)
3)
K. Terzaghi and R. B. Peck: Soil Mechanics in Engineering Practice, John Wiley & Sons, New York, (1948). K. Akai, M. Komatsu and M. Tominaga: "Analysis of Observed Embankment Performance in Terms of Effective Stress ", Proc, of the ASCE Speciality Conference on Performance of Earth and Earth-supported Structures, ASCE, Lafayette, (1972). M. Tominaga, Y. Echigo and M. Hashimoto: " Realtime Construction Control Computer Simulation System ", Paper for 9th International Conference on Soil Mechanics and Foundation Engineering, ISSMFE, Tokyo, (1977).
(748)
Transactions
ISIJ, Vol.
22,
1982
Fig.
6.
A plotted
out
example
on
scale
pit foundation
.
