`Schricker et al.
`
`[19]
`
`[11] Patent Number:
`[45] Date of Patent:
`
`5,561,610
`Oct. 1, 1996
`
`IIIIII|||II|||III||l|||I|||||||lIllllIl|I||||llI||I||||||II|||||||||I|||III
`US00556}610A
`
`4,533,176
`4,749,937
`
`‘H1986 Yamato et al.
`511933 Ishii
`
`231939 Mmushila
`4-391459
`‘I-H939 Berruyer . . . . . .
`4,825,195
`1031990 Laneetal.
`4,961,381
`5.155.468 1nr1992 Stanl
`.
`5,253,923
`1111993 Imame)ete;.1z.fl ......
`5.303.163
`431994 Ebangh et al.
`
` 4,137,053 11n9ss Moore
`
`364l'431.l1
`..... .. 340158?
`.. 364551.01
`
`----- -- 3403577
`. . . . . .. 3401501
`.. 364551.01
`
`364.1550
`
`Exam" —Ed a.rdR. Co"
`P‘
`"'3'
`W
`5"“““°
`"MW
`Attorney, Agent, or Firm—Steven R. Janda; Mario 1. Donato
`
`57
`
`AB
`
`STRACT
`]
`I
`An appamtus is provided for indicating 3 fault condition in
`a machine having a plurality of pammeters each having
`levels being dependent upon machine performance. The
`apparatus includes a sensor adapted to produce an electrical
`signal in response to the level of one of the of parameters;
`and a processor for identifying a trend in the parameter level
`in response I0 the electrical signal. calculating the duration
`and slope of the trend; and detcrlnining whether a first
`warning threshold is exceeded in response to the duration
`and slope of the trend, the firs! warning threshold being :1
`function of durafion and slope_
`
`16 Claims, 10 Drawing Sheets
`
`104.1 1.1
`
`
`
`10-I-.11.2
`
`[54] METHOD AND APPARATUS FOR
`[N])ICA']'[NG A FAULT CONDITION
`
`[75]
`
`Inventors: David R. Schrieker, Dunlap; Satish M.
`sh “ ’Eas P
`'
`b m
`'3 I
`A com’
`°
`0
`.
`.
`.
`.
`[731 A55'3“‘°’°' C°‘"1"““’ 1”" P°°“"= m‘
`
`[211 App]. No.: 263,693
`[22]
`Filed:
`Jnn.30,1994
`
`
`
`Int. Cl.“ ................................................... .. G081} 2.3:'00
`[51]
`[52] US. CL
`..
`364:'55l.I}1; 3401679
`:53] Field of Search "..................................... 340.-'6'.-'9, 680:
`354,I55{}_ 551_01_ 55132
`
`[56]
`
`References Cited
`
`U‘S‘ PATENT DOCUMENTS
`364.-‘S50 x
`11:19:34 Morrow
`611966 Foster .................. 340.1680 X
`H1967 Dubosq et al.
`.... .. 3401679 X
`5.-‘I975 Johnstone ........ ..
`. 235J'151.l1
`3J'l9I"6 Codomo et al.
`.. 3-40-I'l72.5
`6Il9'.-‘S Mitchell
`340152
`1.-‘I930 ‘Morrow
`.. 364.550 x
`911935 Eldridge et al.
`...................... .. 3640424
`
`
`
`Re. 31.750
`3,257,652
`3.298.010
`3.382.305
`3,946,364
`4.093.939
`4,134,205
`4.542.461
`
`
`
`
`
`104-.11.13
`DETERMINE SLOPE AND
`DURATION OF TREND 104-.l1..'5
`OBTAIN WEIGHTING
`FACTOR BASED ON
`
`
`DISTANCE
`OBTAIN §IGNIFlCAI-ICE
`FACTOR BASED ON
`SLOPE AND nummon
`
`"34-11-‘
`
`
`
`COMBINE WEEGHTING FACTOR
`AND SIGNIFICANCE FACTOR
`
`
`
`
`CALCULATE WARNING PARAMETER
`
`IO-1.11.6
`WAR NING
`
`
`LEVEL G R EAT ER
`THAN CAUTION
`
`
`
`m"11'5
`
`TH E E%HO LD
`
`
`
`RESET
`WARNING
`
`WARNING
`LEVEL GREATER
`
`THAN WARNING
`
`THRESHOLD
`
`104.1140
`
`INDICATE WARNING
`
`
`
`Honda Exhibit 1005
`Page 1
`
`'°""""'
`
`
`
`STORE FAULT CONDITION
`
`104.1 1.11
`
`
`
`RETURN
`
`
`
`‘ID-1.11.14
`
`Honda Exhibit 1005
`Page 1
`
`
`
`U.S. Patent
`
`Oct. 1, 1996
`
`Sheet 1 of 10
`
`5,561,610
`
`zoE.o_z2.§oo
`
`>S$Io3._.
`
`5:902
`
`o_zoE8._..._
`
`Honda Exhibit 1005
`Page 2
`
`Honda Exhibit 1005
`Page 2
`
`
`
` U.S.Patent
`
`zotsmzmi
`
`v_z:E§FE
`
`
`
`
`
`$2“:52453mmemm
`
`
`
`102%.5:.»Et<m:\+
`
`I mm
`
`Oct.1,1996
`
`Sago:
`
`o_zoEH:n._
`
`.3
`
`Sheet2of10
`
`/We
`
`on
`
`22R‘-
`
`3:=:
`
`
`
`io9<z<._<zoE8<
`
`5,561,610
`
`mm
`
`Honda Exhibit 1005
`Page 3
`
`elllll 9'
`
`
`
`
`
`
`
`Q.oo._<z4._<zoE8«;8.224:.zoE8<
`
`Honda Exhibit 1005
`Page 3
`
`
`
`
`U.S. Patent
`
`Oct. 1, 1996
`
`Sheet 3 of 10
`
`5,561,610
`
`
`
`READ. MACHINE
`PARAMETERS
`
`
`
`
`
`
`
`IS
`PARAMETER
`TO BE
`TRI-Ib'l?DED
`
`
`
`ARE
`TREND
`
`DEPENDENCIES
`SATIEFIED
`
`
`
`
` STORE
`PARAMETER
`
`
`220
`
`
`
`
`
`HAS
`TIME OR
`
`
`COUNTS BEEN
`-REAQPHED
`
` DETERMINE AND
`
`
`
`STORE TREND
`DATA
`
`222
`
`224
`
`Honda Exhibit 1005
`Page 4
`
`Honda Exhibit 1005
`Page 4
`
`
`
`U.S. Patent
`
`Oct. 1, 1996
`
`Sheet 4 of 10
`
`5,561,610
`
`
`
`
`
`mazumh»4m:o:nommmzszmuzuaomm
`
`com
`
`com:
`
`cow:
`
`3fl1VA GNBHL
`
`Honda Exhibit 1005
`Page 5
`
`Honda Exhibit 1005
`Page 5
`
`
`
`Honda Exhibit 1005
`Page 6
`
`
`
`U.S. Patent
`
`Oct. 1, 1996
`
`Sheet 6 of 10
`
`5,561,610
`
`Honda Exhibit 1005
`Page 7
`
`Honda Exhibit 1005
`Page 7
`
`
`
`U.S. Patent
`
`Oct. 1, 1996
`
`Sheet 7 of 10
`
`5,561,610
`
`:F_'r_E_7... A
`100
`
`101
`
`
`
`
`INITIALIZE DEFAULT FIT
`PARAMETERS AND ZERO
`
`ACCUMULATORS
`
`
` OBTAIN OPTIONAL
`
`OVERRIDES FROM
`DEFAU LTS
`
`102
`
`LOAD X,‘I’ DATA
`PAIRS
`
`1 03
`
`
`
`
`
`104
`
`105
`
`
`
`SEGMENT X.Y DATA
`/ WRITE OUT RESULTS
`
`CLOSE OUTPUT FILES
`
`
`
`
`
`106
`
`Honda Exhibit 1005
`Page 8
`
`Honda Exhibit 1005
`Page 8
`
`
`
`U.S. Patent
`
`Oct. 1, 1996
`
`Sheet 8 of 10
`
`5,561,610
`
`ZERO LRJEMINUS
`STATISTICS
`
`104 1
`
`II-E-5-
`
`
`
`SET X—MINUS
`START TO
`BEGINNING OF
`x,Y DATA
`
`104.2
`
`104.3
`
`THERE MORE
`UNPROCESSED X,Y
`
`PAIRS THAN THE
`
`
`X-PLUS WINDOW
`
`
`
`STORE LAST
`X-MINUS LINE
`
`
`
`DETERMINE
`‘§"1’9‘,ffi'J"§'G
`
`ADD NEW xx
`
`
`PAIR To X-MINUS
`AND X—PLUS
`wmoows
`
`wAIT UNTIL A
`NEW x,v PAIR
`IS RECIEVED
`AND LOAD THE
`X,Y PAIR
`
`
`
`
`
`
`
`
`
`INCREMENTALLY ADD
`NEW POINT T0
`
`STATISTICS AND
`
`CALCULATE NEW
`BEST FIT LINE
`
`
`
`
`
`
`
`FTXSPSLUFTT
`
`
`CRITERIA
`?
`
`Honda Exhibit 1005
`Page 9
`
`
`
`DEFINE X—M|NUS
`AS NEW LINE
`
`DOES
`
`CALCULATE FIT OF
`X-PLUS ON X-MINUS
`PROJECTION
`
`‘O4-7
`
`104,3
`
`Honda Exhibit 1005
`Page 9
`
`
`
`U.S. Patent
`
`Oct. 1, 1996
`
`Sheet 9 of 10
`
`5,561,610
`
`l=_'r_E_E_ V
`
`_s C3 5"“ \I
`
`
`
`FITSTATDENOM =
`XMINUSMSE +
`SLACKCONST *
`XMINUSNARIANCE
`+ BIAS SLACK
`
`
`
`104.72
`
`
`
`
`104.75
`
`
`4 |TSTATDE?NOM > 0
`
`
`
`XPLUS.MSE
`“W” = W961;
`
`104.7.7
`
` RETURN.Ffi
`FAHS (1)
`
` FITSTAT >
`FIT T09-IRANCE
`
`104-.7.5
`
`
`RETURN Ffi
`PASSES (0)
`
`
`104-.7.8
`
`Honda Exhibit 1005
`Page 10
`
`Honda Exhibit 1005
`Page 10
`
`
`
`U.S. Patent
`
`Oct. 1, 1996
`
`Sheet 10 of 10
`
`5,561,610
`
`ZIE‘_lEl_
`
`104.11.1
`
`CALCULATE DISTANCE OF LAST POINT
`
`104.112
`
`104.11/I3
`
`OBTAIN WEIGHTING
`FACTOR BASED ON
`DISTANCE
`
`
`DETERMINE SLOPE AND
`DURATION OF TREND
`
`
`
`OBTAIN SIGNIFICANCE
`FACTOR BASED ON
`SLOPE AND DURATION
`
`
`
`1Q4_11__'5
`
`104-114
`
`104.11.14
`
`COMBINE WEIGI-ITING FACTOR
`AND SIGNIFICANCE FACTOR
`
`104.11.5
`
`CALCULATE WARNING PARAMETER
`
`104.11.6
`
`
`
`WARNING
`LEVEL GREATER
`
`THAN CAUTION
`THRESHOLD
`
`
`104:‘ 1-8
`
`104.11.9
`
`
`
`INDICATE CAUTION
`
`IS
`
`WARNING
`RESET
`N
`LEVEL GREATER
`WARNING
`
`THAN WARNING
`THRES?HOLD
`
`
`
`INDICATE WARNING
`
`STORE FAULT CONDITION
`
`104.1 1.1 1
`
`RETURN
`
`Honda Exhibit 1005
`Page 11
`
`Honda Exhibit 1005
`Page 11
`
`
`
`5,561,610
`
`1
`METHOD AND APPARATUS FOR
`INDICATING A FAULT CONDITION
`
`TECHNICAL FIELD
`
`The invention relates generally to devices for diagnosing
`and indicating fault conditions, and more particularly, to a
`method and apparatus for diagnosing and indicating a fault
`condition in response to the trend in the level of a machine
`parameter.
`
`2
`calculating the duration and slope of the trend, and deter-
`mining whether a first wanting threshold is exceeded in
`response to the duration and slope of the trend, the first
`warning threshold being a function of duration and slope.
`The invention also includes other features and advantages
`which will become apparent from a more detailed study of
`the drawings and specification.
`
`In
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`15
`
`25
`
`For a better understandin of the
`esent invention, ref-
`erence may be made to thge acoomlpranying drawings. in
`Which?
`FIG. 1 is a high level diagrammatic illustration of an
`embodiment of the invention;
`FIG. 2 illustrates a plurality of connections to an elec-
`Ironic module in one embodiment of the invention;
`FIG 3 fllumam an algorithm perfomed by the c}cc_
`tronic module in one embodiment of the invention"
`FIG 4 .11
`.
`’
`,
`'
`1 “mates 3 plural“? of sensor values and besbfit
`“"3 wgmcmi
`FI(_'3. 5 illustrates warning levels as a function of slope and
`d‘l1'3l-‘OD Of 15*’-“dad data?
`FIG. 6 illustrates one instance of a group of data points
`being l-139d l0 0'-313111 3 5113 3331355‘?
`FIG. 7 illustrates a high level algorithm performed in one
`embodiment of the invention;
`FIG_ 3 fllusu-ates 3 portion ofthc algorithm of 1:[G_ 7 in
`more dam];
`-
`-
`-
`-
`m{:_:Gde233]:u:nU:tes a porno“ of the algomhm of FIG" 8 m
`FIG. 1o_illustrates a portion ofthe algorithm of FIG. 8 in
`mm“ detail‘
`
`BACKGROUND ART
`For service and diagnostic purposes, machines are some-
`times equipped with sensors for measuring operating con-
`ditions such as engine RPM, oil pressure, water temperature,
`W051 Pressure, Oil Contamillaliflfl. Blficlfic M0031‘ Wmms
`hvdraulic pressure, system voltage. and the_like- In some
`cases. storage devices are provided to compile a data base
`for later evaluamn of machme pm-olmance and to md In 20
`diagnosis. Service personnel examine the accrued data to get
`a better picture of the causes of the failure or to aid in
`diagnosis. Similarly. service personnel evaluate the stored
`data to predict fixture failures and to correct any pm-,1cm3
`before total component failure.
`In addition, mm stored parameters may be gxamjned by
`service or supervisory personnel to evaluate machine andlor
`operator performance to ensure maximum productivity of
`the machine. These issues are particularly pertinent to over-
`thcrhighway trucks and large work machines such as oft'- 30
`highway mining trucks, hydraulic excavators.
`track-type
`tractors, wheel loaders, and the lilce. These machines rep—
`resent largecapital investments and are capable of substan-
`Sid‘:Lfigifl:2:3;3::£§n°;“§a‘;g5;‘s’:h:h§:f:§r§u:Tnl;°p':“§;;: 35
`in which productivity will be less affected and so minor
`problems can be repaired before they lead to catastrophic
`BEST MODE FOR CARRYING OUT THE
`failures.
`INVENTION
`Similarly, it is sometimes advantageous to accumulate
`.
`.
`.
`.
`parameters only when the machine is in a particular oper- in
`...§,:t*:::..:t ‘;“it.‘..:..t";°:':.“;.t:::“t::;".asst;
`ms tweet tramway
`anal
`.g
`y yddl.
`1
`f
`k oqhi
`12‘
`used during performance evaluation but may alsobe usedin
`ysls.‘ storage’ an
`Sp ay system or a wot mac He
`'
`failure diagnosis and prognosis. For example, the length of
`Employing a complement of on—board and off-board hard-
`.
`.
`.
`.
`.
`.
`ware and software the machine prognostic s stern ll] moni-
`ume spent "1 a pamcum gem while the machine '5 loaded
`tors and derives
`component infomlation and ana-
`may be needed to evaluate machine perfon-Dance"
`13,393 the resulting dam to ind,-cm impending mmpunem or
`The present invention is directed to overcoming one or
`System fanumi
`more of the problems set forth above.
`.,§§.i';.'n“liZ.“n§Z§ §J.?§'."Ei’i.°.‘§.‘3.°.§Z?‘l“.’t.°?.’§.’E“n’?i°f.‘L?E
`DISCLOSURE OF THE INVENTION
`In one aspcc; of me invention, an apparatus is provided 50 machine 12 to a central computer system 16 for analysis. In
`rot indicating a fault condition in a machine having a
`the prefcned emlnodimcnt.
`the data is transfermd by the
`plurality of parameters each having levels being dependent
`Q1-131C0|11|11 53161111? 53'5l3°m back 10 H15 53353-1 Wmpmer
`upon machine performance. The apparatus includes a sensor
`5Y3"-‘J11 15- Altflfflal-‘Veil’. the (3313 i5 l-T3-flsfeiwd l3)’ 3 Benn?“
`adapted to produce an electrical signal in response to the 55
`l¢1°l3h0_n3 5}'5l3°m 01' bl’ Slflfifig 5813 01'! 3 °0mP11l5T C1151‘
`level of one of the parameters; and a processor for identi-
`which is then malbd E0 I116 Central C0fl1PlJl€1' Sill? f0!’ anal!“
`fying a trend in the parameter level in response to the
`515-
`elecn-ical signal, calculating the duration and slope of the
`It should be understood that all aspects of the present
`trend; and determining whether a tirst warning threshold is
`invention could be located on-board the work machine 12
`exceeded in response to the duration and slope of the trend, 59
`thereby eliminating the need for a communication system
`the first warning threshold being afunction of duration and
`14; however, the central computer system 16 allows an
`slope.
`eI1ti.rc fleet to be monitored at a cenn-al location.
`
`45
`
`Subsets of the data are also transmitted to a display
`Inasecond aspect of the invention.,amethod of indicating
`module (not shown) in the operator compartment of the
`afault condition is provided. The method includes the steps
`of sensing a parameter having a level being dependent upon 55 work machine 12 for presentation to the operator in the fonn
`machine performance and responsively producing an elec-
`of gauges and warning messages. During normal operation,
`trical signal,
`identifying a trend in the parameter level,
`gauge values are displayed in the operator compartment.
`
`Honda Exhibit 1005
`Page 12
`
`Honda Exhibit 1005
`Page 12
`
`
`
`5,561,610
`
`5
`
`10
`
`30
`
`SD
`
`55
`
`3
`During out-of~spec conditions, alarms and warning;‘instruc-
`tional messages are also displayed.
`In the preferred embodiment, sensed data is directly
`sampled by sensors 18 of a type well-known in the art for
`producing electrical signals in response to the level of
`operational parameters and includes pulse~width modulated
`sensor data, frequency-based data, five volt analog sensor
`data, and switch data that has been effectively debounced.
`The sensors are connected to an electronic module 20 for
`delivery of the sensor signals.
`In the embodiment of FIGS. 1 and 2. the sensor signals are
`delivered to the electronic module 20 by either direct
`connection of analog sensors, connection by way of an
`RS485 link, or over a datalink governed by SAE specifica-
`tions J1587 and J 1708. A pushbutton is also included to
`trigger the acquisition of a snapshot of data. Connection is
`also provided from the machine battery and key switch to the
`electronic module 2|).
`
`In the preferred errlboditnerrt. the electronic module 2|}
`includes a Motorola 68000 microprocessor, a lower level
`communications board (not shown) of a. type well-known in
`the art, and a memory section 24 including high level flash
`memory and battery backed RAM. 'I‘hc electrortic module
`also includes a pair of RS232 connections, one being avail-
`able for connection to the satellite communications system
`21 and the other being available for connection to an
`ofi—board computer 22 used in download of data and ini-
`tialization of the system. In the preferred embodiment, die
`ofi-board computer 22 is a laptop personal computer.
`To document the performance of the machine andlor its
`major components, performance baselines are stored in an
`array within the memory device located in the electronic
`module 20. These baselines are used during key, repeatable
`performance checks of the machine to help verify machine!
`component health and, as discussed below. are used as
`reference points to determine whether the machine is in an
`operating condition in which machine parameters are to be
`processed and stored.
`A subset of parameters for which trend data is to be
`produced is either predefined or defined via the ofi-board
`computer 22 or the central computer 16. Each parameter
`includes a dependency definition that identifies the condi-
`tions under which data will be stored for trending purposes.
`Typically. the dependency definition is selected to indicate
`the normal operating conditions of
`the machine;
`for
`example, when engine RPM are above a certain level and
`boost pressure is above a predetennincd level. The trending
`definition for each parameter may vary and may be a
`ftmction of several other machine parameters that shall be
`referred to as dependencies. Trend data is gathered and
`stored in memory as the specified dependency definition is
`met over a specified trend period. which is measured either
`in time, such as over a period of ten hours, orin counts, such
`as over a period of ten transmission shifts. Trend data is only
`obtained while the engine is running. Based on the specified
`trend type. the maximum,
`or cumulative value of
`data gathered during this period is then stored as a single
`trend point with counts to determine the average value
`andfor the points available. The determination of whether to
`use the average, maximum, or minimum value to obtain the
`trend point is based on the system designer's decision
`regarding which type of calculation would provide the best
`indication of changes in engine performance or impending
`failures. It should also be understood that multiple values
`could be calculated for the same sensed parameter, i.e. trend
`points could be calculated to indicate both an average value
`and a minimum value for a designated machine parameter.
`
`4
`
`The overall trend is formed by storing a specified number
`of points in the memory device depending on the size of the
`available memory area and the length of the desired histori-
`cal data base.
`
`Tlend data may be reset and the definitions may be
`redefined by the off-board system 22 via one of the com-
`munication ports. For example, if a particular application of
`the machine requires a diferent dependency definition for
`one or more of the sensed parameters, the ofi'—board system
`22 is used to modify the dependency definition by providing
`commands to erase a given array including a given depen-
`dency definition and replace that definition with a new
`dependency definition. Similarly, this function may be per-
`formed by the central computer system 16 via the commu-
`nication system 14.
`In addition to the trend data produced for sensed param-
`eters, it should be understood that calculated values. such as
`net horsepower or driveline torque, may also be Lrended in
`a similar manner. Typically,
`these calculated values are
`determined by the electronic module 20 according to pre-
`determined definitions in response to a plurality of sensed
`parameter signals.
`Referring now to FIG. 3, an algorithm incorporated in an
`embodiment of the invention and executed by the processor
`within the electronic module 20 to perform the above
`functions is now described. In block 212. The electronic
`module 20 determines whether the engine is running.
`Advantageously, the engine is determined to be running if
`engine speed exceeds cranking engine speed. If the engine
`is not running, then the algorithm will not proceed. If the
`engine is running. control proceeds to block 214 where the
`electronic module 20 reads the sensed machine parameters
`from the datalink or other inputs.
`For each of the sensed parameters. the electronic module
`20 determines at block 216 whether that parameter is to be
`processed to provide trend data. If trend data is to be
`provided. the trending definition is retrieved and the depen-
`dency parameters are checked at block 218 to determine
`whether the dependency definition is satisfied. If the depen-
`dency definition is not satisfied, control returns to block 212.
`The dependency definition for each operating parameter of
`interest is defined in terms of other sensed machine param-
`eters. For example,
`the dependency definition for boost
`pressure may be satisfied only when engine rpm is greater
`than a low operating speed and less than a high operating
`speed. when the engine rack setting is greater than a prede-
`termined level, and when the jacket water temperature is
`greater than a predefined operating temperature. That is.
`values for boost pressure are only saved and processed for
`producing trend information when the above conditions are
`satisfied. In this way, all boost pressure values used to
`produce the trend data will have been acquired when the
`engine is in the same general operating condition. It should
`be understood that the actual ranges, minimums. and maxi-
`mums used in the dependency definitions are determined
`empirically to define the operating conditions of interest and
`will vary from machine to machine and application to
`application.
`If the dependency definition is satisfied, control proceeds
`to block 220 where the value of the sensed parameter is
`stored. This process is continued until. at block 22, either
`the time period over which each trend point
`is
`to be
`determined or the number of events for which each trend
`point is to be determined is reached at which point control
`proceeds to block 224 where the electronic module 20
`calculates and stores the trend point. The time period or
`
`Honda Exhibit 1005
`Page 13
`
`Honda Exhibit 1005
`Page 13
`
`
`
`5
`
`6
`
`5,561,610
`
`number of events is selected in response to the designer’s
`desire for precision. die availability of memory space in the
`memory device, and the length of time or number of counts
`required to obtain meaningful trend points. The calculation
`of the trend point may include accumulating the stored
`_
`_
`_
`va_lu_es, selecting the maximum stored value, or selecting the
`ntirurrtum stored value. The calculated trend point is saved
`and the data array for that parameter is then cleared to allow
`for the storage of data for calculation of the next trend point
`for that pammetcr‘
`Trend data obtained by way of the algorithm of FIG. 3 is
`illustrated in FIG. 4. While the illustrated data has a sub-
`stantial variance, straight lines can be fit
`to the data to
`illustrate the general trend of the data by known curve fitting
`techniques, such as the least-squares method. The method of
`determining when a particular straight line trend should be
`terminated and a new line segment defined is described
`below in connection with the algorithm of FIG. 7.
`Based on the slope and duration of the trends illustrated
`in FIG. 4, certain judgements can be made regarding the
`likelihood of impending component or system failure. As
`shown in FIG. 5, functions may be defined in terms of slope
`and duration of particular trends whereby wanting condi-
`tions are indicated in response to either the parameter of
`interest changing at a very high rate for even a small period,
`or at a relatively low rate for an extended period. The area
`indicated by the letter “A" indicates the various combina-
`tions of slope and duration of trends associated with normal
`operating conditions. The areas indicated by the letters “B"
`and "C” indicate the caution region and warning region,
`respectively. That is, normal conditions are indicated if the
`magnitude of either the slope or duration of a particular trend
`is relatively small, but as the magnitude of the slope and
`duration increase.
`the likelihood of a fault condition
`increases.
`
`Turning now to FIG. 6, the method of defining a particular
`line segment associated with a trend in the sensed data is
`described. A first region of data points in a series of data are
`selected and indicated by X—. A second region of data points
`immediately following X— are selected and indicated by X+.
`The sizes of x— and X+ are arbitrary but, in the illustrated
`embodiment, X— consists of 6 data. points and )(+ consists
`of the following 3 data points. In some embodiments, X+
`may include only a single data point; however, multiple
`points are desirable for )(+ to reduce the likelihood that a
`single aberrant point will cause a line segment to be tenni-
`naled.
`
`The data points in X— are used to calculate a best fit line
`using a known technique such as least squared errors. Using
`the derived equation of the line, the data in X+ is tested to
`determine whether the line segment should continue such
`that the data points of X+ are added to the line segment data
`or whether a new line segment shottld be started and the
`previous segment
`terminated. This test is performed by
`using the following inequality:
`Fit-Statistic =
`
`5
`
`to
`
`30
`
`45
`
`SD
`
`MSE(X*.;’* Based o1rX—Re rterriorr
`M3E(X-, 1-) +
`niwiation slack + bias slack)
`
`(Fit [alumna
`
`where: MSE is the mean squared error;
`deviation slack is a bias to the denominator of the fit
`statistic to discount MSE changes when there is a high
`slope to the trend and in the preferred embodiment is
`equal to a slack constant multiplied by the variance of
`X—;
`
`fit tolerance is a constant controlling the tightness of fit of
`the data allowed for a given line segment, in the
`preferred embodiment is equal to 4.8; and
`bias slack is a constant used to prevent a new line segment
`from being formed in the event that the MSE of X— is
`zml In the prfifmod embodiment’ bias 5-lack is set
`aqua} to zero and decision blocks are included in the
`algorithm shown in FIG 9 in handle Simmons where
`the fit Statistic denomjniitor is mu31 to mm
`If the inequality is true. the left most data point in )(+ is
`moved to X-— and a new data point is moved to the right most
`side of X+. If the inequality is false. the line segment is
`terminated and a new X— is defined beginning with the next
`data point following the defined line segment.
`In one embodiment, if the inequality is false, a fit statistic
`base is set equal to the present fit statistic. Data points are
`then continuously added to X— and a new fit statistic is
`calculated when each new data point is added. Data points
`are no longer added to X— when the fit statistic reaches a
`preselected fit poorncss percentage indicative of the num-
`mum ratio of the fit statistic base to the fit tolerance. This
`allows a subset of the data points in X+ to be added to the
`line segment if the trend actually changes in the middle of
`X-t~ rather than at the beginning of the ){+ group of data
`points.
`The algorithms performed by the present invention are
`now described in connection with FlGS. 7 through 10. At
`block 100, the algorithm is started. At block 101, the default
`fit parameters are ‘initialized and any accumulated values are
`set equal to zero. At block 102, the system determines
`whether there have been overrides of the defaults. At block
`
`103, data points are loaded into the computer system by way
`of one of the communication systems. For each of the data
`points, the time at which it was stored and its magnitude are
`arbitrarily chosen as X and Y. This data is next processed at
`block 104 which is described in more detail in connection
`with FIG. 8. Following processing of the data, the output
`files are closed at block 105 and the algotitlun is terminated
`at block 106.
`The functions described in connection with blocks ‘I04
`and 105 describe a system in which the analysis of data
`occurs oll'~board the machine at a central computer 16. In an
`alternative embodiment, the analysis occurs on—board the
`machine in the electronic module 20 on a real-time basis. In
`this case, the results of the processing in block 104 are stored
`but not written out and the function of block 105 is omitted.
`Turning now to FIG. 8, the X— minus statistics are zeroed
`and the X— window is filled with the beginning data points
`at blocks 104.1 and 104.2, respectively. The computer then
`determines whether there are more unprocessed data points
`than the number of points to be included in the X+ window
`at block 1043. If yes, the left most point in X+ is moved to
`the right most position in X— and a new data point is added
`to the X+window at block 104.4. New statistics are dien
`calculated to arrive at a new best fit line for X— using the
`least squared errors technique at block 104.5, although other
`curve fitting tecluiiques are equally suitable. The mean
`squared error of the data points in the }(+ window is then
`calculated at block 104.6 with respect to the best fit line
`derived in block 104.5 to determine the relative fit of the data
`points in }(+ with respect to the line determined using the X-
`data points as compared to the fit of the data points in X—.
`At block 104.7, and as set forth in more detail in FIG. 9, a
`determination is made as to whether the data points in X+
`are a part of the trend described by the X—- line segment or
`whether a new line segment should be defined to describe a
`new trend.
`
`Honda Exhibit 1005
`Page 14
`
`Honda Exhibit 1005
`Page 14
`
`
`
`5,561,610
`
`7
`Block 104-.7.1 is the start of this determination. In block
`1043.2 the fit statistic denominator of the equation dis-
`cussed above is calculated. If tl'1e fit statistic denominator is
`greater than zero at block 104.13, the fit statistic is calcu-
`lated at block 1043.4 and compared to the fit tolerance at
`block 10-1.7.5. If the fit statistic is less than the fit tolerance
`then the fit passes at block 104.13, otherwise the fit fails at
`block 10437.6. If the fit statistic denominator is not greater
`than zero, then the fit fails if the means Squared error of X+
`computed at block 1043.7 is greater than zero, othenvise the
`fit passes.
`Ifthc fit passes, then the warning status is determined at
`block 104.11. If the fit fails, then the data points in X— are
`defined as a new line segment at block 104.8 and the
`warning status is determined at block 104.11. If there are no
`unprocessed X, Y pairs in the )(+ window, then the statistics
`corresponding to the last X- line is stored at block 104.9. At
`block 104.10, the system waits until a new X, Y pair is
`available and loaded and the process continues from 104.4.
`The detailed operation of the warning status determina-
`tion block at 104.11 is set forth in FIG. 10 at block 104.11.].
`At block 104.112 the distance of the last point in X+ from
`the mean of the data points in X- and X+ is calculated in
`standard deviations. Alternatively, the position of the data
`point
`is determined with respect
`to predefined warning
`levels such that an indication is provided of the relative
`distance of the data point from a warning level.
`The slope and duration of the line segment is determined
`at block 104.113 from the equation of the 31- line segment
`obtained using the least squared technique. although other
`curve fitting techniques may he used. At block 10-4.11.4
`based on this slope and duration, the computer obtains a
`significance factor from a look-up table of a type well-
`known in the art. The value of the significance factor
`increases as slope and duration of the trend increase. That is,
`for any given slope, the significance factor increases as
`duration increases; and for any given duration. the signifi-
`cance factor increases as slope increases. The precise values
`used in the look-up table are selected as a matter of design
`choice such that faults are indicated only when there is a
`substantial likelihood that a failure is imminent.
`In one embodiment, the significance factor itself is used
`to determine whether a fault condition is indicated. Alter-
`natively, a warning parameter is calculated at block 10411.5
`by multiplying the significance factor by the distance from
`the mean in standard deviations obtained in block 10411.2.
`In another embodiment, a number weighting factor at block
`104.l1.13 indicative of the relative distance of the data point
`from a warning level as computed in block l04.l1.12 is
`combined at block 104.11.4 with the significance factor to
`produce a warning parameter at block 104.11.5. The mean
`used in block 104.112 may be one of either an exponential
`weighted average of data points, a mathematical average of
`a predetermined number of the most recent data points, a
`mathematical average of all data points stored since the last
`servicing of the machine, or any other similar procedure for
`determining a historical mean.
`The determination of the wanting parameter adds a third
`dimension to the illustration of FIG. 5 such that the warning
`levels are a function of not only slope and duration, but also
`distance from the mean. In this way, a lesser significance
`factor is required to indicate a fault condition when the
`parameter level is a substantial distance from the mean or
`near a warning level.
`The warning parameter is then compared to a caution
`threshold. In the preferred embodiment, if in block 104.11.2
`the data point is below the mean. the resulting distance is
`
`10
`
`20
`
`25
`
`30
`
`45
`
`SD
`
`55
`
`60
`
`expressed as a negative number and similarly, if the slope is
`negative. the resulting significance factor is negative. This
`approach eliminates the false indications of fault conditions
`that may occur when the parameter level
`is low but is
`returning to the mean along a trend having a significantly
`positive slope for a sufficient duration.
`If the warning parameter is less than the caution threshold
`at block 10411.6, then the line segment is in a range similar
`to the two-dimensional area “A" in FIG. 5 which indicates
`a normal operating condition. Any warning that had been
`active is then reset at block 104.113.
`If the warning parameter is greater than the caution
`threshold at block 104.11.6, then the line segment is outside
`the normal operating region. If the warning parameter is
`greater than the caution threshold but less than the warning
`threshold, then the line segment is in a region similar to the
`two—dimensional area “B" in FIG. 5 which indicates a
`caution region as shown at block 10-4.11.8. If the warning
`parameter is greater than the warning threshold as indicated
`at block 10-1.11.9, then the line segment is in a region similar
`to the two-dimensional area “C” i.n FIG. 5 which indicates
`a warning region. This warning is indicated at block
`104.11.10 and stored at block 104-.11.11. As stated above, in
`an alternative embodiment, the significance factor itself is
`compared to caution and warning thresholds and thereby fall
`into the regions identified i.n FIG. 5. If the warning param-
`eter is not greater than the warning threshold. control
`proceeds to block 104.11.11 where the fault condition is
`stored.
`The existence of fault conditions are indicated by any of
`a plurality of available warning means. such as lights or
`