Summary: A quay wall retains soil on one side and holds back water on the other, and the standard that governs its earthquake damage publishes two scales for the same wall. One is the residual horizontal displacement divided by wall height, and the other is the residual tilting angle. Converted into one unit they do not agree, and the gap is widest at the lightest damage grade, where the displacement table stops at 1.5 percent of wall height while the tilting table lets the same wall reach 5.24 percent. A single wall can therefore be graded Degree II by one table and Degree I by the other. The two scales are not two units for one quantity: sliding and rotation are separate failure modes, and each one fills only its own column. This guide converts every published criterion into one unit, sets each field instrument against the column it can actually fill, and explains why on a waterfront the water behind the wall, not the tide in front of it, decides the load.
1. Why a Quay Wall Answers to Two Load Systems at Once
A retaining wall on land carries one load. A quay wall carries two, and the two are measured by different rules. On the soil side the wall resists an earth pressure that depends on the friction angle of the fill and on how far the wall has been allowed to yield. On the water side it resists a hydrostatic pressure that does not depend on the soil at all, because water has no shear strength and therefore cannot enter either the active or the passive state. It simply pushes. That single difference is why a blocked drain, and not a difficult soil, is the most common cause of failure in this class of structure.
Three coefficients make the point in one line. For a soil at a friction angle of 30 degrees the active earth pressure coefficient is 0.333 and the passive coefficient is 3.000, a ratio of nine. The hydrostatic coefficient is 1.000, which is three times the active coefficient and sits between the two soil states. A backfill whose drainage is working is therefore pushed on by about a third of the water column height, and the same backfill with a failed drain is pushed on by three times that, before any soil has moved at all. Published retaining wall practice puts the consequence in one sentence: a blocked drain that lets the water table rise to the surface can more than double the total thrust.
| Thrust source | Coefficient | What sets the coefficient | Value at 30 degrees |
| Retained soil, active state | Ka | the soil friction angle, and how far the wall has moved away from the fill | 0.333 |
| Retained soil, passive state | Kp | the soil friction angle, and how far the wall is pushed in | 3.000 |
| Free water behind the wall | 1.000 | the water level, and nothing else | 1.000 |
Two further loads arrive at the quay face rather than through the fill, and on a working berth both are measured rather than assumed. Cargo, containers, cranes and vehicles stand on the apron and enter the earth pressure diagram as a uniform surcharge, which adds a rectangular block over the full wall height and raises the level at which the resultant of the combined thrust acts. Mooring and berthing arrive as horizontal forces at the bollards and through the fenders, and they are not small. A recent instrumented quay wall in the Port of Rotterdam carries load cells in its ground anchors, water level sensors, tilt sensors, displacement sensors and mooring load sensors at the same time, which is the honest statement of how many separate things can move a quay wall.
The design literature is equally clear about which of these governs. Soil pressure is the governing load acting on a soil-retaining quay wall, and the reason is the loading history rather than the soil. A sheet pile or diaphragm wall is built by dredging in front of it, which removes the passive resistance that was holding it, and then by placing fill behind it, which adds the active thrust that will push it. Both operations are temporary and both are large, and the wall is asked to hold during them. The reference models that predict what the wall will do during those stages are exactly what the instrumented quay walls of the last few years have been built to check, and the published description of one new wall as globally unique for carrying them is itself a measure of how rarely that comparison has been available.
2. The Tide Lag That Only Matters in One Direction
The water next to a quay wall is not a single level, it is two. The level in the harbour follows the tide. The level inside the fill behind the wall follows the tide too, but not at the same speed, because the fill has a finite permeability and the path from the harbour into the fill is long. This is called tidal lag, and its behaviour is asymmetric in a way that decides the design.
As the tide rises, the harbour level climbs faster than the level inside the fill, so the pressure on the seaward face temporarily exceeds the pressure on the landward face. That difference helps the wall. As the tide falls, the harbour level drops faster than the level inside the fill, so the pressure on the landward face exceeds the pressure on the seaward face, and that difference hurts the wall. Published waterfront design guidance states the asymmetry without hedging: the lag in the rise of the water level in the fill is of no concern as regards design of the wall, but the lag in the fall of the water level creates an unbalanced water pressure which must be resisted by the structure. The governing storm is therefore not the high tide. It is the falling tide, with a high water table still standing behind the wall.
Because the size of that lag cannot be measured on every project, the design guidance replaces it with a rule. For an impermeable fill the assumed lag is half the mean tidal range, and the same half range is assumed for a low permeability wall such as sheet piling with few weep holes, because the wall itself then limits the flow. Where the fill is permeable or drained, the assumed lag is zero. Where the fill has an intermediate permeability and the wall itself is permeable, the guidance treats the water standing above low tide as a surcharge rather than as a lag, which is the same load under a different name.
| Drainage condition | Assumed lag | Note |
| Permeable fill, or free drainage | None | weep holes clear and working |
| Impermeable fill, at or below 0.001 foot per minute | Half the mean tidal range | the usual design assumption |
| Low permeability wall such as sheet piling, few weep holes | Half the mean tidal range | the wall itself restricts the flow |
| Permeable wall with gravel chimneys at the weep holes | Water above low tide treated as a surcharge | the fill drains, the load does not leave |
The rule contains a failure mode that no amount of design removes. A weep hole is a deliberate hole through the wall that lets the two water levels equalise, and it works only while it is open. Marine growth, sand and debris close it, and published maintenance guidance for waterfront walls is unusually direct on this point: weep holes must be kept completely clear, and a wall whose drainage has failed will bow, crack its cap or move forward as a whole. On a structure that is inspected from a boat, the weep holes are the part of the asset that is least visible and most decisive.
The practical consequence for a monitoring plan is a boundary that this article will keep throughout. An angle channel reports what the wall has already done. It says nothing about the water level that is doing the pushing. On a quay wall the two are not substitutes, and a port that monitors only the wall is monitoring the result of a cause it is not watching.
3. The Two Damage Scales That Sit in the Same Standard
The reference document for this subject is the seismic design guideline for port structures published in 2001 by the International Navigation Association, known by the acronym PIANC, under the title Seismic Design Guidelines for Port Structures, as the report of Working Group 34 of its Maritime Navigation Commission. It is a performance based document, which means it does not simply say pass or fail. It defines four damage degrees, from serviceable through repairable and near collapse to collapse, and it lets the owner choose the degree the facility has to meet. The damage criteria sit in a chapter of their own, and that chapter contains two scales.
The first scale is the residual horizontal displacement of the wall, normalised by dividing it by the height of the structure. The second is the residual tilting of the wall towards the sea, expressed as an angle. Both are given four rungs, and at Degree I, the lightest, the displacement scale stops below 1.5 percent while the tilting scale stops below 3 degrees. Read as they are printed the two ceilings look unrelated, and the reason they look unrelated is that nobody has converted them.
Convert them and the disagreement appears. A wall whose top has moved horizontally by a distance d over a height H has rotated by the arctangent of d divided by H, so the displacement scale can be read as an angle and the angle scale can be read as a displacement. Doing that turns the three displacement ceilings into 0.859, 2.862 and 5.711 degrees, and it turns the three angle ceilings into 5.241, 8.749 and 14.054 percent. The two sets of numbers were never the same, and they were never meant to be, but they are printed in the same document under the same four grade names.
| Damage degree | Displacement ceiling, and the angle it equals | Angle ceiling, and the displacement it equals | Offset between them |
| Degree I | 1.5 percent, which is 0.859 deg | 3 deg, which is 5.241 percent | 3.49 times |
| Degree II | 5 percent, which is 2.862 deg | 5 deg, which is 8.749 percent | 1.75 times |
| Degree III | 10 percent, which is 5.711 deg | 8 deg, which is 14.054 percent | 1.41 times |
The size of the gap is worth reading carefully, because it does not behave the way a scaling error would. The two scales are offset from each other by about 3.8 percentage points at every rung, which is nearly constant, but as a ratio the offset falls from 3.49 times at Degree I to 1.75 times at Degree II and 1.41 times at Degree III. In other words these are not two units for one quantity with a fixed conversion error. They are closest to agreement where the damage is severe, and furthest apart where the damage is light.
That is the worst possible place for the gap to sit, because the light grades are the ones a monitoring system is installed to catch. Degree I is where an operator wants to know that something has started, and it is the single rung where the two published ceilings differ by a factor of three and a half. A wall whose displacement has reached 1.5 percent of its height is at the top of Degree I on the displacement scale and, at 0.859 degrees, is nowhere near the tilting scale, whose Degree I tolerates 5.241 percent. There is even a narrow window between 5.00 and 5.24 percent of height where the displacement table puts the wall at Degree III, close to collapse, while the tilting table still puts it at Degree I, serviceable. The window is a quarter of one percentage point wide. Its existence is the point.
4. Why One Wall Can Fill One Table and Not the Other
The natural first reading of the previous section is that the standard contains an arithmetic error, and that one of the two scales should be replaced by the other. That reading is wrong, and the reason it is wrong is the reason this article was written. The two scales are not two expressions of one quantity. They are two expressions of two different failure modes, and a real wall can exhibit one without the other.
A gravity quay wall is a mass resting on the seabed and holding back fill. The published analysis of instrumented rigid support structures states that the common deformation modes of such a structure are sliding, tilting and settlement, and that the parameters used to define the damage levels are linked to those modes. Sliding is translation of the whole mass along its base, and it produces horizontal displacement with almost no rotation. Tilting is rotation about the toe, and it produces both, in the fixed ratio that the arctangent gives. Settlement is vertical, and it produces neither, which is why the same guideline carries a third set of criteria for it, expressed as the differential settlement of the apron and the differential settlement between the apron and the land behind it.
Because the failure modes are separable, the damage evidence is separable too, and the largest recorded port earthquake in the reference literature is the cleanest example of a wall that filled one column and not the other. At Kobe in 1995 more than 250 caisson type quay walls were damaged by the Hyogoken-Nanbu earthquake. The published description of that damage says the caisson walls did not collapse or overturn and deformed quite uniformly, maintaining the face lines of the walls almost straight. A wall whose face line stays straight has translated as a body. Its displacements were large and its internal rotations were small, and it therefore had a great deal to report on the displacement scale and very little to report on the tilting scale.
A better documented single case is the block type quay wall at Derince Port in Turkey, which was struck by the 1999 Marmara earthquake at a peak ground acceleration of roughly 0.25 to 0.30 g. The site measurements record 0.50 metres of seaward lateral displacement with no vertical component, while the fill behind the wall settled by 0.5 to 0.8 metres. The wall height was 14.75 metres. Divide 0.50 metres by 14.75 metres and the normalised displacement is 3.39 percent, which places the wall in Degree II on the displacement scale. Take the arctangent of the same ratio and the tilting is 1.94 degrees, which places the very same wall in Degree I on the tilting scale. One wall, one earthquake, two published grades, differing by a whole grade.
The literature does not leave the resolution open. The published analysis of that case states that horizontal displacement was found to be more critical than tilting, and that the acceptable damage level for the wall was obtained by considering the critical condition. That is the practical rule that has emerged from the whole subject: where the two scales disagree, take the more severe one. It is a defensible rule and it is also an admission that the scales were written independently and are not interchangeable.
There is a second consequence of the same separation, and it is the one that matters most to anybody selecting an instrument. If sliding and rotation are separate modes, then a measurement of one of them cannot rule out the other. A tilt channel mounted on the wall reports the angle of its mounting base. If the wall slides as a rigid body, that angle does not change, and the tilt channel reports a healthy wall while the wall has moved half a metre. The reverse statement is also true and is easier to act on: a displacement measurement of zero rules out both modes, because neither mode can occur without moving a point. This asymmetry, and not the sensor’s accuracy, is the reason a quay wall plan cannot be built on tilt alone.
