4 Hazard to Life

Background

4.1 The site of the Organic Waste Treatment Facilities (OWTF) is located at Siu Ho Wan Area 3 which is within the consultation zone of Siu Ho Wan Water Treatment Works (SHWWTW) on North Lantau as shown in Figure 4.1. In accordance with Clauses 3.4.2 of the Environmental Impact Assessment (EIA) Study Brief (ESB 172/2007) [1], a Hazard Assessment (HA) shall be conducted to evaluate the risk to construction workers and operational staff of the Project due to the transport, storage and use of chlorine associated with the operations at SHWWTW.

Scope and Objectives

4.2 According to the technical requirements specified in Section 3.4.2 of the EIA Study Brief [1], the HA is carried out following the criteria for evaluating hazard to life as stated in Annexes 4 and 22 of the Environmental Impact Assessment Ordinance Technical Memorandum (EIAO TM) [2] (Hong Kong Risk Guidelines).

4.3 The objectives of the HA are as follows:

(i) Identify hazardous scenarios associated with the transport, storage and use of chlorine at SHWWTW and then determine a set of relevant scenarios to be included in a Quantitative Risk Assessment (QRA), including the scenario associated with the impact from biogas storage of the Project on the chlorine store of SHWWTW;

(ii) Execute a QRA of the set of hazardous scenarios determined in (i), expressing population risks in both individual and societal terms;

(iii) Compare individual and societal risks with the criteria for evaluating hazard to life stipulated in Annex 4 of the EIAO TM; and

(iv) Identify and assess practicable and cost-effective risk mitigation measures.

Study Approach

Overview

4.4 Due to the proximity and significant increase in population, Lantau Logistics Park (LLP) development would affect the overall risk outcomes. Although the LLP project study has been commenced since year 2004, the EIA process of the project has not yet finalised. There are uncertainties on the population intake during the operation phase of the OWTF. In order to evaluate impact of the SHWWTW on construction and operation stages of the Project as well as to determine whether the two Project stages would increase the SHWWTW risk outcome, the risk assessment covers four scenarios:

1. Year 2008 scenario

The risk of the SHWWTW operation prior to the construction and operation stages of the Project (baseline condition).

2. Year 2011 scenario

The risk of the SHWWTW operation during the construction stage of the Project.

3. Year 2013 scenario without LLP

The risk of the SHWWTW operation after the commissioning of the Project without population of the LLP.

4. Year 2013 scenario with LLP

The risk of the SHWWTW operation after the commissioning of the Project with population of the LLP.

4.5 The hazard assessment consists of the following five tasks:

1. Data / Information Collection: collects relevant data / information which is necessary for the hazard assessment

2. Hazard Identification: identifies hazardous scenarios associated with the operations of the SHWWTW by reviewing historical accident database, such as Major Hazard Incident Data Service (MHIDAS) [9], and relevant similar studies and then determine a set of relevant scenarios to be included in the HA.

3. Frequency Estimation and Consequence Analysis: estimates the frequencies of the identified hazardous scenarios by reviewing historical accident data, previous studies or using Fault Tree Analysis (FTA) and analyses of the consequences of the identified hazardous scenarios.

4. Risk Evaluation: evaluates the risks associated with the identified hazardous scenarios. The evaluated risks are compared with the Criteria for Evaluating Hazard to Life stipulated in Annex 4 of the EIAO TM to determine their acceptability.

5. Identification of Mitigation Measures: practicable and cost-effective risk mitigation measures are identified and assessed as necessary. Risks of mitigated case are then reassessed to determine the level of risk reduction.

4.6 The hazard assessment approach follows the requirements as per the EIAO TM, the EIA Study Brief and the Court of Final Appeal (CFA) ruling [14]. Hazardous scenarios have been identified by reviewing hazardous scenarios developed for similar installations. Hazardous scenarios and frequency adopted in the hazard assessment are confirmed independently using review of historical incidents.

4.7 The Hazard Assessment study for SHWWTW namely the HA report “the Hazard Assessment Study – Final Report, June 1992” under Agreement No. CE 12/91, “North Lantau Water Supply Project” [3] (hereafter refer to as “1992 SHWWTW HA report”) has been reviewed and taken as a reference for the Project.

Hong Kong Risk Guidelines

4.8 The estimated risk levels of the hazardous sources are compared with the risk guidelines stipulated in the EIAO TM Annex 4 to determine the acceptability.

4.9 As set out in Annex 4 of the EIAO TM, the risk guidelines comprise two components as follows:

· Individual Risk Guideline: the maximum level of off-site individual risk should not exceed 1 in 100,000 per year, i.e. 1 x 10^-5 / year

· Societal Risk Guidelines are presented graphically as in Figure 4.2. The Societal Risk Guideline is expressed in terms of lines plotting the frequency (F) of N or more deaths in the population from accidents at the facility of concern. There are three areas shown:

- Acceptable where risks are so low that no action is necessary;

- Unacceptable where risks are so high that they should usually be reduced regardless of the cost or else the hazardous activity should not proceed;

- ALARP (As Low As Reasonably Practicable) where the risks associated with the hazardous activity should be reduced to a level “as low as reasonably practicable”, in which the priority of measures is established on the basis of practicability and cost to implement versus risk reduction achieved.

Study Basis

Data / Information Collection

4.10 The following data / information have been collected for the execution of the risk assessment:

· Population data around the SHWWTW

· Road population (e.g. North Lantau Highway and railways)

· Construction worker population for the Project

· Operational staff population for the Project

· Meteorological data near SHWWTW (including atmospheric stability class, wind speed and wind direction)

Population Restriction around the SHWWTW

4.11 In accordance with the population restrictions endorsed by the Coordinating Committee on Land-use Planning and Control relating to Potentially Hazardous Installations (CCPHI) in the 1992 SHWWTW HA study, population of no more than 300 is recommended for the Area 3 (gas reception station) in which the OWTF site is located. Apart from the OWTF site, 2 bus depots and Siu Ho Wan Vehicle Pound Vehicle Examination Centre and Weigh Station can be found in the Area 3 as shown in Figure 4.3. A site survey is carried out to verify existing population within Area 3. It is estimated that number of workers is about 40 for each bus depot. Population for Siu Ho Wan Vehicle Pound Vehicle Examination Centre and Weigh Station is approximately 10. It is expected the number of onsite staff and visitors in the OWTF is around 45 and 40 respectively. Since visitors would stay at the OWTF for 3 hours and be arranged only 1 or 2 sessions a week, total number of population in Area 3 is estimated 135. The number is below the population restriction. Therefore, the new population associated with OWTF is in compliance with the population restrictions endorsed by the CCPHI in the 1992 SHWWTW HA study.

Population Data around the SHWWTW

4.12 Data from the Population Census of Statistics and Census Department is adopted as far as possible to determine the residential population in vicinity of the SHWWTW. A site survey is conducted by the Consultant on 28 November 2008 to estimate the population at various locations based on observation on-site and/or communication with on-site personnel for the population at locations not covered by the Population Census.^{^[1]} Existing and future population groups are depicted in Figure A of Appendix 4.1.

4.13 For assessment purpose, the operational year of the LLP development is assumed to be 2013 at the earliest. Population data from the LLP study is extracted and adopted in this QRA. Location of population groups for the LLP development is indicated in Figure B of Appendix 4.1.

4.14 Population data adopted in this QRA for all scenarios are tabulated in Appendix 4.1.

Road Population

4.15 Traffic data is obtained from the Annual Traffic Census of Transport Department for calculation of road population. The road population is predicted in accordance with the following equation:

Traffic Population = No. of persons/vehicle * No. of vehicles/hr * Road Length / Speed

4.16 The worst case scenario based on the “bumper-to-bumper” (jammed condition) during peak hours (2 hours a day) is adopted for road population estimation. Annual Average Daily Traffic (AADT) data is applied to calculate daytime and night-time traffic distributions.

4.17 North Lantau Highway is the major source of traffic population. Referring to “The Annual Traffic Census (ATC) 2007” by Transport Department [8], the traffic flow parameter, vehicle classification and average occupancy of vehicles from Lantau link (from Tsing Ma Bridge eastern end at Tsing Yi to Ngong Shuen Au) are used for deriving traffic population. Time periods outlined in Table 4.1 represent the mapping of the Annual Traffic Census data into the time modes adopted in the SHW HA report.

Table 4.1 Time Modes for Temporal Variation of Road Population

Description in Traffic Census	Number of hours in a Day	Time Mode Adopted in 1992 SHWWTW HA Report
Peak	2	Jammed condition
AADT R16/24	16	At capacity condition
Others	6	Light condition

4.18 Traffic population is assumed to be 95% indoor. Average occupancy of 4.2 persons per vehicle is calculated by taking into account traffic mix and according to traffic census data for occupancy of various vehicle classes.

4.19 The jammed and capacity periods are assumed to occur during daytime and in 1 direction only corresponding to daytime AM and PM peaks. This study adopts the worst scenario in calculation of the transient population.

4.20 At the jammed condition, separation distance between 2 vehicles is assumed 2 meters. Assumed length and average occupancy of each vehicle class are tabulated in Table 4.2.

Table 4.2 Assumed Length and Occupancy of Different Classes of Vehicle

Class of Vehicle	Motor Cycle	Private Car	Taxi	Private Light Bus	Goods vehicles		Non Franchised Bus	Franchised Bus
Class of Vehicle	Motor Cycle	Private Car	Taxi	Private Light Bus	Light	Medium & Heavy	Non Franchised Bus	Single- decked	Double- Decked
Vehicle Length (meter)	2.5	5	5	6	7	11	12	12	12
Occupancy[2]	1	1.5	2	7	1.7	1.3	23.7	15.1	43.6

4.21 Traffic population under 3 different road conditions (jammed, at capacity and light conditions) are tabulated in Table 4.3. Population for all scenarios are projected from ATC 2007 data based annual traffic growth rate of 3.94% according to the traffic impact assessment under the same Project. Road population is not provided at different stages of the 1992 SHWWTW HA Study. Thus, no road population from the 1992 SHWWTW HA Study is included in this OWTF Study. Transient population for the minor road Cheung Tung Road is taken as 2.5% population of the North Lantau Highway.

Table 4.3 Projected Traffic Population

Scenario	Traffic Population Along North Lantau Highway (persons per km)
Scenario	Daytime (Working Hours)	Daytime (Non-working Hours)	Night-time
Year 2008	696	393	168
Year 2011	715	441	189
Year 2013	729	476	204

Railway Population

4.22 On-train population of the Airport Express and the Tung Chung Line are included in the assessment. Daily average number of passengers from previous years are used and projected to assessment years. The railway population is calculated according to the following formula:

Railway Population per km = No. of persons per train * No. of trains per hour / Speed

= No. of persons per hour / Speed

4.23 Historical data up to year 2007 on annual number of passengers for both Tung Chung line and the Airport Express [13] are used to project railway population for assessment years 2008, 2011 and 2013 by linear regression.

4.24 Adopting a conservative approach, 16 operating hours per day for both Airport Express and Tung Chung Line is assumed. The average number of passengers in an hour is worked out according to the number of daily operating hours.

4.25 For train speed of 100 km/hr, railway population density in terms of persons per km is obtained as shown in Table 4.4. Geographical extent of the railway population is depicted on Figure A of Appendix 4.1

Table 4.4 Projected Number of Passengers for Railway Lines

Scenario Year	Daily Average Number of Passengers for both Tung Chung LineAnd Airport Express ^a	Total On-train Population (persons per km)
2008	126,600	79
2011	171,700	107
2013	201,800	126

Note:

a Data obtained from enquiry to MTRC and MTRC Annual Report for Year 2007

Population at the Project Site during Construction

4.26 The Project involves construction activities of the proposed waste reception facilities, pre-treatment facilities, digester and post-treatment facilities. According to the draft Working Paper on Contract Options, it is recommended 2 to 3 full time government staff on-site during construction period. The allowable working area is small and therefore the number of worker at the work site would be limited. Onsite population of 60 during construction phase is assumed scattered around the construction site at daytime. Three security staff would be staying at the construction site during night time. During construction phase, population would be exposed outdoor all the time.

Population at the Project Site during Operation

4.27 Biogas production process will be carried out 24-hour a day. Loading / unloading of organic waste and compost will be carried out 14-hour a day in 2 day shifts and 7 days a week. Number of staff for each shift is estimated 45. Truck drivers for delivery of catering waste and loading of compost have been taken into account in the daytime population. Since loading operation will be carried out at daytime only, onsite staff of 5 is assumed for the rest of time (8 hours).

4.28 The Project site will include educational facilities which will locate inside the site office building. Tentative arrangement will be maximum 10 sessions of guided tour in a month during weekdays. Each session will cater for maximum 40 visitors and last for 3 hours.

Table 4‑5 Population for OWTF during Operation Phase

Population Type	Days of Week	Number of Hours in a Week	Number of Persons
Staff	Monday to Sunday	7 days x 14 hours	45
Staff	Monday to Sunday	7 days x 10 hours	5
Visitor	Monday to Friday	7 hours *	40

Note:

* 10 session per month x 3 hours per session x 7 days per week / 30 days per month

Time Modes

4.29 To account for temporal variation of population and operation of SHWWTW, time is divided into 4 time modes as shown in Table 4.6 below:

Table 4.6 Time Modes for Temporal Variation of Population

Time mode	Description	Period	Number of Hours in a Week
1	Weekday day (working hours)	Monday to Friday	5 days x 8 hours
2	Weekday day (non working hours)	Monday to Friday	5 days x 4 hours
3	Weekend day (non working hours)	Saturday to Sunday	2 days x 12 hours
4	Night	Monday to Sunday	7 days x 12 hours

Indoor/Outdoor Population Distribution

4.30 The outdoor proportion of population groups have been estimated by assigning an outdoor ratio to each group. Typical indoor/outdoor ratios for various population categories are listed in Table 4.7.

Table 4.7 Indoor/Outdoor Ratios for Different Population Categories

Population Category	Indoor (Outdoor) Ratio
Residential	0.90 (0.10)
Industrial/Commercial	0.90 (0.10)
Road	0.95 (0.05)
Railway/Bus station	0.00 (1.00)

4.31 Population data adopted in this QRA are tabulated in Appendix 4.1 for existing scenario (Year 2008), construction phase (Year 2011) and operation phase (Year 2013).

Meteorological Data

4.32 Meteorological data is required for consequence modelling and risk calculation. Consequence modelling (i.e. dispersion modelling) requires wind speed and stability class to determine the degree of turbulent mixing potential whereas risk calculation requires frequencies for each combination of wind speed and stability class. Referencing the 1992 SHWWTW HA report, the meteorological data is adopted from the Chek Lap Kok Weather Station. The meteorological data in the 1992 SHWWTW HA report is considered suitable and is used in this QRA. Details of dominant sets of wind speed-stability class combination for both daytime and night-time are given in Table 4.8.

Table 4.8 Wind Direction Frequencies at Chek Lap Kok Weather Station

Day Time	Weather Class (Wind Speed-Stability Class)
Wind Direction	0.5B	0.5D	1F	3B	3E	4D	7D	15D
0 (N)	0.003587	0.001002	0	0.001753	0	0.014445	0.004602	0.001404
22.5 (NNE)	0.002492	0.001256	0	0.001336	0	0.024379	0.005906	0.002602
45 (NE)	0.005434	0.001902	0	0.001333	0	0.013835	0.00184	0.000821
67.5 (ENE)	0.006202	0.002642	0	0.001853	0	0.016694	0.006087	0.004273
90 (E)	0.003381	0.002133	0	0.003033	0	0.039197	0.028908	0.034758
112.5 (ESE)	0.001673	0.001054	0	0.001787	0	0.030858	0.022303	0.020933
135 (SE)	0.000834	0.000692	0	0.000691	0	0.007334	0.003799	0.003725
157.5 (SSE)	0.000923	0.000548	0	0.00064	0	0.003274	0.001501	0.00152
180 (S)	0.000661	0.000813	0	0.001572	0	0.012103	0.002807	0.001093
202.5 (SSW)	0.000714	0.000874	0	0.001321	0	0.01398	0.002054	0.000462
225 (SW)	0.000684	0.000576	0	0.000918	0	0.006211	0.001534	0.000434
247.5 (WSW)	0.00114	0.001147	0	0.001237	0	0.005744	0.001084	0.000529
270 (W)	0.003066	0.001552	0	0.007792	0	0.008339	0.000411	0.000227
292.5 (WNW)	0.006707	0.001678	0	0.010864	0	0.009785	0.000781	0.000436
315 (NW)	0.00676	0.000909	0	0.005209	0	0.01057	0.002837	0.001598
337.5 (NNW)	0.00388	0.00088	0	0.001232	0	0.007488	0.002706	0.00142
Night Time	Weather Class (Wind Speed-Stability Class)
Wind Direction	0.5B	0.5D	1F	3B	3E	4D	7D	15D
0 (N)	0	0.000672	0.004273	0	0.003801	0.009687	0.003086	0.000941
22.5 (NNE)	0	0.000843	0.00755	0	0.01033	0.016349	0.003961	0.001745
45 (NE)	0	0.001275	0.010121	0	0.008169	0.009278	0.001234	0.00055
67.5 (ENE)	0	0.001772	0.014255	0	0.009797	0.011196	0.004082	0.002866
90 (E)	0	0.00143	0.018485	0	0.016089	0.026287	0.019387	0.02331
112.5 (ESE)	0	0.000707	0.014266	0	0.015698	0.020694	0.014957	0.014039
135 (SE)	0	0.000464	0.008112	0	0.00397	0.004919	0.002547	0.002498
157.5 (SSE)	0	0.000368	0.006544	0	0.001835	0.002195	0.001007	0.00102
180 (S)	0	0.000546	0.009247	0	0.006637	0.008116	0.001883	0.000733
202.5 (SSW)	0	0.000586	0.008519	0	0.008168	0.009376	0.001377	0.00031
225 (SW)	0	0.000387	0.005161	0	0.004061	0.004166	0.001028	0.000291
247.5 (WSW)	0	0.000769	0.00608	0	0.002964	0.003852	0.000727	0.000355
270 (W)	0	0.001041	0.005175	0	0.00133	0.005592	0.000275	0.000152
292.5 (WNW)	0	0.001126	0.004468	0	0.001267	0.006562	0.000524	0.000293
315 (NW)	0	0.00061	0.002681	0	0.001717	0.007088	0.001903	0.001072
337.5 (NNW)	0	0.00059	0.002712	0	0.002034	0.005021	0.001814	0.000953

Hazard Identification

Review of SHWWTW Facilities

4.33 The SHWWTW facilities have been reviewed to ensure the latest information has been incorporated into the assessment.

4.34 Gaseous chlorine is greenish-yellow with sharp suffocating odour with boiling point -34°C. It causes acute health hazard and fatality with excessive irritation of the lungs through inhalation. Its physical properties are summarized in Table 4.9. Liquefied chlorine at approximately 90 psi is delivered to and stored at the SHWWTW.

Table 4.9 Properties of Chlorine

Property	Values
Boiling Point	-34°C
Relative Density Gas	2.5 (Air=1)
Relative Density Liquid	1.6 (Water = 1)
Vapour Pressure at 20 °C (bar abs)	6.8

4.35 The maximum storage quantity of the chlorine storage building (chlorine building) at the SHWWTW is 73 tonnes. Compressed chlorine liquid is stored in 1-tonne drums. It is designed to have delivery of up to 12 drums of chlorine liquid every 2 weeks. Each chlorine truck can carry up to 6 drums in each delivery. However, both delivery frequency and delivery quantity have not yet reached the design capacity currently.

4.36 The mechanical ventilation system and the chlorine scrubbing system installed provide controlled air circulation and treatment of air in case of chlorine release. The scrubber system consists of an absorber tower with caustic soda solution as a neutralizing agent. Air contaminated with chlorine gas will be extracted from the plant room, where a chlorine leak has occurred, and neutralized. These systems are designed to prevent chlorine gas escape from the storage area with reliability of 99% in case of release of chlorine.

4.37 Chlorine buildings are constructed by reinforced concrete panels. Integrity of the chlorine buildings would be maintained and all releases would be contained by the chlorine buildings except those caused by aircraft crash according to the 1992 SHWWTW HA.

Review of Previous Similar Studies

4.38 In the 1992 SHWWTW HA report, failure events and the respective hazardous scenarios associated with the chlorine facilities have been identified. Major hazard categories identified are listed below:

· Access Road;

· Container Handling;

· Containers in Storage;

· Connection and Disconnection of Chlorine Containers; and

· Chlorination System.

Review of Identified Hazards

4.39 The release scenarios for the operation of the SHWWTW from the 1992 SHWWTW HA report are reviewed. The SHWWTW is equipped with a “contain-and-absorb system” which works as a container to contain releases inside the chlorine storage building. Releases of chlorine are possible from:

· 1 tonne drums

· Liquid chlorine pipework and fittings (e.g. valves);and

· Gaseous chlorine pipework

4.40 The “contain-and-absorb system” could reduce the hazards to the environment for an internal release inside the chlorine building. If the contain-and-absorb system functions as intended, there will be no significant chlorine hazard to those outside of the chlorine room. In the event of failure, the resulting emission rate will depend on:

· The initial rate of release

· Condition of the absorbed and/or the normal ventilation system

· The integrity of building due to external impact (e.g. Aircraft crash, earthquakes)

4.41 Three main routes of chlorine released to atmosphere were identified in the 1992 SHWWTW HA. They are listed as below:

· Ventilation fails on

· Building collapses

· Scrubber fails, ventilation turns off

4.42 The hazardous scenarios of chlorine release for the SHWWTW operations identified in the 1992 SHWWTW HA report are summarized in Table 4.10. These scenarios are considered suitable and are adopted in this OWTF study except the building collapses. Since the chlorine buildings are constructed by reinforced concrete panels, chlorine releases in earthquakes and spontaneous failures would be contained within the chlorine buildings by referencing to Section 6.2.2 (a) in [3]. Release to the outside atmosphere would be caused by failure of the “contain-and-absorb system” only. Hazardous scenarios are modified and tabulated in Table 4.11.

Table 4.10 Hazardous Scenarios of Chlorine Outflow from the SHWWTW Buildings According to 1992 SHWWTW HA Report

Component	Failure Mode	Hole Size	Initial Release (kg/s)	Release Rate to Atmosphere (kg/s)
				Ventilation Fails On (kg/s)	Ventilation Fails Off, Absorber Fails (kg/s)	Building Collapses (kg/s)
Drum	Spontaneous	Small (6mm)	0.24	0.2	0.03	–
		Large (15mm)	1.7	0.4	0.2	–
		Rupture	1 tonne Rupture	–	–	Instantaneous 1 tonne
	Handling	Small (6mm)	0.24	0.2	0.03	–
	External Impacts	Earthquake	1 tonne Rupture	–	–	Instantaneous 1 tonne
		Aircraft crash, light aircraft crash	2 tonnes Rupture	–	–	Instantaneous 2 tonnes
		Aircraft crash	73 tonnes Rupture	–	–	Instantaneous 73 tonnes
Liquid Line	Spontaneous	6mm	0.24	0.2	0.03	–
	External Impact	6mm	0.24	–	–	0.24
	Evaporators	6mm	0.24	0.2	0.03	–
Fittings	Handling	15mm	1.7	0.4	0.2	–
Fittings	External Impact	15mm	1.7	–	–	1.7
Gasline	Spontaneous	6mm	0.24	0.2	0.03	–
Gasline	External Impact	6mm	0.24	–	–	0.24

Table 4.11 Hazardous Scenarios of Chlorine Outflow from the SHWWTW Buildings Taking Into Account Use of Reinforced Concrete Panels

Component	Failure Mode	Hole Size	Initial Release (kg/s)	Release Rate to Atmosphere (kg/s)
				Ventilation Fails On (kg/s)	Ventilation Fails Off, Absorber Fails (kg/s)	Building Collapses (kg/s)
Drum	Spontaneous	Small (6mm)	0.24	0.2	0.03	–
		Large (15mm)	1.7	0.4	0.2	–
		Rupture	1 tonne Rupture	1.7	1.7	–
	Handling	Small (6mm)	0.24	0.2	0.03	–
	External Impacts	Earthquake	1 tonne Rupture	1.7	1.7	–
		Aircraft crash, light aircraft crash	2 tonnes Rupture	–	–	Instantaneous 2 tonnes
		Aircraft crash	73 tonnes Rupture	–	–	Instantaneous 73 tonnes
Liquid Line	Spontaneous	6mm	0.24	0.2	0.03	–
	External Impact	6mm	0.24	–	–	0.24
	Evaporators	6mm	0.24	0.2	0.03	–
Fittings	Handling	15mm	1.7	0.4	0.2	–
Fittings	External Impact	15mm	1.7	–	–	1.7
Gasline	Spontaneous	6mm	0.24	0.2	0.03	–
Gasline	External Impact	6mm	0.24	–	–	0.24

4.43 Chlorine drums are delivered to the SHWWTW by trucks. In the 1992 SHWWTW HA report, release of chlorine would be caused by road accidents involving vehicle impact, fire and rollover. And spontaneous drum failure during transport was also included. Various failure modes of a chlorine drum and multiple drum failure in a road accident considered in the 1992 SHWWTW HA report are tabulated in Table 4.12.

Table 4.12 Hazardous Scenarios for Onsite Transport of Chlorine

Component	Failure Mode	Hole Size	Mass Released (te)	Release Rate (kg/s)
Drums on delivery truck	Road Accident	6mm	1	0.24
	Road Accident	15mm	1	1.7
	Road Accident	Rupture (1te)	1	Instant
	Road Accident	15mm	4.5	7.65

Review of Historical Accidents Database

4.44 A review of industry incidents, registered in the Major Hazard Incident Data Services (MHIDAS) database is conducted on chlorine gas storage and distribution facilities of similar nature. The MHIDAS database holds details of over seven thousand incidents which have occurred during the transport, processing or storage of hazardous materials which resulted in or it is considered had the potential to cause off-site impact.

4.45 A review of MHIDAS database of the relevant historical incidents of the same genus to SHWWTW has been conducted to confirm if the hazardous scenarios identified are acceptable.

4.46 A search in the MHIDAS using the keywords such as “Chlorine”, “Leakage”, “Water Treatment Works” and “Cylinders” has been conducted to identify incidents involving Water Treatment Process. The following combinations of keywords search are conducted and a total of 26 records have been reviewed.

· Chlorine & Release & Water & Treatment;

· Chlorine & Leakage & Water & Treatment;

· Chlorine & Truck & Fire;

· Chlorine & Truck & Collision; and

· Chlorine & Truck & Impact.

4.47 26 records are identified as the same genus of Water Treatment Plant of this study. These records are studied for further frequencies analysis. Details of each incident are given in Appendix 4.2.

4.48 Table 4.13 summarised types of chlorine incidents recorded in the MHIDAS database.

Table 4.13 Summary of Chlorine Incidents of Water Treatment Plant from MHIDAS

Hazardous Scenario	No. of Cases	Country
Tank/Drum Failure	9	Brazil, Puerto Rico, UK & USA
Pipework Failure	6	Hong Kong, UK & USA
Storage Failure	5	France, Portugal & USA
Others (Transport /Cylinder/Hose Failure)	6	UK & USA

Review of the Preliminary Design of Biogas Production and Storage Facilities at OWTF

4.50 The proposed organic waste treatment facilities are designed to process 200 tonnes of organic waste per day (tpd). Compost and biogas are 2 main products from the waste treatment facilities.

4.51 Biogas is produced continuously within the anaerobic digesters (AD). Specific gas production rates for a typical dry digestion system and a typical wet digestion system are 100 m³/te-input and 125 m³/te-input respectively. Referring to the Working Paper on Technology Evaluation [17], a wet system approach requires a larger volume of digester. On a reasonably conservative approach, potential hazards from the storage of biogas are assessed based on the wet system.

4.52 Most wet digester systems make use of vertical cylindrical shaped reactors. These can be built in either steel or concrete. The walls will be insulated and potentially the inclusion of heat transfer in the walls can be applied. The residence time in the reactor is determined by the applied organic loading rate. For a loading rate of 3 kg/m³/d the required effective reactor volume would amount to 13,300 m³ for a wet system. This could be realised in 5 tanks of 2,660 m³ each (approximately 3,000m³ each). Biogas will be generated continuously from the 5 Anaerobic Digesters with production rate 208 m³/h each tank based on processing capacity of 200 tpd.

4.53 Since the biogas generated from the digesters would be saturated with moisture and contain a small amount of particulate matters and hydrogen sulphide, these biogas would be treated by a biological desulphurisation process to remove the condensate and reduce the concentration of hydrogen sulphide to at least 250 mg/m³. The treated biogas would be compressed and stored in a double membrane gas buffer (Sattler system or equivalent) temporarily at 20-30mbar above atmospheric pressure. Typical setup and operation of the double membrane gas buffer is illustrated in Appendix 4.6. Storage capacity is designed to be 3,500 m³. To account for variation in design, maximum storage capacity is assumed 5,000 m³. The biogas has methane content of 55% to 70% by volume.

4.54 The design of energy recovery system aims to convert the energy contained in the biogas to electricity and heat by the application of cogeneration units (Cogen Units). Biogas for electricity generation would be divided over 3 gas engines, giving sufficient flexibility in operation in combination with maintenance. Biogas will be converted into electricity through generator under normal operating conditions. Only in emergency or under abnormal circumstances, biogas will be diverged to an emergency flare for burning the surplus biogas. The gas holder capacity is designed to cater for 3 – 4 hours production. It allows 1 of 3 gas engines out of operation for 9 – 12 hours without the need for flaring. Potential hazards from the storage of biogas to the SHWWTW are assessed in this QRA. Layout plan for the proposed waste treatment facilities is shown in Figure 2.2.

4.55 Based on maximum capacity 5,000 m³of the gas buffer and gas density of 0.72 kg/m³, the maximum storage amount of the biogas will be equal to 3.6 tonnes. The maximum storage quantity is less than 15 tonnes. The quantity does not exceed the lower threshold quantity for existing Potentially Hazardous Installations (PHIs) for flammable gas and town gas installations in Hong Kong. The proposed waste treatment facilities are not classified as a PHI.

4.56 Biogas is a colourless flammable hydrocarbon gas at atmospheric conditions. It is a mixture of various hydrocarbons and its physical and chemical characteristics have been modelled as a composition of 70-mol% methane and 30-mol% carbon dioxide. Table 4.14 below presents the properties of Biogas. Properties of biogas are very similar to those of Natural Gas (NG). Therefore, the data for NG is presented.

Table 4.14 Properties of Biogas (Natural Gas)

Property	Values
Flammability	Extremely Flammable
Auto-Ignition Temperature	580°C
Flash Points	-188°C
Melting Point	-182.5°C
Boiling Point	-161.4°C
Flammable Limits	5% (Lower) – 15% (Upper)
Vapour Density	0.59-0.72 (air = 1)

4.57 A typical gas buffer tank consists of an external membrane which forms the outer shape of the tank, as well as an internal membrane and a bottom membrane which make up the actual gas space. All three membranes are then clamped to the foundation by means of an anchor ring. A permanently running support air blower provides air to the space between inner and outer membrane and keeps the gas pressure up at a constant level. A non-return valve is installed at the inlet pipe to prevent gas from back-flow. A safety valve is used to prevent the gas buffer tank from overpressure. Gas is discharged through the outlet pipe by regulating the air pressure between the outer and inner membranes.

4.58 Incident review on failure associated with operation of biogas plants as well as storage of biogas and methane is used for identifying any missing scenarios. Since the proposed biogas plant will be operated at atmospheric pressure and moderate temperature, biogas will be stored at atmospheric pressure. Therefore, incident review for methane scenarios is focused on storage of methane gas instead of methane gas production.

4.59 A search in the MHIDAS for material “Methane” has been conducted to identify incidents involving methane gas. A total of 17 records have been reviewed. Only 3 of them are relevant to storage of methane gas. Details of the 3 incidents are given in Appendix 4.2. According to the historical incidents, explosion is the major hazard. However, explosion incidents were reported at confined environment. Moreover, incident record is not found for double membrane type gas holder.

4.60 Since the gas buffer is a spherical double membrane type and is different from column guided water-sealed gas holders, it does not have a gas holder crown. Therefore, tilting of tank top or blown seal failure will not occur in the operation of the double membrane gas buffer tank. However, release of biogas could be from various parts of the gas buffer tank or associated piping and devices. Possible hazardous outcomes include fireball, jet fire, flash fire and Vapour Cloud Explosion (VCE). Hazardous events associated with biogas storage are listed in Table 4.15.

Table 4.15 Hazards Associated with Biogas Storage

Potential Cause	Release Type	Hazardous Outcome
Gas buffer tank	Rupture	Fireball VCE Flash fire
Gas buffer tank	Leak	VCE Flash fire Jet fire
Inlet / Outlet piping (200-300mm diameter)	Rupture / Leak	VCE Flash fire Jet fire
Safety valve	Discharge due to overfilling	VCE Flash fire Jet fire
Pump / non-return valve / flange	Disintegration / Leak	VCE Flash fire Jet fire

4.61 Hazardous events are grouped into release scenarios according to the release size as shown in Table 4.16.

Table 4.16 Release Scenarios

Release Scenario		Release Size
Gas buffer tank	Rupture	3,600 kg
Gas buffer tank	Leak (junction at inlet / outlet pipe)	300mm hole size
Inlet / Outlet piping Safety valve Pump / non-return valve / flange	Full bore rupture	300mm hole size
Inlet / Outlet piping Pump / non-return valve / flange	Leak	30mm hole size (10% of pipe diameter)

Biogas Storage Impact on the Chlorine Store of SHWWTW

4.62 Impact of biogas storage on the chlorine facilities at the SHWWTW is evaluated by investigating the potential structural damage to chlorine store, chlorine pipeline and chlorine drums.

