Fueling system vapor recovery and containment performance monitor and method of operation thereof

Abstract

A method and apparatus for monitoring and determining fuel vapor recovery performance is disclosed. The dispensing of liquid fuel into a tank by a conventional gas pump nozzle naturally displaces a mixture of air and fuel ullage vapor in the tank. These displaced vapors may be recovered at the dispensing point nozzle by a vapor recovery system. A properly functioning vapor recovery system recovers approximately one unit volume of vapor for every unit volume of dispensed liquid fuel. The ratio of recovered vapor to dispensed fuel is termed the A/L ratio, which should ideally be approximately equal to one (1). The A/L ratio, and thus the proper functioning of the vapor recovery system, may be determined by measuring liquid fuel flow and return vapor flow (using a vapor flow sensor) on a nozzle-by-nozzle basis. The disclosed methods and apparatus provide for the determination of A/L ratios for individual nozzles using a reduced number of vapor flow sensors. The disclosed methods and apparatus also provide for the determination of fuel dispensing system vapor containment integrity, and the differentiation of true vapor recovery failures as opposed to false failures resulting from the refueling of vehicles provided with onboard vapor recovery systems.

Claims

1. A system for monitoring vapor containment in a liquid fuel dispensing facility having a fuel storage system that provides fuel for dispensing by at least one fuel dispensing unit, comprising: a pressure sensor operatively connected to the fuel storage system to measure pressure in the fuel storage system at a location within the fuel storage system comprised from the group consisting of a fuel storage tank vapor space, a fuel storage tank vent pipe, a pipeline system, and a vapor return pipeline; and a monitor operatively connected to the pressure sensor, wherein the monitor is adapted to monitor a pressure level in the fuel storage system, and decay of the pressure level, over a period of time, and determine if the fuel storage system has a vapor leak based upon the monitored pressure level and a model estimating a rate of change of pressure adjusted for ullage volume of the fuel storage system.

2. A method of monitoring vapor recovery in a liquid fuel dispensing facility having a plurality of fuel dispensing points connected to a main fuel storage system by a liquid fuel supply pipeline and a vapor return pipeline, said monitoring method comprising the steps of: determining at multiple times an amount of vapor flow through the vapor return pipeline; determining at multiple times an amount of liquid fuel dispensed through the liquid fuel supply pipeline; and for each of the plurality of fuel dispensing points, determining a plurality of ratios of vapor flow to dispensed liquid fuel for said fuel dispensing point based on amounts of vapor flow through the vapor return pipeline at the first determining step and amounts of liquid fuel dispensed through the liquid fuel supply pipeline at the second determining step; for each of the plurality of fuel dispensing points, determining a first proportion of the ratios determined for said fuel dispensing point at the third determining step that are below a first predetermined threshold; determining whether a predetermined statistically significant variability exists in the first proportions of ratios among the fuel dispensing points of the plurality of fuel dispensing points, wherein the statistically significant variability corresponds to existence of a vapor block condition in the vapor return pipeline; and further comprising a method of determining vapor recovery system failures associated with a single fuel dispensing point among the plurality of fuel dispensing points, said method of determining vapor recovery system failures comprising the steps of: for each of the plurality of fuel dispensing points, determining a plurality of the ratios of vapor flow to dispensed fuel; for each of the plurality of fuel dispensing points, determining a second proportion of vapor flow to dispensed fuel ratios that are below a second predetermined threshold; determining a third proportion of vapor flow to dispensed fuel ratios below the second predetermined threshold for the plurality of fuel dispensing points; and for each fuel dispensing point of the plurality of fuel dispensing points, determining existence of a block in vapor flow at the fuel dispensing point based on a comparison of the second proportion of vapor flow to dispensed fuel ratio of the fuel dispensing point to the third proportion of vapor flow to dispensed fuel ratios.

3. The method of claim 2, wherein the first predetermined threshold and the second predetermined threshold are the same.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention will be described in conjunction with the following drawings in which like reference numerals designate like elements and wherein:

(2) FIG. 1 is a schematic view of a fueling system vapor recovery performance monitor in accordance with an embodiment of the present invention.

(3) FIG. 2 is a schematic view of a fueling system vapor recovery performance monitor in accordance with another embodiment of the present invention.

(4) FIG. 3 is a graph used to convert vapor leakage rates based on ullage pressures.

DETAILED DESCRIPTION OF THE INVENTION

(5) A first embodiment of the invention is described in connection with FIG. 1, which shows a vapor recovery and containment monitoring system for use in a liquid fuel dispensing facility 10. The dispensing facility 10 may include a station house 100, one or more fuel dispenser units 200, a main fuel storage system 300, means for connecting the dispenser units to the main fuel storage system 400, and one or more vapor (or air) flow sensors (AFS's) 500.

(6) The station house 100 may include a central electronic control and diagnostic arrangement 110 that includes a dispenser controller 120, dispenser current loop interface wiring 130 connecting the dispenser controller 120 with the dispenser unit(s) 200, and a combined data acquisition system/in-station diagnostic monitor 140. The dispenser controller 120 may be electrically connected to the monitor 140 by a first wiring bus 122. The interface wiring 130 may be electrically connected to the monitor 140 by a second wiring bus 132. The monitor 140 may include standard computer storage and central processing capabilities, keyboard input device(s), and audio and visual output interfaces among other conventional features.

(7) The fuel dispenser units 200 may be provided in the form of conventional gas pumps. Each fuel dispenser unit 200 may include one or more fuel dispensing points typically defined by the nozzles 210. The fuel dispenser units 200 may include one coaxial vapor/liquid splitter 260, one vapor return passage 220, and one fuel supply passage 230 per nozzle 210. The vapor return passages 220 may be joined together before connecting with a common vapor return pipe 410. The units 200 may also include one liquid fuel dispensing meter 240 per nozzle 210. The liquid fuel dispensing meters 240 may provide dispensed liquid fuel amount information to the dispenser controller 120 via the liquid fuel dispensing meter interface 270 and interface wiring 130.

(8) The main fuel storage system 300 may include one or more main fuel storage tanks 310. It is appreciated that the storage tanks 310 may typically be provided underground, however, underground placement of the tank is not required for application of the invention. It is also appreciated that the storage tank 310 shown in FIGS. 1 and 2 may represent a grouping of multiple storage tanks tied together into a storage tank network. Each storage tank 310, or a grouping of storage tanks, may be connected to the atmosphere by a vent pipe 320. The vent pipe 320 may terminate in a pressure relief valve 330. A vapor processor 340 may be connected to the vent pipe 320 intermediate of the storage tank 310 and the pressure relief valve 330. A pressure sensor 350 may also be operatively connected to the vent pipe 320. Alternately, it may be connected directly to the storage tank 310 or the vapor return pipe 410 below or near to the dispenser 200 since the pressure is normally substantially the same at all these points in the vapor containment system. The storage tank 310 may also include an Automatic Tank Gauging System (ATGS) 360 used to provide information regarding the fuel level in the storage tank. The vapor processor 340, the pressure sensor 350, and the automatic tank gauging system 360 may be electrically connected to the monitor 140 by third, fourth, and fifth wiring busses 342, 352, and 362, respectively. The storage tank 310 may also include a fill pipe and fill tube 370 to provide a means to fill the tank with fuel and a submersible pump 380 to supply the dispensers 200 with fuel from the storage tank 310.

(9) The means for connecting the dispenser units and the main fuel storage system 400 may include one or more vapor return pipelines 410 and one or more fuel supply pipelines 420. The vapor return pipelines 410 and the fuel supply pipelines 420 are connected to the vapor return passages 220 and fuel supply passages 230, respectively, associated with multiple fuel dispensing points 210. As such, a vapor return pipeline designates any return pipeline that carries the return vapor of two or more vapor return passages 220.

(10) The AFS 500 is operatively connected to a vapor return pipeline 410. A basic premise of the system 10 is that it includes at most one AFS 500 (also referred to more broadly as vapor flow sensors) for each fuel dispenser unit 200. Thus, the AFS 500 must be operatively connected to the vapor return system downstream of the vapor return passages 220. If such were not the case, the system would include one AFS 500 per nozzle 210 which violates the basic premise of the invention. Each AFS 500 may be electrically connected to the monitor 140 by a sixth wiring bus 502.

(11) In order to determine the acceptability of the performance of vapor recovery in the facility 10, the ratio of vapor flow to dispensed liquid fuel is determined for each fuel dispensing point 210 included in the facility. This ratio may be used to determine if the fuel dispensing point 210 in question is in fact recovering an equal volume of vapor for each unit volume of liquid fuel dispensed by the dispensing point 210.

(12) In the embodiment of the invention shown in FIG. 1, each dispensing point 210 is served by an AFS 500 that is shared with at least one other dispensing point 210. Mathematical data processing (described below) is used to determine an approximation of the vapor flow associated with each dispensing point 210. The amount of fuel dispensed by each dispensing point 210 is known from the liquid fuel dispensing meter 240 associated with each dispensing unit. Amount of fuel (i.e. fuel volume) information may be transmitted from each dispensing meter 240 to the dispenser controller 120 for use by the monitor 140. In an alternative embodiment of the invention, the dispensing meters 240 may be directly connected to the monitor 140 to provide the amount of fuel information used to determine the A/L ratio for each dispensing point 210.

(13) Each AFS 500 measures multiple (at least two or more) dispensing point return vapor flows. In the embodiment of the invention shown in FIG. 1, a single AFS 500 measures all the dispensing point vapor flows for the facility 10. In the case of a single AFS per facility 10, the AFS is installed in the single common vapor return pipeline which runs between all the dispensers as a group, which are all tied together into a common dispenser manifold pipe, and all the main fuel storage tanks as a group, which are all tied together in a common tank manifold pipe. Various groupings of combinations of feed dispensing point air flow's per AFS are possible which fall between these two extremes described.

(14) With reference to a second embodiment of the invention shown in FIG. 2, it is appreciated that multiple AFS's 500 could be deployed to measure various groupings of dispensing point 210 vapor flows, down to a minimum of only two dispensing point vapor flows. The latter example may be realized by installing one AFS 500 in each dispenser housing 200, which typically contains two dispensing point's 210 (one dispensing point per dispenser side) or up to 6 dispensing points (hoses) in Multi-Product Dispensers (MPD's) (3 per side). The vapor flows piped through the vapor return passage 220 may be tied together to feed the single AFS 500 in the dispenser housing.

(15) As stated above, the monitor 140 may connect to the dispenser controller 120, directly to the current loop interface wiring 130 or directly to the liquid fuel dispensing meter 240 to access the liquid fuel flow volume readings. The monitor 140 may also be connected to each AFS 500 at the facility 10 so as to be supplied with vapor flow amount (i.e. vapor volume) information. The liquid fuel flow volume readings are individualized fuel volume amounts associated with each dispensing point 210. The vapor flow volume readings are aggregate amounts resulting from various groupings of dispensing point 210 vapor flows, which therefore require mathematical analysis to separate or identify the amounts attributable to the individual dispensing points 210. This analysis may be accomplished by the monitor 140 which may include processing means. Once the vapor flow information is determined for each dispensing point 210, the A/L ratios for each dispensing point may be determined and a pass/fail determination may be made for each dispensing point based on the magnitude of the ratio. It is known that the ratio may vary from 0 (bad) to around 1 (good), to a little greater than 1 (which, depending upon the facility 10 design, can be either good or bad), to much greater than 1 (typically bad). This ratio information may be provided to the facility operator via an audio signal and/or a visual signal through the monitor 140. The ratio information may also result in the automatic shut down of a dispensing point 210, or a recommendation for dispensing point shut down.

(16) The embodiments of the invention shown in FIGS. 1 and 2 may provide a significant improvement over known systems due to the replacement of the multiple AFS's 500 (one per dispensing point, typically anywhere from 10 or 12 up to 30 or more per site) and their associated wiring with a single, or fewer AFS's 500 (about as many or less, depending upon dispensing point groupings).

(17) With reference to the embodiments of the invention shown in both FIGS. 1 and 2, the mathematical analysis performed in the monitor 140 is designed to find correlations between aggregate vapor volume measured during AFS 500 busy periods and individual dispensing point 210 dispensed liquid fuel volume readings. The analysis is done separately for each AFS 500 and it's associated dispensing point group (two or more dispensing point's). The end result is a set of estimated dispensing point A/L ratios, one ratio per dispensing point. After a group of AFS 500 busy period data records are accumulated, a series of mathematical steps accomplish this beginning with a simple, 1-variable function solution and ending with more complex function solutions until all ratios are determined. If a ratio can be determined in an earlier step, it is not necessary to estimate it in a subsequent step (it can be set as a constant in later steps to simplify computation of any remaining unknown ratios). The sequence of solvable function types are:

(18) Type 1: A single linear function with one unknown for any AFS busy records with only 1 active dispensing point.

(19) Type 2: Two linear functions with two unknowns for any pair of similar AFS busy records with 2 (identical) active dispensing point's (two simultaneous equations with two unknowns).

(20) Type 3: Three or more linear functions each with two or more unknowns for any remaining (unsolved) set of AFS busy records (at least as many functions as unknowns).

(21) Each AFS 500 busy period data record is formed after the AFS becomes idle by recording the aggregate vapor volume, A, and the individual metered liquid volumes, L.sub.m, where the subscript, m, denotes the dispensing point or meter number. This number ranges from 1 to M total meters. Idle detection can be done by various means, including:

(22) 1) the monitor 140 can track reported dispenser meter 240 start/stop events from the dispenser controller 120, the dispenser current loop wiring 130, or directly from the liquid fuel dispensing meter 240; or

(23) 2) the Automatic Tank Gauging System 360 can provide main fuel storage tank 310 liquid fuel levels to the monitor 140 for detection of static level conditions (no ongoing dispensing) in all the storage tanks 310.

(24) The latter method (No. 2) can be used if it is desired that all AFS's 500 be idle prior to forming AFS busy data records. In the case of a single AFS 500 per facility 10 (shown in FIG. 1), this method can always be used.

(25) The simple form of the relationship between A, L, and the A/L ratio, R, for an AFS busy record with one (1) active dispensing point is:
A=L.sub.mR.sub.m

(26) so the simple solution for function type 1 is:
R.sub.m=A/L.sub.m
where R.sub.m is the estimated A/L ratio for active dispensing point (meter), m.

(27) In the more general case, each AFS busy period data record, n, has a measured aggregate vapor volume, A.sub.n, and the individual metered liquid fuel volumes, L.sub.nm, where the first subscript, n, denotes the data record number and the second subscript, m, denotes the dispensing point or meter number as before. The record number, n, ranges from 1 to N total records.

(28) The generalized form of the relationship between A.sub.n, L.sub.nm, and R.sub.m for multiple-dispensing point records is:
A.sub.n=L.sub.n1R.sub.1+L.sub.n2R.sub.2+L.sub.n3R.sub.3+ . . . L.sub.nmR.sub.m

(29) In the case of a pair of similar busy records with 2 active dispensing point's (same 2 dispensing point's in both records) the relationships are:
A.sub.1=L.sub.11R.sub.1+L.sub.12R.sub.2
A.sub.2=L.sub.21R.sub.1+L.sub.22R.sub.2

(30) so the solutions for functions of type 2 are:
R.sub.1=(A.sub.1L.sub.22A.sub.2L.sub.12)/(L.sub.11L.sub.22L.sub.12L.sub.21)
R.sub.2=(A.sub.2L.sub.11A.sub.1L.sub.21)/(L.sub.11L.sub.22L.sub.12L.sub.21)

(31) Functions of type 3 can be solved as a least squares problem using standard matrix arithmetic.

(32) Example record data set with subscript notation:

(33) TABLE-US-00001 n A.sub.n L.sub.n1 L.sub.n2 L.sub.n3 etc . . . L.sub.nM 1 18 0 12 6 etc . . . 0 2 33 10 15 0 etc . . . 8 3 21 7 0 0 etc . . . 14 etc . . . N 18 0 0 18 etc . . . 0

(34) For the entire data set, the matrix relationship is:

(35) $\left[\begin{array}{c}{A}_1\\ A_2\\ A_3\\ \mathrm{.Math.}\\ A_n\end{array}\right]=\left[\begin{array}{cccc}{L}_{11}& L_{12}& \mathrm{.Math.}& L_{1m}\\ L_{21}& L_{22}& \mathrm{.Math.}& L_{2m}\\ L_{31}& L_{32}& \mathrm{.Math.}& L_{3m}\\ \mathrm{.Math.}& \mathrm{.Math.}& \mathrm{.Math.}& \mathrm{.Math.}\\ L_{n1}& L_{n2}& \mathrm{.Math.}& L_{nm}\end{array}\right]\left[\begin{array}{c}R1\\ R2\\ \mathrm{.Math.}\\ R_m\end{array}\right]\mathrm{or}A=\mathrm{LR}$

(36) The solution for the ratio vector, R, is:
R=(L.sup.TL).sup.1L.sup.TA
where the first term is the inverse of the transposed nm matrix, L, times itself which results in an mm matrix, the middle term is the transposed matrix, L, which is an mn matrix, and the last term is the vector A of length n, all of which results in the vector R, of length m (one A/L ratio per meter).

(37) This approach can provide good estimates of the true A/L ratios, even with excessive variability (noise) in the sensor readings. More records result in better estimates for a given level of variability but there must be at least as many records as unknowns for minimal performance.

(38) Dispensing point ratio solutions are based on the simplest function type possible. As a data set is processed and ratio solutions are determined, they are in turn used to simplify solutions for remaining records in any record set. As an example, if two records exist in a set, one of type 1 (a single active dispensing point busy period), and a second with two active dispensing points, one of which is the same dispensing point as in the first record, the first record is solved directly as a type 1 function and it's ratio result is used to simplify the function for the second record. This produces a second type 1 function.

(39) Example Records (2):

(40) TABLE-US-00002 n A.sub.n L.sub.n1 L.sub.n2 1 5 10 2 19.5 12 15

(41) Initial Functions:
A.sub.1=L.sub.12R.sub.25=10R.sub.2
A.sub.2=L.sub.21R.sub.1+L.sub.22R.sub.219.5=12R.sub.1+15R.sub.2

(42) Solve first, substitute solution in second to simplify:
5=10R.sub.2R.sub.2=5/10=0.5
19.5=12R.sub.1+15R.sub.219.5+12R.sub.1+15*0.5=12R.sub.1+7.5

(43) Solve second as a type 1 function:
19.5=12R.sub.1+7.512=12R.sub.1R.sub.1=12/12=1.0

(44) This simplification method is used at each step of the data set solution process:

(45) Step 1: Form simple (1-dispensing point) or generalized function forms for each record.

(46) Step 2: Solve all Type 1 functions.

(47) Step 3: Substitute solutions from prior step into remaining set of functions.

(48) Step 4: Reduce all functions to simpler forms and repeat from step 2.

(49) Step 5: Find and solve any Type 2 function pairs.

(50) Step 6: Substitute solutions from prior step into remaining set of functions.

(51) Step 7: Reduce all functions to simpler forms and repeat from step 2.

(52) Step 8: If possible, solve remaining functions as a Type 3 least squares problem.

(53) Step 9: If step 8 is not possible, wait for more data records to solve the remaining functions.

(54) Alternatively, replace the 9-step sequence with steps 8 and 9 alone. This approach has the benefit of always averaging or reducing the effects of variability in the sensor readings.

(55) The various embodiments of the invention discussed herein may also be used to detect vapor recovery equipment failures. Stage-II vapor recovery equipment failures can have two distinct effects on patterns of A/L ratios. The failures are determined by identifying these patterns in the solved ratio set. The first type of failure involves a dispensing point nozzle 210, a hose 212, or vapor return passage 220 path restriction, or a vacuum assist pump failure which blocks or reduces air-vapor flow. The above solution methods may be used to identify this type of failure by identification of one dispensing point with a consistently lowered ratio.

(56) The second type of failure that can occur involves a dispensing point 210 with a defective air valve which does not close properly to block reverse vapor flow (i.e. out of the nozzle) when the dispensing point is idle. In such a case the ratio for the defective dispensing point will not be affected because when the dispensing point is active, the vapor flow is normal. However, when idle, vapors from other active dispensing points can be pushed past the defective air valve, out of the leaky dispensing point nozzle, and into the atmosphere. The active dispensing point(s) AFS 500 may or may not register the amount of lost vapor, depending upon whether the leaking dispensing point is part of the AFS group (won't register) or not (will register). If not, the idle AFS 500 will register reverse vapor flow. In that case, the leaking dispensing point can be detected by the reverse flow signal when it should be idle.

(57) Using the above solution methods described in connection with the first and second embodiments of the invention, when the leaking dispensing point is a member of the active AFS 500 group it results in lowered ratios for all dispensing points in the group except for the leaking dispensing point. Also, the lowered ratios vary depending upon the number of active dispensing point's during each busy period. When more (good) dispensing point's are active in an AFS 500 group, the lost vapor effect is shared in the solution, resulting in less depression of the individual ratios. Furthermore, if only part of the vapors escape to the atmosphere, the effect is reduced, resulting in less depression of the individual ratios. Accordingly, a post-solution analysis may be conducted on the ratio patterns to determine the likely failure type, active dispensing point restriction or idle dispensing point leak.

(58) A third embodiment of the invention concerns the use of a single vapor pressure sensor 350 (same as CARB requirement) to actively determine the tightness of the overall vapor containing elements of the facility including the fuel storage system 300, (which includes the vent pipe 320, pressure relief valve 330, etc.), the vapor return pipelines 410, the vapor/liquid splitter 260, the vapor return passages 220, the dispenser hose 212, the nozzle 210, etc. The vapor pressure sensor 350 may be connected anywhere in the fuel storage system 300 or the pipeline system 400, which includes but is not limited to the storage tank 310 vapor-space, the common vapor return pipeline 410, or the storage tank vent pipe 320. The vapor pressure sensor 350 may be used periodically to actively measure the leakage of vapors from the overall system instead of constantly measuring for leakage amount.

(59) The method in accordance with the third embodiment of the invention may be carried out as follows. The monitor 140 may be connected to and access pressure readings from the vapor pressure sensor 350. The monitor 140 controls the active test which is initiated by determining an idle period during which none of the dispensing units 200 are in operation (similar to the A/L detection method using either dispensing meter events or ATGS tank levels). The idle condition may be continuously monitored and the test aborted if any dispensing units go into operation during the test. During the idle period the vapor pressure sensor 350 is used to determine the pressure in the system (i.e. the pressure in the storage tank 310). If the pressure is not adequately negative (vacuum) for the test, the vapor processor 340 may be turned on to draw a negative pressure in the storage tank 310 as it processes vapors. If the vapor processor 340 is used, the monitor 140 may be used to monitor the vapor pressure readings until they become adequately negative, typically 2 or 3 inches w.c. Once the vapor pressure is adequately negative, the vapor processor 340 may be turned off. Thereafter the vapor pressure sensor 350 readings may be monitored during the remaining idle time. If the system is adequately tight, the negative pressure readings should hold or degrade only slowly. If the negative pressure degrades too rapidly toward zero, the monitor 140 may indicate that the system has failed the leakage test. A pass/fail threshold is used to make this determination. It can be set as a percentage of the initial negative pressure amount based on the desired detection sensitivity and should be related to the amount of air inflow detected relative to total storage tank 310 vapor space (ullage volume).

(60) In an alternative of the third embodiment of the invention, a single or multiple AFS's 500 located in the common or multiple vapor return pipeline(s) (same as A/L detection equipment) may be included to conduct an improved active test for system tightness. While a pressure sensor 350 alone suffices for conducting a tightness test, AFS 500 readings can add to the amount of information available to augment test sensitivity and confirm the tightness condition or help locate the source of a leak. Any air inflow from a leak point will register as flow on the AFS(s) 500. Flow and flow direction are a general indicator of the location of the source of incoming air (which dispensers and/or tanks/vents). Note that the AFS 500 readings are generally the more sensitive indicator of vapor recovery system tightness failure since negative pressure degradation is small due to the small amount of air inflow over seconds or minutes of time relative to the generally large storage tank vapor-space volumes. For significant negative pressure degradation, the amount of air inflow needs to be a significant portion of the storage tank vapor-space volume which can be in the thousands or tens of thousands of gallons.

(61) The optional AFS(s) 500, and dispenser controller 120, dispenser current loop 130, or optional ATGS 360 are connected to the monitor 140 which acquires and processes the data from the devices to conduct the tightness test and also controls (on/off) the vapor processor 340. Note that only one vapor pressure sensor 350 is needed for multiple storage tanks 310 as long as they share a common vapor recovery system so that their vapor spaces are connected (piped) together.

(62) In another alternative embodiment of the invention, the ATGS 360 may not be required to conduct an active test for system tightness. In this case, the idle state of the vapor recovery system during which the tightness test is conducted must be determined by (lack of) fueling meter 240 activity and a precise estimation of leak rate is not possible since tank 310 vapor ullage space volume is not known. Instead, a general pass/fail indication can be provided when the pressure decays at a preset rate during a test period.

(63) In yet another embodiment of the present invention, the systems shown in FIGS. 1 and 2 may be used to conduct an improved passive vapor containment test. This test uses pressure in the vapor containing elements of the facility, barometric pressure, and ullage space measurements to calculate the change in pressure over time for the vapor containing elements of the facility. This calculation, which is not usually based on data collected when the facility is operating at 2 to 3 inches w.c., may then be normalized to indicate leakage rates for a facility held at 2 to 3 inches w.c.

(64) This passive method may be initiated by monitoring the pressure of the main fuel storage system 300 or any vapor containing element of the facility 10 between fuel dispensing periods with the pressure sensor 350. Pressure data derived from sequential groupings of monitored pressures and ullage determinations derived from the ATGS 360 readings are recorded at periodic intervals by monitor 140. The derived recorded data permits the determination of rate of change of pressure, p.sub.rate, versus time, t, obtained from a linear regression model:
p=p.sub.rate.Math.t
within each interval, the main storage system 300 total ullage volume, V.sub.ullage represented by the sum of the individual storage tank 310 ullage volumes, V.sub.ullage:
V.sub.ullage=v.sub.ullage1+v.sub.ullage2+ . . . +v.sub.ullageN for tanks 1 to N
where
v.sub.ullage=(tank capacity)(volume of fuel in tank)
obtained from the ATGS 360, and the average pressure, p.sub.avg over each interval:
p.sub.avg=(p.sub.1+p.sub.2+ . . . +p.sub.N)/N for pressure samples 1 to N in the interval
are recorded if the correlation to the linear model is acceptable, generally based on high correlation between pressure with respect to time and the model.

(65) Upon collection of a daily sample of such records, the product of pressure rate and the total ullage volume, p.sub.rate.Math.V.sub.ullage, is sorted by the associated average pressure, p.sub.avg, and grouped into equally spaced average pressure ranges. A collection of averages of the products, (p.sub.rateV.sub.ullage).sub.avg, within each group:
(p.sub.rateV.sub.ullage)avg=((p.sub.rate.Math.V.sub.ullage).sub.1+(p.sub.rate.Math.V.sub.ullage).sub.2+ . . . +(p.sub.rate.Math.V.sub.ullage).sub.N)/N
for products 1 to N in each group. The midpoints of the average pressure ranges, p.sub.mid, within each group are used with a linear regression model to estimate the rate of change of pressure times ullage volume, P.sub.rateV at a selected test pressure, p.sub.test, of, say, 2 inches of water column, if the correlation to the linear model:
P.sub.rateV=(P.sub.rateV).sub.slope.Math.p.sub.test
is acceptable generally based on high correlation between the average products, (p.sub.rateV.sub.ullage).sub.avg, with respect to midpoint pressures, p.sub.mid, and the model. A typical graph of this model for a tank system is shown in FIG. 3. It is noted that the curve must cross the origin which indicates no rate of change of pressure, thus no leakage, when there is zero pressure drop across any leakage path, since for leakage to occur a pressure driving force is needed regardless of ullage volume.

(66) The regression yields the slope coefficient, (P.sub.rateV).sub.slope, which is used to calculate the estimated pressure times ullage volume, P.sub.rateV at a selected test pressure, p.sub.test, of, say, 2 inches of water column at which a leakage test failure rate can be defined, similar to the standard CARB TP-201.3 test procedure. In other words, if there is a leakage path and if the pressure in the ullage space of the tank system is set to 2 wcg (water column gauge) (above ambient pressure), the tank will leak at the estimated rate, v.sub.rate, of:
v.sub.rate=P.sub.rateV(p.sub.test)/p
where p is the absolute pressure in the tank ullage space, typically 410 wca (water column absolute) (assuming ambient is 408 wca). This can be interpreted to mean that the rate of volume vapor loss from a leaking tank is equal to the proportional rate of change of absolute pressure times the total ullage volume. Note that p.sub.test is a gauge pressure (referenced to ambient) and p is an absolute pressure (referenced to a vacuum). This relationship is derived from the ideal gas law, which governs the relationship between pressure, p, and volume, v, in an enclosed space at low pressures and temperatures:
p.Math.v=n.Math.R.Math.T
where n is moles of gas, R is the universal gas constant, and T is absolute temperature. Replacing n with mass per molecular weight (MW):
p.Math.v=m.Math.R.Math.T/MW
Rearranging terms and replacing constant terms with k:
m=k.Math.p where k=v.Math.MW/(R.Math.T)
Rate of mass loss due to a leak from an enclosed space is found by forming the relationship of the difference between the ending and starting mass divided by starting mass and the time period of the loss:
(m2m1)/m1.Math.t=(k.Math.p2k.Math.p1)/k.Math.p1.Math.t
m/(m1.Math.t)=(p2p1)/p1.Math.t
m/(m1.Math.t)=p/p1.Math.t
m/t=p.Math.m1/p1.Math.t
The last form describes the rate of mass loss as a function of starting mass times proportional pressure change rate over the test period. To find volume loss rate, relate mass and volume by mass density, p:
=m/v or m=.Math.v so m1=p1.Math.v and mass loss: m=.Math.v
Substituting in above equation:
.Math.v/t=p.Math.p1.Math.v/p1.Math.t
Assuming mass density does not change appreciably:
v/t=p.Math.v/p1.Math.t where 1
Dropping the subscript and using notation for volume loss rate, v.sub.rate:
v.sub.rate=p.Math.v/p.Math.t
which can be interpreted to mean that the volume loss rate is the proportional change of pressure times volume per unit time. But part of this expression is the calculated value derived from measurements in the above section:
v.sub.rate=P.sub.rateV/p where v.sub.rate=p.Math.v/t at the selected test pressure, 2 wcg

(67) Using the above example, the volume leak rate, v.sub.rate, is:
v.sub.rate=P.sub.rateV/p=6000/410=14.6 CFH or cubic feet per hour at 2 wcg

(68) As described above, in yet another embodiment of the invention, the system may also perform a method of distinguishing between true vapor recovery failure events and ORVR equipped vehicle refueling events. Identifying a false vapor recovery system failure due to refueling an ORVR-equipped vehicle may be accomplished by applying standard statistical concepts to a group of dispensing or refueling events from all the dispensing points 210 at a dispensing facility 10 to identify true failed vapor collection dispensing points as opposed to failed tests due to ORVR vapor-blocking activity.

(69) There are two assumptions that may be made as a predicate to determining true failed vapor collection: (1) that ORVR and non-ORVR activity occurs somewhat randomly amongst all the dispensing points; and (2) that average ORVR activity does not reach 100% of all refueling events (a maximum of 80% can be assumed). Given these assumptions, a group of vapor collection event A/L measurements taken from all the dispensing points 210 at a dispensing facility 10 may be used to make the following determinations:

(70) 1) Determine if the proportions of failed (close to zero A/L) and non-failed events are statistically different at individual dispensing points relative to their expected proportions, due to the activity of ORVR vehicles, derived from all the dispensing points; and

(71) 2) Determine if the proportion of the failed (close to zero) events at each dispensing point are statistically different from the proportion of the failed events derived from all the dispensing points, which are largely due to the effect of ORVR vehicles.

(72) As a result of these determinations, the A/L ratio measurements may be used to test for blockage or leakage caused vapor recovery failure, with a mix of ORVR and non-ORVR vehicle activity.

(73) On a regular (e.g. daily) basis, each dispensing point 210 may have a number of A/L determinations associated with it. It is presumed that there are k dispensing points 210 and the i.sup.th dispensing point has n.sub.i A/L ratio determinations. Let X.sub.i be the number of A/L determinations for dispensing point I that indicate a zero or blocked A/L ratio. The assumption is that fueling an ORVR vehicle will result in a zero or blocked A/L ratio. The total number of A/L determinations for the site is:
n=n.sub.i
and the total number of zero A/L ratios is:
X=X.sub.i

(74) An overall test can be conducted to determine whether there are any significant differences in the proportion of A/L ratios indicating blocked vapor flow among the dispensing points 210. This can be accomplished using a chi-squared test on the table of data from the k dispensers:

(75) TABLE-US-00003 Dispenser 1 Dispenser 2 Dispenser k Total Number X1 X2 . . . Xk X blocked Not blocked n1 X1 n2 X2 . . . nk Xk N X Number n1 n2 . . . nk N
The chi-squared statistic is given
by: X.sup.2=(O.sub.iE.sub.i).sup.2/E.sub.i
where O.sub.i is the number observed in each cell of the table and E.sub.i is the expected number in that cell. The data in the cells indicate the number of A/L ratios that indicate a blocked condition for each dispensing point and the number of A/L ratios indicating a not blocked condition for that dispenser. The expected number blocked ratios for dispenser I is:
E.sub.i1=n.sub.i(X/N)
and the expected number of not blocked ratios for dispenser I is:
n.sub.iE.sub.i

(76) The summation is carried out over 2 k cells. This statistic is compared to the critical value from a chi-squared table with k1 degrees of freedom. If it is significant, there is evidence that the dispensers have different proportions of blocked A/L ratios, so that one or more would appear to be blocked on at least an intermittent basis.

(77) In turn, an individual test can be performed for each dispenser. This tests whether each dispenser has a proportion of zero A/L ratios that exceeds the overall proportion for the station. The following equation may be used to compute the overall proportion of zero A/L ratios for the period:
P=X/N
The following equation may be used to compute the proportion of zero A/L ratios for each dispenser:
p.sub.i=x.sub.i/n.sub.i
From the foregoing calculations, it may be concluded that there is evidence that dispenser I is blocked if:
p.sub.i>P+z.sub.(0.16/n.sub.i).sup.1/2
where z.sub. is the upper percentage point from a standard normal distribution. If a 1% significance level is desired, z.sub. is 2.326, for example, (or 1.645 for a 5% significance level). The number 0.16 in the formula results from assumption of the most conservative case; that 80% of the vehicles are ORVR vehicles. Once a truly blocked dispensing point is detected, an audio or visual signal may be provided by the monitor 140 to indicate this condition. Truly blocked dispensing points may also be automatically shut down as a result of such detection.

(78) It will be apparent to those skilled in the art that various modifications and variations may be made in the preparation and configuration of the present invention without departing from the scope and spirit of the present invention. For example, various combinations of the methods described above may be implemented without implementing the full system shown FIGS. 1 and/or 2. Thus, it is intended that the present invention cover the modifications and variations of the invention.