The 72MW Linhekou hydroelectric project is located in China’s Shanxi province, in the middle of the Yellow river valley, west of the Taihang mountains. The project features an RCC arch dam with peak height of 96.5m. The bottom thickness of the crown cantilever section is 27.2m, the elevation at the dam crest is 515m and the arc length of the dam crest is 311m. The dam’s RCC volume is approximately 224,600m3, which is roughly 70% of the project’s total concrete volume. The contractor began to place RCC on 23 December 2001, at a compacted thickness of 30cm, and dam construction was completed on 25 June 2003. The dam began to retain water on 27 October 2003 and electricity generation began on 1 May 2004.

On 8 August 2002, a horizontal crack at el 4551m was found on the upstream face of the dam. Water pressure tests and bored dam concrete cores indicated that the crack was along the surface between the eighth and the ninth placing layer of the 11th placing lift, which is from el 452.7m to el 455.7m. The crack had developed from the dam upstream face to the downstream face, and had a total cracking area of nearly 1000m2, with a width less than 0.3mm. After careful observation, indoor tests and field experiments – measures including chemical grouting into the crack surface, sealing the crack boundary and installing anchor steel piles through the crack – were carried out twice for the arch dam.

An expanding problem

After discover of the crack, which may have appeared previously but remained undiscovered, another horizontal fracture was found at the same elevation on the dam’s downstream face. The elevation of the upstream crack was 455.1m (length approximately 40m), with the downstream one at el 455.26m (length approximately 20m), owing to the RCC layer inclining upstream slightly. A simple water pressure test was carried out on 16 August 2002, resulting in water leaking out of both the upstream and downstream cracks, and water splashed out from several other locations. Also, Ca(OH)2 emitted from both cracks. After chemical grouting was undertaken for the first time on 21 October 2002, it was found that water still leaked out of the cracks, and the cracks were extending toward the crown cantilever. This clearly indicated that the first chemical grouting was not successful.

Crack treatment for the second time began on 20 November 2002. During the drilling of chemical grouting holes, especially when high-pressure winds drove water into the grouting pipe system and the crack surface, the cracks continued to further extended toward the crown cantilever, which was evident in the wet cracking trace. After surveying, it was found that the upstream crack at el 455.1m extended from the right abutment to a location 19.95m right from the reference plane (as shown in Figure 1), and the downstream crack at el 455.26m extended from the right abutment to a location 28.6m right from the reference plane. The crack ran through the upstream face to the downstream face and the total crack area was about 1000m2. This damage is an infrequent occurrence in RCC dams, and required a very difficult treatment process to fix the problem.

Preliminary analysis

According to laboratory test results – along with monitoring recordings from the engineer, the raw material quality, concrete construction quality, RCC physical and mechanical parameters and the results of bored dam concrete cores – the RCC quality of the 11th placing lift was no different to other lifts and met the design requirements. The time interval between the eighth and the ninth placing RCC layer also satisfies the requirements laid out in Construction Specification of Roller Compacted Concrete DL/T 5112-2000.

It is stated in the records of Linhekou project’s dam construction that when placing the 11th RCC lift (from 6 to 10 May 2002), water leaked from several tie-ins of the buried cooling PVC pipes, and then the cooling pipes under el 461.0m were blocked with sand grout on 30th August 2002. During the period from 6 May to 30 August 2002, the placed RCC was sometimes cooled by water of about 6-9°C from a cooling water pool (located at el 530m) and sometimes by natural river water (from 10-25 May, the river water temperature is about 15-24°). The water head from the water pool to the crack surface is roughly 75m.

Therefore the water pressure at the leaked tie-ins of the cooling pipes is very high, with powerful water splitting action. In addition, there is a temperature drop between the cooling water and the just-placed RCC mix – for example, in June 2002, the cooling water temperature was 6-9° and the water temperature at the exits of the cooling pipes was 10-18°, a temperature drop of about 4-9° at that time. The strength of the just-placed concrete is disadvantageously low at this time, as the water leaking out of the cooling pipes will, with the help of the high water head, dilute and sweep away the cementious materials along the placed RCC layer, then deteriorate the bond quality of the interface between RCC layers. Damage can occur from a weak interface, even from the initial crack from the tensile stress caused by the temperature drop.

