DIVISION OF.

LIBRARY SERVICES Enhancing teaching, learning and research

COMMONWEALTH OF A U S T R A L I A
Copyright Regulations 1969

WARNING
*

This material has been reproduced and communicated to you by or on behalf of Central Queensland University pursuant to Part VB ofthe Copyright Act 1968 (the Act). The material in this communication may be subject to copyright under the Act. Any further reproduction or co.munication of this material by YOU may be the subject of copyright protection under the Act.

Do not remove this notice,

Mathematical Evaluation for Controlling Hazards

William T. Fine

Two significantneeds have been recognized in order to facilitate efficient control of hazards for the purpose of accident prevention. One is the need for a method to determine the relative seriousness of all hazards for guidance in assigning priorities to preventive effort. The second need is for a method to measure the justification of the estimated cost of the conternpla.ted corrective action. To fulfill these needs, the author gives a formula that calculates the risk of a hazardous situation and gives a numerical evaluation to the urgency for remedial attention to the hazard. Calculated risk scores are then used ta establish priorities for corrective effort. An additiormi formula weighs the estimated cost and effectiveness of any contemplated corrective action and gives a quantitative estimate of the justification of the cost.

A problem that frequently faces many safety personnel is to determine how serious each known hazaxcl is, and to decide to what extent he should concentrate his resources to correct the situation. Normal safety row tines such as inspections and investigations usually produce a varying list of hazards that are too numerous to be corrected all at once. Decisions must be made as to which are the most urgent. On costly projects, higher managemenk often asks whether the particular risk justifies the cost of the work required to eliminate it. Since all budgets are limited, it is necessary to decide whether costly projects to eliminate hazards are really worthwhile. The question of whether an expensive engineering project is justified is usually answered by a general opinion that may be little better than guesswork, Unfortunately in many cases, the decision to undertake any costly correction of a hnzsrd depends to a great extent on the saEesnznns7zip of safety William T,Pine is chief of the Safety Depart-

ment, Naval Ordnance Laborntory, Silver Spring, Maryland.

personnel. Because of insuficient information, the cost of correcting a very serious hazard may be consiclered prohibitive by management, and the project postponed; or due to excellent selling jobs by safety personnel, elaborate engineering or construction jobs may be approved when the risks involvecl really do not justify them, This article presents a formula for calculating the risk due to a hazard. I t quantitatively evaluates the potential severity of a hazardous situation. Use of the formula will provide R logical system for safety personnel and management to set priorities for nttention to hazardous situations. This does not imply in any way that a cost, no matter how great, is not worthwhile if it will prevent an accident and save a human life. However, one must approach accident prevention with reason and good judgment. The 1naxirnum possible benefit must be derived from any expenditure for safety. When an analysis results in a decision that the cost of certain measures to eliminate n hazard is not justified, this does not mean that the hazard is not serious and may be ignored.

Decenaber 1g71 /Volume 3/Nu-mhar 4

157

rt. does 1 1 m n that, 1)asecl on an evaluation of the controlling factors, the return on the investment or the amount of acciclent prevvntion bencfit is below the standards estahlisbed. The nmount of money involved will provide greutcr safety benefit if used to alleviate other higher-risk hazards that this system will identify. As for the hazard in question, less costly preventive lneasures sltoulcl be sought.
CALCULATING RELATIVE RISK SCORES

For the purpose of this presentation, three i~nport:mtdefinitions are: a. Hnzard: m y unsafe condition or potential source of an accident. Examples are: an unguarded hole in the ground; defective brnkes on a vehicle; a deteriorated wood ladder; a slippery road. b. Ilazurcl-eusnt: the combination of a Ixmrcl with some person or activity that could start an accident sequence. Examples of hazard-events are: a person walking througll a field that contains an unguarded we11 opening; a person not wearing eye protection in an area especially hazardous to eyes; a person driving a vehicle that has defective brakes; a man climbing up a defective ladder; a person driving a vehicle on a slippery road. c. AccZdent sequence: the chain of occurrences tlmt begins with a 3azard-event and ends with the consequences of an accident. The expression a calculated risk is oftell used for situations where work is to be done without proper safety precautions. Usually such work is done without any mathematical measurement of hazards. Using the fonnuln below, the risk is actually calculated and, thus, more meaningfulIy quantified. Such quantification is achieved by considering the potential consequences of an accident, the exposurs factor, and the p~*oljability factor, The fornlula is : R=CxExP where R represents the risk score C represents the consequences E represents the exposwe factor P represents the probability factor

Conseqttencas refer to the most probable results of a potential accident, including injuries and property damage. The rating selected depends on nn appraisal of the entire situation surrounding the hazard and accident experience. Factor l in Table 1 gives degrees of consequence ranging froltl minor to catastrophic. A nurnericaI rating nssociated with each level appears in tl1e column at the right. If an identified hazard has tlw potential of producing n catastrophe involving numerous fatalities or over $1,000,000 dnmnge, its numerical value in the formula will be 100. If, as is more common, an identified hazard c m produce a disabling injury or dalmge up to $1,000, it will have a rating of 5. The rest of the values may be read directly from the remaining items rrnder Factor 1. Exposure refers to the frequency of occurrence of the hazard-event ( a conlbination of a hazurd with a person or activity that can start an accident sequence). Factor 2 in Table 1 gives the various levels of exposure and the numerical rating associated with each level. Selection of the appropriate exposure level is based on observation, past experience, and knowledge of the activity concerned. Events that occur continuously or many times daily receive a rating of 10 while events that are only remoteIy possible receive a rating of 0.5. Probability refers to the likelihood that, once the hazard-event occurs, the complete accident sequence will follow with the necessary timing and coincidence to result in the accident and consequences. This is determined by careful consideration of each step in the accident sequence all the way to the consequences. Factor 3 in Table 1gives the various probability levels and the related ratings. Factors 4 and 5 will be discussed later,
EXAMPLES OF RISK SCOIiES

Use of the above forn1ula cm be demonstrated using four real, but widely different, examples. Tha diversity is intended to illustrate tke broad app1icabili.t~ the method. of
Exumple no. 1involves a quarter-mile stretch of two-lane road used by both vehicles
Journal of Safety Research

158

TABLE 1 RATINGS FOR RISK CALCULATION AND COST JUSTIFICATION

FACTOR
CLASSIFICATION
AATINQ

1. Consequences (Most probable result of the potential accident)

2. E

osure ?The frequency of occurrence of the hazardevent )

Catastrophe; numerous fatalities; damage over $1,000,000; major disruption of nctivities Multiple fatalities; damage $400,000to $1,000,000 Fatality, damage $100,000 to $500,000 Extremely serious injury (amputation, permanent disability); damage $1,000 to $100,000 Disabling injury; damage up to $1,000 Minor cuts, bruises, bumps; minor damage Hazard-event occurs: Continuously (or many times daily) Frequently ( a proximntely once daily) Occasionally ( ?rom once per week to once per month) Unusually (from once per month to once per year) Rarely ( it has been known to occur) Remotely possible (not known to have occurred) Complete accident sequence: Is the most likely and expected result if the hazard-event tirkes place Is quite possible, not unusual, has an even 50/50 chance Would be an unusual sequence or coincidence Would be a remotel possible coincidence Has never happene after Inany years of exposure, but is conceivably possible Practically impossible sequence (has never happened) Over $50,000 $25,000 to $50,000 $10,000 to $25,000 ~1,000 $10,000 to 100 to $1000 25.00 to dl00 Under $25.00 Hazard positively eliminated, 100% Hazard reduced at least 75% Hazard reduced by 50% 75% to Hazard yeduced by 25%to 50% Slight efEect on hazard (less than 25%)

1 0
3
1
0.5

6
2

3. Probability

accident sequence will follow to completion )

( Likelihood that

10
6 3

4. Cost fnctor (Estimated dollar cost of proposed corrective action )

01 . 10 0 4 3 2
1 0.5

0.5

5,-

De ee of correction 6 e g r e e to which hazard will be reduced)

l
3 4
6

and pedestrians departing or entering the grounds. There is no sidewalk, so pedestrians n frequently walk i the road, especially when the grass is wet or snow covered. There is little hazard to pedestrians when all the traffic is going in one direction only. The hazard occurs, however, when vehicles are going in both directions and passing by each other. The vehicles require the entire width of the road, and pedestrians must then walk on the grass alongside the road, An accidental fatality could occur if a pedestrian steps into the road, or remains in the road, at a point where two vehicles are passing. Three steps are required to determine a risk score for this situation, One must list the accident sequence of events that could result in the undesired consequences, select
December 1971/Volume 3/Number 4

the value for each element in the formula, and perform the achlal calculation, The accident sequence might involve the following seven events, 1. It is a wet or snowy day, making the grass along the road wet and uninviting to walk on. 2. At quitting time, a line of vehicles and some pedestrians are leaving the grounds, using this road. 3, One pedestrian walks on the right side of this road and he is oblivious to the traff~c. (The hazard-event. ) 4, Although traffic is one way out at this time, one vehicle comes from the opposite direction causing the outgoing t r d c line to move to the right edge of the road. 5, The pedestrian on the right side of the

X 59

road fails to observe the vehicles, and he retnains in the road. 6. The clriver of one vehicle fails to notice the peclestrian and strikes him from the
I*el\l-.

7. The pcdestrinn is killed. Given tlw above series of events, the components of the forluda are supplied below, The consequcnce is :l fatality; therefore, C=25 (See Trtble 1). Tn relation to exposure, the hazard-event is the pedestrian remaining in t h e road and hilviI1g an ;lttitucle o f disregurd. This type of person appot~rs occmionnlly. Therefore, E d . Factor 2 in Table 1 gives the approxilnnte frequency of occurrence. The proba1)ility of 2111 events of the accident sequence following the hazard-event is: conceivably possible, &hough it has never happened in many years. The reasoning is ns follows: events 4, 5, S; and 7 are individutdTy unlikely, so the combination of their occurring sinlultancously is extremely remote, Event 4 is unlikely because traffic is one way at quitting time. Event 5 is unlikely because a number of drivers would undoubtedly sound theirhorns ancl force the pedestrians attention, Event 6 is unlikely because most drivers are not deliberately reckless. Event 7, a fatality, is unlikely because vehicle speeds are not great on the rond, and the most likely case would be a glancing blow and minor injury, Not even a minor injury has ever been reported in the real situation. In view of the above, the probability ( P ) is equal to 0.5. The third step involves substitution of values into the formula and the performing of the simple arithmetic. R=CxExP=25~3~0.5=37.5 The reader is cautioned that the risk score of one case alone is meaningless. Additional hazardous situations must also be calculated for comparative purposes nncl to develop a definite pattern,
Example no. 2 deals with an actual situation that occurred several years ago, Fifty (50) compressed air hoses were in use for general cleaning purposes j.n a machine shop. They were being used without proper pressurereduction nozzles at various pressures, some up to 90 pounds per square inch, This caused potential eye hazards, although eye protec160

tion w a s worn by most men. The most probable consequence of this l x m r d was a seriotts eye injury. Applying the same three steps as in the first example, the accident sequence is: 1. Many nlachine operators use compressed air streams to blow metal chips from work, 2. Most employees occasionally relnove their safety glasses while still in the hazitrdous nrea. (The haznrd-event, ) 3. Onc employee who is not wewing eye protection walks past a machine while an air hose is being used. 4. A mcatal chip blows into the employees (?ye. 5. A serious eye injury results, Given these circumstances, the consequence is rz disabling eye injury, so C=5. The exposure is directly obsewabIe. The haznrd-event, an employee removing his eye protection while still in an eye haza1a OLlS d area, occurs many times daily. E=10. The probabiIity of the total accident sequence is judged to be quite possible, so P=6. The calculation then is: R=CxExP=5xlOx6=300 After deriving the above two risk scores, plus two more, perspective will be gained by relating the scores to a larger context of risk scores calculated independently.

Example no, 3 involves a more complex situation, A 12,000 gaIlon propane storage tank is subject to two I~azards. One hazard is the fact that the tank is located beside a well travelled road. The road slopes and is occasionalIy slippery clue to rain, snow, or ice. It is possible that a vehicle (particularly a truck) could lose control, leave the road, strike, and rupture the tank, causing a propane gas explosion and fire that could destroy several buildings. The consequences might amount to $200,000 damage plus a fatality. The second hazard is the tanvs location close to ultra-highly compressed air lines and equipment. A high pressure pipeline explosion could result from a malfunctioning safety valve, a human error in operating the equipment, damage to a pipeline, or other causes, Blast or flying debris could conceivably strike the propane storage tank, rupture it, and cause it to explode with the
Journul of Safety Research

silme
111

Conse(pc.!nccs as the runnway vehicle.

tllis cilsc, there ilre two hazards; the evilltli1tio11 is dol^ i n two V;\rts, 0 1 1 ~for ei\cll of the huznrcls, ancI thc total scores arc!

i~clderl, Considering just the situation involving the runaway vehicle, the sequence might I,e as follows: 1. hlfi111y vehicles t~rc: clriven C ~ O W the hill Imiclc the storage tank. 2. The road has suclclenly l~ecolne sIippery due to m uncspcctecl freezing rain. 3. One truck starts to sliclt! on tllc slippery road as it descends tlw hill. (Thc hsmrdevent, ) 4. The driver loses his steering control at a point when he is uphill from the tank i111d moving toward it. 5. rrhcvehicles brakes fail to stop it from sliding. 6. The vehicle heads directly townrcl the tank, 7. The vehicle strikes the tanlc with enough force to rupture it and permit the propane gas to begin to leak. 8. A spark or flame, perhaps fro111 a vehicle fire, ignites the propane. 9. An esplosion ancl conflagration occur. 10. Building and equipment clamage is $200,000 and one man is killed, Here, the consequences involve one fatality and n loss of $200,000, Therefore, C=%. The hazarci-event that would start the ncciclent sequence is the truck starting to slide on the road. This has happened rarely, Therefore, the exposure factor ( E ) is equal to 1. To decide on the likelihood that the complete accident sequence will follow the occurrence of the Bazard-event, consider the probability of each event: Event 4: The loss of steering control at the precise point in the road appronching the tank is possible but would be U ~ I L I S L I ~ ~ . Event 5: Once the vehicle started to slide, if the road were ice-covered it would be expected that the brakes would fail to stop the slide. Event 6: The vehicle heading directly toward the tank is highly unlikely, Monlentum would cause the vehicle to continue stmight down the road. Event 7: The likelihood of the vehicle striking the tank with great force, and squarely, is extremely unlikely.

If a vehicle wcrc? sliding on an ice-covcrecl stlrfilce toward the til11lc, it odd be easily divcrtecl from its direction of travel by a 11ulnl)rr of o1)structions between the road ancl the tank, When roads are slippery, travel is cmtailcd m ~ clrivers are cautioned to l clrivc slowly, A slow rate of speed would be wMwly to produce enough force to clan~age tlw tank. The shape and position of the tank are such that : vehicle would tend to glance I OR of it. However, events 8, g, and 10 nrc likely to follow if event 7 takes place, In sunx~1ary,the highly unlikely nature of most of the events from 4 through 7 gives a net probability of almost one in ndlion. It has never happened, but it i conceivnble. s Therefore, P=0.5. The calculation for the first risk is: RI = 25 x 1x 0.5 = 12.5 The seconcl hazard is treated similarly, The list of events is: 1, Normal daily activities involve operation of e q u i p ~ e n t and pressurizing of pipelines, some of which are in the vicinity of the propane storage tank. 2. A pipeline containing air compressed to 3000 pounds per scpnre inch, approximately 50 feet away from the storage tank, has become deteriorated or damagecl. (Thc hazard-event.) 3. The pipeline bursts. 4. Metal debris is thrown in all directions by the blast and several pieces strike the propme tank with sufficient force to rupture the tank. 5. Propane starts to leak from the tank, 6. A spark ignites the propane fumes. 7. The propane and air mixture explodes. 8. Building damage is $200,000 ancl one inan is killed, The pipeline hazard would rtlte a C=%. In regard to the hazard-event, high pressure air lines have occasionally been neglected or clamaged. The frequency of such occurrences is considered UIILISLI~~. Therefore, E =2. An estimate is nlnde of the probability that a damaged pipeline will explode and that the explosion will occur close enough and with enough blast for flying debris to strike the propane tank with great force. Several bursts have occurred in the past years, but none has damaged the propane
161

Decamber 1Q71/Volume 3/Number 4

ing of these action categories cleptmds 011 the sound judgn~cnt of persons who are knowleclgc?t~blc about Iocal conditions, Su& a locally clevelopect Risk Score Sumn1av and Action S h e d is a most useful device. When IlecessiIry and advisable, the list ~ ~ be presented to management for top level uwareness m d approval of thc uction colt ~ m n The beneficial uses of this listing also . include thc following: I k ~ n ~ p Z a 4 deals with n less scvcre but m. 1. It establishes priorities for attention by probably no re CO~IIIIOII type of situation. both safety and management personnel, There are approximately 100 householdHLIZM~OUS situations appear in the order type refrigerators in use in which various of their importance. The position on the kincls of chemicals are stored, Many of these list for any item may be lowered by correfrigerators are not sparkproof. Flammt~ble rective action. The corrective action may voIatile solvents storecl in non-sparkproof be clirectecl toward the possible conserefrigerators could leak, vaporize, contact quences, the exposure, or the probability electrical sparks, m d result in an explosion factor. or fire. The most likejy results would be 2, It provides guiciance to indicate urminor injuries uncl possible damage estigency of newly discovered hazards. For mated at $200. The event sequence might each new hazardous situation, compute be: the risk score. Its urgency is indicated by 1, Various kinds of cllemicals are placed the action category in which its risk score and stored in apprnximately 100 refrigerfalls, In particular, the list would serve as ators, (This is no1ma1 practice.) guidance for stopping a job when a highly 2. Occasionally ff amlnabIe volatile solvhazardous situation is noted in an essential ents are placed in a non-sparkproof refrigoperation. If the risk score is above the erator, (This is a violation of safe pracupper critical line, the jot must be stopped tice; the hazard-event.) until some corrective action can be taken 3. A solvent container leaks or the cover to lower at least one of the three factors is not adequately tightened. to get the score into a less urgent category. 4, Fumes reach an electrical spark. 3. The list provides a means of setting 5. Fumes explode and cause $200 damgoals and objectives for the safety proage. gram, other than or in addition to, those For this situation the consequence rating is based on accident statistics. For example, set at 5. Since the safety procedure is frea safety program can be rated or safety quently violated the exposure rating is set at accolnplishlnent demonstrated by the 6. The probability of the accident sequence number of cases for which corrective acfollowing the violation and resulting in the tion ~ R been taken, resulting in the cases S accident is considered i.emotely possible. being placed in less urgent categories. A Therefore, P d . goal could be to have no hazardous situaThis produces the following calculation: tions above the lowest category, The R=Sx6x1=30 safety status of an organization can be Using the above procedures, the risk indicated by the number of items in each scores for an additional 22 hazardous situcategory at any given time. ations were calculated. In Table 2, these 26 With reference to Table 2, and a11 of cases appear in order according to the magthe examples used, the author hastens to nitude of their risk score. This is, in essence, point out th.at very few of the listed hazarda ranking of hazards in the order of the relaous situations presently exist at the Naval tive seriousness of their risks. The risk scores Ordnance Laboratory, the l o a l e where this are divided into three groups according to program originated. Most of the severe cases the urgency of action required. The positionwere selected for expediency from previous
162

1 1

Journal of Safety Research

TABLE 2
RISK SC O R E S U MM A RY A N D ACTION

smn
ACTION

nIYK SCOnE

1500

750
750
450

375
300

Immedinte correction required. Activity should be discontinued until llnzard is reduced.

270 200

180
150

l50
135
DO
~~

Urgent. Requires ttttention as soon as possible.

Compressed oxygen cylinder standing rnxectlred near wdl, little traffic or movement. Pedestrinn and hmd-ctrrt traffic t l t blind corner i n hilli~ay shop huilding. of Oxygen nnd itcet lene c linders stored to ether, cnps on, good ventiihon, ireproof surrounfings. Inadequate I~andrniling dong olltsicla stairway, occtlsionnl use every day. Large ropune storage tank in busy area: vel~icle tra&, and high prcssurc nir operations. Both pedestrian? and vehicles using snmc road. Road not nlwnys wick enough for Iwth. Chemicals stored in non-s arkproof refrigerators, 0cc:lsiondfy including f ammahle volatile liquids. Broken sidewalk, occnsionnl pedestrian trnffific, holes and loose concrete. Persons near explosives lmilding, within range of possi1)lc missiles; stlfc. procedures in huilding. Portn1,le vacwm pump lacking helt guard. Pmnp moved around occmionally hy several employees, Machinist using heavy file without file l~andle, in daily use, Worklnan using hnmrver with loose head, in use daily for odd johs.

85
60
45
40

37.5
37.5
30
Hazard should be eliminated without delt~y, situation but is not an emergency,

30
25

18
18

18

years experience, were hypothetiad, or combined experience and hypothesis, It is recommended that potential users of this system compile workable lists as soon as possible from their past experience, It is helpful to generate hypothetical (but reasonable) situations to be used RS guidance for comparative evnluntion of new hnzards t w they occur,
Dcccmber 1971/Volume 3/Nurnber 4

COST JUSTIFICATION

Once a hazard has been recognized, appropriate corrective action must be tentatively plnnned and its cost estimated. A slight expansion of the formula used to determine risk scores can be used to measure justificntion; i.e., to determine if the esti163

nuztetl cost is justified, The formula is: CXEXP IcCFxDC c , E , nlld P 1.rfcr to collseqllellces, esposurc, ;unci prohbility as before; CF ;m1 DC refer to cost f ~ c t nncl degree of correction. The ~ ~ r cost factar is a nwasure of the estimated dollur cost of the proposed corrective action. Fnctor 4 in Tr\ble 1 gives the values and their rcspcctivc? ratings. Tllc dcgree of correction is ill1 clstimnte of the clcgree to which tI-l(!proposcd corrective action will eliminate or allevinte tl-ta hazard, forestnll the hazardevent, or interrupt t h c ~nccident sequencc. The estitnate will be i>nsccl on cxpcrience nncl knowledge of the activity concerned. The cltlssificittiotl ancl the associated ratings nppear in Tablc 1 as Factor S. They rnnge from slight effect on the llazard to total elimination ( 100%). When the required vnltws are obtained, they nre placed in the formula to determine the nun1erical value of a countermeasures justification score. For any justification score of 10 or more, the expenditure will bc considered justifiecl, For a score less than 10, the cost of the contemplated corrective action is not justified. This critical justification scorey has been arbitrarily set at 10, hut is based on experience, judgment, ancl the current budgetary situation. Extended experience at a particular organization may dictate elevation or lowering of the critical score. Accident experience, local budgetnry situations, and n general appraisal of the organizations safety status would enter into such a change.

EXAMPLES OF JUSTIFICATION SCORES

The same examples used to illustrate the scores can serve here as calculation of ~ i s k well, Example no, 1 involved the hazard of pedestrians ancl vehicles using the same roadway. A possible corrective action to recluce this risk would be the construction of a sidewalk alongside the road, The estimated cost is $1,500, The I formula is then used to cletermine whether this expenditure is justified. The C, E , and P components are given in the earlier discussion as %,3, and 0.5, re,l64

spectivclly. Since the estimated cost is $1,500, CF=3, The degrec.1 of correction accon1plished by building n parallel sidewalk is judgecl to be at least 75% not 100%; but therefore, DC=2. (See Table 1, Factor 5.) The calculations tlwn are: 25 S 3 S 0.5 --:37.5 - 6.25 J= :3xz 6 Since J is lcss .than 10, the conclusion is that the cost is not justified. It is important to note that tho lack of suflicient justification c*vnlllatestho situation from the safety viewpoint an/!/. h4anagement might feel that thcrc is ildclecl justification for morale or other pruposes. Additionnlly, since the risk score is still a substantial 37.5, other less costly corrective measures should be sought, This might include improved administrative controls to enforce one-way traffic measures, Y ~ C ~ Lspeed, and encourage pedestrians t o I C ~ use another exit gate. This will reduce the risk score by reducing both exposure and prohbility. Examplc no. 2 involved the hazard due to compressed air heing used in a shop wjthout proper pressure reduction nozzles. The proposecl corrective action is installation of proper pressure reducing nozzles on the 50 air hoses, at a cost of $8.00 each, or $400. To determine justification for the expenditure, values for each element in the formula are required. The values of C, E , and P as discussed before are 5,10, and 6 respectively. The cost of the corrective action is $400, so CF=2. The corrective action will reduce the hazard by at least 50%, DC=3. Substiso tuting in the formula: 5x10~6 - 300 -J = 2 x 3 - 6 - 50 Since J is well above 10, the cost of installing pressure reduction nozzles is strongly justified, Example no. 3 was concerned with the cloubly hazardous location of t h e 12,000 galIon propane storage tank, The proposed corrective action is to relocate the tank to n place where it will be less likely to be damaged by any external source, at an estimated cost of $16,000, The values of C, E , and P are set at 2 5 1 , and 1.5respectively, w.ith the t w o hazards combined. Since the cost of relocntion is $16,000, CF=4. In even the best location available, there remained a remote

Journal of Safetv Research

TABLE 3 WORKSHEET FOR RISK AND COST JUSTIFICATION CALCULATION

h, Problem:

B. Locntion:

Hazard # Remedy #

c. Sequence of event! or factors necessary for accident:

1. 2,
3.

4, 5,

6.

7. D. Fonnuln Factors: C Consequence E Exposure P Prol~nbility CF Cost Factor DC Degree of Correction

Rating

E, Circle one: The estimated cost of corrective action:

is juSti6ed
is not justifled

F. Name

Date

possibility of damage to the tank, so DC=2. The J formula would then read: 25 x 1x 1 s --37,5

I=

4x2

- 4.7

I3ased on the established criteria, the cost

of relocation is not justified, I t is emphasized

that the conclusion reached does not mean that the hazard is of little or no significance. The risk score is still 37.5, and this remains of appreciable concern. Since the potential consequences of an accident are quite severe, effort should be expended to reduce the risk, by reducing either the eqosure or the probability, or devising another less costly corrective action. In this case, an ndditionnl steel plate bnlrier could be erected to protect the tank from t h e compressed air activities, and one or two strong posts in the ground could minimize danger from the road. Thus, the risk score and the probaDocember 1Q71/Volume 3/Numbar 4

bility of serious damage to the tank would be considerably lessened at nominal cost, Examnple no, 4 related to the hazard of non-sparkproof refrigerators being used for the storage of chemicals, The proposed correction is to place warning signs on all such refrigerators cautioning against their use for storing volatile solvents. The application of such decals along with administrative controls would probably reduce violations by 50 to 75%.Since the cost of warning decals for 100 boxes would be $87, the following calculation is appropriate: 5 x 6 ~ 1 30 - 10 J=
- I_-

1x3

Accorclhg to the criteria, the cost of the signs is justified. The case requires further review since the effectiveness of the corrective action is set at only 50 to 75%.Before installation of the warning signs, the risk
I65

as remotely possible, Two remotely posscore wi~s x 6 x l = 30.After installation, 5 sible occurrences might be regarded as the (sposurc fwtor is reduced considerably conccivnbly possible. Endeavor to be conso that h! ncw risk score ( 5 x 2 x 1) is equJ sistent, Consider otzhy the consequences to 10. This is it rciatively low risk, not of an chosen ttbove, If lesser or greater consecnwrgcncy wture, lmt ulso not to be coinquences we studied, additional computaplc tely ignored. Longer range solutions tions must be mnde since both the conseshould be consiclcred, such as insuring that quences and probability ratings will be only sparkproof ref]-igeratorsrue prchased i n thc future l u ~ tJli~t I when repairs are 11~td~: different. Given all the required values, the 011 box, it is idtcrecl to lwcome sparkproof, risk scores ancl justificntion scores can be For convellience in uncicrtal<ing11 hazard ga-terxted. analysis to rletcwnine risk i ~ n d justification When there are alternative corrective scores, a worksheet appenrs a s Table 3, measures that have acceptable justification In sltmnary, one must stnte the problem scores, the one that is most desirable from briefly and consider the most IikeIy consethe safety viewpoint is the one that produces quences of an accident rclated to 11 specific the greatest reduction in the risk score, Thus, Ilazarcl. Then one must review all fnctors for &c11 alternative, it is useful to assume carefully, preferably 011 the scene. After listthat the corrective measures are in effect and ing the actual step-by-step sequence of then recompute the risk score. Selection of evcants that wilI most likcly produce the concorrective actions may, of course, be influsequences chosen, a decision must be made enced b y nonsafety considerations such RS (IS to the most appropriate corrective action, the size of the investment required, the relaGiven this, an estimate of the cost of impletive effects on morale, aesthetics, efficiency, mentation and a juclglnent about the degree convenience, and ease of implementation, of correction that will be accomplishecl are One should not be a slave to the justificanzade, When alternative coL1lltLrllleasUres tion formula. A highly hazardous situation we possible, the cost factor and degree of may exist for which 110 corrective action hazard recluction steps are repeated. gives an acceptable justification score. In The hazard-event, RS definecl earlier, is such a case, whatever corrective action is specified ancl its frequency of occnrrence is necessary to reduce the risk should be taken, notecl. If there is doubt because a value falls regardless of the justification score, The between two ratings, select t~ proportional fornmln is for gzlicktnce only, and it pertains value in between; i.e., interpolate. To deonly to safety justification. Use of the fortermine the probability rating, consider the mula should result in significant improvelikelihood of each event in the accident ment in some of the methods used by the sequence. For examnple, if two unusual cosafety engineering profession, plus substanincidences are required, this coulcI be judged tiaI monetary swings.

Journalof Safety Research