Exception processor system

Abstract

An exception processing system for use in conjunction with manufacturing facilities, and automated manufacturing cells in particular is provided. The exception processor is adapted to receive alarms from a cell controller indicating that an unplanned event or exception has occurred in cell operation. The exception processor implements an automated recovery procedure that responds to the alarm, corrects the exception, and returns the cell to normal operation. The exception processor also statistically monitors cell operation in order to avoid exceptions before they occur, and to provide better control over cell processes.

Description

TECHNICAL FIELD

The invention disclosed here generally relates to error detection in automated manufacturing facilities, and more particularly, to systems or subsystems that have an automated capability to process or manage errors in such facilities, and return them to normal operation.

BACKGROUND INFORMATION

The degree of automation in manufacturing has continuously risen over the years as a result of new developments and improvements in computer technology. The driving force behind automation has been, and continues to be, the desire on the part of manufacturers to reduce labor requirements in a cost-effective manner, and to improve quality. To better realize these goals, most modern-day manufacturing facilities utilize automation to some degree, particularly those which are involved in the mass production of parts.

For example, numerically controlled machines or work stations have been commonplace in many manufacturing facilities for years. A more recent trend in automation is the development of manufacturing "cells" where a cell controller, or central computer, automatically coordinates a network of individually programmable work stations that carry out the various stages and processes required to produce a given part, or a set of related parts.

An automated "cell" generally manufactures parts by following predefined plans that deterministically order how each part should be made, generally as a sequential series of steps that are implemented at individual work stations. In addition to coordinating the work stations, the cell controller continuously verifies the completion of process techniques and steps from one work station to another. This is generally accomplished via a combination of sensor technology and microprocessor-based communications from individual work stations. The cell controller automatically detects an error, or an "exception," in the part's manufacture, and interrupts further operation of any one or more work stations in the cell until the error is corrected.

It is believed that, prior to the development of the present invention, error correction in manufacturing cells (hereinafter referred to as "exception processing") has chiefly been accomplished via manual means. That is to say, although the cell controller may have the capability of identifying the nature of an error, it does not have the capability of informing the human operator as to why or how the error occurred.

By way of illustration, the cell controller may indicate to the operator that a part is not properly positioned at some stage in the cell. However, it does not have the capability of informing the operator as to whether or not the improper position was caused by a materials problem (e.g., the part has an inherent dimensional defect), or an equipment problem involving an electronic or mechanical failure in some portion of a work station. It has always been left to the operator to discern the nature of the problem, and then fix it. In some cases, the operator has the capability to make the fix himself. In other cases, the operator calls in outside maintenance people having expertise beyond the operator's capabilities.

Past experience has shown that when an exception occurs in a manufacturing cell, as much as 80% of the total time required to fix or correct the exception is spent in simply identifying the cause of the exception, or in other words, identifying what is wrong. The remaining 20% is spent in physically implementing corrective action. For this reason, an automated tool having the capability of more efficiently ascertaining the cause of exceptions, and nothing else, would result in significant cost savings to the manufacturer.

As a person skilled in the art would know, automated manufacturing is evolutionary in nature. The degree of automation has progressed to higher and higher levels with new improvements in technology and an ever-expanding knowledge base of learned experience In reality, there are probably no fully automated manufacturing cells in existence today. In other words, virtually all "automated" cells are semi-automatic in the strictest technical sense, because all require, to varying degrees, some form of monitoring or control by human operators. In many cases, this may be no more than a person monitoring a CRT. In certain cases, it may involve a person physically handling materials during some stage of cell operation.

As will become apparent, the invention disclosed here provides yet another evolutionary advance in automated manufacturing cells that has heretofore been unknown in the art. First, it provides a tool that significantly improves the efficiency of both detecting and correcting exceptions in manufacturing cells, and correspondingly reduces the costs associated with past methods of manually correcting exceptions. Second, and just as important, it provides a tool that anticipates exceptions, and maintains better process control over a cell's operation. The end result is that the invention improves the overall quality of the parts produced by the cell.

SUMMARY OF THE INVENTION

The invention is an exception processing system that is designed to monitor operation of a manufacturing facility, and to both correct and anticipate unplanned events during the operation of such facility It is particularly well-suited for use in conjunction with automated manufacturing cells that generally carry out a plurality of preplanned steps and procedures to make parts. Such steps and procedures are implemented by a plurality of individual work stations whose respective operations and functions are coordinated by a central computer, which is commonly called a cell controller.

The exception processing system includes a computer that is operatively connected to the controller of the facility. The computer receives information from the controller, and sends information back to the same. For example, the controller is generally operable to provide an alarm when an unplanned event, or an exception, occurs in the operation of the facility. The exception processor is programmed or ably adapted to receive and acknowledge such alarm. The exception processing system's computer is further programmed to request data updates or information from the controller concerning part-making steps and procedures under its direct control, some of which may be requested for the purpose of correcting an unplanned event in response to a controller alarm, while other information may be periodically requested as part of the exception processor's ongoing monitoring and statistical process control over the facility. The exception processor's statistical process control involves continuously logging into memory, or a statistical knowledge base, any corrective action taken to correct an unplanned event, and any ongoing statistical monitoring that would permit analysis of cell operation for anticipating future exceptions. Statistical process control also involves logging the value of the parameter used to determine whether the process implemented by the cell is within control limits.

The exception processing system includes a set of preprogrammed recovery procedures, wherein each procedure is adapted to provide a certain preplanned pattern of possible corrective actions for returning the manufacturing facility to normal operation after a given unplanned event. Recovery procedures employ a combination of heuristic and deterministic techniques. The computer automatically accesses and implements at least one of such recovery procedures in response to a cell controller alarm.

Part of implementing a recovery procedure may involve requesting further data from the cell controller and/or facility sensors. It may also include requesting information from a human operator. With respect to the latter, the exception processing system has a user interface, preferably in the form of a conventional CRT and keyboard, that is connected to the computer for enabling it to prompt the input of certain information from the operator. The CRT is also used to display data such as drawings, instructions and documentation to the operator to assist him in performing recovery, repair or maintenance functions that require human intervention.

The computer is programmed by an "object-oriented" program which facilitates its adaptation to different automated manufacturing facilities. This type of programming method defines various portions of the exception processor system into "objects" which operate independently relative to each other, and process and respond to messages according to their own internal behavior and internal data. The objects as disclosed here are portioned into an exception processor handler object: an operator screen object: a recovery log object; a communication interface object; a protocol object; a message object; and a statistical processing object.

The exception processor handler object is essentially the brains of the exception processing system and defines how the system responds to given controller alarms, and defines the system's statistical process control over the manufacturing facility. The operator screen object provides the user interface (e.g. CRT and keyboard) with the exception processor system. The recovery log object records alarms or exception conditions and makes records of recovery procedure steps that are implemented in response to alarms. It also records the values of operator inputs and other parameters observed by the system's computer, regardless of whether or not such information is created as a result of implementing a recovery procedure, or as a result of the system's ongoing statistical monitoring of the facility. Any action taken by the operator or by the system's computer is also recorded by this object. The communication interface object links or operatively connects the exception processor's computer to the manufacturing cell. The protocol and message objects are linked to the communication interface object in that they respectively define digital communication protocols between devices, and the type of messages that are sent from one device to another.

If the facility is of a type that utilizes sensors to monitor manufacturing steps and procedures, the computer is operatively connected to such sensors for monitoring facility operation. Some facilities have independent work stations with their own built-in self-test capability. In such cases the system's computer may be operatively connected to such components, to implement work station self-tests in accordance with certain recovery procedure instructions.