As for the crack development after the first treatment, this may have been the result of the tensile stress caused by the internal concrete temperature drop. For example, the thermometer named T341 (located in the middle of the #3 and #4 induced joint at el 460m, 3.4m away from the dam upstream face) recorded a temperature drop of roughly 13° in October 2002. Another reason for the crack could be that high pressure water and winds used for cleaning the crack surface during the first treatment produced stress concentration at the crack tip, making the crack extend further, toward the crown cantilever.

It can be proposed that water leaking out of the cooling pipes, the water splitting action induced by cooling water head and the temperature drop from cooling the RCC too early are the possible causes of forming the weak plane and the horizontal crack along the RCC. The tensile stress caused by the internal temperature drop and the external excitations of high-pressure water and wind used for crack treatment may then have sped up the development of the crack.

Due to the limitation of preliminary discussion, however, the above deductions on the cause of the crack formation are based only on the qualitative analysis – any final conclusions should be made through further research.

Comprehensive treatments of the cracks

First treatment

According to the requirements of design specification for concrete arch dams SL282-2003 on the tensile strength of dam concrete and the characteristics of arch dam stress distribution, the first treatment measures included 184 drilled holes for anchor steel piles through the cracking surface, 84 drilled holes for chemical grouting and sealing the external boundary of the cracks.

According to the seepage resistance requirements of the arch dam, a long narrow slot was cut along the upstream cracking trace, then a waterproof material named SR and the epoxy sand grout were filled into the slot. Finally, a layer of epoxy vitreous texture was glued on the crack face.

The first treated region is about 890m2, i.e. from the right abutment to a location 26.75m at the upstream face and 10m at the downstream face left, away from the #3 induced joint. The chemical material used is the epoxy resin LPL, the viscosity of which is relatively large.

Second treatment

Because the initial crack treatment did not solve the problem, to ensure the seepage resistance and improve the tensile strength of the cracked RCC layer and to control the crack development, further crack treatment was carried out. Treatment measures included 150 drilled holes for chemical grouting, 19 drilled holes for installing anchor steel piles and 43 spare drilled holes for chemical grouting, placed in the intact region between the crack front line and the reference plane.

The treatment area was at the dam top face, at el 470.7m. Among the measures for the second attempt, six rows of chemical grouting holes were drilled. The first row was 2m away from the dam’s upstream face, the second row was 2m away from the first row, and each row after that was 3m apart. The holes are 16m deep and reach the plane that is 40cm under the crack surface, each has a diameter of 56mm. The anchor steel piles used are 18.6m deep and are placed to restrict the crack from developing; the grouting pipe system of the spare grouting holes was extended and collected in the observation chamber at el 478m. To meet the seepage resistance requirements of the arch dam, after beating away the improper filler for the first time, a longer and wider slot was cut along the upstream cracking trace, then the SR plastic filler and the epoxy sand grout were carefully filled into the slot. A layer of epoxy vitreous texture was finally glued on the crack surface, as shown in Figure 2.

The second treated region is about 1000m2 long, from the right abutment to 19.55m at the upstream face and 28.6m at the downstream face, right away from the reference plane. The chemical grouting material is a kind of epoxy resin named ZH798, which was successfully applied in the foundation treatment of the project, and is particularly suitable for crack or joint treatments with a width less then 0.3mm. Epoxy sand grout should be filled into the slot after driving away the water that is in the chemical grouting pipe systems and the cracking surface; chemical grouting can be carried out only after the strength of sand grout filled into drilled holes for anchor steel piles and the strength of the epoxy sand grout filled into the slot are more than 50% of the design strength.

Treatment procedures

The construction procedure for the crack treatment the second time was as follows: tidy up the dam top face (el 470.7m) for drilling holes; drill holes for chemical grouting and installing anchor steel piles; at the same time, dig the slot along the upstream cracking trace; tidy up the treatment working surface after drilling holes; clean the drilled holes for installing anchor steel piles with high water pressure; drive away the water from the drilled holes with high pressure wind; install the anchor steel piles then fill and compact sand grout into the interspace of the holes.

