CHAPTER 51 THE ECONOMIC VALUE OP A NEW BREAKWATER ARMOUR UNIT 'DOLOS' E.M. Merrifield and J.A. Zwamborn ABSTRACT The Dolos, a new type of armour unit which closely resembles a normal ship's anchor, was developed and tried out under field conditions on the main breakwater of East London harbour. Since these full-scale Dolosse proved very successful, tests were made in a wave channel to compare the stability of Dolosse with other known types of armour blocks. The test results showed that the Dolos is outstandingly stable, and since manufac- ture and random placing of Dolosse offers no particular difficulties it is concluded that in many cases the use of Dolosse in armour layers may lead to more economical solutions for rubble mound breakwater and shore protec- tion works. INTRODUCTION Rubble mound breakwaters are normally protected against damage from storm waves by a cover layer of very heavy armour units or breakwater blocks. If natural rock blocks were to be used for this purpose, the re- quired unit weight may be in the order of 40 tons and more. Rocks of this size are difficult to obtain and almost impossible to handle on any large scale. It is, therefore, quite understandable that harbour design engi- neers and research workers alike have done their utmost to develop smaller concrete blocks which, due to their particular shape, would form an inter- locking cover layer of much higher efficiency. As a result many different types of blocks have been developed, varying in geometric shape from the simple rectangular or cubular block to highly complicated shapes such as tetrapods and hezapods. A new type of armour unit for breakwaters and coastal protection works, named 'Dolos', was developed by the senior author. A number of Dolosse were tried out under field conditions on the main breakwater at the Port of East London, South Africa. The results of these full-scale tests appeared promising and it was decided that, in order to obtain more comparative data, Dolosse, rectangular blocks, tetrapods and tetrahedrons be tested compre- hensively in the wave channel of the Council for Scientific and Industrial Research in Pretoria. System Harbour Engineer, South African Railways and Harbours Administration, East London, South Africa. Head, Hydromechanics Research Department, South African Council for Scientific and Industrial Research, Pretoria, South Africa. 885 886 COASTAL ENGINEERING Armour units are normally dumped at random. Due to their particular shape it is possible, however, to pack the Dolosse in a regular pattern. Tests were, therefore, made with both randomly dumped and regularly packed Dolosse, although it was realized from the start that it would be extremely difficult, if not impossible, to realize the latter in practice. EXPERIENCES WITH DOLOSSE IN PRACTICE Construction on the main breakwater at East London commenced in August 1873> with the tipping of rubble on the foreshore. However, little advance was made until the first 25-ton* rectangular block was placed into the sea in March 1876. Thereafter, the breakwater was constructed as a mound formed by rectangular blocks weighing from 15 to 30 tons each, topped with a 36-ft wide concrete cap reaching to 16 ft above LWOST306 and a seaward parapet of 5 ft 6 ins high. By 1884, 1,500 ft of breakwater had been completed and the structure was ended off with a round head. Between 1911 and 1917 the breakwater was extended a further 776 ft using 40-ton rectangular blocks placed at random while the end portion was raised to 19 ft above LWOST. In 1935 the third and final stage of construction commenced. The breakwater was extended by a further 1,000 ft, also to 19 ft above LWOST using 33-ton blocks. This work was completed in 1939 and the breakwater is now 3,276 ft long. DEVELOPMENT OP THE «D0L0S' The seaward face of the breakwater was at one time protected with a random layer of 35-ton rectangular blocks over a length of 1,000 ft on the seaward end of the breakwater and with 41-ton blocks over the remainder. During 1944 a severe storm breached the breakwater some two hundred feet from the end, carrying away a considerable number of 33-ton protective armour blocks. The breakwater was repaired and the whole seaward face pro- tected to a height of 24 ft above LWOST with 41-ton rectangular blocks placed at random to an approximate slope of li horizontally to 1 vertically. In I963, i.e. nineteen years afterwards, it was estimated that the outer half of the breakwater had lost at least fifty per cent of its seaward ran- dom block protection, while a few sections were almost stripped bare to the original mound core. It was, therefore, evident that the existing rectan- gular 41-ton armour blocks did not provide a stable protection and, if the high costs of replacement were to be brought down to a reasonable figure, some other type of armour block would have to be used. Consideration was given to various known types of specially shaped blocks but, due to restrictions (patent rights) and the costs involved, it was decided rather to develop some other original form. Wooden models were * 1 Ton « 2,000 lb. ** Low Water Ordinary Spring Tide. ECONOMIC VALUE 887 made of numerous block shapes based on the idea, firstly, that they should form a cover layer with a high void to solid ratio, to facilitate dissipa- tion of wave energy and, secondly, that each block should be linked with others to form a knitted composite structure, rather than a loose group of individual blocks. Moreover, the block should have enough mechanical strength to withstand the rigours of rough handling when being placed on the breakwater, and the shape should be such that the blocks can be manu- factured economically. The shape that seemed to satisfy these requirements best was the 'Dolos1, an anchor shaped block with dimensions as shown in Figure 1. The name 'Dolos' (plural 'Dolosse') was given to the block because of its South African association. The name refers to the knuckle bones of a sheep or goat, used by children as toy oxen in the old trek (pioneering) days, and also to the small bones used by African witchdoctors for divining. Packing and placing tests of the wooden models on various slopes were carried out and it was found that due to the anchor shape of the Dolos one leg always hooks into the underlayer, while due to the legs being tapered towards the ends the blocks are wedged tightly between other blocks, thus forming a good interlocking structure. Preliminary tests were also carried out to determine whether a more economical result could not be obtained by laying the blocks to pattern. However, it soon became evident that, in practice, the task of laying to pattern on rough slopes, battered by an ever-moving sea, would be virtually impossible. It was then decided to manufacture some full-size Dolosse and to test these blocks on the East London breakwater. MANUFACTURE OF DOLOSSE At this early stage no laboratory tests had been carried out, but it was nevertheless decided to select a size of Dolos that was less in weight than that which would have to be used for other well known types of blocks and, at the same time, would be large enough to interlock with the remain- ing rectangular blocks on the breakwater face. The selected size was an eleven-foot high block (h «= 11 ft) weighing 19$ tons. The waist was slight- ly thicker than the dimension 0.3 h which is shown in Figure 1. It was brought to a round figure of 3 ft 9 ins (i.e. O.34 h). The slight thicken- ing of the waist for the larger sizes of Dolosse is considered a reasonable adjustment to cope with the higher stresses in the concrete during handling. The hexagon cross-section, shown in Figure 1, was preferred to a cir- cular one for ease of making the shuttering and extracting the Dolos from the mould. In practice, this section is near enough to a circular one to prevent undesirable concentrated flow, resulting in high run-up, and reflec- tion of wave energy on large flat surfaces. The moulds were built up of 3/l6-inch thick mild steel plate panels flanged and ribbed around all edges and bolted together. These casings are fixed permanently in one position with their lower halves in a pit and with the upper surfaces left open to receive the concrete mix (see Figure 2). 888 COASTAL ENGINEERING The Dolos is lifted from its mould within 18 to 24 hours after cast- ing, depending on air temperatures. In order to remove the cast Dolos, one section of the mould on the horizontal leg is folded hack and the upper section is removed in one piece, while two vertical joints on the lower vertical leg are merely loosened to break the suction (see Figure 2). This system of removing the casting from the mould shortly after pouring con- crete considerably reduces the required number of moulds and, consequently, the size of the casting yard. The following concrete mix (by volume) was used: 1 Portland cement 2.66 stone, i inch to dust ) 1 Slagment 2.66 stone, -f inch to •£ inch) graded aggregate 2 Sea sand 4.00 stone, 1^ inch ) In order to ensure an initial strong resting toe at the bottom of the vertioal leg, the first mix poured into the mould has cement substituted for the slagment. Sufficient water is used to provide a stiff workable mixture, which is compacted with a small pencil vibrator. The mixture is a strong one, but this is considered necessary in order to develop a high mechanical strength in the Dolos and in order to minimise chemical and abrasive attack on the concrete. Slagment was originally used because it is cheaper than cement and presents less storage problems. Since the South African Railways Research Laboratories have recently thrown some doubt on the good properties of slagment when used under alternating wet and dry con- ditions in the sea, Portland cement will, in future, replace slagment. However, Dolosse placed on the breakwater two years ago have as yet shown no signs of chemical deterioration. HANDLING AND PLACING OP DOLOSSE A frame consisting of three pieces of scrap rails (80 lb per yard) tack welded together is placed along the central axes of the three legs of the Dolos mould (see extreme right Figure 2). Two steel rope lifting loops are wound around the central rail while the ends project out of the mould providing lifting eyes after casting the block (see Figure 2). In this way it is possible to lift the blocks out of the moulds only one day after casting. A study of eye-bolts cast into old blocks and concrete structures at East London harbour many years ago had shown that the metal has only corroded to slightly below the concrete surface. No damage had been suf- fered by the concrete when the cover around the protruding steel was thick enough. It is, therefore, felt that corrosion of the lifting loops pro- truding from the Dolosse will cause no significant damage to the concrete. The freshly cast Dolosse are carefully placed in a nearby curing yard and left there for seven days. Thereafter they are closely packed in the final curing yard and left for a minimum period of 21 days (total minimum curing time 28 days). The Dolosse are finally transported onto the breakwater in railway trucks and placed by a 40-ton capacity travelling Titan crane having a max- imum reach of 65 ft. The blocks are placed over the existing 41-ton rec- tangular blocks to an average slope of about 1§ to 1. The lifting loops ECONOMIC VALUE 889 are not used for this operation but the Dolosse are slung around their middle sections by means of an ordinary wire rope sling with a trip hook fixed at one end. DOLOSSE PLACED ON THE EAST LONDON BREAKWATER A small number of the 19§-ton Dolosse were placed in a line (not inter- locked) on a section of the foreshore near the root of the breakwater to test the individual characteristics of the blocks. They were subjected to breaking waves up to 18 ft in height and, although only seated on small loose round boulders, they moved very little by swinging sideways and tend- ing to "dig in". They showed no tendency to roll or glide away as happens to rectangular blocks. By the end of 1965 approximately 450 Dolosse had been placed at random around the end of the breakwater and along a short section of its seaward face (see Figure 3)« I* was found during the first onslaught of a severe storm that Dolosse, which were not completely stable yet, moved into more secure positions and a general "settling down" of the Dolosse occurred, forming a permanent and better packed group. After this initial settling no subsequent movement has been observed and the blocks have now withstood the severest storms, with estimated wave heights of up to 25 ft, of two winters, while during the first winter (1964), five 41-ton rectangular blocks were swept over the breakwater cap, at a section where there was no Dolos protection. During a storm or 'heavy seas', and particularly when the wind is blowing in the same direction as the waves, it is quite impossible to tra- verse the breakwater due to large amounts of water splashing over the top, and due to strong clapotis. On one occasion when the waves were estimated to be of the order of 20 ft high, the only manner in which the light at the end of the breakwater could be reached was by means of a steam locomotive. At the round head, which is protected by Dolosse, it was possible to walk about the breakwater deck with perfect safety, and only a light spray brought over by wind was experienced (see Figure 4)« No damage of any sort, including erosion, has been observed in any of the Dolosse over a period of two years and, although many blocks fell and slid four to five feet during placing, none of them suffered any damage except for minor chipping of the edges, DESIGN CRITERIA FOE BREAKWATER COVER LATERS A schematic cross-section of a rubble mound breakwater is shown in Figure 5« The main body or core of the breakwater may consist of normal quarry run material. This core is covered by rocks of various sizes (so called 'underlayers') over which armour units forming the final cover layer are placed. In Figure 5» the required rock weights as given by Hudson1 are all expressed as a proportion of the equivalent block weight (W ) of the e armour units. The equivalent block weight is defined as the weight of quarry stone which provides the same protection as the particular armour unit (having a weight W) to be used. Although Hudson's approach is quite 890 COASTAL ENGINEERING acceptable when using known armour units, it will become clear later that in the case of Dolosse it may be better to define the size of the stone in the underlayer as a proportion of the actual weight (W) of the Dolos. In order to arrive at an economic breakwater cover layer design, factors such as design wave height, stability of blocks, porosity of the cover layer, shape factor of the blocks and wave run-up should be taken into account. These factors are dealt with in more detail in the following sections. DESIGN WAVE HEIGHT Figure 6 is a typical diagram for the Cape Town area showing the fre- quencies of occurrence of deep sea maximum wave heights (H ) for five 0 max directions. The frequency of occurrence lines are based on just under one year's records collected by the Division of Sea Fisheries research vessel Africana II in deep sea, using the N.I.O. accelerometer type wave recorder. It is realized that the recording period is short, but since no better in- formation on waves in South African waters is available at present, the lines shown in Figure 6 are the only basis for design (at least for the West and South coast of South Africa) until such time as more wave data be- come available. Similar wave data are being collected for other places on the South African coast at present. While Figure 6 refers to deep sea wave heights, the design wave height (H) for a particular location on the coast is easily determined from these deep sea wave characteristics by using the well known refraction analysis^ and after taking into account the effect on wave height of the reduced water depth in front of the breakwater*. REQUIRED BLOCK WEIGHT AND STABILITY FACTORS The required weight of an individual armour unit may be determined from the following formula given by Hudson1: L 5 H (1) Kp A5 ootg <X where W is the block weight, Y the specific weight of the armour unit, e H the design wave height, A the relative density of the block ( A « ( Y - ^ )/J where jf is the specific weight of water), o<. the slope angle a (see Figure 5) and Kj the stability factor. For the cases where no damage is allowed at all the stability factor (Kj) is defined by equation (l) when H is the wave height at which damage just starts. Stability factors for the no-damage and no-overtopping criteria as given by Hudson1 are shown in Table I. These values are reported to apply only to the trunk of the break- water (not for breakwater heads) and where the waves do not break just be- fore the structure. Moreover, since the influence of factors such as ir- regularity of waves, methods of placing the units and permeability of the rubble mound structure are all combined in the single parameter Kj) it is necessary to use some care when applying model Kj) values for prototype design. Based on a very limited amount of full-scale field data Hudson ECONOMIC VALUE 891 suggests a minor adjustment of KJJ values for full-scale block design. TABLE I. STABILITY FACTORS (K ) ACCORDING TO HUDSON D K D Method of Armour Unit placing Recommended for Model Values full-scale random, double Quarry stone 3 3 layer Tetrahedrons H tt 5.5 Tetrapods H tl 8 8 Hexapods n -n 10 9 uniform, single Hexapods 22 layer Faape et al have shown that the stability factor can be expressed as a function of the damage. Much larger values for KJJ are found to be appli- cable when a few per cent of damage is considered acceptable. In this case the cost for the required maintenance will have to be weighed against extra capital investment when using larger armour units to arrive at the most economical design. POROSITY, THICKNESS OP COVER LAYER AND REQUIRED NUMBER OP BLOCKS The porosity (p) is defined as the percentage voids of the total volume of the cover layer. A high porosity of the armour layer is beneficial since wave run-up as well as the total concrete volume required in the cover layer are reduced. The thickness (r) of an armour cover layer may be defined as: r - n C V1/3 (2) where n is the number of layers, C a shape factor which is related to the packing density of the blocks, and V the volume of the block. The required number of blocks (N) to cover a unit area is then found from« . -P.) - /3 n 0 (1 v 2 (3) Since the number of blocks required to cover a given area of the breakwater slope is proportional to the shape factor (C), low values of C should be aimed at in block design. 892 COASTAL ENGINEERING WAVE RUN-UP The wave run-up (R) determines the crest height of a non-overtopping breakwater (see Figure 5). High porosity results in a reduced wave run-up. Block shape also affects wave run-up. MODEL TESTS Tests were made in the outdoor wave channel of the Council for Scien- tific and Industrial Research. This channel is 4 ft wide, 3.5 ft deep and has a total length of 111.5 ft, the effective length (distance between wave paddle and model breakwater) being about 90 ft. Waves are generated by a paddle which is driven by an electric motor through a variable speed hydrau- lic transmission. In front of the wave generator is a wave filter which absorbs, to a large extent, waves reflected by the model. Wave heights of between 4 and 14 inches and wave periods of between about 0.5 and 5 seconds could be produced with the available equipment. Three different sizes of model Dolosse (weighing 993» 427 and 185 gr. respectively), two sizes of rectangular blocks (1,262 and 929 gr.), model tetrapods (834 gr«) and tetrahedrons (594 S1*) were tested in the wave channel?. The three types of Dolosse were not exactly geometrically simi- lar. The values given in Figure 1 are the mean dimensions of the three types of Dolosse and, in fact, they agree closely with the geometry of the medium size ones. The thickness-to-height ratio was 0.34 for the large size, 0.27 for the small size and 0.31 for the medium size Dolosse. The large sizes were thus relatively heavier whereas the small ones were about 25 per cent more slender than the large ones. Due to this, a slight dif- ference in behaviour regarding stability could be expected. TEST CONDITIONS All armour units were tested on a slope of 1 in 1.5 and were generally put down in two layers dumped at random on an underlayer of quarry stone. The weight of the underlayer stone was £ of that of the medium size Dolosse or 0.1 of that of the large size rectangular blocks. The cover layer reached from 1 ft below to 1 ft above mean water level. At lower levels quarry stone, having a weight of about twice the medium Dolos weight, was used in the primary cover layer. In the case of the Dolosse, tests were also made with the blocks placed on a regular pattern as a single layer. Two typeB of armour units were tested simultaneously side by side in the flume. Wave heights were increased in steps of about 2 ins from 4 ins to 14 ins, each step constituting a test run. Separate series of tests were carried out for wave periods of 1.2, 2 and 3 seconds. The water depth in front of the model breakwater was 2.5 ft to still water level in all cases. "Damage" was assessed in the main tests in terms of the movement of a block over a distance greater than 2 ins (called 'damage'). This concept of damage was later broadened to include those blocks which rocked to and fro to such an extent that structural damage would probably occur and the ECONOMIC VALUE 893 blocks would be lost effectively for wave absorption (called 'total damage'). These features were recorded by visual observation during* each test and a check was provided by taking photographs before and after each test. Per cent damage was calculated in terms of the total number of blocks placed on the face. Waves produced by the mechanical wave generator were of the regular type comparable, to some extent, with regular swell in nature but not with storm waves. Due to the great depth in front of the structure waves only broke on the model breakwater itself. Hudson-*- found that for the shallow water case, when waves break just before the structure, somewhat lower stability factors than the standard values (Table I) must be applied. Since the aim of the present study was to compare the behaviour of the Dolosse with that of other armour units only, it was considered acceptable to limit the tests to the deep water oase using regular waves. However, for full-scale application in a particular situation possible effects on block stability of irregular waves (wave spectrum) and shallow water should be taken into account. RESULTS OP STABILITY TESTS The stability factor which is typical for a particular type of block follows from equation (l) viz.: - a3- ( ) 4 V hJ cotgc< For a particular armour unit having a volume V and a relative density A placed on a breakwater face of slope c< , KD is thus proportional to H^. For the 0%-damage case the value of H was taken to be the wave height at which damage just started (comparable with Hudson's no-damage case, see Table I). In addition to the Of or no-damage values for Kp as defined above, one can also define Kj) values for x^-damage (x/> 0), which are, of course, associated with higher waves (height H )>H). In these cases H X x must be substituted in equation (l) to obtain Kj) . The test results for 'total damage' are summarised in Figure 7 where KD values are plotted against per cent 'total damage'. An important point is how the test results compare with previously published ones. Test re- sults obtained at Delft for cubes and tetrapods, as reported by Paape et al^, are therefore also shown in Figure 7. The results obtained in Delft for cubes are seen to be in very close agreement with the CSIR's tests on rectangular blocks. The same holds for tetrapods for the lower (and thus the more important) percentages of total damage. It is, therefore, con- cluded that the agreement with previously published results is quite satis- factory, bearing in mind the possible minor differences in test conditions (e.g. initial packing of the blocks) inherent in this type of investiga- tion. Hence the test results for the new Dolos block may be relied upon with confidence. Hudson1, in Table I, refers to KD values for the 0^-damage case only. Comparable values for Kj extracted from the CSIR's results are shown in 894 COASTAL ENGINEERING Table II. TABLE II. CSIR'S STABILITY FACTORS (K-p) FOR THE C$-DAMAGE CASE Method of *D Armour Unit placing Damage Total damage random, double Rectangular blocks 2.5 2.3 layer Tetrahedrons n n 1.5 1.2 Tetrapods it n 6.5 2-5 Dolosse M II 40 24 uniform, single Dolosse 25 20 layer The KD values of Table II for tetrahedrons and tetrapods are much smaller than those given by Hudson (Table I). However, the 0$-damage stage is extremely difficult to decide upon. It is, therefore, quite possible that at the stage which Hudson selected as 0?&-damage some minor damage had, in fact, taken place. For instance it is seen from Figure 7 that the values of Kp for l^-dam8ge compare very well with Hudson's values. This clearly demonstrates the deficiencies of the no-damage criterion and emphasizes the importance of determining the actual damage which takes place for each par- ticular wave height**. The validity of the model Kj values for random Dolosse (Kj • 24) is confirmed by the experience with the 19§-ton Dolosse on the East London breakwater head. These blocks withstood, without moving, 25-ft high waves and with y * 150 lb/ft', A « 1.34 (seawater) and cotg o< = 1.25 this a means a Kj) value of 19.6 or more. Since a slightly smaller stability factor may be expected for the breakwater head compared with the trunk of the breakwater the agreement with the model value is considered good. It is clear from Figure 7 that Dolosse are much more stable than the other types of blocks. Since the required block weight is inversely proportional to Kj) the high values of Kj) for Dolosse mean that smaller individual units may be used for a particular design wave height. The rapid increase of the Kj value after a few per cent of damage empha- sizes the strong tendency of the Dolosse to interlock, thus forming a semi-monolitic cover layer of great stability. Although the Dolosse packed to pattern seem to compare very favourably with other blocks dumped at random and, in fact, the zero-damage KD values were found to agree closely with the Kj value for uniformly placed hexapods (Dolosse, KJJ = 25 to 20, hexapods K5 > 22, see Tables I and II), a serious disadvantage of the packed Dolosse was found to be that once damage has started the coherence of the structure is lost and total failure results. This makes the use of packed Dolosse much more risky compared with any type of armour unit dumped at random and since placing to pattern on real break-