4.63 Biogas rises and dilutes rapidly due to its buoyancy when it is released to the atmosphere. In case of instantaneous release of biogas, immediate ignition near the release source would lead to fireball. Vapour cloud explosion (VCE) would occur when vapour cloud is trapped between facilities and is ignited. Potential structural damage to the SHWWTW by a fireball and a vapour cloud explosion are caused by thermal radiation and overpressure respectively. Assessment criteria for the thermal radiation [10] and overpressure [5]&[7] effects are adopted and shown in Table 4.17.

Table 4.17 Assessment Criteria for Biogas Hazards

Outcome

Effect

Assessment Criteria

Damages

Fire

Thermal radiation intensity

37.5 kW/m²

/ Jet flame / fireball

process equipment damage

VCE

Overpressure

0.2 bar

damage to heavy machinery

Fireball

4.64 For immediate ignition of an instantaneous release of a flammable material, a fireball will be formed. Instantaneous ignition of a certain mass of fuel (flammable gas) results in explosion and fire of hemispherical shape. Heat is evolved by radiation. The principal hazard of fireball arises from thermal radiation. Due to its intensity, its effects are not significantly influenced by weather, wind direction or source of ignition. Sizes, height, shape, duration, heat flux and radiation will be determined in the consequence analysis.

Jet fire

4.65 Jet fire is typically resulted from ignition of continuous discharge of a flammable material. Major concerns regarding jet fire are jet flame and the heat radiation effect generated from the jet flame. Horizontal release is applied to capture the worst scenario.

Flash fire

4.66 Flash fire is the consequence of combustion of gas cloud resulting from delayed ignition. The flammable gas cloud can be ignited at its edge and cause a flash fire of the cloud within the Lower Flammable Limit (LFL) and Upper Flammable Limit (UFL) boundaries. Major hazards from flash fire are thermal radiation and direct flame contact. Since the flash combustion of a gas cloud normally lasts for a short duration, the thermal radiation effect on buildings and facilities near a flash fire is limited.

Vapour Cloud Explosion (VCE)

4.67 VCE can occur when a flammable vapour is ignited in a confined or partially confined situation. The early explosion occurs at the source of the release. TNO Multi-Energy model is used for detailed consequence analysis which accounts for confinement effect by specifying confined strength and confined volume.

4.68 The effective hazardous distances are quantified by DNV’s PhastRisk v6.54. Multi-energy model [6] in the PhastRisk is used for estimation of overpressure in vapour cloud explosion. Confined strength is an input parameter in the model for calculation of overpressure. In order to apply the multi-energy model, guidance suggested by Kinsella [12] is adopted to determine the confined strength via the determination of the blast strength class. Kinsella [12] divided blast strength category into 12 categories. Blast strength category is a combination of ignition strength, obstruction, existence of parallel plane confinement / unconfinement. “Blast strength category” is used for determining the blast strength class. “Blast strength category” 1 represents high in ignition strength, obstruction and confinement. The lower blast strength category is, the higher the blast strength class. The highest blast strength class 10 is equivalent to detonation of TNT explosive. Thus, high blast strength class implies high initial overpressure. Hazard distance in a VCE increases with the increase in initial overpressure. Blast strength category 3 (equivalent to confined strength between 5 and 7) is estimated based on the following assumptions,

· High Obstruction –50% volume blockage ratio

· Existence of parallel plane confinement – vertical walls

· Low ignition strength – ignition sources such as spark (mechanical or electrical), flare stack, hot surface

4.69 Considering the 300mm diameter hole size scenario, the whole content of the buffer tank releases to the atmosphere in less than 10 minutes. It is assumed that the amount of gas in a VCE is the same as the rupture scenario. For the 30mm equivalent hole size leak scenario, it is assumed the amount of gas in a VCE is equivalent to 10 minute discharge without being noticed. The hazard distance for VCE in the rupture scenario (the worst case) is indicated on Figure 4.5. Hazard distances obtained from consequence modelling for identified release scenarios are tabulated in Table 4.18. The hazard distances show hazard of biogas storage does not have impact on the chlorine facilities at the SHWWTW while the distance between the gas buffer tank and the chlorine building is more than 150m.

4.70 Duration for a flash fire and fireball is very short. Although a flash fire spreads up to 116m from the gas holder, the flash fire sustains only for seconds. The perimeter of the fire reduces rapidly when biogas is ignited and consumed in the fire. On the other hand, the fireball duration only lasts for 6 seconds in the gas buffer tank rupture event according to modelling results. Therefore, thermal radiation due to flash fire or fireball would not cause damage to facilities or buildings at the SHWWTW. While maximum flame length of 31m from the gas holder is estimated for jet fire events, the flame does not reach the SHWWTW. Thus, jet fire event would not pose fire risk to the SHWWTW facilities or chlorine store room.

Table 4.18 Hazard Distances

Release Scenario		Release Size	Outcome	The Worst Hazard Distance (Weather Class)
Gas buffer tank	Rupture	3,600 kg	Fireball	40m
			VCE	85m
			Flash fire	116m (15D)
Gas buffer tank	Leak	300 mm dia.	VCE	85m
			Flash fire	25m (1F)
			Jet fire	31m (15D)
Inlet / Outlet piping Safety valve Pump / non-return valve / flange	Full bore rupture	300 mm dia.	VCE	85m
			Flash fire	24m (1F)
			Jet fire	33m (15D)
Inlet / Outlet piping Safety valve Pump / non-return valve / flange	Leak	30 mm dia.	VCE	26m
			Flash fire	2.9m (1F)
			Jet fire	5m (15D)

Frequency Estimation

Review of Generic Frequencies

4.71 Generic frequency is estimated based on the historical incidents review identified the accidents that involved the use of chlorine in water treatment works and the road transport of chlorine, the generic accident frequency can be estimated through the information of the number of water treatment works involved, the operating period and the total number of accidents occurred within the operating period.

4.72 The generic frequencies estimated in the 1992 SHWWTW HA report has been compared with the historical accident frequencies. Details on historical incidents extracted from MHIDAS can be found in Appendix 4.2. The objective of the comparison is to confirm the appropriateness of adopting event frequencies, which were derived in the 1992 SHWWTW HA report, for this QRA.

4.73 The historical accident frequency of a WTW is estimated to be 1.01 x10^-4per plant-year. In the 1992 SHWWTW HA report, the sum of frequencies of all failure cases is 2 x10^-3 per plant-year.

4.74 The comparison of the results concludes that the frequencies used in the 1992 SHWWTW HA report are more conservative than the estimated values from the historical incident. Therefore, the frequencies in the 1992 SHWWTW HA report are considered appropriate for this QRA by adopting a conservative approach. Details of generic frequency estimation are given in Appendix 4.3.

Summary of Release Frequencies of Siu Ho Wan WTW

4.75 If the contain-and-absorb system functions as intended, there will be no chlorine hazard outside the chlorine building. As the actual scrubber installed at the SHWWTW is of a type with a higher reliability of 0.99, failure probability of 0.01 is adopted in the analysis. This is apportioned equally between the 2 cases “Ventilation Fails On” and “Absorber Fails, Ventilation Off”. Taking into account the 0.01 failure rate of the contain–and-absorb system, the failure frequencies for chlorine release to the atmosphere are listed out in Table 4.19.

Table 4.19 Summary of Release Frequencies to Outside of Building

Component	Failure Mode	Hole Size	Contain & Absorb Failure Frequency (per year)
			Vent On	Building Collapse	Vent off, No Absorption
Drum	Spontaneous	Small	4.02E-05	-	4.02E-05
		Large	3.07E-05	-	3.07E-05
		Rupture	2.92E-06	-	2.92E-06
	Handling	Small	1.35E-05	-	1.35E-05
	Earthquake	Rupture (1 drum)	3.65E-06	-	3.65E-06
	Aircraft crash	Rupture (73 drums)	-	4.20E-08	-
	Aircraft crash	Rupture (2 drums)	-	7.70E-08	-
Liquid Line	Spontaneous	6mm	1.50E-07	-	1.50E-07
	External Impact	6mm	-	5.50E-05	-
	Evaporators	6mm	2.00E-06	-	2.00E-06
Fittings	Handling	15mm	2.96E-06	-	2.96E-06
	External Impact	15mm	-	5.50E-05	-
Gasline	Spontaneous	6mm	6.00E-08	-	6.00E-08
	External Impact	6mm	2.50E-08	-	2.50E-08

4.76 Estimates of failure frequencies and representative failure size case for each failure during transportation are presented in Appendix H.II of the 1992 SHWWTW HA report. Onsite road transport failure frequencies of the mitigated case (i.e. using route 2) are presented in Table 4.20 below.

Table 4.20 Onsite Road Transport Failure Case Frequencies

Failure in road accidents	Small Leak, 0.24kg/s (per year)	Large Leak, 1.7kg/s (per year)	Catastrophic failure (per year)	Large Leak of 4.5te, 7.65kg/s (per year)
Failure in Road Traffic Accidents
Roll Over	1.53E-06	6.74E-07	-	-
Crushed at rear	6.62E-08	-	-	-
Crushed at side	1.12E-07	-	-	-
Vehicle Fire	5.16E-08	-	-	-
Tanker Fire	-	-	-	3.17E-09
Other Failures on Road
Spontaneous	2.56E-06	1.99E-06	1.90E-07	-
Load Shedding	-	2.27E-05	-	2.52E-06
Truck Fire	-	-	-	1.62E-07
Total	4.32E-06	2.54E-05	1.90E-07	2.69E-06

4.77 From Section 6.2.2 of the 1992 SHWWTW HA report, reinforced concrete panels were used instead of brick panels in the construction of the chlorine drum store and duty area. The integrity of the drum store in the event of an instantaneous rupture of 1-tonne drum due to either spontaneous failure or earthquakes can be maintained. All releases other than those caused by aircraft impact would be confined by the building and would only lead to an emission to outside atmosphere in the event of failure of the contain-and-absorb system. This measure is also postulated to be 1.7 kg/s continuous release due to earthquakes for the mitigated cased in the 1992 SHWWTW HA report.

Consequence Analysis

4.78 The dispersion model used in the 1992 SHWWTW HA report is based on the Cox and Carpenter dense cloud dispersion model. The implementation of the Cox and Carpenter model illustrated in the World Bank Hazard Analysis software packages (WB) was used for analysis.

4.79 Detailed results of the consequence analysis conducted for this risk assessment are shown in Table 5.1(a) – (l) of the 1992 SHWWTW HA report, which tabulate the effect zones associated with various end points of the hazardous outcomes considered. The risk consequence results are listed out in Appendix 4.4. Dispersion model results given in Table 5.1(a) – (l) of the 1992 SHWWTW HA report are represented by:

d: downwind travel (meters);

w: maximum crosswind travel (meters); and

m: distance to maximum width (meters), w, occurs.

4.80 These parameters are utilized within the risk integration software (ToxicRisk) to define the footprint area of the hazard. Number of people affected is calculated together with meteorological data and population distribution data.

4.81 A Lethal Dose (LD) contour is described by 2 semi-ellipses, Figure 4.4. One ellipse has major axis with length m and another ellipse has major axis with length (d-m). Both ellipses have minor axis with length w. LD contours are modelled by 16-point polygons. The number of points is sufficient to have close approximation to the LD contours. Error due to this approximation is minimized.

4.82 The protection factor 0.9 is applied to indoor population. Thus, probability of fatality for indoor population is assumed 10% of outdoor population.

4.83 Escape factors are applied to outdoor population by considering a person able to escape to indoor in continuous release events. Probabilities of successful escape are assumed 0%, 20% and 80% when a person is at LD90, LD50 and LD3 fatality zones.

4.84 Having taken into account both protection and escape factors, probabilities of fatality are adjusted under different circumstances and tabulated in Table 4.21.

Table 4.21 Probability of Fatality

Fatality Zone	Probability of Fatality (outdoor)	Indoor			Outdoor
Fatality Zone	Probability of Fatality (outdoor)	Escape Factor	Protection Factor	Probability Fatality	Escape Factor	Protection Factor	Probability Fatality
Instantaneous Release Events
LD90	0.90	0	0.9	0.090	0	0	0.90
LD50	0.50	0	0.9	0.050	0	0	0.50
LD3	0.03	0	0.9	0.003	0	0	0.03
Continuous Release Events
LD90	0.90	0	0.9	0.090	0	0	0.9000
LD50	0.50	0	0.9	0.050	0.2	0	0.4100
LD3	0.03	0	0.9	0.003	0.8	0	0.0084

Risk Summation

4.85 This section presents the risk results which are derived by combining the frequency of hazardous outcome events with the associated consequences. Risk summation was conducted using in-house software ToxicRisk.

Individual Risk

4.86 Since hazardous events for all 4 scenarios are the same, individual risk contours due to the SHWWTW for existing, construction phase and operation phase are shown in Figure 4.6. These contours express the risk to a hypothetical individual present outdoors 100% of the time. As seen from the figure, the 10^-5 per year contour falls entirely within the plant boundary. Individual risk from the plant is therefore considered acceptable in accordance with the Hong Kong Risk Guidelines which state that the individual risk off-site should not exceed 10^-5 per year.

Societal Risks

4.87 Societal risks for all four scenarios are plotted on Figure 4.7. For both construction and operation phases, societal risks fall into the lower “ALARP” region. Figure 4.8 shows societal risks for “OWTF population only” and for “overall population” for the Year 2013 scenario without LLP Development.

4.88 Another measure of societal risk is expressed in terms of off-site Potential Loss of Life (PLL) value. PLL can be calculated by summing up fN pairs. A sample listing of fN pairs for Year 2011 Scenario, which are relevant to transport events and OWTF population, is given in Appendix 4.5. Results of PLL for each scenario are summarized in Table 4.22.

Table 4.22 Summary of PLL

Population Group	Potential Loss of Life, PLL (per year)
Population Group	Year 2008	Year 2011	Year 2013 w/o LLP	Year 2013 w LLP
Background Population	4.96E-05	5.14E-05	5.26E-05	5.90E-05
OWTF	-	1.71E-05	6.76E-06	6.76E-06
Concrete batching plants	6.59E-06	-	-	-
Overall population	5.62E-05	6.85E-05	5.94E-05	6.58E-05

4.89 In comparison with the Year 2008 baseline scenario, societal risk at operation stage of the OWTF is slightly higher than the baseline scenario. Currently, the OWTF site is occupied by 2 concrete batching plants with estimated population of 40 (each plant is estimated of 20 persons). Comparison of risk outcome due to change of population properties is made by referring to Table 4.22. PLL for the OWTF operation is at the same magnitude as the current one for concrete batching plants. Although the maximum population at the OWTF is higher than the population of concrete batching plants, 40 of the maximum population come from visitors and their presence only occupies a small fraction of time in a year. Therefore, risk contribution of the OWTF is similar to the existing concrete batching plants.

4.90 As a whole, the OWTF is a minor contributor to the overall risk outcome. Operation of the OWTF accounts for 11% and 10% of the total PLL for the Year 2013 scenario without and with LLP Development respectively. Minor contribution of the OWTF is depicted on Figure 4.8 in which societal risk curve for “OWTF population only” is below the societal risk curve for “overall population”.

4.91 However, PLL at the Project site during construction phase accounts for 25% of the total PLL. As the societal risk falls into the ALARP region, mitigation measures should be proposed to reduce the risk to as low as reasonably practicable.

Risk Mitigation Measure Identification and Analysis

4.92 The societal risks of the SHWWTW during both construction and operation phases of the proposed Project fall into the lower “ALARP” region of the Risk Guideline. Therefore, mitigation measures are identified and analysed. Candidate mitigation measures in the Route 16 EIA study [16] have been reviewed. Mitigation measures, which are feasible and necessary for this study, are listed out in Table 4.23.