5. The Deflection Limit That Is Written Before Any Earthquake
The damage scales above apply to a seismic event, which for a given berth may not happen for a century. Day to day the wall is governed by a much tighter limit that has nothing to do with earthquakes, and the two limits have to be kept apart in the specification or they will be applied to each other by accident.
British practice for the design of maritime works, published as BS 6349, sets the maximum horizontal displacement accepted for a vertical wall under quasi-permanent load combinations at one three-hundredth of the total height of the wall, written as H over 300. On a wall 14.75 metres high that is 49 millimetres. The same number expressed as a slope is 0.333 percent, which is 0.191 degrees. Compared with the seismic damage scale that the same kind of structure is graded on, this is a factor of 4.5 tighter at the first rung, and the comparison is not a contradiction because the two are different kinds of number.
The H over 300 limit is a serviceability limit. It exists to protect the things the wall carries rather than the wall itself: surface water drainage that must keep flowing the right way, buried services, crane rail alignment, the apron pavement, and the berthing operations that need a predictable face. The seismic damage scale is a residual limit. It exists to record what was left behind after a rare event, and it is deliberately loose because a port authority that must choose between a repairable berth and no berth at all will usually take the repairable berth. Both numbers are correct, and a specification that mixes them will either over-constrain the earthquake case or under-constrain the drainage case.
A third group of criteria, for the apron, sits between those two and protects a different asset again. The published limits for the apron are a differential settlement of 3 to 10 centimetres across the apron, a differential settlement of 30 to 70 centimetres between the apron and the land behind it, and a residual tilting of the apron towards the sea of 2 to 3 degrees. These are not limits on the wall. They are limits on the surface that the crane and the trucks use, and they are satisfied or not independently of what the wall does.
| Criterion group | What it protects | Published limit | Failure mode it catches |
| Vertical wall, quasi-permanent | drainage, services, crane rail alignment | H over 300, which is 0.191 deg at 14.75 m | serviceability deflection |
| Residual displacement, seismic | the wall as a retaining mass | below 1.5 percent of height at Degree I | sliding |
| Residual tilting, seismic | the wall as a retaining mass | below 3 degrees at Degree I | rotation |
| Apron, seismic | pavement and the land behind it | 3 to 10 cm differential, 2 to 3 deg tilt | differential settlement |
An actual design shows how far a competent plan can sit below all of them. On a permanent tied-back quay wall in the Aqaba New Port, the calculated normalised residual horizontal displacement was about 1 percent for the more frequent earthquake level and about 1.8 percent for the rarer one, both inside the Degree I ceiling of 1.5 percent except in the rarer case. The residual tilting of the apron was calculated at 0.4 degrees and 0.55 degrees against an allowance of 3 degrees, so the design used roughly an eighth of the tilting budget. The differential settlement on the apron was calculated at 0.05 and 0.10 metres against an allowance of 0.03 to 0.10 metres, which is the criterion that came closest to being binding. The one operational consequence the designers accepted for the more frequent event was that the crane rail on the capping beam might have to be relocated, and they made that cheap in advance by widening the recess that holds it.
6. What Each Field Instrument Can Actually Fill In
A quay wall plan has to answer both columns of the damage table and the apron criteria as well, and no single instrument fills all of them. The division of labour is set by what each device physically reports, not by which one is more accurate.
| Field instrument | What it reports | Damage column it can fill | What it cannot establish |
| Tilt sensor on the capping beam | the angle of its mounting base | the tilting column | whether the wall also translated |
| Total station on prisms, or GNSS | the displacement of a point | the displacement column | rotation, unless two heights are read |
| In place inclinometer in a casing | horizontal displacement against depth | the displacement column, with depth | anything above the top of the casing |
The tilt sensor is the only one of the three that needs no line of sight, no borehole and no survey crew, which is why it is the instrument that can be left on a wall for years. Its reading is a difference between two points on the structure itself, so it reports the rotation of the wall rather than the movement of the wall, and that distinction is exactly the one that made the two scales disagree. A single tilt sensor will fill the tilting column honestly for as long as it stays fixed. It will fill the displacement column only by an assumption, and the assumption is that the wall is rotating about its toe, which is the mode a sliding wall does not have.
Two tilt sensors at different heights on the same wall section do something a single unit cannot, because the pair separates the two modes. A rotation about the toe produces a linear increase in horizontal displacement with height, and the angles read at two heights are consistent. A rigid translation produces identical displacements at every height and no angle change at all. Reading a pair rather than a single unit is how a tilt installation is made to report translation as well, and it costs one extra unit per section.
The second half of the selection question is the scale of the numbers involved, and here the answer is genuinely reassuring rather than a sales argument. The published accuracy of a high accuracy wireless tiltmeter is plus or minus 0.005 degrees, and its resolution is 0.001 degrees. Against the lightest seismic tilting ceiling of 3 degrees, the accuracy is 0.17 percent of the criterion and the resolution is 0.03 percent of it. Against the tighter serviceability limit of 0.191 degrees, the accuracy is 2.6 percent and the resolution is 0.5 percent. The zero point temperature drift, which is the term that decides the tightest angular criteria in structural monitoring, moves a reading by 0.040 degrees over a 40 degree Celsius swing on one published model and 0.080 degrees on another, which is 21 and 42 percent of that same serviceability limit.
So on a quay wall the instrument is not the limiting term anywhere, and it is worth saying so plainly. The same zero drift is worth 140 percent of the tightest published criterion for a storage tank settlement case, which is one part in a thousand or 0.057 degrees, and yet it is worth well under half of the tightest quay wall criterion, because the criteria for a wall of this kind are written in degrees of the order of one. What limits a quay wall plan is not the sensor. It is the choice of which published criterion applies, the definition of the reference the wall is measured against, and whether anybody is watching the water that is pushing it.
7. The Water Pressure That Two Authorities Assume Opposite Ways
The previous sections treated the water behind the wall as a static problem with a lag. During an earthquake it stops being static, and this is where the published design assumptions and the measured behaviour part company in a way that a monitoring engineer should know about, because the two answers point in opposite directions.
The conventional assumption, associated with the port design guidelines of the same association, is that the pore water pressure behind the wall is at its maximum when the wall is in the active phase, meaning when the fill is pushing the wall outwards, and at its minimum when the wall is in the passive phase. That assumption is convenient because it makes the worst case coincide with the phase that already carries the largest earth thrust. A centrifuge test programme on a quay wall with partially stabilised backfill reported the opposite in every test case. At the active states, where the inertia force acts on the fill in the direction of the wall, the recorded pore water pressure reached minimum values rather than maximum ones.
The mechanism proposed for that reduction is not mysterious. The wall moves seaward, the fill behind it extends laterally, and that lateral extension relieves the total stress in the fill, which under partially undrained conditions allows the pore pressure to fall. The published study also records that the pore pressure is likely to go negative through this dilatancy. The measured consequence is the sentence that matters: the reduction in pore water pressure caused the total active earth pressure to be even smaller than the hydrostatic pressure in many cases.
An honest reading of that result has to carry its own limits, and the same paper states them. The model backfill was crushed rock chosen for a grain size to wall height ratio compatible with the prototype, and its permeability may not have been high enough to represent the prototype sized stone. Because centrifuge similitude requires the model soil to match the prototype permeability to reproduce dissipation, the undrained conditions may have been imposed too strictly and the pore pressure variation may have been exaggerated. A companion test run with fresh water, equivalent to making the backfill fifty times more permeable, did show smaller within-cycle pore pressure changes. The qualitative finding that pore pressure reaches a minimum in the active phase survived that check.
Two further observations from the same work are worth carrying into a monitoring plan. The first is that the oscillation of the quay wall lagged behind the oscillation of the stabilised soil, which means the uncemented fill between them was being laterally extended and its stress relieved during the active phase. The second is that stabilising the fill close to the wall reduced the active earth pressure, and that this reduction was much more pronounced at depth. Both point the same way: the pressure that a wall experiences is a function of how the fill and the wall move relative to each other, and that is not a quantity that a design assumption can settle in advance.
8. Mounting on a Structure That Is Wet, Salted and Working
A quay wall is one of the harder structures to instrument, and the difficulties are not mainly thermal. The installation sits in the splash zone, where wetting and drying cycles carry chlorides into any crack and where corrosion proceeds faster than anywhere else on the structure. The mounting height is chosen against the tide rather than against the convenience of the installer, because a sensor below high water is a sensor that will be serviced from a boat or not at all. The wall is also struck and loaded continuously, by fenders under berthing, by bollards under mooring and by cranes on the rail above, so the base detail has to survive a mechanical environment that a storage tank or a bridge pier does not present.
The layout follows the failure modes rather than the wall length. Because differential movement concentrates at the joints between wall sections, a monitoring plan that places sensors at even intervals along a long wall and none at the joints will spend most of its budget away from the places that move. A published design for a new deep water quay wall makes this explicit: a structural joint along the diaphragm wall and the capping beam was found to be necessary so that differential horizontal displacements could develop during a seismic event without transferring between sections, and the joint is exactly where the monitoring belongs. On the horizontal plane four control points are normally placed at the corners of the crest of a caisson, although three would strictly be enough to define the movement of that plane.
Where the wall is long enough that cabling is impractical, the field has settled on wireless nodes at the sensor locations with a small number of gateways collecting them, and this is the same architecture the company supplies for other long linear assets. An excavation monitoring scheme on a 659 metre quay wall in northern Scotland used twenty four in place inclinometers installed in the piles at regular intervals along the front and anchor walls, at depths of 15 to 35 metres depending on the area, with the capping beam monitored separately by automated total stations on prisms so that the two measurements could be compared against each other. The stated reason for choosing in place instruments over manual inclinometer surveys was access: a manual survey needs a crew and a clear borehole every time, while an in place string records without a visit. That is the same argument that applies to tilt sensors on a capping beam, and it is the argument that decides most real plans.
One boundary should be stated before the instrument is chosen. A tilt sensor on a capping beam sees the wall from above. It cannot see a deep seated slip surface passing under the wall, it cannot see the condition of the toe, and it cannot see below the dredge line. Where those are the governing risks, a borehole string is the instrument that answers them, and it answers a different question at a different cost. The two are complements in the same plan rather than alternatives in a comparison, and a plan that contains only the tilt sensors has chosen which question it will not answer.
9. The Port Where Half the Quay Walls Are at the End of Their Life
The reason this subject has moved from academic to commercial in a single decade is age. The Port of Rotterdam has stated that approximately half of its quay walls will reach the end of their design lifetime in the coming years. The problem is not that the walls are visibly failing. It is that the design models cannot say how much life is actually left, because the deformations that decide it are small, cumulative and hidden underwater.
The response has been to instrument new walls far more heavily than the old design codes ever required, and the reference case is the deep sea quay wall in the Amaliahaven basin in Rotterdam. It is 1.8 kilometres long with a soil retaining height of 29 metres, carries an 18.5 metre wide relieving platform on top of a combined wall and driven cast in situ pile substructure, and is anchored with screw injection anchors. The instrumentation covers five open ended tubular piles and ten driven piles with fibre optic strain sensors and inclinometer casings to measure bending moments, axial forces and deformation along the pile length; twenty two ground anchors measured along their full length with fibre optic strain sensors, with load cells on eight of them as an independent check; and additional sensors for water level, tilt, displacement and mooring loads. The published purpose is to compare what the wall actually does against what the design model predicted, across tidal effects, seasonal temperature change, mooring loads and crane loads.