An important advantage of the invention is that it provides an exception processing system that has the capability of automatically determining the cause of exceptions, and thus significantly reduces the amount of time required to correct the exception. The recovery log object's ability to log exception processing activities enables operators to investigate how the system arrived at certain conclustions, and provides statistics on exceptions that can be used to subjectively and quantitatively define the progress of manufacturing cell processes over defined time periods. Further, information from the operator can be input into the processor's knowledge base for later use by others. Unlike normal computer systems, the exception processor is programmed to accept "don't know" or "unknown" as valid answers to its queries, and thereafter continue toward a solution taking into account unknown variables.

The invention as summarized above, will become better understood and more fully explained upon consideration of the following description which is to be taken in conjunction with the drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

In the drawings, like reference numerals and letters indicate like parts throughout the various views, unless indicated otherwise. The invention is designed to be used in conjunction with automated manufacturing systems involving complex steps and procedures under the control of a central computer. The description of these kinds of systems is particularly well-suited to the use of flow diagrams. Accordingly, many of the figures are in such form, where a generalized system and subsystems or substeps thereof are interrelated by reference numerals that incorporate decimal point notation. As the skilled person would immediately understand, the numeral in front of the decimal point is common to both a general system and its substeps. Numerals that follow the decimal point are unique to particular substeps.

FIG. 1 is a schematic view illustrating the components of an exception processor system in accordance with the invention, and illustrates how the exception processor system is functionally and operatively connected to a manufacturing cell;

FIG. 2 is another schematic view of an exception processor system in accordance with the invention, and further illustrates the functional communication links between the exception processor system and the manufacturing cell;

FIG. 3 is a pictorial view of a manufacturing cell for the automated fabrication and assembly of floor panels, and is typical of the kind of cell that would use the exception processor system disclosed here;

FIG. 4 is a flow diagram illustrating the various steps and procedures of the manufacturing cell shown in FIG. 3;

FIG. 5 is a flow diagram illustrating substeps of the "material loading" step shown in FIG. 4;

FIG. 6 is a flow diagram illustrating substeps of the "ditch" step shown in FIG. 4;

FIG. 7 is a flow diagram illustrating substeps of the "drill/route" step shown in FIG. 4;

FIG. 8 is a flow diagram illustrating substeps of the "manual clean prior to pot" step shown in FIG. 4;

FIG. 9 is a flow diagram illustrating substeps of the "potting" step shown in FIG. 4;

FIG. 10 is a flow diagram illustrating substeps of the "cure" step shown in FIG. 4;

FIG. 11 is a flow diagram illustrating substeps of the "potting plane" step shown in FIG. 4;

FIG. 12a is the first in a series of two views, is a flow diagram illustrating substeps of the "manual clean" step shown in FIG. 4;

FIG. 12b is the second in a series of two views, and illustrates substeps of the "manual clean" step shown in FIG. 4;

FIG. 13 is a flow diagram illustrating substeps of the "insert install" step shown in FIG. 4;

FIG. 14 is a flow diagram illustrating substeps of the "cure" step shown in FIG. 4;

FIG. 15 is a flow diagram illustrating substeps of the "part mark" step shown in FIG. 4;

FIG. 16 is a flow diagram illustrating substeps of the "final route" step shown in FIG. 4;

FIG. 17 is a flow diagram illustrating the substeps of the "remove parts" step shown in FIG. 4;

FIG. 18a is the first of a series of two views, and is a flow diagram illustrating a detailed recovery procedure in accordance with the invention, for correcting an exception relating to bar code errors in the manufacturing cell of FIGS. 3 and 4, and relates to recovery decisions called for in FIGS. 5, 6, 7, 8, 10, 11, 12a, 13, 15 and 16;

FIG. 18b is the second in a series of two views illustrating a bar code exception recovery procedure;

FIG. 19 is a flow diagram illustrating still another recovery procedure for correcting an exception in the system shown in FIGS. 3 and 4 that relates to a sheet position error, and relates to recovery decisions called for in FIGS. 6, 7, 8, 10, 13, 15 and 16;

FIG. 20a is the first of a series of two views, and is a flow diagram illustrating still another recovery procedure for correcting an exception in the system of FIGS. 3 and 4, such exception relating to incomplete processing of a complete nest of parts, the recovery procedure further relating to recovery decisions called for in FIGS. 6 and 7;

FIG. 20b is the second in a series of two views illustrating a recovery procedure for correcting an exception that relates to incomplete processing of a complete nest of parts; and

FIG. 21 is a schematic diagram illustrating the interface protocols between the invention and the system shown in FIG. 3.

BEST MODE FOR CARRYING OUT THE INVENTION

I. Introduction

Referring now to the drawings, and first to FIG. 1, shown generally at 10 is an exception processor system (hereinafter "exception processor") in accordance with the invention. The exception processor 10 is shown operatively connected to a manufacturing facility (hereinafter "manufacturing cell") which is indicated generally at 12.

Briefly, manufacturing cells typically include a plurality of work stations 14A-14H which cooperatively implement a plurality of preplanned steps and procedures for making parts. Typical work stations in modern-day manufacturing cells are often operated by computerized numerical control "CNC" methods or programmable logic controllers (PLCs). Use of these acronyms is shown in FIGS. 1 and 2, for example.

A cell controller 16, which is typically a computer, interfaces to each work station 14A-14H, and to cell sensors 18. The cell controller 16 coordinates work station activities, including verification of process techniques, and modifies machine motions at individual work stations. Any action taken by the cell controller 16 is generally taken as a result of preprogrammed procedures in combination with data provided to the controller 16 by any one or more work stations 14A-14H, or by cell sensors 18.

The above represents a generalized or generic description of manufacturing cells, and automated cells in particular. A more specific example of a manufacturing cell is illustrated at 12 in FIG. 3, and will be described in greater detail later on, but in the context of using the exception processor 10.

The exception processor 10 is preferably implemented as a stand alone unit that is operatively connected to the cell controller 16, and to the cell sensors 18. It includes a conventional microprocessor-based computer 20 that has a user interface 22 consisting of a CRT monitor 22A and a keyboard 22B. The preferable hardware components making up the computer and interface 20, 22 are described in a later section of this description.

The computer 20 is programmed to implement a certain recovery plan or procedure 24 in response to an exception alarm received from the cell controller 16. Any given recovery procedure 24 is taken from a library of recovery procedures, and is selected on the basis of the type of exception indicated by the cell controller 16. In some cases, an operator may note something going wrong with cell operation even though an alarm 10 has not been initiated by the cell controller 16. In such cases, the operator may directly initiate a heuristic recovery procedure by requesting advice from the exception processor 10.

Recovery procedures in accordance with the invention involve a combination of deterministic and heuristic procedures designed to automatically determine the cause of a given exception, and to propose a solution as indicated at 26, 28, 36 in FIG. 1. Recovery procedures are at least partially deterministic in the sense that they set forth a logical series of potential corrective steps to be implemented by the exception processor 10, until determination of cause, and solution to the exception, is achieved. Part of such task may require obtaining further data from the cell controller 16, as indicated schematically at 30, or from the cell sensors 18, as indicated schematically 32. It may also require the operator to input certain information requested or prompted by the exception processor 10, as indicated at 34. The latter may involve requesting the operator to manually perform or implement certain tasks as part of a recovery procedure

There are generally three types of problems necessitating recovery. One is a process problem involving some fault in processing a part. The second is an equipment failure, such as a machine breakdown. The third is a material problem involving a defect in a piece of material that is being processed by the cell.

Recovery procedures 24 are obtained from vendor-specified recovery procedures, or they are developed by interviewing experienced operators and maintenance people having knowledge as to the causes of exceptions or breakdowns in automated cells, and how such exceptions are typically fixed. From these interviews, programmable recovery procedures are developed.

By way of example, a person having experience maintaining an automated cell or similar system will typically follow a certain logical path in diagnosing the root cause of an exception. Such path was developed from his or her past experiences in solving similar exceptions. Experience-based manual diagnostic procedures are very similar in most respects to flow chart diagrams that outline the logic for software programs. Hence, they are well-suited for adaptation to a programmable computer, where the computer is programmed to follow predefined steps that are outlined or developed from past maintenance experience, and thereby automatically determines the cause of exceptions. Once the cause has been identified, the corrective procedure required to solve the exception is often apparent.

Recovery procedures 24 utilize both forward and backward chaining methods. As the skilled person would know, backward chaining involves working backward from a known conclusion to determine whether or not certain actual parameters match assumed parameters that would justify the conclusion. Forward chaining involves working forward from known parameters or values to reach all conclusions that are logically consistent with the known parameters.

As mentioned above, exception processor recovery procedures 24 involve a combination of deterministic and heuristic methods. Deterministic methods are very much based on known cause and effect relationships. Heuristic methods are based on learned experience. They may or may not come into play in combination with determinsitic methods.

In some cases, purely deterministic recovery plans are inadequate to correct exceptions, but heuristic plans may provide the solution. In other words, sometimes a problem arises whose cure is known on the basis of prior learned experience, although why the cure works is unknown. Maintenance people that have experience with automated facilities over an extended period of time will have their own personal memories of these kinds of situations which, in some cases, occurred repeatedly. This kind of information is stored in the exception processor by way of interviewing maintenance people, and by way of the exception processor having its own unique capability of making use of operator experiences.

Heuristic experience is implemented in exception processor recovery procedures 24 as suggestions to the user as to possible courses of correction action, sometimes after a deterministic plan has been tried and failed. Other times, heuristic procedures are implemented either before or during some portion of a deterministic one.

The exception processor 10 also has the capability of avoiding future exceptions via ongoing statistic analysis or "statistical process control" of cell operation, as indicated at 38. The exception processor 10 logs the time, data and nature of each exception, and the corrective procedure or solution implemented to solve each exception. This is stored in a computerized knowledge base 40. The exception processor 10 continuously requests certain data updates from the cell controller 16, and from the cell's sensors 18. This information is also stored in the processor's knowledge base 40.

The exception processor 10 uses the data in its knowledge base 40 to continuously analyze cell operation, and to statistically determine past and future trends in cell operation. Over a period of time, the exception processor 10 may determined, statistically, that one type of exception occurs far more frequently at a given work station 14A-14H than others, suggesting an ongoing problem that may or may not be permanently fixed, depending on its nature. If it is fixable, then the overall result is that the exception processor 10 has functioned to improve the efficiency and quality of overall cell operation. If it cannot be permanently fixed, then at least the exception processor will have identified the problem in a way so as to anticipate like exceptions in the future, including the time they are likely to occur, so that preventative action can be taken and down-time minimized.

Along the same line, certain cell processes can be continuously monitored by the exception processor 10 in order to determine if each process was completed within a certain predefined design tolerance. Although a given process may be within tolerance, there may be a statistical deviation within the range of what is acceptable that indicates a situation progressing toward an exception.

A more specific illustrative example, and one that is directed to the floor panel system shown in FIG. 4, is that at some point in the cell a measurable quantity of adhesive is applied to a part during an "insert installation" procedure. The specific item measured is the diameter of a ring of adhesive, which is acceptable if it falls within a certain range. The exception processor determines, on a statistical basis, whether or not the ring diameter is increasing or decreasing from one application to the next, albeit all may remain within an acceptable tolerance. Such statistical trending serves to indicate that the device applying the adhesive is possibly going out of adjustment or otherwise has a mechanical problem, and enables fixing the problem before the size of the adhesive exceeds tolerance, thus avoiding an exception.

