ANNEX G      

 

G-1      Layout of Tank Farm

 

 

G-2      Outer Bund Wall (Type 1 and 2)

 

There are two types of outer bund walls and their details are shown below.

 

 

 

 

 

G-3      Simulation of Potential Fuel Splash from the Area within the Inner Bund to the Emergency Access Road within the Tank Farm and from the Emergency Access Road over the Outer Bund to the Adjacent Grassed Area

 

Objectives: The objectives of the trial were to simulate bund overtopping events and to demonstrate how far fuel spilled within the inner bund could potentially reach if it (1) splashed from the Emergency Access Road over the outer bund onto the adjacent grassed area and (2) then splashed from the area within the Inner Bund to the emergency access road (and thereafter into the stormwater drain).  The trial used water discharged through a fire hose to simulate spilled fuel.

 

Scenario 1 – Water splash splashing from inner bund the Emergency Access Road over the outer bund onto the adjacent grassed area between site boundary and Scenic Road (Picture G3.1).

 

Water was discharged for 10 minutes via a fire hose connected to a fire hydrant. The flow rate from the fire hose was estimated at between 1000 – 2000 litres per minutegiving a water jet velocity of between 15m/s to 30m/s.  This is consistent with the splash velocity calculated at inner bund wall. situated about 1 meter outside the inner bund.  The discharge rate was about 1000 liter/min (i.e. about 1m³ per minute). The water jet was discharged at an angle of approximately45^o pointing over the outer bund to give a maximum distance of water trajectory towards Scenic Road (Pictures G3.2 and 3.3) and reaching about 25m into the grassed area.

 

Based on observations, the The water fell on grassed area approximately half way to Scenic Road (Picture G3.5). The water soaked into the ground and then drained back into an open drainage channel alongwhich parallels the tank farm boundary fence (Picture G.3.4).within and along the grassed strip.  An inspection of Scenic Road confirmed that no water splash reached the road. the grassed area to determine how far the water reached confirmed that the water reached approximately half of the width of the grassed area and no water reached the Scenic Road.   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

G3.1 Grassed area between site boundary and Scenic Road

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

G3.2 Simulation of water splashing from the inner bund the Emergency Access Road over the outer bund onto the adjacent grassed area between site boundary and Scenic Road

 

 

 

 

 

 

G3.3 Water splashing over the outer bund toward Scenic Road

 

 

 

 

 

 

 

 

G3.4 After water splashes over the outer bund some is absorbed by the grassed area with most draining to the open channel drainage along site boundary

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

G3.5 Inspection after liquid splashes over the outer bund confirmed that water reached approximately half of the width of the grassed area but not Scenic Road

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Scenario 2  – Water splashing from the area within the inner bund to the emergency access road within the tank farm

 

At the east end of the Administration building, water was discharged eastwards at ground level via a fire hose onto the emergency access road for 10 minutes (the discharge rate was about 1m³ per minute). The hose direction was then changed to south, i.e. along the emergency access road) for another 10 minutes.  The result demonstrated that water collected near and drained into the gulley provided along the emergency access road within the tank farm. The distance from the gulley to the cable car is about 200m

 

 

 

 

 

 

 

 

 

 

G3.6 Water splashing from the area within the inner bund to the emergency access road

within the tank farm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

G3.7 Water spreading area

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

G3.8 Water spreading area

 

 

 

G-4      Estimation of Frequency of Aircraft Accident

 

The probability of aircraft incident at the tank farm  is very unlikely.  This is supported by an analysis using the methodology adopted for assessing the aircraft crash frequency in the Hazard Assessments studies of Water Treatment Work (WTW) in Sha Tin. In the study, the methodology was that of Phillips (1987) and the model took into account specific factors including the ‘target’ area (in this case, area of Tank Farm), the radial distance from the runway end (R, in km) and its angle from the runway axis (q). The empirical function is shown below:

 

F = {Crash Rate} x {f(R, q)} x {Proportion of flights in specified direction} x

{Proportion of flights on specified runway} x {Number of flights} x

{Target area of the Tank Farm}

 

Where f(R, q) = 0.23 exp (-R/5) exp (-|q|/5)

 

According to the information obtained from the Civil Aviation Department (CAD), the 2 parallel runways (commonly called the North Runway and the South Runway) are both used for arrival and departure in a segregated mode. Based on the New Airport Master Plan, Environmental Impact Assessment Update, the proportion of flights in a specified direction on specified runway is 0.55 and 0.45 for aircraft landing from the west and landing from the east respectively.

 

 

An Approach Crash Rate (1.2 E-08) has taken into account the most recent data held by the US National Transportation Safety Board for aircraft accidents involving US carriers and findings of Table H1 of the Sha Tin WTW report is adopted for the estimation of crash frequency in different directions. Calculation details are shown in the following table using the following data:

 

Target area of Tank Farm inner bund = 26038 m² = 2.6E-02 km²

Proportion of flights landing from the West = 0.55

Proportion of flights landing from the East = 0.45

Assumed Proportion of flights on specified runway = 0.5

Number of flights = 98423 (based on 2001 statistics)

 

 

	Direction	R (km)	q (Degree)	f(R, q)	Frequency (per year)
Southern Runway	From East	3.24	40	4.04E-05	2.79x10-10
	From West	1.45	20	3.15E-03	2.66x10-8
Northern Runway	From East	4.12	58	9.24E-07	6.39x10-12
	From West	2.92	40	4.30E-05	3.63x10-10

 

The calculation shows that the frequency of a crash is 2.7x10^-8 per year (ie on the southern runway approaching from the west).

 

It must be stressed that the site of the tank farm  is surrounded by  buildings and a hill 68m high on the south east quadrant which provides a shielding  that reduces further  the chance of a direct hit on the tanks farm by aircraft. In addition, the aircraft would needs to fly through a narrow gap between  the hill and surrounding building in order to hit the tanks. Furthermore, aircraft movements  at HKIA are carefully controlled by air traffic controllers who will not permit any low flying aircrafts from flying over or close to the tank farm, thus further reducing the possibility of aircraft striking the tanks. 

 

Considering the location of the tank farm and the taking off directions, it is very unlikely that the tank farm would be hit by an aircraft.t.Therefore, the potential hazard of the tank farm being struck by aircraft during taking off is not considered any further.

 

Taking the above factors into account and based on the best  judgement, the frequency of an aircraft accident resulting in a direct hit on  the farm is  below 1x10^-9 per year. which is the frequency cut-off value in the Risk Criteria shown in Annex G6.     Therefore, the potential hazard of the tank farm being struck by aircraft is not considered any further.

 

G5       Frequency of Tilt Angle

 

 

 

(Remarks: only data with pool diameters greater 10m were extracted for analysis)

 

 

G5.1 Relevant data extracted from HSE Contract Research Report (Ref 296/1996)

 

 

 

 

 

As shown in the above table, the maximum recorded tilt angle is 60^o, thus any fire at distance greater than L(m) should pass over the cable cabin and L is determined by

 

           tan 30^o = (43m/Lm)

 

where   43m is the height of the cable cabin

            30⁰ = 90^o – maximum recorded tilt angle

L is the horizontal distance between the fire and cable cabin

 

Thus L = 75m

 

L = 75m

 

 

G-6      Determination of Individual and Societal Risk Guideline (EIAO-TM Annex 4)