Table 4.23 Feasible Risk Mitigation Measures

Item	Mitigation Measures	Remarks
Construction Phase
A.1	Suspension of construction work during chlorine deliveries	According to WSD, maximum number of deliveries for chlorine drums between years 2007 and 2009 is 6 times a year. The incurred cost is minimal. This option is further assessed in Cost Benefit Analysis.
A.2	Enhance emergency response arrangements, e.g. provision of visual and audible alarms, training, safe refuge, emergency and evacuation plan etc.	Provision of a means of alerting construction workers in case of chlorine release is considered an effective mitigation measure. Therefore, it is recommended to implement as a good practice. A safe refuge for 60 persons requires 10 or more 20L compressed air cylinders by taking into account rescue operation of FSD. It is not practical to store and maintain this number of air cylinders on site. An evacuation plan is a practical measure to facilitate a timely and effective response to a chlorine release. Training should be provided for an effective emergency response in case of release.
A.3	Fence around the site boundary facing SHWWTW chlorine store	According to the works of Meroney (1991) on vapour barriers in Lees’ [10], a solid fence barrier with 3m – 12m height has near field concentration reduction factor between 2 and 9. Moreover, wind tunnel tests in the reassessment study for the Sha Tin Water Treatment Works found that cloud height of 4m was resulted in 1-ton release. For the OWTF site with a solid fence 3m along the site boundary facing the SHWWTW and within 150m downstream, concentration reduction factor between 2 and 3 can be obtained by interpolation of Meroney’s modelling results. Based on consequence modelling results, LD50 contours reach the OWTF boundary in most release case. Considering concentration reduction factor of 3, a chlorine cloud is diluted to 30% of the original concentration passing the fence. Having applied probit equation for chlorine (TNO 1992), probability fatality is reduced to less than 1%. Therefore, this option is considered effective reducing fatality due to significant chlorine dilution. Although it is most effective to construct a fence near the release source, construction of the fence at the SHWWTW boundary would cause interference to works for the integration of Siu Ho Wan and Silvermine Bay Water Treatment Works. Besides, this arrangement affects operation of the SHWWTW and has to be agreed by WSD.
Operation Phase
B.1	Site office as far away as possible from the SHWWTW chlorine store; avoid windows or openings on facades facing the SHWWTW	This option has already been implemented into in the preliminary design of the layout plan. The effect has been accounted for in the risk modelling. Although it is not further assessed, it is considered a good design practice.
B.2	Fence around the site boundary facing SHWWTW chlorine store	According to the works of Meroney (1991) on vapour barriers in Lees’ [10], a solid fence barrier with 3m – 12m height has near field concentration reduction factor between 2 and 9. Moreover, wind tunnel tests in the reassessment study for the Sha Tin Water Treatment Works found that cloud height of 4m was resulted in 1-ton release. For the OWTF site with a solid fence 3m along the site boundary facing the SHWWTW and within 150m downstream, concentration reduction factor between 2 and 3 can be obtained by interpolation of Meroney’s modelling results. Based on consequence modelling results, LD50 contours reach the OWTF boundary in most release case. Considering concentration reduction factor of 3, a chlorine cloud is diluted to 30% of the original concentration passing the fence. Having applied probit equation for chlorine (TNO 1992), probability fatality is reduced to less than 1%. Therefore, this option is considered effective reducing fatality due to significant chlorine dilution. Although it is most effective to construct a fence near the release source, construction of the fence at the SHWWTW boundary would cause interference to works for the integration of SHWWTW and Silvermine Bay WTW. Besides, this arrangement affects operation of the SHWWTW and has to be agreed by WSD.
B.3	Enhance emergency response arrangements, e.g. provision of visual and audible alarms, sheltering place, training, emergency and evacuation plan etc.	Installation of alarms and provision of training to workers of contractors and onsite personnel improve the effectiveness of emergency plan. Training should be provided to ensure all onsite personnel know what action should be taken and where sheltering place is. It is considered a good practice. A safe refuge for 85 staff and visitors requires 10 or more 20L compressed air cylinders by taking into account rescue operation of FSD. It is not practical to store and maintain this number of air cylinders on site. An evacuation plan is a practical measure to facilitate a timely and effective response to a chlorine release. Training should be provided for an effective emergency response in case of release.

Analysis of Mitigation Measures

4.93 In this study, the cost effectiveness is assessed by Cost-Benefit-Analysis (CBA) using calculation of the Implied Cost of Averting Fatality (ICAF) for each mitigation measures identified. The ICAF is calculated using the equation as follows by taking into account the reduction in potential loss of life (PLL) using calculation

ICAF =	Cost of Mitigation Measure
	(Reduction in PLL Value x Design Life of Mitigation Measure)

4.94 The ICAF can be compared with the value of life (proposed to be HK$33M in this study) to determine whether the implementation of the identified mitigation measures is reasonably practicable.

4.95 Design life of mitigation measure for construction phase is taken as 2 years according to the construction program. Design life of mitigation measure for operation phase is assumed 15 years.

4.96 Mitigation measures are quantified and cost estimation is provided in Table 4.24.

Table 4.24 Quantification of Risk Mitigation Measures and Cost Estimation

Item	Mitigation Measures	Quantification / Cost Estimation
Construction Phase
A.1	Suspension of construction work during chlorine deliveries	Quantification No fatality is assumed for the OWTF construction site in all transport related events. Cost Estimation According to information from WSD, maximum number of chlorine deliveries between years 2007 and 2009 is 6 times per year and each delivery takes half days to 1 day to complete. It implies maximum 12 days delay on the construction program. Construction workers will not be paid during suspension of work. Cost is estimated incurring from salary for security personnel. Estimated cost is HK$ 4,800 (= 2 workers x $200/shift x 1 shift/day x 6 days/year x 2 years) throughout the construction period.
A.2	Enhance emergency response arrangements, e.g. provision of visual and audible alarms, training, safe refuge, emergency and evacuation plan etc.	Quantification In most release scenarios ( with release rate ≤1.7kg/s), LD90 cannot reach the OWTF site while LD50 and LD3 contours have downwind distances of 280m and 625m. Applying probit equation for chlorine (TNO 1992) and escape time, probability fatality is found to be less than 1% for evacuation from the OWTF site. It is assumed no fatality for OWTF population is assumed in all events except aircraft crash events and chlorine transport event with release rate 7.65kg/s. Cost Estimation Cost mainly involves installation of alarm system and provision of training to construction workers. The total cost is estimated within HK$15,000.
A.3	Fence around the site boundary facing SHWWTW chlorine store	Quantification As fatality probability at OWTF could be reduced to less than 1% in most case with the fence, no fatality for OWTF population is assumed in all events except aircraft crash events. Cost Estimation Solid fence with minimum height 3m will be erected for boundary facing (200m length) the SHWWTW during construction phase for security purpose. The incurred cost is mainly due to maintenance of those damaged sections, for example in typhoon. Such fence could be made of wooden panels and the maintenance cost is estimated HK$200 per m. It is assumed 10% of the fence requires maintenance per year. Maintenance cost of HK$8,000 is estimated for 2-year construction period.
Operation Phase
B.1	Site office as far away as possible from the SHWWTW chlorine store; avoid windows or openings on facades facing the SHWWTW	Quantification This measure has already built into the preliminary design. No further risk reduction in further assessment. Cost Estimation No extra cost is incurred.
B.2	Fence around the site boundary facing SHWWTW chlorine store	Quantification As fatality probability at OWTF could be reduced to less than 1% in most case with the fence, no fatality for OWTF population is assumed in all events except aircraft crash events. Cost Estimation Fence for construction phase could be retained for operation phase. No additional setup cost would be incurred. The maintenance cost is estimated HK$200 per m. It is assumed 10% of the fence requires maintenance per year. Maintenance cost of HK$60,000 is estimated for 15-year design life of the OWTF.
B.3	Enhance emergency response arrangements, e.g. provision of visual and audible alarms, sheltering place, training, emergency and evacuation plan etc.	Quantification In most release scenarios (with release rate ≤1.7kg/s), LD90 cannot reach the OWTF site while LD50 and LD3 contours have downwind distances of 280m and 625m. Applying probit equation for chlorine (TNO 1992) and escape time, probability fatality is found to be less than 1% for evacuation from the OWTF site. It is assumed no fatality for OWTF population is assumed in all events except aircraft crash events and chlorine transport event with release rate 7.65kg/s. Cost Estimation Negligible cost as there should be a fire alarm system and fire drill at the OWTF site.

4.97 Mitigation measures are evaluated independently. Reduction in PLL for each mitigation measure is calculated according to the quantification method shown in Table 4.24. Results are tabulated in Table 4.25. Since the site office building of the OWTF has already been moved as far away as possible from the SHWWTW chlorine store in the preliminary layout design, calculation for Option B.1 is not available.

Table 4.25 Quantification of Risk Mitigation Measures and Cost Estimation

Mitigation Measures	Reduction in PLL (per year)
Construction Phase
A.1	1.13E-05
A.2	1.24E-05
A.3	1.66E-05
Operation Phase
B.1	-
B.2	6.55E-06
B.3	4.93E-06

4.98 To evaluate the justifiable expenditure on risk mitigation measures at this risk level, Maximum Justifiable Expenditure (MJE) are determined as follows.

MJE =

Decrease in PLL Value (per year) x Value of Life (HK$) x operating life time of construction works (years) x aversion factor

4.99 The aversion factor indicates the level of aversion to accidents causing large numbers of fatalities (HSE [15]). Aversion factor of 20 (Maximum Aversion Factor for risks at the upper region of the Risk Guidelines) is proposed to adjust the Value of Life to reflect people’s aversion to high risk. This is a conservative factor adopted even though the FN curves located at the low “ALARP” region. With this factor applied, the adjusted Value of Life of HK$660M will be adopted.

4.100 ICAF and MJE for the selected mitigation measures are calculated and tabulated in Table 4.26.

Table 4.26 Estimated ICAF and MJE for Selected Mitigated Measures

Mitigation Measures	ICAF (HK$ M)	Adjusted Value of Life (HK$ M)	MJE (HK$)	Cost (HK$)	Justified
Construction Phase
A.1	213	660	14,875	4,800	Y
A.2	603	660	16,408	15,000	Y
A.3	241	660	21,865	8,000	Y
Operation Phase
B.1	0	660	-	0	Y
B.2	611	660	64,824	60,000	Y
B.3	0	660	48,816	0	Y

4.101 All mitigation measures shown in Table 4.26 are classified as justified.

4.102 Details of the good practice measures for reducing risk level are listed as follows,

· The number of workers on site during construction stage should be kept as the level as assessed in this report.

· Construction works should be suspended when delivery of chlorine takes place.

· 3m high fence should be constructed along the boundary facing the SHWWTW as shown in Figure 4.10.

· Emergency evacuation procedures should be formulated and the Project proponent should ensure all workers on site should be familiar with these procedures as well as the route to escape in case of gas release incident. Relevant Departments, such as Fire Services Department (FSD), should be consulted during the development of Emergency procedures. Diagram showing the escape routes to a safe place should be posted in the site notice boards and at the entrance/exit of site. A copy of the latest version emergency procedures should be dispatched to Tung Chung Fire Station for reference once available.

· The emergency procedures should specify means of providing a rapid and direct warning (e.g. Siren and Flashing Light) to construction workers in the event of chlorine gas release in the SHWWTW.

· The construction site officer should establish a communication channel with the SHWWTW operation personnel and FSD during construction stage. In case of any hazardous incidents in the treatment works, operation personnel of SHWWTW should advise the site officer to inform construction workers to proceed with emergency procedure. The site agent should appoint a Liaison Officer to communicate with FSD Incident Commander on site in case of emergency.

· Introduction training should be provided to any staff before carryout construction works at the Project site.

· Periodic drills should be coordinated and conducted to ensure all construction personnel are familiar with the emergency procedures. Upon completion of the drills, a review on every step taken should be conducted to identify area of improvement. Prior notice of periodic drills should be given to Station Commander of Tung Chung Fire Station (at contact numbers Tel: 2988 1898 or 2988 8733 and Fax: 2988 1688). Joint operational exercise with FSD and SHWWTW is recommended.

4.103 The following safety measures are recommended in the design of the layout plan in attempting to minimize the impact of a chlorine release on the personnel in the Project site during operation phase,

· The site office should be close to the western boundary of the Project site (referring to the preliminary layout plan on Figure 2.2) and away from the Siu Ho Wan WTW’s chlorine store as far as possible;

· 3m high fence should be constructed along the boundary facing the SHWWTW as shown in Figure 4.10.

· Emergency evacuation procedures should be formulated and the Project proponent should ensure on site staff should be familiar with these procedures. Diagram showing the escape routes to a safe place should be posted in the site notice boards and at the entrance/exit of site. A copy of the latest version emergency procedures should be dispatched to Tung Chung Fire Station for reference once available.

· The emergency procedures should specify means of providing a rapid and direct warning (e.g. Siren and Flashing Light) to personnel on site in the event of chlorine gas release in the SHWWTW.

· The OWTF site officer should establish a communication channel with the SHWWTW operation personnel and FSD. In case of any hazardous incidents in the treatment works, operation personnel of SHWWTW should advise the site officer to inform personnel on site to proceed with emergency procedure. The site agent should appoint a Liaison Officer to communicate with FSD Incident Commander on site in case of emergency.

· Periodic drills should be coordinated and conducted to ensure all on site personnel are familiar with the emergency procedures. Upon completion of the drills, a review on every step taken should be conducted to identify area of improvement. Prior notice of periodic drills should be given to Station Commander of Tung Chung Fire Station (at contact numbers Tel: 2988 1898 or 2988 8733 and Fax: 2988 1688). Joint operational exercise with FSD and SHWWTW is recommended.

Mitigated Scenarios

4.104 Having the recommended mitigation / safety measures in place with quantification method as given in Table 4.24, risk outcomes for the mitigated scenarios are obtained and presented in societal risk curves as shown in Figure 4.9. The FN curves indicate that the recommended mitigation measures can reduce the risk outcome as low as reasonably practicable and the societal risk complies with the risk guidelines stipulated in Annexes 4 and 22 of the EIAO TM.

Conclusion

4.105 A hazard assessment has been conducted following the criteria for evaluating hazard to life as stated in Annexes 4 and 22 of the EIAO TM (Hong Kong Risk Guidelines). The assessment has reviewed and evaluated hazardous scenarios from SHWWTW to both the construction and operation stages of the proposed organic waste treatment facilities.

4.106 Hazardous scenarios associated with the chlorine storage and on-site transport of chlorine in the SHWWTW and the respective occurrence frequencies have been identified and confirmed by reviewing of historical hazardous incident database to ensure hazards and its respective occurrence frequencies are suitable for use.

4.107 Potential hazards from the biogas storage to the chlorine storage at the SHWWTW have been identified. The hazards have been evaluated and would not affect the storage, use or transport of chlorine in the SHWWTW.

4.108 A QRA expressing population risks in both individual and societal terms has been conducted to evaluate the risk to population in the vicinity of the SHWWTW. This assessment considered the future growth of nearby population.

4.109 From the results of the QRA, individual risk at the Project site is found to be at the order of 1E-06 per year which is acceptable in comparison with the criteria (1E-05 per year) in the Risk Guidelines.

4.110 With respect to societal risk, the FN curves indicate that the risk level falls marginally into “ALARP” region in the baseline scenario and both construction and operation phases (Year 2008, Year 2011 and Year 2013).

4.111 Cost effective mitigation measures have been identified and demonstrated by CBA. Mitigated scenario for both construction and operation phases have been assessed. Risk outcomes for mitigated scenarios are reduced to as low as reasonably practicable.

4.112 Safety measures for both construction period and operation stage are recommended in order to ensure that the personnel in the project site would understand the emergency procedures when they work in vicinity of the SHWWTW. Safety measures have already built into the layout plan of the proposed waste treatment facilities to safeguard onsite personnel.

4.113 In conclusion, the risks during construction and operation of the proposed organic waste treatment facilities are considered to be reduced to as low as reasonably practicable with the implementation of the “Good Practice” measures during construction period and operation stage.

References

[1] Environmental Impact Assessment Study Brief "Organic Waste Treatment Facilities, Phase I" (ESB 172/2007).

[2] Hong Kong EPD, issued under Section 16 of the Environmental Impact Assessment Ordinance. Technical Memorandum on Environmental Impact Assessment Process, Annexes 4 and 22.

[3] PHH/SWK Water Consultants Joint Venture (1992). The Hazard Assessment Study – Final Report, June 1992, Agreement No. CE 12/91 “North Lantau Water Supply Project”.

[4] Health and Safety Executive (HSE), Safety Report Assessment Guide: Methane Gas Holders (2001).

[5] Center for Chemical Process Safety (CCPS) of the American Institute of Chemical Engineers, Guidelines for Consequence Analysis of Chemical Releases (1989).

[6] TNO Yellow Book, Methods for the calculation of physical effects resulting from releases of hazardous materials (CPR 14E).

[7] Approved EIA Report “Development of a Biodiesel Plant at Tseung Kwan O Industrial Estate" (EIA-156/2008).

[8] Hong Kong Transport Department, the Annual Traffic Census 2007.

[9] Major Hazard Incident Data Service (MHIDAS).

[10] Lee’s Loss Prevention in the Process Industries, 3rd Edition, 2005.

[11] Yoshitomo Inaba, Tetsuo Nishihara, Study on explosion characteristics of natural gas and methane in semi-open space for the HTTR hydrogen production system, Nuclear Engineering and Design 232 (2004) 111–119.