Wait for three days after the sand grout sets, then clean the drilled holes for chemical grouting with high pressure water and wind mixture; install the grouting pipe system, including an injection pipe and a vent pipe for each grouting hole, then fill and compact sand grout into the interspace of the holes; seal the crack upstream and downstream boundary and install vent pipes along the cracking trace; wait for another three days after the sand grout is set in the grouting holes, then clean the cracking surface with high pressure water; carry out water pressure test to determine the maximal grouting pressure; drive away the water in the grouting pipe system and cracking surface by high pressure wind after sealing the crack boundary and installing four inch (10.16cm) vent pipes on the crack upstream and downstream face.

Finally, drive away the water existing in the grouting pipe system and in the crack surface by pressing the acetone; drive away and reclaim the acetone by the high pressure wind; carry out the chemical grouting through grouting pipes on the dam top face; carry out the chemical grouting through vent pipes on the crack upstream and downstream face; wait for several more days after the slurry of epoxy resin ZH798 has set; carry out water pressure test for checking the treatment effects and continue to place RCC above the treatment working face.

According to the test results, the initial viscosity was about 16MPa under the environment with a temperature of 20°, and about 33-55MPa for the ordinary mix proportion after 12 hours. Thus, as long as the grouting pressure and the net grouting time could be guaranteed, the viscosity of ZH798 is suitable for the horizontal crack along the placing RCC layer.

Other factors

Of course, grouting effects are not only related to crack width. Other factors need to be taken into consideration, such as viscosity of grout slurry, grouting pressure, net grouting time and grouting equipment. For a crack with a specified width, selecting chemical grouting material, grouting equipment, grouting pressure, net grouting time and criterion of ending grout in accordance with crack width, crack area and crack orientation is very important to the treatment process. Moreover, grouting pressure, net grouting time and criterion of ending grout should be decided on the basis of field tests and adjusted along with the actual site construction.

Flow control pump special for chemical grouting was adopted for the second time, the outage and the grouting pressure can be adjusted in accordance with the actual requirements, and it is especially suitable for the fine large scale horizontal crack. An injection pipe and a vent pipe are placed for each grouting hole, and the injection pipe is connected to the grout pump with high pressure grouting pipes.

Field tests were carried out before grouting on a large scale to check the acetone pressure test and each procedure of the chemical grouting. Maximal grouting pressure during the chemical grouting is 1.2MPa. Water, a mixture of water and acetone, acetone, a mixture of acetone and grouting slurry, till pure grouting material slurry outflow in turn from the downstream crack and the downstream checking hole for consolidation grouting in the right abutment, verified that the effects of using acetone for driving away water and using grouting slurry for driving away acetone were very helpful for the crack treatment.

Effects of crack treatments

Grout consumption of the crack

In order to ensure the workability of the grouting slurry, the viscosity of the slurry was checked every 2-3 hours. According to original recordings, most of the initial viscosity varied from 8-12MPa, the final viscosity for each chemical grouting hole is less than 20-25MPa. During the whole process of chemical grouting, the grouting pressure is steady, the pressure value and the net grouting time met the design requirements.

As mentioned, the treated region the second time is about 1000m2. The grout consumption of the grouting holes on the dam top face is 5176.2 litres, of which the consumption of the grouting pipe system was 325.3 litres, the wasted consumption was 176.4 litres, consumption under grouting pressure (including leakage out of the crack) was 3674.5 litres and the net consumption forced into the crack by pressure was about 830.5 litres. Grout consumption through the vent pipes on the upstream crack was 32.7 litres, and through the downstream vent pipes it was 72 litres. The total net grout consumption was 935.2 litres – so it is obvious that a lot of chemical grout material slurry was filled into the crack.

Concrete cores were bored from the dam top face at el 515m in August 2003 – the longest concrete core is 10.57m. Some cores in the treated region were used to check the effects of the crack treatment; the crack section was fully filled with grout material.

Bond performance between grout material ZH798 and RCC

In December 2002, in order to check the bond strength between ZH798 and the RCC, RCC specimens 90 days old were glued with ZH798 slurry after the splitting tensile test. After curing the glued specimens for either 28 or 90 days, bond strength of the glued specimens was tested. The splitting tensile strength of the glued specimens aged 28 days was about 2.29-2.56MPa; those with an age of 90 days were about 2.26-3.23MPa. Judging from the splitting sections of glued RCC specimens, most of them ruptured from the intact RCC, but not from the glued splitting section, indicating that concrete surfaces can be bonded well by ZH798.

Indoor bond test is different from the actual chemical grouting. When boring the dam RCC cores located in the crack region, many cores broke at the cracked section; this implies that the bond strength of the actual chemical grouting is less than the tensile strength of intact dam RCC.

Results of water pressure test

Four holes were bored in the treated region for a water pressure test in August 2003. The test results show that the water permeability of the test segment from el 456m to el 454m (the crack is located at el 455.1m) is 0.00312 Lu, 0.00284 Lu, 0.0053 Lu and 0.0366 Lu respectively, meeting the design requirements. Before the first filling of the dam, water pressure tests were carried out for the spare grouting pipes in the observation chamber, and the results indicate that there is no further crack development. In addition, the treatment effects were verified by the dam retaining water during the flood season in July 2003, its first filling since October 2003, and the fact that it has generated power since May 2004. There is no leakage in the treated section, and the observation data indicates that the dam operates safely after initial filling.

Experiences and lessons

• It is inevitable that RCC will have to be placed in high temperature conditions when repairing a large-scale RCC dam, and buried cooling plastic pipe is one effective temperature control measure. But the tie-ins of cooling pipes should be processed carefully to avoid damage from the compact rollers and leakage out of the pipes. For example, vulcanisation is the best way to connect tie-ins.

• Select the right time to begin the RCC cooling in accordance with the setting time of the RCC mix and the temperature variation of the RCC. Thus, even if there is some damage in the cooling pipes, the induced leakage cannot lead to large-scale damage to the RCC layer, as long as the strength of the placed RCC is high enough at that time.

• There are many kinds of chemical grout materials for concrete crack treatment available. The chemical grout material should be selected in accordance with the different characteristics of the actual crack. For example, the epoxy resin LPL was successfully applied in the vertical crack treatment of China’s Three Gorges project, but failed in the horizontal crack treatment of Linhekou arch dam at the first attempt because the high viscosity of the slurry, the little outage and the low grout pressure of the grouting pump were not appropriate to the fine large scale horizontal crack. However, ZH798 was successfully applied in the crack treatment on the second attempt, as its viscosity can be adjusted, the outage of the grout pump is large and the grout pressure can be very high – all these are needed for fine large scale horizontal crack. So, selecting grout material, grout equipment and grout parameters in accordance with the crack width, the cracking area, the crack orientation and other characteristics of the crack is vital to crack treatment.

• It is necessary to select appropriate grouting time in accordance with the air temperature and the RCC temperature variation. Grouting in low temperature conditions with steady temperature stress in the dam and without crack development can bring more success. However, it is usually difficult to make such a measured decision, since waiting to perform crack treatment just prolongs the construction time and delays the initial generation time, as the crack’s condition gets worse.

• It is important to ensure the crack surface is clean before chemical grouting by avoiding the pollution from wastewater and powder from drilling holes – this can improve the bond quality between chemical grout material and concrete. For example, chemical grouting should be carried out before installing anchor steel piles; the pollution to the crack surface can be lessened and the bond strength improved.

• Placing necessary instrumentation on the cracks to observe the operation period of the project can help engineers learn lessons for the future.

At the present time, the Linhekou dam crack treatment work has proved successful. However, the final conclusion can be made only after a longer period of operation. Although costs were high for the crack treatment, it provided lessons for other RCC dam construction and valuable suggestions for similar crack treatment.

Author Info:

Chen Guanfu, Zhou Jianping and Zhao Quansheng are with China Hydropower Engineering Consulting Group Company, Beijing, China. E-mail: