Summary

Japan is located in one of the most active crustal movement zones in the world and has heavy rainfall. Thus, to secure the safety and reliability of railway operation, it is important to reduce disasters resulting from slope failures along railways. To predict the occurrence of slope failure with high precision is a very difficult problem, so the railway companies in Japan generally take some kinds of countermeasures to reduce slope disasters. One is to protect slope surfaces with artificial faces like concrete facing wall to prevent failures. The other is the so-called operation control like stopping the trains when the current rainfall amount is likely to exceed the limit to cause failures. To make these countermeasures more efficient and economical, we have completed a new estimation method. And we continue to research various subjects and to develop a new method of evaluating the effectiveness of slope protection works against rainfall or a new slope protection method. In this paper, we report, aiming at the reduction of slope disasters due to rainfall, on the present countermeasures taken by railway companies in Japan against disasters and the state of relevant research and development by Railway Technical Research Institute(RTRI).

Keywords :

disaster reduction, slope failure, slope protection work, operation control of trains, statistical analysis, critical rainfall, estimation standard

1 . Introduction

Japan is located in one of the most active crustal movement zones in the world with topographic and geological characteristics of complexness and brittleness. In addition it has heavy rainfall and frequent earthquakes. Thus, disasters occur every year along railways, resulting from slope failures. To secure the safety and reliability of railway operation over the country, it is important to reduce disasters resulting from slope failures along railways. In this paper, we report, aiming at the reduction of slope disasters due to rainfall, on the present countermeasures taken by the railway companies in Japan against disasters and the state of relevant research and development by Railway Technical Research Institute(RTRI).

2 . Natural condition and slope disasters in Japan

(1) Natural condition of Japan

The Japanese Islands are located in the region where four large plates collide each other, generating big stress. The Islands have been formed by this stress and resulting from the remarkable upheaval of the land through Quaternary period, steep mountainous areas in the stage of maturity occupy about 70% of the Islands. About 60% of Japan is constituted of new age strata in Cenozoic era unlike the continents mainly constituted of pre-Mesozoic strata, and its geological structure is complex with a lot of faults and foldings. So the ground has been broken in deep part and there are many volcanos and frequent earthquakes. In addition Japan belongs to the monsoon climate zone with heavy rain and heavy snow. The average precipitation(the total amounts of rainfall and snowfall) per year is about 1700mm, which is about two times those of Europe and America. And because of narrow land and steep mountains, the rate of precipitation run-off from the ground surface is high amounting to four times those of Europe and America.

Thus, Japan is placed under unfavourable natural conditions with rapid change of ground surface by weathering and erosion, and this is a large factor causing frequent occurrences of slope failures.

(2) Actual conditions of slope disasters

The construction of railway network in Japan was started in the latter half of 19th century and the majority of the network was completed in the first half of 20th century. Fig.1 shows the structural composition of the defunct Japanese National Railways. Especially about 90% of the narrow-gauge lines are in embankments and cut slopes which are generally thought vulnerable to heavy rainfall. The number of natural disasters along railways was in excess of 8000 per year about 30 years ago, but after that, it has largely been reduced approximately to 1000 per year in these 10 years(about half of them obstructing operation). We think this is an effect of the countermeasures tirelessly being continued. Fig.2 shows the details of natural disasters which recently occurred(The Japanese Geotechnical Society,1997). Slope disasters, namely cut slope collapse, embankment collapse and rockfall, occupy half of all failures and most of them are caused by rainfall. Although total number of disasters has dropped, the tendency of an increase in slope disasters' ratio has appeared. This means that the outflow of rainfall water has changed with a regional exploitation advancing along railways as a result of economic growth of Japan.

Thus, to secure the safety and reliability of railway operation in Japan, it is important to reduce disasters resulting from slope failures along railways.

3 . Reductive method of slope disasters

For the reduction of slope disasters, it is important to predict when, where and what kind of slope failure may occur. If this prediction could be done precisely, effective countermeasures could be carried out beforehand. But a slope failure is a natural phenomenon and its occurring mechanism has not been clarified enough, so it is difficult to predict it precisely. To prevent the accident due to slope disasters, the following ways have been practiced in Japan and the status

quo is that several of them are combined for the safety of railways.

(1) Slope protection works

There is a way to protect or strengthen unstable slopes in advance. The general way is : covering slope faces with concrete, setting the concrete-frame or the concrete-block pitching on slope faces for the protection from erosion and the reduction of infiltration into the slope. To strengthen the slope, the piles or anchors are used frequently. The best way to prevent the slope failure must be the protection of slope surface with these mechanical reinforcements, but it is rather hard to construct works for treating all slopes along railways because it needs a great investment. Thus, it depends on the order of priority given to a specific unstability.

(2) Operation control of trains

The operation control is a way of stopping trains or regulating speed of trains in case the current rainfall amount is likely to exceed the limit to cause slope failures. It has been continuously performed since the defunct Japanese National Railways. The indices dictating operation control are the accumulated rainfall[R] and hourly rainfall[r]. The former is an index for total amount of rainfall and the latter is one for the intensity of rainfall, and, in general, we use a combination of these indices. Fig.3 is an example of operation control being specified and operation control values being set stepwise. In the operation control, the most important thing is how to determine the control values. In the present practice the values are determined experimentally by referring to the past disaster data.

(3) Measurement and detection of slope failure

For the landslide and rockfall which are less related to rainfall, the detection of a foreboding sign of failure (displacement or vibration of the ground surface, change of underground water, acoustic emission from breaking rock etc.) to predict the failure or the setting of sensors along railway tracks to catch its falling can become an effective alternative to countermeasure works. We had an experience of forecasting the time of occurrence of slope failure with high accuracy(Yamada et al., 1970). It was a success enabled by the measurement of ground surface displacement based on the theory of Saito(Saito, 1968). Although this method is applicable for a large scale failure with slow movement, it is difficult to apply to a small scale failure and a rockfall with rapid movement and rare foreboding sign. And setting the sensors and doing measurements are necessary at a large costs and manpower, so these methods are applicable to limited cases.

4 . Research and development at RTRI

(1) Estimation of slope against rainfall

In the countermeasure of slope disasters, the most fundamental problem is how to predict the adequate amount of rainfall that can unstabilize slopes. If it could be done more precisely, the works of slope protections could be carried out more reasonably and the operation control of trains could be performed more efficiently during rainfall. This means that the protection works can be performed beginning from a weaker slope more economically, and that with operation control we can avoid too many train-stops induced by setting lower criterion for rainfall. The risk estimation method for slope failure due to rainfall was tried about 25 years ago(Japanese National Railways, 1974). Although this method has been used widely, it involved several problems such as its low accuracy and the little relation between the index calculated from this method(the result is presented as daily rainfall) and the amount of rainfall used for the actual operation control.

..We have completed a new estimation method recently (Okada et al., 1994, Sugiyama et al., 1995). First we gathered a lot of disaster data due to rainfalls along railways (about 150 embankment collapses and 200 cut slope collapses), and investigated the collapse site to get more detailed data. Next we did statistical analyses based on the theory of slope stability. A potential of slope stability during rainfall is generally expressed as follows:

As the objective variables we selected an accumulated rainfall[R] and an hourly rainfall[r] which are used for the actual operation control, and the objective variable is defined as [R] multiplied by [r], which have the powers(constants m and n) respectively as follows:

S=Rm rn

As the explanatory variables, we selected topographic and geological conditions, profiles of slope, soil strength and water catchment situations, which can be investigated and measured on the site.

The statistical analyses were done with both embankment collapses and cut slope collapses. In the analysis of cut slope collapses we divided them into two groups, that is surface collapses and deep collapses, because the shapes of collapse and items for collapse were different. The risk estimation standard for embankment resulting from analyses gives Table 1. We also make similar risk estimation standards for surface collapse and deep collapse of cut slope. Then we call the estimated result [Rmrn] "critical rainfall". The powers in the standards are given in Table 2. Flow chart of estimation for cut slope collapse is given in Fig.4. In the case of cut slope, it is necessary to discriminate between surface collapse and deep collapse by the discrimination value.

Three kinds of critical rainfalls obtained by the estimations are represented by non-linear curves on an accumulated/hourly rainfall graph, shown in Fig.5. If the curve is drawn in the upper right, the slope is more stable to rainfall, and if the actual rainfall comes beyond the curve, the probability of slope failure will become higher. In this figure, operation control value is also plotted and you will be able to understand that our estimation method is useful to the actual operation control. We have verified the accuracy of this method using the data of slope failures which recently happened.

This newly proposed estimation method can be performed on personal computers at present, and already in the practical use with the railway companies. It is surely expected to be able to contribute to the reduction of slope disasters.

(2) Evaluation of slope protection works against rainfall

The slope protection has not been designed by calculation, so it is not made clear how much rainfall the slope protection will bear. For this reason the proposed estimation method does not include such factors as the effect evaluating of slope protection works. But a lot of slopes along railways have been covered by protection works, so evaluation of slope protection effect against rainfall is an important thing.

Now we are carrying out rainfall tests using a model slope and a numerical simulation to get the basic performance of slope protection works. As seen in the concrete-block pitching on embankment face, shown in Fig.6, it is clear that : (a)When the penetration of rainfall through slope face is completely prevented, the total amount of rainfall, which causes collapse of an embankment, can be increased to about four times that when no cut-off is done. (b) For 50% cut off under the same condition as (a), the allowable amount of rainfall can be increased to 1.5 times (Sugiyama et al., 1994).

Further we are making clear the performance of concrete-frame. We think, by taking these results into account in the new slope estimation method, more precise estimation of slope will be make possible.

(3) Other research and development

We continue to research various subjects in the field of slope disaster reduction. Major subjects taken up are as follows.

Development of new slope protection work

More economical and effective slope protection work is expected to reduce slope failures and rockfalls. We are studying the applicability in Japan of flexible rockfall protection barriers using ring nets, which were developed in Europe.

Development of more adequate rainfall index for operation control

Slope failure is induced not only by a rainfall at present but also by an antecedent rainfall. Effective rainfall is an index which is able to represent the antecedent rainfall, and we are studying application of this index for operation control(Muraishi et al., 1995).

‡BDevelopment of slope evaluation method using remote sensing

?@There are a lot of slopes along railways and the information on wide areas is necessary for the effective management of the slopes. So in Japan aerial photographs taken by airplane or helicopter have been used to grasp the conditions of slopes and their surroundings. With further progress in this traditional method we have proposed a new method for the extraction of a slope at which the collapse is likely to happen. It is a method using a digital image analysis to be combined with remote sensing data and GIS(Geographic Information System)(Noguchi et al., 1989,1991).

5 . Concluding remarks

The present status of rainfall disasters investigated and the countermeasures taken along railways in Japan and the slope disaster R&D undertaken by RTRI are outlined. The slope failure is a natural phenomenon and it is impossible at present to predict its occurrence precisely and to prevent it perfectly. However by improving the forecast technique and providing adequate countermeasures, the safety and reliability of railways can be secured. Hereafter we would like to continue R&D in the field of slope disaster reduction and to initiate a technical information system between researchers of different countries taking this opportunity.

BIBLIOGRAPHY

1)Japanese National Railways(1974)　1)Japanese National Railways(1974)　: "Standard for structure maintenance (in Japanese)", Japan Railway Civil Engineering Association

2)Muraishi, H., Sugiyama T. and Kagawa, S.(1995)　2)Muraishi, H., Sugiyama T. and Kagawa, S.(1995)　"A study of hazard index for disaster in terms of effective rainfall(in Japanese)", RTRI Report, Vol.9, No. 3, pp.7-12.

3)Noguchi, T., Setojima, M., Okada, K. and Muraishi, H.(1989):"Extraction of topographic and geological information on slope land along railways using aerial photo and geographic information", Proceedings of IGARSS'89,Vol.3, pp.1669-1673.

4)Noguch, T. and Sugiyama, T. (1991)　4)Noguch, T. and Sugiyama, T. (1991)　: "Evaluation for disaster susceptibility of slope along railway image overlay method(in Japanese)", RTRI Report, Vol. 5, No.10, pp.11-18.

5)Okada, K. and Sugiyama, T. (1994):"A risk estimation method of railway embankment collapse due to heavy rainfall", Structural Safety 14, Elsevier, pp.131-150.

6)Okada, K., Sugiyama, T., Muraish, H., Noguchi, T. and Samizo, M.(1994): "Statistical risk estimating method of rainfall on surface collapse of cut slope", Soils and Foundations, Vol.34, No.3, pp.49-58.

7)Saito, M. (1968)　7)Saito, M. (1968)　: "Research on forecasting the me of occurrence of slope failure (in Japanese)", Railway Technical Research Report, No.626.

8)Sugiyama, T., Okada, K., Muraish, H., Noguchi, T. and Samizo, M.(1995): "Statistical rainfall risk estimating method for a deep collapse of cut slope", Soils and Foundations, Vol.35, No.4, pp.37-48.

9)Sugiyama, T. and Muraishi, H. (1994)　9)Sugiyama, T. and Muraishi, H. (1994)　: "Durability against rainfall of embankment provided with slope protection(in Japanese)", RTRI Report, Vol.8, No.7, pp.13-1

10)The Japanese Geotechnical Society (1997)　10)The Japanese Geotechnical Society (1997)　: "Rept of ground disaster due to rainfall (in Japanese)".

11)Yamada, G., Kobashi, S., Kusano, K. and Kubomura,K.(1970)　11)Yamada, G., Kobashi, S., Kusano, K. and Kubomura,K.(1970)　: "Collapse of Takabayama Tunnel on the Iiyama Line by landslide(in Japanese)", Railway Technical Research Report, No.706