That work is new enough that its first finding is about construction rather than service. The published observation is that during dredging the strains begin to accumulate and then stabilise after the dredging stops, which confirms in a real wall what the load sequence above predicts: the temporary stage is the one that moves the structure, and the completed structure is more stable than the site that built it. An older and more expensive lesson comes from a Spanish port, where four caissons of a quay wall failed by deep sliding through the base of the rockfill berm, partly because silty clay had not been cleaned out of the dredging trench and partly because loads above the design elevation were applied during the preloading stage. The recorded settlement of that wall progressed from zero through 25, 50, 60, 70 and finally 80 centimetres over roughly nine months, which is what a slow failure looks like in a record that is being read.
The same body of monitoring work carries a warning about direction. On the instrumented gravity quay walls reviewed for that study, the start of backfilling consistently induced a rotation in the caisson that displaced the crest landwards, and only once the backfill was complete did the crest normally displace seawards. A monitoring plan that begins during construction and treats every seaward trend as damage will misread the first stage, and one that begins after construction will have no record of the landward excursion that preceded it. The same review records that only two of thirteen monitored quays had failed, and that for both, back analysis showed horizontal movements much larger than the design predictions. That is the sentence that justifies the sensors.
ZC Sensor’s own port record comes from the Yangshan deep water port, the second largest port in the world after Shanghai, built on islands at the mouth of Hangzhou Bay. The works at Shengsi County in Zhoushan used artificial fill reinforcement combined with natural deposition of sea sediment and sand to create about 100,000 square metres of land, and because the seabed sediments vary in stiffness, soil collapse was a recognised risk during the fill. ZCT-CX100 in place inclinometers were used to monitor the lateral displacement of the soil caused by the dynamic collapse process, in three sections 25 metres below the seabed, each 25 metres deep and carrying eight probes, placed at the shore and at 350 and 700 metres from land. Before construction the sections were read once a day, and during the dynamic compaction the readings were taken at higher frequency so that the spacing of the compaction points and the compaction energy could be corrected in time. That is a fill and ground improvement project rather than a quay wall structure, and it is described here as such. The company supplies the instruments, publishes no design role, and on a project of this kind the instrument is a small part of the engineering.
Frequently Asked Questions
- Can a tilt sensor on its own confirm that a quay wall is safe?No, and the reason is structural rather than a matter of accuracy. A tilt sensor reports the angle of its mounting base, so a wall that has slid seawards as a rigid body can move a large distance while the angle stays flat. The reference damage criteria for port structures include a normalised displacement column and a residual tilting column, and a tilt channel can only fill the second. A wall plan needs a displacement measurement as well, either from prisms or GNSS, from a borehole string, or from a pair of tilt sensors at different heights that can separate translation from rotation.
- Which of the two damage criteria should be applied, displacement or tilting?Where both are available, apply the more severe one, which is the rule that published analysis of a real port earthquake settled on. Where only one is available, say which one and why. The two scales in the same guideline are offset from each other by about 3.8 percent of wall height at every rung, so they are not interchangeable, and the disagreement is largest at Degree I, the lightest grade and the one a monitoring system exists to catch.
- Why does the water behind the wall matter if the wall was designed for the tide?Because the design assumes the water behind the wall can get out. The tide on the harbour side falls faster than the level inside the fill, and it is that falling direction, not the high tide, that leaves an unbalanced pressure pushing the wall seawards. Design guidance handles this by assuming a lag of half the mean tidal range for an impermeable fill, and by assuming no lag where drainage works. A weep hole that has been closed by marine growth changes which of those two assumptions applies, and that is a maintenance question rather than a design question.
- How often should a quay wall be read, and what triggers a closer look?The interval follows the operations rather than the calendar. Instrumented port projects record higher frequency readings during dredging, excavation, backfilling and dynamic compaction, and reduce to a daily or longer interval once the site is quiet, because the construction stage is what moves the structure. The events that justify densifying the record afterwards are a dredging campaign in front of the wall, a change in what is stored or handled on the apron, a blocked or repaired drainage system, and any single reading that has moved beyond the noise the same sensor produced in its first month.
TAG: