CA2173794C - Apparatus and method for testing for valve leaks by differential signature methods - Google Patents

Abstract

A portable computing device is used for acoustic testing of valve systems for valve leaks. Transducers are applied to appropriate locations in the valve system to obtain sound signals, which are fast Fourier transformed into valve signatures. Multiple signatures are displayed simultaneously and are compared automatically by the system to derive a recommended result and manually to make determinations of valve leaks by the differential signature method or the like. The signatures and the user's determinations are stored on a hard drive. Various panels are displayed to give the user step-by-step instructions for performing the test. Thus, training is simplified, use of paper is vastly reduced or eliminated entirely, and the results can be ported to the client's office computers.

Description

APPARATUS AND METHOD FOR TESTING FOR VALVE LEAKS
BY DIFFERENTIAL SIGNATURE METHOD
FIELD OF THE INVENTION
This invention relates to acoustic detection of leaks in valves carrying fluid and more particularly to an apparatus and method for applying hardware and software to automate the process of acoustic detection of leaks in such valves.

Systems for carrying gases or liquids typically cont'vn numerous valves which are subject to developing leaks. One method of detecting leaks in valves of such systems, the differential signature method, is described in the following documents Dimmick et al, "Ultrasonic leak detection cuts valve maintenance costs," Power Engineering, August, 1986;
AVLA"' Acoustic Ikrlve Leak Analyzer Operator's Manual, ~ 1989 Leak Detection Services, Inc.; and AVLA'"Acoustic I~blve Leak Analyzer Test Engineer's Guide, ~ 1989 Leak Detection Services, Inc.
The differential signature method works by measuring acoustic signatures defined as the amplitude of ultrasonic signals emanating from a test point as a function of frequency or of time foc a given frequency. Typically, three acoustic signatures of amplitude as a function of frequency are taken, one at the valve and two (called the background signatures) at positions on a pipe upstream and downstream from the valve. The background signatures are taken upstream and downstream of the valve to prevent errors due to inaccurate human measurements and omissions of vital data. The differences between the signature at the valve and the background signatures indicate the presence or absence of a leak and its specific severity. For a non-leaking valve, these differences should be small.
The ratios of the signature at the valve and the upstream and downstream signatures in decibels at the dominant frequency, which is the frequency at which the amplitudes have peak values, indicate the severity of the leak. The hardware for this method has typically included a device for taking the acoustic signatures and plotting such signatures on graph paper with pens of different colors. The user has then had to manually determine the presence and severity of the leak on the basis of the plotted signatures.

C~~ 1~7:~ 1~~4 The conventional method described above has the following disadvantages.
First, it results in a massive amount of paper. Second, as the hardware performs no analysis of the data, and as neither the hardware nor the plots indicate the procedures for determining the presence and severity of a leak, the conventional method requires extensive training of the user, many manhours, and countless paper plots.
SUMMARY OF TIII= INVrNTION
It is an object of the invention to provide an apparatus and method for automating the process of acoustic detection of leaks in valves carrying fluids.
1t is a further object of the invention to provide such an apparatus and method while eliminating the need to print out hard copies of the acoustic signatures.
It is a further object of the invention to provide such an apparatus and method that require little training for proper use.
It is a further object of the invention to provide such an apparatus and method that offer the advantages of elimination of bulky paper records, faster data collection, faster and automated repeatability confirmation, easier data manipulation and analysis, integration of planning, recording, analyzing, reporting, and record-keeping, ease of upgrading and modification by simple software changes, ease of tailoring the system to a specific client's needs, and reduction of training and skill requirements for data collection.
It is a further object of the invention to provide.such an apparatus and method that offer the advantages of digital collection, analysis and conclusion recording;
a data collection unit and accessories including a software-controlled, hardware FFT module to convert transducer signals to spectra for electronic transmission to a computer; a data collection and in-plant analysis software module; the ability to use an office PC to develop and maintain databases of clients' valve systems; the ability to use a generic office PC to develop standardized test procedures, to permit more detailed comparison and analysis of the signatures, and to prepare reports with a minimum of human intervention to maintain historical records of individual valves in clients' plants; and to enable survey clients to view test data and conclusions, but not to change them.
It is a further object of the invention to provide a computer based valve signature recorder having a programmed function key interface which guides the user through the planning, testing and analysis of a valve leakage survey, to provide an accurate and organized apparatus for managing multi-series data on five hundred or more valves, and to provide a method of analyzing valves with such an apparatus, which method automatically handles many details of instrument operation which could fmstrate the non-electronics oriented operator.

~1 737 9~
To achieve these and other objects, the invention includes an apparatus comprising transducer means for feceiving sounds from the valve system and for converting the sounds into electrical signals; transform means, such as fast Fourier transform means, for receiving the electrical signals, for transforming the electrical signals by a transform such as a fast Fourier transform to produce sound signatures, and for outputting digital data representing the sound signatures; means for comparing at least two of the sound signatures to derive a recommended result regarding the valve leaks; and digital storage means for storing the digital data representing the sound signatures and a conclusion or judgment derived from the data and the io recommended result. The conclusion includes the user's judgement on whether there is a leak, the economic need for repair, and the quantitative severity of the leak if any. The invention also includes a method comprising the steps of attaching a plurality of transducers to the valve system; operating the plurality of transducers to receive sounds from the valve system and to convert the sounds into electrical signals; fast Fourier or otherwise transforming the electrical signals to produce sound signatures and outputting digital data representing the sound signatures;
comparing at least two of the sound signatures to derive a recommended result regarding the valve leaks; and storing the digital data representing the sound signatures and the conclusion or judgement (as defined above) in a digital storage medium.
In a broad aspect, then, the present invention relates to an apparatus for performing a 2 o process of testing a valve system for a valve leak, the apparatus comprising: transducer means for receiving sounds from the valve system and for converting the sounds into electrical signals;
transform means for receiving the electrical signals, for computing a transform of the electrical signals to produce at least two sound signatures, and for outputting digital data representing the at least two sound signatures; and computation means for (i) receiving said digital data and 2 5 comparing said at least two sound signatures to obtain a comparison result, and (ii) automatically deriving a recommended result regarding whether said valve leak exists from the comparison result.
The comparison result may be a ratio of amplitudes of said at least two of the sound signatures at a selected frequency which is taken from the group consisting of a manually 3 o selected frequency and an automatically selected frequency.
Moreover, the comparison result may be a ratio of foot mean squares of weighted amplitudes of said at least two of the sound signatures, the root mean squares being calculated over a predetermined frequency range.
Furthermore, comparison result may be a root mean square, over a predetermined 3.°~ frequency range, of weighted ratios of amplitudes of said at least two of the sound signatures.
In the apparatus of the present invention, the amplitudes of the sound signatures at a dominant frequency may be summed to derive a sum; and the comparison result for each valve may be a ratio of an amplitude of that valve's sound signature at the dominant frequency to the sum.

3(a) The apparatus of the present invention may further comprise digital storage means for storing the digital data representing the sound signatures and a conclusion derived from the recommended result and said at least two sound signatures.
The apparatus of the present invention may further comprise display means for displaying said at least two sound signatures and said recommended result. Such display means may also display a difference calculated from said at least two sound signatures.
In the apparatus of the present invention the computation means may comprise an interface means for receiving a user's input for at least one of planning, testing, analysis and reporting of 1 o a test survey. Such interface means may comprise a touch screen.
The apparatus of the present invention may further comprise connection means for outputting the digital data representing the sound signatures and the conclusion to an external computer. Such connection means may comprise a port for a floppy drive.
The apparatus of the present invention may further comprise digital storage means for i5 storing an information database and a group of test structures for said process of testing.
The computation means may comprise control and display means for guiding the user in understanding the process of testing. The computation means may also comprise selection means for controlling the interface means to enable the user to make choices required for the process of testing. Such selection means may enable the user to select a parameter and may control the 2 o transducer means and the transform means so that the sound signatures are produced in accordance with the parameter selected by the user. Such parameter may be frequency range. Such parameter may be bandwidth. Such parameter may be transform type. Such parameter may be averaging type. The selection means may further control the apparatus to collect the sound signatures in accordance with the choices made by the user. The interface means may display one or more of 2 5 the sound signatures simultaneously and the user may use the interface means to indicate a portion of one of the one or more of the sound signatures; the interface means may generate a touch signal and outputs the touch signal to the control means, which changes a scaling factor by which the one or more sound signatures are displayed on the interface means in accordance with the touch signal.
The interface means may have a plurality of manners of actuation by the user, and may comprise 3 0 help means for displaying, when the user performs a predetermined one of the plurality of manners of actuation, a context-sensitive descriptive message. The apparatus may also comprise digital storage means for storing a plurality of instructions for operation of the computation means. In this case, the transform means may comprise storage means for storing a plurality of instructions for operation of the transform means.
3 5 In another broad aspect, the present invention relates to a method for testing a valve system for valve leaks, the method comprising: (a) positioning at least one transducer on the valve system to receive at least two different sounds from a valve in the valve system and to convert the sounds into electrical signals; (b) transforming the electrical signals to produce sound signatures and outputting digital data representing the sound signatures; (c) comparing at least two of the sound signatures, based on said digital data, to obtain a comparison result; and (d) deriving a 3(b) ~ ~ ~ 3 ~ 9 ~
recommended result regarding valve integrity from the comparison result and displaying the recommended result to a user.
In the method of the present invention, the comparison result may be a ratio of amplitudes of said at least two of the sound signatures at a selected frequency which is taken from the group consisting of a manually selected frequency and an automatically selected frequency. . -.~-- ~-, In the method of the present invention the comparison result may be a ratio of root mean squares of weighted amplitudes of said at least two of the sound signatures, the root mean io squares being calculated over a predetermined frequency range.
In the method of the present invention the comparison result may be a root mean square, over a predetermined frequency range, of weighted ratios. of amplitudes of said at least two of the sound signatures.
In the method of the present invention, amplitudes of the sound signatures at a dominant ~.5 frequency may be summed to derive a sum; and the comparison result for each valve rnay be a ratio of an amplitude of that valve's sound signature at the dominant frequency to the sum.
The method of the present invention may further comprise storing the digital data representing the sound signatures and a conclusion derived from the sound signatures and the recommended result in the digital storage medium.
2 o The method of the present invention may further comprise (e) outputting the digital data representing the sound signatures and the conclusion to an external computer.
Step (e) may comprise writing the digital data representing the sound signatures and the conclusion onto a machine readable medium.
The method of the present invention may further comprise a step of storing an 2 5 information database and a set of plans which are standard test structures or modified test structures derived from the standard test structures, said plans also including automated analysis structures. The method of the present invention may also include step (e) comprising (i) compressing the digital data representing the information database, plans, sound signatures and conclusion into a compressed file and (ii) writing the compressed file onto a machine readable 3 0 medium. In the method of the present invention valves in the valve system may be grouped into items and given item numbers, each of the item numbers being identified by a primary (group) number and a sub number, a primary (base) item having a system-specific function and all other items having a function supporting the primary item. Each set of items grouped under a primary item may be assigned to one of said plans. Moreover, all items may be compiled into an 3 5 ordered database having information categories including items, plans and identification information. Furthermore, the method of the present invention may comprise repeatedly searching the ordered database according to any of the information categories for selecting an order of group testing and separating results of such searching by test completion. The method may further comprise a step of displaying instructions which guide the user to comprehend and 4 o conduct the method for testing. The method may also comprise a step of enabling the user to t r j'?~ c.

3(~) '~ ~~7 9~
make choices required for the method of testing. The step of enabling may comprise enabling the user to select a parameter, the sound signatures being produced in accordance with the parameter selected by the user. The parameter may be frequency range. The parameter may be bandwidth. The parameter may be transform type. The parameter may be averaging type.
In the method of the present invention the sound signatures may be collected in accordance with the choices made by the user. When one or more of the sound signatures are displayed simultaneously and the user indicates a portion of one of the one or more of the sound signatures, a scaling factor by which the one or more sound signatures are displayed may be 1o changed. The method may further comprise a step of displaying, when the user makes one of the choices in a predetermined manner, a context-sensitive descriptive message. In the method of the present invention, the step of displaying instructions may comprise:
(i) prompting the user to secure the valve system in a proper configuration; (ii) in a case of positive user response, proceeding with the method of testing; and (iii) in a case of negative user response, displaying x5 a warning message; thereby promoting user safety and precluding invalid signatures. Such method may further comprise prompting the user to perform an operation selected from the group consisting of: (i) placing the transducers on a portion of the valve system for testing; (ii) reseating the transducers; (iii) saving a signature when the signature is determined to be good;
and (iv) correcting a situation which causes an invalid signature.
Furthermore, the method of 2 o the present invention may also comprise selecting a highest signature of two or three contiguous signatures; marking the highest signature with a RMS difference repeatability test result; and determining whether the RMS difference repeatability test result conforms to a predetermined repeatability criterion; and if the RMS difference repeatability test result does not conform to the predetermined repeatability criterion, the step of displaying instructions may comprise 2 5 instructing the user to reseat the at least one transducer. The method of the present invention may further comprise: (i) prompting the user to remove and reseat the transducers and to repeat signature collection; and (ii) calculating a RMS difference repeatability for verification of proper transducer placement. The method may further comprise directing the user into test and analysis completion of an item group in a proper order. In such a case, the method further comprises 3 o permitting the user to advance to a position in the plan selected by the user while maintaining full test completion prompting. In the method of the present invention, step (c) may comprise:
(i) displaying two or more signatures; (ii) receiving a user input representing a selection of a first signature and a second signature from the two or more signatures displayed and a location;
and (iii) determining a difference between the first and second signatures at the location selected 3 5 by the user; and (iv) displaying the difference obtained in step (c)(iii).
Step (c) may further comprise determining the recommended result based on the difference and displaying the recommended result. The method may also comprise temporarily storing one or more recommended results until the user accepts them or enters conclusions different from the recommended results. Step (c)(i) may comprise automatically selecting pairs of signatures for 4 o display, comparison and result recommendation.
n 3(d) In the method of the present invention step (d) may comprise correcting the sound signatures for sound attenuation, said sound attenuation being a function of frequency and distance.
In the method of the present invention step (a) may comprise automatically prompting the user (i) to ensure that predetermined valves in the valve system are open or closed, (ii) to indicate that the predetermined valves have been opened or closed as prompted, (iii) to attach the transducers to predetermined locations on the valve system, and (iv) to indicate that the transducers have been attached to the predetermined locations.
i o The sounds are not received until the user has indicated in step (a) that the predetermined valves have been opened or closed as prompted and that the transducers have been attached to the predetermined locations.
In such a method, the user is automatically prompted to perform steps (a) and (b) a plurality of rimes for a plurality of configurations of opened and closed valves in accordance ~.5 with a sequence stored in the automated interface device; step (b) may comprise fast Fourier transforming the electrical signals to produce the sound signatures and displaying the sound signatures to the user; and step (c) may comprise automatically calculating a ratio of said at least two of the signatures and automatically determining, in accordance with a stored rule, the comparison result in accordance with the ratio; and step (c) may further comprise prompting the 2 o user to input the conclusion on the basis of the sound signatures displayed in step (c) and the recommended result. Step (c) may comprise receiving a user input indicating which of the sound signatures are to be compared and deriving the ratio from the sound signatures selected by the user.
In the method of the present invention, step (a) may include positioning the transducer 2 5 at a plurality of positions on the valve system, or positioning the at least one transducer in one or a plurality of positions on the valve system and controlling the valve system to provide a plurality of differential pressures across the valve system.
In the apparatus of the present invention, the at least two sound signatures may comprise sound data for at least one frequency greater than or equal to 100 kHz.
3 o In the method of the present invention, the sound signatures may comprise sound data for at least one frequency greater than or equal to 100 kHz.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will now be described more fully with respect to the drawings, in which Fig. 1 shows a block diagram of the apparatus according to the invention;
35 Fig. 2 shows an OK window in the user interface of the apparatus according to the invention;
Fig. 3 shows a YFS/NO window in the user interface of the apparatus according to the invention;
f.

3(e) Fig. 4 shows a multiple/choice window in the user interface of the apparatus according to the invention;
Fig. 5 shows a numeric window in the user interface of the apparatus according to the invention;
Fig. 6 shows an alphanumeric window in the user interface of the apparatus according to the invention;
Fig. 7 shows a CONTROL PANEL in the user interface of the apparatus according to the invention;

LPL I ~~.~~~1~

Fig. 8 shows a LIS'C PANEL in the user interface of the apparatus according to the invention;
Fig. 9 shows a MAKE PANEL in the user interface of the apparatus according to the invention;
Fig. 10 slows a SIGNATURE PANEL in the user interface of the apparatus according to the invention;
Fig. 11 shows a REVIEW PANEL in the user interface of the apparatus according to the invention;
Fig. 12 shows an INFO PANEL in the user interface of the apparatus according to the invention;
Fig. 13 shows an ANALYZE PANEL in the user interface of the apparatus according to the invention;
Fig. 14 slows a CONCLUDE PANEL in the user interface of the apparatus according to the invention;
Fig. 15 shows a tree structure for the use of the various panels in the user interface of the apparatus according to tl~e invention;
Fig. 16 shows the first part of a flow chart of tire initiali~ltion of the apparatus and the operation of the CONTROL PANEL;
Fig. 17 shows a flow chart of the "PICK ITEM" operation;
Fig. 18 shows the second part of the flow chart .of the initialization of the apparatus and the operation of the CONTROL PANEL;
Fig. 19 shows a flow chart of the "LOAD ALL" operation;
Pig. 20 shows a flow chart of the operation of the LIST PANEL;
Figs. 21 and 22 show a flow chart of the operation of the MAKE PANEL;
Figs 23, 24A, 2413, 25 and 26 show a flow chart of the operation of the SIGNATURE PANEL;
Figs. 27 and 28 slow a flow chart of the operation of the REVIEW PANEL;
Fig. 29 shows a flow chart of the operation of the INTO PANEL;
Pig. 30 shows a flow chart of the SAVE RESULTS operation within the INFO
PANCL;
Figs. 31-33 show a flow chart of the operation of the ANALYZE PANEL;
Fig. 34 shows a flow chart of tire operntion of the CONCLUDE PANEL; and Figs. 35-37 show a flow chart of the operation of a digital signal processor used in the invention.

L~a~ i ~ ~ ~~
DETA1LCD DrSCRIp I'ION Or 'fIlr PRrrrRRPD CM)30D1MrNT
A presently working system according to the invention will now be described with respect to the block diagram of dig. 1. Apparatus 100 includes two transducers, channel A
transducer 102 and channel B transducer 104, for taking sound readings from the valves.

5 Signals from these two transducers go into, for example, fast Fourier transformer (rFT) module 106 or the like, where they are fast Fourier analyzed to yield acoustic signatures.
CPU 108 controls storage of these sound signatures on hard drive or other storage device 112 and the display thereof, as well as receipt of user input, on touch screen or other user interface device 110. The apparatus can exchange data with other devices through port 114 for a floppy drive or the like. These components as provided in the presently working embodiment will tae described in further detail below.
The presently working system takes the form of a portable, battery-operated laptop personal computer (PC) with data collection software, a hardware FFT module, spare batteries, transducers and otlrcr accessories, all packed in a rugged case Ibr in-plant use.
Separate, off-line battery chargers and a miniaturized printer can be provided.
The working system measures 12.5 x 10 x 2.5 inches and weighs 10 pounds. A 12 volt, 2.3 amp hour, lead acid gel cell battery with quick exchange access powers the apparatus. A small rocker on/off switch on the lower right-hand side of the system constitutes the only external control on the system.
The display is a 6 x 8 inch backlit liquid crystal display (LCD) with video graphics array (VGA) resolution (640 x 480 dots). A transparent, pressure sensitive panel covers the full LCD. Moderate finger pressure registers a touch and release with its horizontal and vertical position. Thus, the touch of a certain firnction displayed on the LCD
replaces a dedicated push-button on the system. Other constructed embodiments of the invention have included a conventional monitor and a mouse rather than a touch screen. Those skilled in the art who have reviewed this specification will understand which user interface devices are useful for which purposes.
'Iwo BNC connectors on tl~e upper right-band side of the system provide Channel A
and Channel B transducer inputs. The Channel A input, toward the top, is used for most applications. Standard 100 kohm input impedance amplifiers are optimized for use of the transducers on almost any valve signature. Automatic gain controls, invisible to the operator, replace the external attenuation or gain switch critical to other instruments.
The Channel A and Channel B transducer inputs go to the hardware FFT module 106.
This module includes an integrated circuit for effecting the fast Fourier transform of the C~~ i ~ ~~~~'~

signals received through the transducer inputs under control of code stored in an EEPROM
or the like.
The operating program and test data are stored on an internal 80 Mbyte hard drive.
Five connectors on the left hand side of the system allow the system to be used as a 38GSX DOS-compatible computer. The serial port (COM1), external CRT, and keyboard connectors match most available peripherals. The floppy-disk connector is not standard, but is designed to be used with a 3'h" drive powered by the internal battery. A
backup drive unit may be used in place of the floppy-disk drive; such backup drive units are commercially available and will not be described in detail here.
The transducers are e.g. accelerometers specially selected for their response to acoustic emission signals in the 10-200 kHz frequency range. Commercially available matched pairs of two types of transducers are preferably used. When used with the differential signature method, the transducers should be used in the factory-matched pairs.
The preferred transducer is the most sensitive, especially at frequencies near 25 kHz, and can be used without a standoff at temperatures ranging from -100°F
to +350°F. 1t loses sensitivity if the cable length exceeds 3 meters. A second type of usable transducer is about 20 dD less sensitive and can be used to record signals that would overrange the apparatus if recorded with the preferred transducer. The second type of transducer has a temperature range which is -100°F to +350°I: It is most sensitive at frequencies near or above 50 kHz and can be used at cable lengths up to 19 meters without loss of signal strength.
The choice of transducers depends on the situation encountered, but once chosen, the same transducers, cable, and standoff must be used throughout a test.
Otherwise, the results of various measurements cannot be compared to determine whether a valve leaks.
The transducer standoff is an 8-inch long by 3/e-inch diameter steel rod with a ~h-inch diameter ball end. The other end is threaded, allowing the transducer to be attached by the 10-32 mounting stud. The ball end mates with the spherical indentation in the disk. A small amount of coupling compound should be applied at the transducer-rod interface and the ball-disk interface to assure a good acoustical contact.
The transducer standoff may be used as a hand-Meld probing tool to move quickly from one valve location to another or as an insulator to keep the heat of a hot valve or pipe from damaging the transducer.
If acoustic measurements on steam valves and, piping or other high-temperature structures are desired, the transducer standoff must be used to keep the transducer cool. The transducer must not be allowed to exceed its temperature limits, or it will become permanently damaged or destroyed. A small amount of coupling compound is required at r C~~1.7~194 all metal-to-metal interfaces. No excessive pressure is required if the surfaces are free of dirt and mated properly. The transducer end of the standoff should be held while the measurement is made.
The transducer cables are shielded, high-temperature coaxial cables with BNC
S connectors at one end and microdot connectors at the other end. Normally, one 15-ft cable is supplied with each transducer. Other lengths are available.
The software program resides on the root directory of the hard drive mentioned above. Thus, it can be upgraded as needed with no need to change a ROM chip or the like.
1'he external floppy drive referred to above can be used for this purpose.
l0 In oaer to develop a clear comprehension of the complex analyzer program, the building blocks will first be described, and then a step-by-step program operation will be described. In this description, the valves are grouped as a plant, a unit, an item, a sub-item, a plan or a series. The unit is the highest group and typically represents several days' work.
An item is a particular assemblage of valves being tested and Is designated by a whole 15 number. Decimal numbers indicate sub-items (particular valves or the like).
A plan is a test structure defined in terms of a previously prepared assemblage of valves or a modification of a previously prepared assemblage of valves; a plan may represent a single valve or several. Fewer than a hundred plans can be used to cover almost all actual situations. A
series is a set of signatures for all valves in an item. In a plan, any assemblage of valves 20 can be defined as a system (e.g., preheat system, main steam system, auxiliary steam system, or lobe oil system). Parts of a system can be further defined as subsystems.
1'he display and touch screen form a flexible function-key interface for directing all operator-instrument interaction. Every function key is an independently named and located touch zone on the display screen. In most cases, each function key is identified by a word 25 or symbol enclosed in a finger-sized rectangle. The exceptions are selection and display changes, which are located directly on a signature. Any time an activated function key is touched, the rectangle will be highlighted until the function key is released.
If the finger pressure is removed while within that rectangle, the appropriate command or character is recognized by the program. Then, the display changes to reflect that selection. Typing the 30 first character of any function key on the external keyboard will also produce the same response. A mouse or other pointing device could also be used to actuate the function keys, the working environment permitting.
Function keys can produce tlu~ee types of responses on the display. Some trigger an action which results in a visible change to their own display. Others produce a pop-up 35 window within the display requesting a single key, numeric or alphanumeric response. The LA~I~~1~4 g third type actually changes the whole display to a new panel. This panel will have a different set of function keys on it.
Pop-up decision windows, as shown in Figs. 2-6, are used whenever the program requires the operator to make an individual selection. There are five types of selection windows: the OK window of Fig. 2, the YESINO window of Fig. 3, the multiple-choice window of Fig. 4, the numeric window of Fig. 5, and the alphanumeric window of Fig. 6.
The OK window announces a particular situation and waits for the user to acknowledge it.
The YES/NO window asks a particular question and waits for the user to press "YES" or "NO". The multiple-choice window presents a key for each of the allowed entries. After one of the entries is pressed, the window is removed. The numeric window asks the user to type in a number selection. Of the various keys available in this window, "
< " removes a digit, "RET" sends tire number to the program while removing the window, and "ESC"
cancels any entry and removes the window. A default selection is shown in parentheses.
The alphanumeric window requests a name or comment. Of the various keys available in this window, " < " removes a character, "RET" sends the character line to the program while removing the window, and "ESC" cancels any entry and removes the window.
If the entry has already been started when this window appears, the previously entered part is displayed and can be extended or corrected.
To simplify organization and execution of a valve leakage survey, the program breaks down the task into manageable subtasks as much as possible. The instrument presents on the display only functions necessary to the task at hand. This task-related group of functions on the display is called a task panel. The various task panels will be described with respect to Figs. 7-14, which show the panels themselves, and Fig. 15, which shows a tree structure for the use of the panels.
There are three levels of panels. Movement between panels in the tree structure of Pig. IS is vertical. Operation always starts at CONTROL PANEL 700 of Fig. 7 at the top and proceeds down and then back up. A complete valve leakage survey entails preparation, valve testing and result reporting. This procession shows up as the left to right ordering of the panels: "LIST," "MAKE PLAN," "SIGNATURES," and "REVIEW," shown in Figs.
8-11 as 800-1100, respectively.
Program operation begins automatically when the power switch is turned on.
Initialization takes about two seconds. The unit, item and plan are loaded from a status file so that the user will have access to those settings last used. Operation then proceeds from the top, namely, the control panel.
The CONTROL PANEL has two main purposes. First, it defines the operating environment of the test at hand. Second, it provides function keys 702-710 to direct the user C~~ ~ r'? 1~~
into one of the four second-level display panels, namely, "LIST," "MAKE PLAN,"
"SIGNATURES," and "REVIEW," or to the "QUIT" command to quit analyzer operation, respectively.
All operating data are filed under the 1UNIT directory. The use of this directory S Billows survey work to proceed on two or more independent groups of valves without clearing and reloading for each. Valve grouping cari be any level desired, such as plant, unit, system or subsystem as those skilled in the art will appreciate.
The unit key 712 is used to select the unit name, which is used as the directory on the hard drive to work with; in the DOS directory tree, this directory is located directly under \UNIT. This entry is rarely changed. One unit's work usually covers many days.
Pressing this key lists the units already established. 'The previously active unit is highlighted. The user touches the unit name desired. Item loading will follow, and the control panel will return, displaying the results of the item load.
The correct unit should be verified at the top of the control panel every time. An incorrect selection can waste much work by using the wrong list and plans and mixing up signatures with those of a different unit.
The item key 714 pops up a numeric window to select the next primary item. The default next item (in the list) is shown in parentheses. After a number is selected, the series is reset to zero. If the item is being selected for the very first time, the primary item subdirectory will be created. Then tl~e loading sequence is executed.
The plan key 716 lists the plans available within the particular unit. Most selection of plans should be done during building of the valve list and subsequently designing any new plans. The previously active plan is highlighted. The user touches the plan desired. 1f a change of plan is accepted, the program proceeds to reload the primary item signatures. In the illustrative example given in the figure, plan 20, which has five valves, is shown in piping diagram 701. 1n table 724 in the lower right, items .0 through .4 are shown as closed by X, open by O, or traps by T in each of steps 1 through 4. In this notation, parentheses indicate that a valve is under pressure, while slashes indicate that a valve is not under pressure, and d indicates that a differential signature (to be described below) is needed for that valve.
The series key 718 pops up the numeric window to accept entry of a series number between 0 and 15. The IBighest series which holds signatures is shown. The user enters a number between 0 and 15. If this number,is more than 1 higher than the highest series used, it will be reduced to prevent skipping a series. If a valid change of series is accepted, the program proceeds to reload primary item signatures. On the series key, "series" is spelled "xeries" so that the user can select either the series key or the signature key by typing the C ~ ~ ~i 7 .:~ ~ ~ 4 first letter. Of course, other conventions could be used, such as the underlined letters well known to users of graphical user interfaces. As shown, the current series is series 0. Series 1 has a separate set of signatures and may involve a different test, e.g., after repairs rather than before repairs.
S Only signatures saved under the selected series will show up as saved. The user should not mistake an incorrect series for a missing signature situation.
The frequency key 720 selects a frequency range, wherein 200 klz is the standard frequency range used for valve leakage signatures. A touch of this key immediately changes the frequency to the next selection of those available, 1001200/400/800 khz.
Under each 10 setting, any signature displayed and saved covers from 10% to 100% of the frequency range.
The frequency is automatically set as signatures are loaded. The frequency cannot be changed after at least one signature is saved in a series. rrequency may be changed by advancing to the next series.
The average key 722 is used to select the average variable. There are two purposes of the average variable. The standard amplitude vs frequency signature display uses average as the length of time in seconds that the signature is averaged each time GO
is pressed.
Average = 0 means that only a single reading is used for the signature.
Touching this key causes the value to alternate between 0, 1 and 2. The second purpose of this variable is for setting tloe length of time that is displayed in a time sweep signature (amplitude vs time).
The list, make, signature, and review keys 702-708 switch, respectively, to the list, make, signature, and review panel analyzer operations, which will be explained below. 'The use of the signature key 706 presupposes the selected primary item, plan, series, frequency range, and average.
The quit key 710 terminates program operation. The analyzer returns to the DOS
' prompt and can then be turned off by the power switch.
Any time the item, plan or series is changed, the program automatically takes the following load sequence: it (1) searcljes for the primary item subdirectory on the disk and (2) loads any signatures already saved that fit the selected plan and series.
Saved signatures are marked by a * in the plan.
Signature loading may lead to one of the following error messages:
"ERROR: Plan 1 used. Restore it?" If all signatures are saved under a different plan, that plan can be restored by a "yes" to this question.
"Warning: Multiple Plans Used." If some signatures were saved under a different plan, this note will pop up.
"Signatures Don't Match." If the plan has been modified and some signatures saved do not match the entry in the plan, this note will show.

"Latest Sigs 'Taken in Series I ." IF there are signatures saved of higher series, the highest series taken will be noted to help prevent mixing new signatures with the old series.
The list panel 800 is a means to interrogate the valve list and display items meeting any selected set of criteria. The valve list is an ordered database having information csitegories (fields) including items, plans and identification information, and is organized as a hierarchical and relational data structure.' First the operator places an entry in any number of the fields of the search definition line 802. When the desired criteria have been entered, the operator presses the search key 804 to initiate the search and display the results. Finally, tire operator selects the item desired for collection and analysis by touching the item. It will be displayed at the top of the screen. The operator presses the TEST IT key 806 to return to the control panel with the selected item.
All the fields except two use an alphanumeric text search. The search word can be any combination of letters, numbers, and certain puncrtiation characters, but no spaces. 'Che allowed length is the size of the field displayed. A match requires that the complete search , word be found anywhere within the corresponding field of a particular item.
Only the ilem and elevation fields use a numeric upper/lower limit search. The operator must enter both an upper limit and a lower limit to search by these fields. A match requires that the corresponding field entry for a particular item fall between the upper and lower limits.
'lv enter a search criterion in a particular field, the operator touches that field in the search define area (lines 2-5). Touch the defined key 805 and the alphanumeric or numeric window will pop up. The operator then enters the word (or number) desired and presses RLTURN. RETURN with no entry removes any search criteria from that field. The ESC
key 816 ignores any change of entry.
One example of a broad search is an entry in just the elevation field to assist in survey ordering by all valves located together.
1'he search key initiates a search of the valve list. Cach item in the list is taken field by field and is rejected if any field search word is not found within the corresponding field of that item. Only if all the fields match the search criteria is the item displayed in the found list.
A broad search might return many items in the found list. These will be separated into two sections with a dotted line between them. The items above the dotted line are not yet tested while those below the clotted line have already registered a conclusion. Therefore, as the list panel is repeatedly used in the course of a survey with constant search criteria, items will be selected from the top selection to go through collection and analysis and will then show up on the bottom section.

'' V 7 v~ ~l 7 ~~ ,: ~J ~+

To select an item, the user touches the item number. The user examines the highlighted item line to verify the selected item as desired. If not, the user simply touches another item or redefines the search to produce a different found list.
After a particular item has been identified (highlighted), the TEST IT key will take that item number and return to the CONTROL PANEL just as if the operator had pressed the ITEM function and typed in that number. It will be ready to go to SIGNATURES.
' The ESC key returns to the CONTROL PANEL with no change of item.
The up and down arrow keys move the cursor from the search definition area through the found list. The left and right arrows move the cursor horizontally only within the search definition area.
1'he MAKr PANCL 900 provides all the tools necessary to prepare a plan to test any primary item and its associated sub items. The analyzer does not place any constraints on the plan design other than a maximum of 15 valves and 15 steps (rows and columns). The number of sub items in the plan cannot be decreased to fewer than the number showing under List. The analyzer automatically places the optimum lineup entry, X
(closed) or 0 (open), to the left of each signature entry, (X) or /0/ or IXI or Xd. In this notation, parentheses indicate that a valve is under pressure, while stashes indicate that a valve is not under pressure, and the d indicates that a differential signature is taken.
"T" is a time signature (infr~r). Some keys have a dual purpose in preparing the plan in accordance with whether valves or piping is selected first and highlighted. The differences in purpose will be explained in greater detail below.
The valves key 902 are pressed to enable the following firnctions of the MAKE
PANEL. They include all operations on the plan and on valves in the layout drawing. One valve in the drawing will be highliglUed.
7b place an entry into the plan table at the position of the plan cursor, the user presses one of the entry keys 904, namely, X, O, (X), lOI, IXI, Xd and T. If the cursor is on a sub item or a step, there is no effect. The various valves in the plan are indicated by decimal numerals, with .0 being the most important valve.
Pressing the MOVE key 906 allows the highlighted valve in the layout drawing to be moved. The arrow keys are next pressed to move the valve to the desired location; when done, the user presses the ESC key 908. The other keys do not function while MOVE is in process (MOVE key highlibhted).
' If the plan cursor is on a sub item, and there are fewer than IS sub items in the plan, a new sub item row will be opened in the plan when the INSERT key 910 is pressed. If a sub item row was previously deleted, it will automatically fill in this inserted row. If the r ~ ~,~~ ~ ~s ~y4 cursor is on a step number and here are fewer than 15 steps, a new step column will be opened up.
1f the plan cursor is on a sub item when the DELETE key 912 is pressed, that whole sub item row will be deleted, and the rows below it will move up. An INS can be used to restore an incorrectly deleted sub item row. If the cursor is on a step number, that step column will be deleted and any columns to the right of it will move over.
The four arrow keys 914 move the plan cursor around within the plan. If the cursor is on step number at the top of the plan, moving right beyond the last step will add another step up to 15 max. If the cursor is on a sub item number at the left of the plan, moving down below the last sub item will add another sub item to a maximum of 15. If the cursor is on a sub item number at the left of the plan, pressing the left arrow will bring up the alphanumeric window requesting entry of the plan valve description. It is noted that the plan will scroll horizontally and vertically to accommodate the limited display size, and also that the highliglned valve in the layout drawing will track the sub item that the plan cursor is on The piping key 916 enables the MOVE, INS and DEL functions to assume their secondary purposes of operating on pipe sections in the layout drawing. One pipe, if any, in the drawing will be highlighted. These secondary purposes will now be described.
Pressing the MOVE key allows the highlighted pipe of the layout drawing to be moved. Next, the ARROW keys are pressed to move the pipe to the desired location, and finally, the user presses ESC when done. No other keys.function while MOVE is in process (MOVE key highlighted).
After pressing the INSEItf key to insert a pipe, an ARROW key must be pressed to determine the pipe orientation, horizontal or vertical. The highlighted pipe will appear to the right of the drawing. Next the ARROW keys are used to place it as in MOVE, and finally, the user presses ESC when done.
When the delete key is pressed, the I~igl~liglUed pipe, if any, in the drawing will be deleted.
Any ARROW key causes the highlighting to move among the pipe elements in order to select the pipe section which the user desires to move or delete.
When the plan design is complete, the SAVE key 918 is pressed to save the plan on the disk drive. The display first lists the plans available within the present unit. Then it pops up the alphanumeric window to accept a name entry. Name entries can be any unique combination of up to four characters (frequently the names are just orderly numbers to ease record keeping). If the name selected is already in use, the YesINo window asks for verification to overwrite chat plan. After the plan is saved, the analyzer leaves the MAKE
PANEL and returns to the CONTROL PANEL.

~~~~ i ~.5~y~

Actuating the CSC key causes the analyzer to leave the MAKE PANEL and return to the CONTROL PANCL. If changes have been made, a YES/NO window requests confirmation to restore the plan as last saved, and then return to CONTROL.
The purpose of the plan is to provide the organized stricture for testing and analyzing a particular group of valves. Most valve groups will fit into a standard pre-tested and optimized plan thereby reserving the MAKC PANEL for preparation for an odd group or for restructuring testing under restricted access conditions.
The remaining panels are used in valve testing and analysis. Their use will now be described.
One of the most important features of the analyzer is the semi-automated signature collection process. The SIGNAfURI: 1'ANI?L 1000, shown in l~ig. 10, leads the operator tlu-ough each pressure and background signature recommended in the plan denoted by valve diagram 1001 and W ble 1003. This leading follows the cycle of collecting all signatures in one step and then moving right in the plan table to the next step. A step column in the plan table represents all items (valves) residing in the indicated condition, as well as the type of signature to be collected at each item. Witlr reference to rigure 9, X =
closed valve; O =
open valve; (X) is closed valve pressure amplitude versus frequency signature;
/O/is open valve background amplitude versus frequency signature; /X/ is closed valve background amplitude versus frequency signature; Xd is combined closed valve pressure plus two background amplitude versus frequency signatures; (T) is closed pressure amplitude versus time signature; and ITI is either closed or open background amplitude versus time signature.
Moving to a different step in tire plan dictates changing the condition (and thus pressure by opening or closing one or more valves in a plan) of one or more valves. The automated collection process thereby gathers pressure and the background signatures on all valves required using a minimum of valve openings and closings. Normally it starts with the plan cursor highliglUing the first valve and first step. As shown in step 1, valve .0 is closed, and valves .2, .3 and .4 are open. 1f they are not already~in this state the operator must open or close them as needed to satisfy the condition indicated. Then a time signature is taken by hitting GO and then SAVC. The rest of the steps proceed similarly.
If beginner operation is selected, a valve alignment procedure is displayed on startup and each time the cursor moves between steps. A YES/NO window announces each valve the operator should set up and its position. If all the announcements For the present step are answered "YCS," the analyzer proceeds as normal. If any are answered "NO", the setup announcements for that step are stopped, but a warning will issue on each GO.
The instrument advances the cursor to the first signature in the current step and then issues the message to place the transducer on the indicated valve and to hit GO.

~~~'1 73 X94 Tlre user hits GO and the instrument responds with the message to hit GO again until repeatability (as described below) is satisfied. When it is satisfied, the instrument prompts flue user to hit SAV)r, then the user bits SAVC. The instnrment advances the cursor to the next signature in that step, and the above process is repeated for every signature in the step.
5 The instnrment advances the curser to the next step and returns to the valve alignment procedure described above.
One signature can be collected for each (X), 101, IXI and (T) and ITI element in the plan. A differential element, Xd, requires three signatures. To collect these, whenever the plan cursor is on an Xd, an extra group of keys 1002 shows up just above the plan table.
10 This group of keys takes the form shown for the single transducer differential analysis by default. In this case, the user will be prompted to position the transducer and hit GO and SAVC for each of the three locations, as described above. If the user wishes to perform a two-transducer differential analysis, he bits the MODC key, and then keys for the two-transducer differential analysis appear. In this case tire prompting asks the user to place two 15 transducers at appropriate locations rather than just one, so that two signatures can be taken simultaneously.
The Go-Save routine is completed for each location by either manually touching valve, or one of the three keys 1002 or letting the analyzer automatically advance through them. The saved * in the plan table represents the valve signature of an Xd set while the up and down signatures saved are marked by the * on the corresponding keys.
Signatures displayed on the SIGNATURC PANEL are scaled by the ATTCN
(attenuation) selling. To shrink down the display, the user may touch anywhere in the lower half of the signature box. 7b enlarge the display details, the user touches the display anywhere in the upper half of the box. The ATTCN changes in steps of 5 dD from -10 to 60 dD. A change of 5 dB eduals a magnification of 1.78, while 60 dB indicates a magnification of 1,000. The absolute signal amplitudes are determined by adding the A1'TCN value to the approximated dB nupber of floe peak as read off the left side of the signature. 1n the signature panel shown in Fig. 10, llre AI"1'CN setting is 35 dB.
The MODC key 1004 cycles through the available modes for a certain plan element.
Mode changes the function of the BW key 1006, whose use will now be described.
Under modes A, B, and A-B, the MODC key switches the display smoothing bandwidth between 1% and 3% of the frequency range. In time sweep mode, the display ~shpws amplitude vs time. This amplitude is an average over a fixed frequency band having a width of 10% of the frequency range. When the BW key is pressed, the display switches G~~ ~ l ~~ ~~~-to amplitude vs frequency. The transducer is placed on the trap (a type of valve, as those skilled in the art will appreciate) and the display is touched anywhere to read a signature.
Next, the time sweep center frequency is chosen by touching the signature at an appropriate area of high signal. The dark bsrr under the signature indicates the frequency band to be averaged. Tlre Esc key 1008 is pressed.when the band selected is correct. The display will revert back to time sweep mode with the selected center frequency displayed under the BW
key.
The CLR key 1010 removes any signature started by GO but not SAVED, allowing the GO procedure to be restarted. It is used whenever the initial GO on a valve produces a non-repeatable signature due to a bad application of the transducer to the valve.
The GO key 1012 starts G0, which is the central function of the whole analyzer.
The most time is spent here actually testing valves. Any time the plan cursor highlights a signature type, namely, one of (X), lOI, IXI, Xd, and 'f, one of these messages is displayed under the CLR and GO keys and GO can be pressed.
"Place transducer, on item 1.0 and press GO." When GO is pressed, the analyzer acquires the signature and displays it unless one of these messages appears:
"Not settled;
retake" or "CHECK: Resent transducer on Item 1.0 and press GO."
"Good repeatability; press Save." All signature elements in the plan except for T
(time sweep) employ a repeatability test requesting more than one GO press.
The initial GO
at a particular plan element displays as a solid line. Subsequent GO readings show as a dotted line and are used only for repeatability checking. If it passes, a SAVE
is requested.
If it does not pass, the check message is repeated. If the initial reading was not good (cannot be repeated), CLR must be pressed to retake an initial.
GO on a time sweep starts a slow time trace of amplitude (see DW). The length depends on the average set in the control panel. Tirne sweep signatures do not request a repeatability check because of their time dependent nature. Save is requested when done.
If the present element already had a saved signature, indicated as "*" in the plan, it will be noted as "SAVED" on the signature, and will display as a solid line until GO
acduires a new signature.
Each signature must be manually saved to keep it for further analysis by pressing SAVE after a GO has produced a repeatable signature. The solid line (initial GO reading) on the display will be stored on the disk drive under the proper file name.
The save function may produce one of the following messages:
"Not stable, save anyway?" If the GO function could not produce a repeatable signature, a SAVE can be carried out with a Y)=S to this question.

.~ ~ ~I ~ u~ ~' 4 "This series has data saved, Do you want to overwrite?" If a previous signature was saved at this Sub Item and Step, this YF.SINO window will request confirmation to overwrite. "Do you want to save to the next series?" If it is not desired to overwrite (lose the previous signature), then a YES answer to this window will automatically advance to the next series. The plan wilt reflect the saved status of all signatures in the new series. The series number is printed just above the plan.
1f there is a signature already saved at the present sub item and step, the ERASE key 1014 will ask for a YES/NO confirmation and then erase the signature from the disk drive.
This is used if the user does not plan to overwrite an improperly taken signature.
SAVE also automatically advances the cursor through the plan to the next uncompleted signahire. First it checks for the next signature to complete the present step.
If no more are needed the cursor moves to the next step, follows the valve alignment prompting procedure and continues to check that step as above. 1f all signatures are completed, the message is display. "Plan Completed".
The INFO key 1016 switches the analyzer down to the INFO PANEL.
The ANALYZE key 1018 key switches the analyzer down to the ANALYZE
PANEL. This is commonly used after the plan is filled in with as many signatures as are reduired to obtain a recommended result and/or to draw a conclusion. The ANALYZE
function does not have to be used immediately after taking the signatures on the items in one plan. A primary item can be reelected in the control panel and this function accessed at any time.
The ESC key 1008 moves the analyzer up to the CONTROL PANEL.
The LEFT/RIGIIT ARROW keys 1020 above the plan move the plan cursor between steps of the plan. If the signatures elements in that step have not been saved, a warning message will appear. If beginner mode is enabled, the alignment check is stepped through each time.
The UP/DOWN keys 1022 below the plan move the plan cursor between the valves (sub items) of the plan, The valve layout drawing reflects both the position (valve) that the plan cursor rests on and the state of each valve at the cursored step in the plan. Closed valves (i.e. X, (X), IXI or Xd elements) are filled in the drawing. Open valves (i.e. O or l0l elements) are not filled in.
The REVIEW PAN)rI~ 1100 of Fig. 11 provides an overview of survey completion atul outputs to feed the work to other databases.
The initial panel lists all primary item numbers having any signatures taken.
If any results have been recorded, the date of the last will be displayed next to the item number.

The review item key 1102 pops up the numeric window to enter an item number.
If RETURN is pressed, the display switches back to CONTROL PANEL and then to signatures using the entered item. A more direct way to do this is simply touch the item number on the screen.
The LIST ALL KEY 1104 is pressed to see a full listing of completed signatures, results and comments. Those shown are only tl:e first series. First, on the left is the item number followed by columns of plan elements, each representing one completed signature.
These elements are displayed as in the plan table except for uncompleted and non-signature elements not shown. Next, the extra series column is marked if signatures have been taken in additional series. That is followed by the results and then the comments that have been entered.
When the output key 1106 key is pressed, all results and comments are compiled into a file named RESULTS.CMP. 'this data can be directly imported into PARADOX for reporting.
When the ZIP key llU8 is pressed, all work for the unit is zipped into one file for backup purposes. This includes VALVE 1.1ST, PLANS, SIGNATURES, RESUL'1~ and INFORMATION FILIrS. 1'he file name is <Unit>.EXE. This file is a self extracting archive, as those familiar with archived files will know. After backing up with the ZIP key, the operator can QUIT from operation of the program and then use the DOS COPY
to copy this file to a 3.5 inch floppy disk. 'The use of this size .floppy disk, of course, is given as an illustrntive example; those skilled in the art wlro have reviewed this specification will understand what other machine readable media (e.g., computer readable media) can be used and under what circumstances.
Restoration onto another computer is done by creating the proper directory for the UNIT, copying the zipped file to it, and typing the UNIT name, i.e.:
<Unit> Id to where < Unit > is the name of the UNIT which the survey covers. For example, if the unit name is LAB2, the self-extracting archive will be named LAB2.EXE, and the command to restore the data will be LAB2 Id lo.
The ESC key 1110 returns the display to the CONTROL PANEL.
Fig. 12 shows the INFO PANCL 1200. As can be seen from this figure, the INFO
PANEL allows display and input of information about each valve.
Fig. 13 shows the ANALYZr I'AN)L 100. By comparing the signatures, the user can determine the existence and severity of leaks by the differential signature method described above. It is also possible to automate (described below) determination of the ~A~ i %S X94 dominant frequency and thus to automate signature comparison to obtain a recommended result, which the user can then accept or reject as a conclusion.
Pig. 14 shows the CONCLUDC PANRL l4tX). 1n this panel, column 1402 identifies each item. Column 1404 identifies the result for each item, where L = leaking, L2 =
probably leaking, T = tight, T2 = probably tight, NA = not applicable, and NT
= not tested. It is emphasized that the CONCI.UDC PANIrL is a visual representation of the analysis performed by the device of tire invention. This analysis is deemed a recommended result. An operator may draw a conclusion from the displayed result and accept the result as is or override the result. Whatever decision is made by the operator, his decision is conclusive but, of course, derived after evaluating the recommended result.
The operator's conclusion is entered through multiple-choice window 1410. Column 1406 gives the estimated size of each leak, small, medium, or large. Column 1408 allows a free-form comment for each valve including the user's conclusion.
Now that the user interface has been described, the procedures followed by the program which controls the apparatus will be detailed with reference to the flow charts set forth in Pig. 16-34. Some of these procedures can be implemented in off the-shelf business software, such as PARADOX for data management, GXCEL or QUATTRO PRO for numerical analysis, and WORD for report preparation.
The operation of the apparatus from switching on the power to selecting one of the lbur panels directly underneath tire CONTROL PANEL is shown in the flow chart split between rigs. 16 and 18. When the apparatus is powered on (step 1602), the status file is loaded (step 1604), and the item is picked in the PICK ITrM operation (step 1608; to be described in fiuUher detail below with reference to rig. 17). The user can then pick a unit number (steps 1610-1616), in which case the item number defaults to 1 (step 1618), or can pick an item number (steps 1620-1624), in which case the series number defaults to 0 (step 1G2G), either way, the PICK I'fLM opersrtion begins again (step 1628). The user can also pick a plan (steps 1630-1636) or a series (steps 163.8-1644). Zither way, the LOAD ALL
operation begins (step 1646; to be described in further detail below with reference to rig.
19). Once the PICK ITCM or LOAD ALL operation finishes, the user is returned to the control panel. The user can also pick a frequency (steps 1802-1810) or an average (steps 1812-1820). The user may also press QUIT (step 1822), in which case the program is exited (step 1824). Once any of these procedures except QUIT is done, the user may go into LIST
(steps 1826 and 1828), MAKC (steps 1830 and 1832), SIGNATURC (steps 1834 and 1836), or R>JVICW (steps 1838 and 1840).
The PICK ITCM operation will now be described with reference to Fig. 17. When this operation is called (step 1702), it is checked whether the item number designated is in C~~ ~l .~r~:l the list (step 1704). If not, the previous item is kept (step 1714), and control is returned to the operation that called the PICK ITEM operation (step 1716). If the item number designated is in the list, the item is set (step 1706), the valve information is loaded (step 1708), the plan per list is set (step 1710), and the LOAD ALL operation is commenced (step 5 1712). Once the LOAD ALL operation is finished, control is returned to the operation that called the PICK ITEM operation.
The LOAD ALL operation will now be described with reference to Fig. 19. Once this operation is called (step 1902), for all subiu~::~s in the plan, ttre valve conclusions and external conditions are loaded (step 1904), and the lunitliteml directory is searched (step IO 1906). )each file name is checked to see whether it is for a signature file, and the highest signature series is held (steps 1908-1912). If the file is not the correct series, and there are more files, the search continues (steps 1914 and 1936), and if the series selected is too high, the series is reset (steps 1938 and 1940). If the file is the correct series, the file is read to see whether the signature should he loaded into the plan or whether there is a signature error IS (steps 1916-1934), and steps 1936, 1938, and (if needed) 1940 are rerun. If the series selected is not too high, the plan is checked (steps 144-1960), and control is returned to the operation that called the LOAD ALL operation (step 1962). There is also a provision to load a plan (step 1942) if no plan is selected or if a different plan was used.
The operation of the LIST PANEL will be described with reference to Fig. 20.
20 When this panel is called (step 2002), it is displayed (step 2004), the last search definition is displayed (step 2006), and the working item is displayed (step 2008). Now, the user may define a search (steps 2010-2022), perform a search (steps 2024-2036), select the working item (step 2038), and then press either DO-IT to use the selected item (steps 2040 and 2042) or ESC to keep the previous item (steps 2044 and 2046). Then the user is returned to the control panel (step 2048).
The operation of the MAKE PANEI. will be described with reference to Figs. 21 and 22. When this panel is callec! (step 2102), it is displayed (step 2104). The user can now move valves or pipes (steps 2106-2114), insert piping (steps 2120, 2122, 2116, and 2118), insert valves (steps 2120-2136), delete piping (steps 2138-2142), delete valves (steps 2138, 2140, and 2144-2154), make entries (steps 2202-2206), use the arrow keys to move the cursor to a specific pipe section or valve (steps 2208-2216), or select piping or valve mode (steps 2218-2228). Then the user saves the plan (2230-2232) or escapes to restore the previous plan (steps 2234-2236); either way, the user is returned to the CONTROL PANEL
(step 2238).
The operation of the SIGNATURE PANEL will be described with reference to Figs.
23, 24A, 24B, 25 and 26. When this panel is called (step 2302), it is displayed (step 2304).

~~~ll~l~4 It is determined whether there is a signature entry at the cursor in the plan (step 2306). If so, it is determined whether the signature is already saved (and marked in the plan with an asterisk) (step 2308). In accordance with this determination, the user is guided through the steps of taking a signature (steps 2310-2314), or the saved signature is displayed (step 2316).
When the DW button is hit (step 2318), the type of signature entry is determined (step 2320).
If the signature is a frequency signature, the user is given the option of switching display smoothing between 1 % and 3 % (step 2322). : ~ tae signature is a time signature, various options are provided for displaying the frequency signature (steps 2324-2338).
When the "GO" button is pressed (step 2402), the user is guided through the steps of collecting a fi-equency signature (steps 2408-2452) or a time signature (steps 2454-2466).
An important function of the SIGNATURE PANCL is the determination of repeatability of signature taking. The measurement of valve acoustic noise by a transducer is dependent upon good coupling of the noise into the transducer. Although it is impossible to verify maximum coupling with 100% certainty, experience has shown that the maximum attainable coupling fUC a particular valve is repeatable to better than 1 dB.
This attainable signal is then a function of the transducer contact force, angle and coupling compound used.
Sensitivity to each of these conditions shows greater than 60 db variation in level. Thus, the only condition which produces a stable level, assuming human control, is the maximum attainable coupling.
The operation of the SIGNATURE PANEL automatically performs the following repeatability measurement on every signature collection, while providing the user with the needed instruction messages to carry out the user operations indicated:
1. The user seats the transducer on the valve and hits GO (step 2402). The first signature produced at a valve is displayed as a solid line (S1) (step 2448).
2. The user lifts the transducer, reseals it on the valve and hits GO again.
This second signature is displayed as a dotted line (S2) (step 2430).
3. The system computes the following:
Repeatability =~E(S1(f)-S2(f))2 /~E(S1(f))2 where S1(f) is the magnitude function of frequency and lJ is a sum over the 10% to 100%
frequency range (step 2436).
4. The user goes back to step 2 to retake another S2 and compare it to S1 if the repeatability is not less than 10% (1 c!B) (step 2446).
If the repeatability is less than 10% (1dB), the signature is good. If not, the user goes back to step 2 or 3 according to the message (depending upon the 2nd level ~~~1~~~~y~

repeatability). This repeatability criterion is adjustable. Prior processes required human judgment, which varied from operator to operator.
The second level repeatability check is a test (not conclusive) of the variation of the signature with time while the transducer is held in one position. Each time GO
is hit, either two or three signatures are taken quickly. If prey do not repeat by the same criteria as above, the highest is kept but marked as unstable.
The repeatability will come out OK only if that first signature was stable. If it does not, either the cause is poor transducer couplinL at SI, or the system is unstable (background noise or pressure). An appropriate message i~ displayed leading the operator to a good signature or a conclusion of time varying instability. In that case, a differential or time signature might be able to support a conclusion.
Anytime a signature is displayed, the user can change the attenuation by pressing the upper or lower half of tire signature (steps 2468-2480). After a signature is collected with the GO button, the user can choose to save the signature (steps 2502-2524) or to erase il IS (steps 2526-2542). If the user needs to skip some signatures or complete them in a different ' order, he can use the arrow keys to move the cursor to a different item and/or step in the plan (steps 2610-2626). 1n addition, the user c.:, choose the INFO PANEL
(steps 2602 and 2604) or the analyze panels (steps 2606 and 2608), or to escape and return to the control panel (steps 2628-2634).
The operation of the REVIEW PANEL will be described with reference to Figs. 27 and 28. When this panel is selected (seep 2702), it is displayed (step 2704).
It is possible to review an item by number (steps 2706-2716) or to list all items (steps 2718-2748). It is also possible to output information (steps 2802-2814), to zip the unit into the self extracting archive described above (steps 2816 and 2818), or to escape to the control panel (steps 2820 and 2822).
1'he operation of the INFO PANI~I. will be described willr reference to Fig.
29.
When this panel is selected (step 2902), it is displayed, including the valve information, results table, and external flags (step 2904). At this time, the CONCLUDE
PANEL may be selected (steps 2906 and 2908), or the user may move to the next sub-item, if there is one (steps 2910-2916). 'The arrow keys may be used to move the cursor to different series in the result table (steps 2918 and 2920). When the externals field is touched, (step 2922), the externals pointer is moved to the field touched (step 2924), and the externals flag is toggled on/off (step 2926). When the escape key is touched (step 2928), the results are saved (step 2930; saving results wi~l be explained below with reference to Fig. 30), and operation is returned to the control panel (step 2932).

%l ~~

The operation of saving results will be explained with reference to Pig. 30.
When this operation is called (step 3002), it is determined whether there have been any result changes (step 3004), and if so, the user is asked whether they are to be saved (step 3008).
If there are no result changes, or if the user elects not to save them, control is returned to the previous operation (step 3006). If the user elects to save the result changes, each sub-item is saved in turn (steps 3010-3016), arid control is returned to the previous operation (step 3018).
The operation of the ANALYZE PANEL will be described with reference to Figs.
31-33. When this panel is called (step 3102), it is determined whether the cursor is on a differential signature (step 3104). If so, the procedure of Pig. 33 will be followed, as explained below. 1f not, the signatures to display are marked based upon the cursor position in the plan (step 3106), and the ANALYZE PANEL is displayed (step 3108). The user can select items with the up arrow (steps 3110-3118) or with the down arrow (steps 3120-3130) or may go to the Results panel (steps 3132 and 3134).
When the SELECT button is pressed (step 3202), the user can select a sub-item with the numeric keypad (steps 3204-3212). With the DIC' key, the user may display the difference between any two signature windows jn a third window (steps 3214-3218). The 1NF0 key calls the INTO PANEL (steps 3220 a:;d 3222). The left and right arrows move signatures left and riglU to view other steps (steps 3224 and 3226). The ESC
key saves results and returns to the control panel (steps 3228-3232).
The operation of the ANALYZE PANEL in the case of a differential signature will lie explained with reference to Pig. 33. In this case, the upstream, valve, and downstream signatures are displayed on the top row, while the valve-upstream and valve-downstream differentials are displayed on the mid row (step 3302). Then, the user can go to the CONCLUDE PANEL (steps 3304 and 3306) or select a sub-item and a step with the numeric keypad (steps 3308-3318), in which case the selected signatures are displayed on the bottom row (step 3320). The user can also escape to the CONTROL PANEL (steps 3322-3326).
ST~NATURE COMPARISON Tl?CI1NIOUP.S
The software performs valve leakage analysis by signature comparison. Valve leakage analysis is usually a complex issue. Three factors must be resolved.
First, leakage noise has a significant but predictable coupling between valves in a system.
Second, multiple leak sources must be identified. Third, external background noise sources (e.g., pumps, condensers) can mask leak noise. Therefore, there are times when changing valve openings/closings are not sufficient or possible to positively identify leak sources and sizes.
In these cases, the following differential methods are used.

t;~~ ~l i S ~~4 The differential methods automate the general method of signahrre comparison where two signatures are subtracted in one display to give better visual resolution of amplitude comparison and eliminate the human errors of the previous manual processes.
The operator analysis and also the automatic analysis then take into account the noise coupling function as follows. Noise is attenuated as it travels down a pipe from one valve to the next. This coupling is a function of frequency with attenuation at a frequency given in d13 per length as measured in units of pipe diameters. The attenuation increases approximately in ftmction of the square root of frequency. Por example, the attenuation between two valves ten pipe diameters apart mibht lie l0 d13 at 120 kIIz but .-;nly 1 dI3 difference at 20 kIIz. Thus, an analysis should place more weiglU on the observed difference at the higher frequencies.
The following three differential methods start with the most general and move to the more complicated and specific.
(1) Signature Comparison Method Any plan has the option of displaying the difference between two signatures in the Analyze Panel. I-Ieretofore these signatures were recorded on different sheets of paper for visual analysis by the operator, and the difference was neither calculated nor displayed.
Certain simple plans have an automatic additio.. :.~ Difference displays; in the case of only a few valves, the plans store information indicating which signatures are to be used. In addition, the D1r k,:y can manually display the difference between any two signatures already showing on the Analyze Panel: Difference = Sigl - Sig2. The user touches the first signature, then the second signature, then the Dif Key, and finally, the location to place the difference display. This method produces an exact difference between signatures at different valves and/or different steps.
Certain standard plans have an automatic setup to display several difference signatures. ror example, the simplest is Plan 1. It displays pressure-background both across the valves and across the steps. An operator can quickly judge the results upon entering the Analyze Panel.
(2) One Transducer Differential Signature Differential signatures are used where it is not possible to remove pressure from a valve to obtain a background comparison signature. This method employs a single transducer and takes tluee separate signatures to compare a valve with background noise on either side of it. It assumes that the signatures are stationary over the time of collection.
The three signatures are sequentially taken on the valve, upstream of it and downstream of it. Up and Down can be on another valve or just the pipe. Although a greater separation of up, valve and down locations is desired, ,,.:..:'::1 comparisons can be made just off the valve onto the pipe.

~~ awl i'~ ~9~
This set of three signatures is marked by one Xd in a plan which then brings up three extra keys above the plan. After collecting the signatures, the user hits Analysis with the plan cursor on the Xd. A dedicated Analysis Panel will come up showing the three signatures on the top line and two difference displays, Valve-Up and Valve-Down, on the 5 middle line. The system will then automatically produce a recommended result indicating whether or not there is a leak by performing one of the analysis procedures described below.
(3) 'Itvo Transducer Differential Method In certain cases where the valve condition cannot be changed and background conditions are unstable, signatures must be taken simultaneously at two locations selected 10 from the valve and at up and down locations relative to the valve. This method prevents false readings due to a ;,hinge in conditions between signatures. Tlu~ee transducers would be difficult to manage simultaneously, but only two at a time are required to produce valid signatures.
When the cursor is on an Xd in the Signature Panel, the user hits Mode to change 15 to two channel collection. Two keys wiil appear instead of three above the Plan with the V
U key highlighted. The user places the channel A transducer at the valve location and the channel B transducer at the Up location and hits Go. After saving this, the highlighted key will automatically switch to V D. The user repc;ats the Go with the channel B
transducer placed at the down location and saves this. The ,;;peatabilily is complicated in this method 20 due to the necessity to reseal both transducers properly.
The V-V key IS 111 Opt1011 key used to acquire a differential signature with both transducers located on the valve. If taken, the V-V signature is used to apply a transducer sensitivity correction function to the channel D part of llre differentill signatures in the display and automatic anllysis as follows:
25 C~Y(f) = S2~Y(f)/S l ~~(f) both transducers on Valve C2'(f) = S2(f) * CY~(f) applied to channel B at Up and Down Analyze Panel display 1nd automatic analysis .proceeds almost the same as in the one transducer method above, except both valve signatures (fibm V U and V D) appear on the top line, the up and down signatures appear on the middle line, and the difference signatures on the bottom line.
AUTOMATIC PROCEDURES FOR ANALYSIS QF SIGNATURES
Automatic analysis, My which the invention automates a process which was previously performed manually, will now be described with reference to examples which are meant to be illustrative ratter than limiting. For instance, the equations used are examples and specific numeral citations can be used to fit the situation at land.

~~~175~94 2ti The most important factor' in proper application of automatic analysis is the selection of signatures upon which to base recommendations. One assumption is that the plan has been properly designed to represent all the valves affecting pressures and leak noise contribution. Another is that the proper valve alignment has been established to produce pressure and no pressure conditions called for in the signature type. The next most important factor is the amplitude criteria used to quantify leakage. Another optional factor is tire frequency weighting function applied to each signature as it is fed into the amplitude criteria calculations. Usually, the weig!rting function emphasizes the higher frequencies, i.e., W(f) _ ~2 * f/ Range (unity mid range). 'this is due both to lower transducer sensitivity with increasing frequency and to the acoustic ;,ttenuation with frequency de-emphasizing background noise. An operator selected semi-automatic method described below adds a band select limiting to the weighting, i.e., W(f) = 0 below f~ -.OS*Range and above fb +
OS*Range.
1n operator-selected two-signature analysis, each time the operator uses the Dif key in the Analyze Panel, a ratio R is calculated of the two signatures chosen.
1'he total band Differential Signature Entry Analysis (r, ASAmax and result where ASA =
acoustic signature amplitude, which is the sum of attenuation and relative amplitude) is displayed in the upper right corner of that difference display. 'I'Iris display is automated; previous processes required that the operator record attenuation and calculate the ASA.
The Band key and then any peak on the difference display are hit to replace the analysis with a band limited analysis at tire dominant frequency chosen.
One main advantage of organizing valves into standard plans is to apply proven standardized analysis methods to them. As the methods are proved, they can be added to specific plans by the program. They could include any number of different comparisons as described above. The method of ratio calculation is individually selectable for each comparison, and the frequency weighting function can be applied or ignored for each. Then the result dependency for each item is selected as tlje maximum/minimum of one or more of the comparisons and the limit criteria are m~~iCed if necessary.
This analysis technique preprogrammeu into the plan can still be temporarily overridden in the Analysis Panel by the use of the Band key calculating a dominant frequency ratio.
Differential signahrre entry analysis is a specific preset example of the above and uses p6int-by-point ratio criteria (explained below) to calculate R for both Valve-Up and Valve-Down. 1'he lower of these and the Valve ASAmax are taken to apply to the Absolute Result Specification.

~A~1731~4 When a group of parallel valves under matched conditions cannot be operated, a direct comparison analysis can be done to calculate relative leak rates among the group. The most dominant frequency is selected over the group. The ratio of each signature amplitude to the sum of all the signature amplitudes at the dominant frequency is the proportionate leak rate.
Three kinds of ratio criteria will now be explained. The signatures to be used can be selected manually or automatically as needed. Por example, if adjacent signatures in a row or column are valve and background signatures, they can be automatically selected. In single-frequency ratio criteria, at an operation-selected frequency f, the ratio is calculated as:
R = S1(f) / S2(f) The dominant frequency is selected automatically at the highest peak of a frequency weighted signature. If the highest peak is found at the lowest frequency, the next highest is selected. Manual selection with the Band key can be anywhere; however, it is again automatically maximized within 8% of the frequency touched. At the selected frequency, the ratio is R = S 1 (f) / S2(f) In RMS ratio criteria, the ratio of RMS is calculated thus:
2O SRMSt ='~E(S1(f) * W(fj)Z I E W(f) SRMSZ = ~Fr(S2(f) * W(~)Z / E W(~
R = SRMSt / SRMS2 In point-by-point ratio criteria, the RMS of the ratio function is calculated thus:
C(f) _ (S1(f) - S2(fj) l (S1(f) + S2(f)) R=E(C(n*A(~)/EA(~
V ='J (E(C(~2 * A(1)) - (E (C(n * A(n))2/ l: A(~ ) / E A(~~ where R is the weighted signature difference ratio (-1 < R, < 1);
V is the variance of the weighted signature difference;
S1(f) is the magnitude function of frequency and ~ is a sum over the 10%
to 100% frequency range for each f where S1(f) or S2(fj > 12 dB ASA; and A (fj is a second weighting function to discount lower amplitude frequencies and to take into account the sound attenuation (which is a function of frequency and distance) of a linear system:

~A~ ~i 7:'. ~~4 A (fj = 1 (S1(fj+S2(f)) + (Slmax+S2max) > .1, = 10 * (S 1 (f) +S2(f)) / (S 1 ma:; -I-S2max) otherwise.
This approach bypasses the W(f) function.
The ratio, once calculated, may be used to determine the existence and severity of a leak thus:
Large leak: R ~ 25 dB
Medium leak: 25 dB > R ~ 10 dB
Small leak: 10 dB > R >_ 0 dB
Tight: R < 0 dB
Absolute result specification uses R and V thus:
Tight: R < O, V < .2 Leak: R > .3, V < .2 Otherwise, operator judgment is required. If a leak is found, the size of the leak is given automatically from the maximum ASA of the smoothed valve signature:
Large: ASAmax ~ 55 dB
Medium: 55 db > ASAmax Z. 35 dB
c Small: ASAmax < 35 dB
ror steam traps, a time survey analysis is required. The system takes a time sweep signature of average amplitude over a 10% ba~:;:;u:lth centered at a selected frequency. The frequency can be selected when the operator hits BW. A standard frequency spectrum I appears by itself. The user touches a peak on the signature, and that frequency is selected.
When Go is hit, a time signature records average amplitude for either 30, 60, or 120 seconds (set by Avg on Control Panel).
Automatic analysis first identifies open and closed states on the pressure time signature. Maximum amplitude (Amax) is calculated as the average of the highest 5% of points and minimum amplitude (Amin) is the average of the lowest 10% of points. If the ratio Amax/Amin is less than 5 dB, it is called stuck shut. If a background (no pressure time or frequency signature) is completed on the trap, it can be further classified. The average amplitude of the pressure (Pav) and ba:.i;ground (Bav) signatures and the ratio of PavIBav are calculated. If the ratio is greater than 20 dB, it is called stuck open. If the ratio is less than 5 dB, it is called stuck shut.
If the trap is not stuck shut, any point above (Amax+Amin)/2 is called open while points below are strut. The result is classified as:

.. Cl~~ 173 i ~4 railed Amin/Bav > 3 dB
or length of all Open periods < 5 sec or time between start of Open periods > 15 Satisfactory otherwise Any time the . Analyze Panel calculates a recommended result, it is displayed automatically in the CONCLLJDI: I'AN)rL~but marked as recommended (flashing).
When the user touches a recommended result in the CONCLUD1? PANCL, it becomes an accepted result which could be changed manually by touching it again. If not accepted before returning to the CONTROL PAN13L, all recommended results are lost.
lU 'fhe operation of the CONCLUD>J PANrI~ will be explained with reference to Pig.
34. Cssentially, machine-recommend insults are displayed which can be accepted or rejected by the user in formulating his conclusion. When this panel is called (step 3402), the insults table is displayed (step 3404), with the item, result, size, and comment for each item under the primary item. The user may enter a conclusion, namely, whether a leak is judged to exist (steps 3406-3410), the estimated size of the teak (steps 3412-3418), or a continent (steps 3420-3424). The user may also fill any items below the cursor not having a result with a copy of the cursor item line (steps 3426 and 3428) or escape to the control panel (steps 3430 and 3432).
A useful feature of this invention is the use of multiple prompting levels.
Signature collection for leakage requires a complex plan procedure.with as much information collected as possible. Only after much experience can an operator recognize unnecessary steps in specific circumstances. Schedule pressures usually dictate maximum testing speed.
1'hcrefore, the system has two types of process t::ompting two user selectable levels. Some are displayed at either level and others only at full prompt level. They help to expedite and minimize training required.
The system remains at the process prompting level set until reset by the following method. The user quits from the Control Panel to go to DOS and restarts the system adding the appropriate argument as follows:
AVLA/b for Full prompt Level AVLA/e for Cxperienced Level The first type of prompting is the status mes~wb~s displayed on the left above the piping diagram. Most of these show at bath levels. Instmctions for signature collection follow this order:

~;~,~ ~ l~~y~+
Place Transducer on Item 5.1 hit Go CI-I)rCK: Reseat Transducer on Valve 5.1 and hit GO
Good Repeatabili:.; '..it Save Plan Completed 5 Other messages include:
Not Settled Retake ' Not A Signature Step No Data Acquired The full prompt level adds some other messages 10 Step 1 alignment not completed The second type of prompting uses the YCS/NO window to ask a question and receive a response. Although it is used widely through the panels to help prevent making a mistake and losing data, this description mainly details its use in signature collection at full prompt level. Valves are grouped under a primary item with a plan testing stmcture. A
15 plan will provide a signature entry for every combination of valve pressures (open/closed) necessary to resolve any possible source of leakage.
An inexperienced operator needs to follow a fixed procedure of valve opening/closing and signature collection. Actually, the most important reason is to observe plant safety and prevent open flow paths. Tire system aids this by leading the operator through each detail 20 of the process of valve alignment, selling the user which valves to open and close and which signatures to take in each step, in addition to the above-described signature collection messages. Upon starting and after each step in the plan is completed, a sequence of questions show in the YI:S/NO window as in the following example:
Close Item 5.1 - READY?
25 Open Item 5.0 - READY?
Open Item 5.2 - RCADY?
Note that proper plan design (not overriding the automatic entry placement in the MAKE
panel) will place valve closings in a non-signature step preceding valve openings, thus automating the valve alignment process and signature entries. The system will automatically 30 move the plan cursor, ask the valve openingslclosings and request necessary signature collection. Advancement through the plan is first up/down and then left to right. If the operator bypasses a signature, the system tries to return to it before advancing to the next Step. If the operator tries to move to a different Step with the arrow keys, a YESINO
question requests confirmation:
Step N01' completed, are you sure?

LA~~~ l~~'~~

The system remembers when the user performs actions out of the predetermined order and prompts the user to perform all necessary actions. Thus, even when the user varies the order, all necessary actions are performed; otherwise, some might be skipped.
A related feature, and one of particular interest to inexperienced users, is the help function. If the user is not sure how to use any particular key, he can hold down the key for at least one second to receive a descriptive, context-sensitive help message. If he presses and releases the key within one second, the key performs its normal function.
Of course, other actuations could be used, such as the use of a right mouse button for the help message and a left mouse button for normal operation. Alternatively, a help key, analogous to the F1 key used in many standard applications, can be provided. Battery life indication and a brightness control for the display can be built into the help feature.
Finally, the operation of digital signal processing performed in FFT module 106 will be explained with reference to Figs. 35-37.
When the FFT module is initialized, the following defaults are set (step 3502):
analysis = FFT, gain = mid-scale, range = 200 kHz, and frequency band = 92 kI-Iz. Then a user command is awaited (step 3504).
At this stage, the user can reset the analysis mode (steps 3506-3510) or the sample rate (steps 3512-3516) or issue a wakeup command (step 3518). In the case of the last option, "U" is sent (prep 3520), and the FFT module samples and processes analog data (step 3522, to be explained below with reference to Fig. 36) and awaits any command (step 3524).
Now the user can set the band (steps 3526-3530), sum over the band (steps 3532-3536), have processed data sent from the A channel (steps 3542 and 3544) or the B channel (steps 3538 and 3540), set the filter and channel (steps 3546-3550), or put the module into idle mode.
The step of sampling and processing analog data will be explained with reference to Fig. 36. When this step is called (step 3602), in accordance with the number of channels set (step 3604), the module samples either 16,384 points over one channel (step 3608) or 8,192 points over two channels (step 3606). Then, in accordance with the process type selected (step 3610), the module performs a FFT (steps 3612-3620), an autocorrelation (step 3624), or an envelope calculation (steps 3626-3632). If the amplitude peak is over or under range, the gain may be decreased or increased automatically as required (steps 3634-3642), and the original process is returned to (step 3644). Previous processes required manual judgment and recording of the proper gain.
The step of sending the processed data, as in steps 3540 and 3544 of the operation of the I=FT module, will be explained with reference to Fig. 37. When this operation is called (step 3702), the gain is sent (step 3704), and it is determined whether the signal amplitude is within range (step 3706). If so, every two magnitude points are averaged, and CAS y !~5!~~4 data from 10% to 100% of the range are sent (step 3712). If not, it is determined whether the signal amplitude is under range at the maximum gain (step 3708) or over range at the minimum gain (step 3710). If either of these conditions is met, step 3712 is executed. If not, step 3712 is not executed. The original process is returned to (step 3714).
The above description is meant to be illustrative rather than limiting. Those skilled in the art who have reviewed this specification will understand that other embodiments may tie made that fall within the scope of the invention. ror example, the apparatus may be equipped with a bar-code reader for use in plants that have bar-coded valves.
Also, a color screen can be used to superimpose signatures. Instead of fast Fourier transforms, other transforms may be used, such as lime domain transforms, correlation, convolution, and band-limited time correlation. Instead of software, firmware or the like could be used to guide the user and to perform analysis. Those skilled in the art who have reviewed this specification will readily understand how to make the modifications appropriate for their purposes without departing from this invention.

Claims (66)

1. An apparatus for performing a process of testing a valve system for a valve leak, the apparatus comprising:
transducer means for receiving sounds from the valve system and for converting the sounds into electrical signals;
transform menus for receiving the electrical signals, for computing a transform of the electrical signals to produce at least two sound signatures, and for outputting digital data representing the at lest two sound signatures; and computation means for (i) receiving said digital data and comparing said at lend two strand signatures to obtain a comparison result, and (ii) automatically deriving a recommended result regarding whether said valve leak exists from the comparison result.

2. An apparatus as in claim 1, wherein the comparison result is a ratio of amplitudes of said at least two of the sound signatures at a selected frequency which is taken from the group consisting of a manually selected frequency and an automatically selected frequency.

3. An apparatus as in claim 1, wherein the comparison result is a ratio of root mean squares of weighted amplitudes of said at least two of the sound signatures, the root mean squares being calculated over a predetermined frequency range.

4. An apparatus as in claim 1, wherein the comparison result is a root mean square, over a predetermined frequency range, of weighted ratios of amplitudes of said at least two of the sound signatures.

5. An apparatus as in claim 1, wherein:
amplitudes of the sound signatures at a dominant frequency are summed to derive a sum; and the comparison result for each valve is a ratio of an amplitude of that valve's sound signature at the dominant frequency to the sum.

6. An apparatus as in claim 1, further comprising digital storage means for storing the digital data representing the sound signatures and a conclusion derived from the recommended result and said at least two sound signatures.

7. An apparatus as in claim 1, further comprising display means for displaying said at least two sound signatures and said recommended result.

8. An apparatus as in claim 7 wherein the display means also displays a difference calculated from said at least two sound signatures.

9. An apparatus as in claim 1, wherein the computation means comprises an interface means for receiving a user's input for at least one of planning, testing, analysis and reporting of a test survey.

10. An apparatus as in claim 9, wherein the interface means comprises a touch screen.

11. An apparatus as in claim 1, further comprising connection means for outputting the digital data representing the sound signatures and the conclusion to an external computer.

12. An apparatus as in claim 11 wherein the connection means comprises a port for a floppy drive.

13. An apparatus as in claim 1, further comprising digital storage means for storing an information database and a group of test structures for said process of testing.

14. An apparatus as in claim 9, wherein said computation means comprises control and display means for guiding the user in understanding the process of testing.

15. An apparatus as in claim 14, wherein the computation means comprises selection means for controlling the interface means to enable the user to make choices required for the process of testing.

16. An apparatus as in claim 15, wherein the selection means enables the user to select a parameter and controls the transducer means and the transform means so that the sound signatures are produced in accordance with the parameter selected by the user.

17. An apparatus as in claim 16, wherein the parameter is frequency range.

18. An apparatus as in claim 16, wherein the parameter is bandwidth.

19. An apparatus as in claim 16, wherein the parameter is transform type.

20. An apparatus as in claim 16, wherein the parameter is averaging type.

21. An apparatus as in claim 15, wherein the selection means further controls the apparatus to collect the sound signatures in accordance with the choices made by the user.

22. An apparatus as in claim 15, wherein, when the interface means displays one or more of the sound signatures simultaneously and the user uses the interface means to indicate a portion of one of the one or more of the sound signatures, the interface means generates a touch signal and outputs the touch signal to the control means, which changes a scaling factor by which the one or more sound signatures are displayed on the interface means in accordance with the touch signal.

23. An apparatus as in claim 15, wherein the interface means has a plurality of manners of actuation by the user, and wherein the interface means comprises help means for displaying, when the user performs a predetermined one of the plurality of manners of actuation, a context-sensitive descriptive message.

24. An apparatus as in claim 14, further comprising digital storage means for storing a plurality of instructions for operation of the computation means.

25. An apparatus as in claim 24, wherein the transform means comprises storage means for storing a plurality instructions for operation of the transform means.

26. A method for testing a valve system for valve leaks, the method comprising:
(a) positioning at least one transducer on the valve system to receive at least two different sounds from a valve in the valve system and to convert the sounds into electrical signals;
(b) transforming the electrical signals to produce sound signatures and outputting digital data representing the sound signatures;
(c) comparing at least two of the sound signatures, based on said digital' data, to obtain a comparison result; and (d) deriving a recommended result regarding valve integrity from the comparison result and displaying the recommended result to a user.

27. A method as in claim 26, wherein the comparison result is a ratio of amplitudes of said at least two of the sound signatures at a selected frequency which is taken from the group consisting of a manually selected frequency and an automatically selected frequency.

28. A method as in claim 26, wherein the comparison result is a ratio of root mean squares of weighted amplitudes of said at least two of the sound signatures, the root mean squares being calculated over a predetermined frequency range.

29. A method as in claim 26, wherein the comparison result is a root mean square, over a predetermined frequency range, of weighted ratios of amplitudes of said at least two of the sound signatures.

30. A method as in claim 26, wherein:
amplitudes of the sound signatures at a dominant frequency are summed to derive a sum; and the comparison result for each valve is a ratio of an amplitude of that valve's sound signature at the dominant frequency to the sum.

31. A method as in claim 26, further comprising storing the digital data representing the sound signatures and a conclusion derived from the sound signatures and the recommended result in a digital storage medium.

32. A method as in claim 26, further comprising (e) outputting the digital data representing the sound signatures and the conclusion. to an external computer.

33. A method as in claim 32, wherein step (e) comprises writing the digital data representing the sound signatures and the conclusion onto a machine readable medium.

34. A method as in claim 26, further comprising a step of storing an information database and a set of plans which are standard test structures or modified test structures derived from the standard test structures, said plans also including automated analysis structures.

35. A method as in claim 34, wherein step (e) comprises (i) compressing the digital data representing the information database, plans, sound signatures and conclusion into a compressed file and (ii) writing the compressed file onto a machine readable medium.

36. A method as in claim 34, wherein valves in the valve system are grouped into items and given item numbers, each of the item numbers are identified by a primary (group) number and a sub number, a primary (base) item having a system-specific function and all other items having a function supporting the primary item.

37. A method as in claim 36, wherein each set of items grouped under a primary item is assigned to one of said plans.

38. A method as in claim 37, wherein all items are compiled into an ordered database having information categories including items, plans and identification information.

39. A method as in claim 38, further comprising repeatedly searching the ordered database according to any of the information categories for selecting an order of group testing and separating results of such searching by test completion.

40. A method as in claim 34, further comprising a step of displaying instructions which guide the user to comprehend and conduct the method for testing.

41. A method as in claim 40, further comprising a step of enabling the user to make choices required for the method of testing.

42. A method as in claim 41, wherein the step of enabling comprises enabling the user to select a parameter, and wherein the sound signatures are produced in accordance with the parameter selected by the user.

43. A method as in claim 42, wherein the parameter is frequency range.

44. A method as in claim 42, wherein the parameter is bandwidth.

45. A method as in claim 42, wherein the parameter is transform type.

46. A method as in claim 42, wherein the parameter is averaging type.

47. A method as in claim 41, wherein the sound signatures are collected in accordance with the choices fade by the user.

48. A method as in claim 41, wherein, when one or more of the sound signatures are displayed simultaneously and the user indicates a portion of one of the one or more of the sound signatures, a scaling factor by which the one or more sound signatures are displayed is changed.

49. A method as in claim 41, further comprising a step of displaying, when the user makes one of the choices in a predetermined manner, a context-sensitive descriptive message.

50. A method as in claim 40, wherein the step of displaying instructions comprises:
i) prompting the user to secure the valve system in a proper configuration;
ii) in a case of positive user response, proceeding with the method of testing; and iii) in a case of negative user response, displaying a warning message;
thereby promoting user safety and precluding invalid signatures.

51. A method as in claim 50, further comprising prompting the user to perform an operation selected from the group consisting of:
i) placing the transducers on a portion of the valve system for testing;
ii) reseating the transducers;
iii) saving a signature when the signature is determined to be good; and iv) correcting a situation which causes an invalid signature.

52. A method as in claim 40, further comprising:
selecting a highest signature of two or three contiguous signatures;
marking the highest signature with a RMS difference repeatability test result;
and determining whether the RMS difference repeatability test result conforms to a predetermined repeatability criterion; and wherein, if the RMS difference repeatability test result does not conform to the predetermined repeatability criterion, said step of displaying instructions comprises instructing the user to reseat the at least one transducer.

53. A method as in claim 40, further comprising:
i) prompting the user to remove and reseat the transducers and to repeat signature collection; and ii) calculating a RMS difference repeatability for verification of proper transducer placement.

54. A method as in claim 40, further comprising directing the user into test and analysis completion of an item group in a proper order.

55. A method as in claim 54, further comprising permitting the user to advance to a position in the plan selected by the user while maintaining full test completion prompting.

56. A method as in claim 40, wherein step (c) comprises:
i) displaying two or more signatures;
ii) receiving a user input representing a selection of a first signature and a second signature from the two or more signatures displayed and a location; and iii) determining a difference between the first and second signatures at the location selected by the user; and iv) displaying the difference obtained in step (c)(iii).

57. A method as in claim 56, wherein step (c) further comprises determining the recommended result based on the difference and displaying the recommended result.

58. A method as in claim 57, further comprising temporarily storing one or more recommended results until the user accepts them or enters conclusions different from the recommended results.

59. A method as in claim 58, wherein step (c)(i) comprises automatically selecting pairs of signatures for display, comparison and result recommendation.

60. A method as in claim 26, wherein step (d) comprises correcting the sound signatures for sound attenuation, said sound attenuation being a function of frequency and distance.

61. A method as in claim 26, wherein:
step (a) comprises automatically prompting the user (i) to ensure that predetermined valves in the valve system are open or closed, (ii) to indicate that the predetermined valves have been opened or closed as prompted, (iii) to attach the transducers to predetermined locations on the valve system, and (iv) to indicate that the transducers have been attached to the predetermined locations;
the sounds are not received until the user has indicated in step (a) that the predetermined valves have been opened or closed as prompted and that the transducers have been attached to the predetermined locations;

the user is automatically prompted to perform steps (a) and (b) a plurality of times for a plurality of configurations of opened and closed valves in accordance with a sequence stored in the automated interface device;
step (b) comprises fast Fourier transforming the electrical signals to produce the sound signatures and displaying the sound signatures to the user; and step (c) comprises automatically calculating a ratio of said at least two of the signatures and automatically determining, in accordance with a stored rule, the comparison result in accordance with the ratio; and step (c) further comprises prompting the user to input the conclusion on the basis of the sound signatures displayed in step (c) and the recommended result.

62. A method as in claim 61, wherein step (c) comprises receiving a user input indicating which of the sound signatures are to be compared and deriving the ratio from the sound signatures selected by the user.

63. A method as claimed in any one of claims 26 - 62, wherein step (a) includes positioning a pair of transducers at two locations on the valve system and controlling the valve system to provide a plurality of differential pressures across the valve system.

64. A method as claimed in any one of claims 26 - 62, wherein in step (a) a single transducer is positioned sequentially on a valve, upstream of the valve, and downstream of the valve to obtain three signatures for comparison of the signature on the valve with the background noise beside it.

65. An apparatus as in claim 1, wherein the at least two sound signatures comprise sound data for at least one frequency greater than or equal to 100 kHz.

66. A method as in claim 26, wherein the sound signatures comprise sound data for at least one frequency greater than or equal to 100 kHz.