Referring now to FIG. 2, the functional relationship and interconnections between the exception processor 10 and the manufacturing cell 12 will now be explained in greater detail. As the skilled person would know, automated manufacturing cells have a generic architecture. As will be explained in greater detail below, the programmable aspects of the exception processor are implemented via "object-oriented" programming, which is a form or technique of programming that would be familiar to those having a high level of expertise in computer science or computer programming. This form of programming makes the exception processor 10 easy to adapt to virtually any manufacturing cell 12.

For example, the cell 12 illustrated in FIG. 3 is an automated floor panel fabrication and assembly cell operated by a VAX 3800 (trademark) cell controller. In order to operatively connect the exception processor 10 to such controller, it is necessary to interface the exception processor with the controller via an applications software package 42. As the skilled person would know, this involves simply interfacing the data protocols of the exception processor 10 with that of the of cell controller 16.

FIG. 2 illustrates the functional communication productivity between the exception processor 10 and the cell controller 16. The cell controller outputs alarms to the exception processor parser 44, as schematically indicated at 46. This informs the exception processor that an exception has occurred somewhere in the operation of cell 12. The exception processor acknowledges the alarm, as shown at 48, and takes over control of the cell 12.

Every alarm has its own unique code corresponding to a particular exception. The exception processor 10 fetches a recovery procedure 24 from its memory that is directly responsive to the particular alarm code indicated by the cell controller 16. In some cases, the cell controller 10 outputs a collection of alarms, all of which have a common root cause. In such case, the exception processor 10 recognizes a pattern that caused the alarms, and calls up a recovery procedure designed to address the root cause instead of calling up separate recovery procedures for each alarm.

As it implements a recovery procedure 24, part of the procedure may require a further request of data from the cell controller 16 as shown at 50, 52, 54. Periodic data updates may also be requested, as indicated at 60, as part of statistical monitoring. Data requests and data transmissions are communicated to and from the cell controller 16 via the applications program 42.

Depending on the particular recovery procedure, the exception processor may also request an individual work station to run its own diagnostic tests as shown at 54, 56, 58. Many PLCs or work stations have built-in diagnostic equipment enabling certain self-tests for hardware or software faults which, in some cases, can create exceptions to cell operation. Depending on the nature of the recovery procedure, self-tests may be requested and the resulting diagnosis may provide the cause of the exception, and suggest the solution required.

Having provided the above general description as to how the exception processor 10 works, attention is now directed to FIGS. 3-21 which illustrate implementation of the exception processor 10 in an automated floor panel fabrication and assembly cell 12 that has been implemented by The Boeing Company at its Spokane, Wash. manufacturing plant.

II. The Floor Panel Manufacturing Cell

The cell 12 shown in FIG. 3 is a system designed to produce floor panel parts for aircraft. Briefly, floor panel parts have a sandwich construction consisting of a Nomex (trademark) or aluminum honeycomb core, with thin sheets of fiberglass or graphite on each side of the core. The cell has a plurality of different work stations that are equivalent to the work stations 14A-14H indicated in FIG. 1. The first station is a material loading station, indicated generally at 62. There, a person loads sheets 64 onto a jig 66. Each sheet 64 is tilted upwardly into a vertical position. A set of three buffer carts 68, 70, 72 move transversely along rails 74, whose positions are obtained from encoders on the cart, to move sheets 64 from one work station to the next.

Sheets are automatically taken from cart 68, 70, 72 by frames 77 at each work station that move longitudinally along tracks 78. Such frames 77 clamp onto a particular sheet in accordance with all controller instructions when it is transferred from a cart 68, 70, 72. With the exception of manual work stations, each work station is a conventional X,Y station operable to move a sheet in two directions adjacent a fixed spindle.

The middle buffer cart 70 first transfers a sheet 64 from the material loading station 62 to a partial router work station, indicated generally at 80. Each cart has a plurality of numbered slots for holding more than one sheet 64, and the cell controller 66 keeps track of which sheet is in what slot. After partial routing, the sheet 64 is moved to a cleaning station 82 and then brought back to a potting station 76 for applying a potting compound. It is then sent to a curing station 86 followed by a planing station 76 to level the potting. It then goes back to the partial router station 80 for drilling holes. After drilling, the sheet 64 is moved to another station 84 for the application of adhesive and installation of inserts. Thereafter, the sheet 64 is moved into a curing station 86. This is followed by marking parts on the sheet as shown at 88, and final routing, as shown at 90. After final routing, parts are manually removed from the cell 12 as shown at 92.

Referring now to FIG. 4, there is shown a flow chart diagram of the FIG. 3 cell described above. FIGS. 5-17 provide a more detailed description as to the nature of the procedures which take place at each work station. In order to better facilitate understanding of how FIG. 4 relates with subsequent Figs., conventional reference numeral indications have been dispensed with. Instead, each block shown in FIG. 4 is identified by a particular numeral. For example, the "material loading" station, which is generally indicated by reference numeral 62 in FIG. 3, is indicated by number "10" in FIG. 4, and will hereinafter be designated as follows: "(10)". Other work stations or procedures in the FIG. 3 cell are indicated by sequential numbers, beginning with (10) at the material loading station, and ending with (22) at the final router work station (which is indicated generally at 90 in FIG. 3).

Referring now to FIG. 5, the various steps and procedures that take place at the material loading station (10) are illustrated. First, the operator logs a sheet onto a CRT which is dedicated to that particular work station. The cell controller 16 downloads information into the CRT, followed by the operator obtaining and inspecting the sheet (10.3, 10.4). If the sheet is acceptable, it is then loaded into the work station or work cell (10.7). If it is rejected, then it is either scrapped or reworked (10.5). If scrapped, then the sheet goes to salvage (10.6). If repairable (10.8), it is reworked and then loaded into the work cell (10.7). Loading the sheet is followed by a wand bar code procedure (10.9) which would be familiar to the skilled person.

The cell controller verifies the sheet by matching its bar code (10.10). If it matches, then the sheet is accepted and transferred from the material loading station to a buffer cart (10.12, 10.13), all of which is controlled by the cell controller 16. If the bar code does not match, then an exception occurs which is indicated to the exception processor 10 via an alarm from the cell controller (10.11).

FIGS. 18a and 18b illustrate a recovery procedure in accordance with the invention for a bar code error. First, the work station goes into feedhold, as indicated at 92 in FIG. 18a. This is followed by the work station sending an error message to the cell controller, as shown at 94. The cell controller displays an error message 96 that causes the operator to put the work station into a semi-automatic mode 98.

The exception processor prompts the operator to visually inspect the bar code label, as shown at 100. If it is readable, then the operator is asked to verify the person-readable code with cell data, as shown at 102. If there is no match, then the exception processor initiates a cell audit 104. If there is a match, then the operator is asked to conduct a wand test strip 106. If this test turns out positive, then the operator is instructed to clean the label, as shown at 108, and retry the bar code read 110. If this is successful, then the exception processor returns the work station to normal operation, as schematically indicated at 112, 114, 116.

If the wand test strip was negative, then the exception processor indicates to the operator that the bar code reader must be repaired. This is shown at 117 in FIG. 18a. Such instruction could in turn implement still another recovery procedure designed to assist the operator in repairing the bar code reader by himself, or in at least identifying the exact nature of the problem so that he knows what kind of maintenance expertise is required to fix the problem. Such other recovery procedure is not disclosed here.

In the event the bar code label is not readable, as shown at 118, the exception processor prompts the operator and requests information as to Whether or not the bar code label is dirty 120. If it is, the exception processor instructs the operator to clean it, as shown at 122. After cleaning, if the label is readable, then the exception processor instructs the operator to jog and retry the read which, if successful, returns the cell to normal operation, as shown at 124, 126.

In the event the label either was not dirty, or the jog and retry operation did not work, the exception processor instructs the operator and/or cell controller to return the sheet to the manual work station, as shown at 128. There, a determination is made as to whether or not the label can be reworked, as shown at 130. If not, the part is scrapped as shown at 132. If it can be reworked, then a new label is printed, and is inspected and applied if it is good. These procedures are respectively indicated at 134, 136, 138 in FIG. 18b. Thereafter, the cell is returned to normal operation.

The various steps and procedures involved in the "ditch" step (12) in FIG. 4 are illustrated in FIG. 6. This procedure occurs at the partial routing and drilling work station 80 shown in FIG. 3. First, a sheet is transferred from a buffer cart (12.1). Then, a sliding frame 77 of the cell indexes and clamps the sheet (12.2). Indexing and clamping is followed by a verification of sheet position (12.3). if the sheet is not in proper position, then a sheet position exception occurs and the cell controller sends an alarm to the exception processor 10 calling for the implementation of a recovery decision (12.4).

A preferred recovery procedure for such exception is illustrated in FIG. 19. There, the material handling controller 144 first informs the cell controller that a material handling error exists, as indicated at 146. The operator or exception processor acknowledges such error, as indicated at 148, and takes over control of the work station, as indicated at 150. The exception processor first instructs that the sheet be moved from the work station back to a buffer cart, as shown at 152. If this is successful, then the sheet is returned to the work station, as shown at 154, and the sheet position is once again verified. If verification is a success, then the exception processor returns control to the work station as indicated at 156.

If the sheet was not successfully moved back from the work station to the buffer cart then the exception processor instructs the operator to take the work station off line, as shown at 158. The operator is instructed to visually inspect the problem, as shown at 160, and to determine whether or not he or she can fix it. If "yes," then a fix is attempted, as indicated at 162. If the fix is successful, the work station is returned to automatic, and the error is cleared, as shown at 164.

A visual inspection may indicate that the sheet was damaged or is not otherwise in good condition. If so, then it is scrapped, as shown at 166. If it is in good condition, then operation of the work station is returned to the cell controller.

If the operator determines that he or she cannot fix the problem, then a maintenance person is called in, as shown at 168. One advantage to the exception processor 10 described here is that the operator will be pre-informed by the recovery procedure as to the nature of the problem, and can communicate that to the appropriate maintenance person. This makes it easier for the maintenance person to anticipate what tools and/or parts may be required to make the fix, and results in lowering the down-time of the work station and/or the cell as a whole.

Referring again to FIG. 6, after the cell has recovered from the sheet position error, it then verifies the bar code (12.5). If the bar code is rejected, then a bar code exception occurs and the recovery plan previously described (see FIGS. 18a and 18b) is implemented. If the bar code is accepted, then the work station routes the ditch per MCD (machine control data) parameters, and determines whether or not all parts in the nest are complete (12.8). If not, then still another exception is created which calls for the recovery plan illustrated in FIGS. 20a and 20b. There, the exception processor queries the operator and/or the cell controller as to whether or not the exception may be attributed to a machine fault 170, tool outage 172, or bad MCD data 174.

If the exception appears to be a machine fault, then the exception processor requests whether or not the operator can correct the problem. If not, the operator is instructed to call maintenance to correct the error, as shown at 176, 178 in FIG. 20a. If the part was not damaged after correction, then normal operation is restored, as shown at 180.

If the operator was capable of fixing the problem, then the same procedure is followed. However, if the part is determined to be damaged, then a decision is made as to whether or not the part can be skipped, which is indicated at 182. If the part can be skipped, then the work station is instructed to skip it, as shown at 184, and normal cell operation is restored, as shown at 186. If the part can't be skipped, then the sheet is removed and scrapped as shown at 188, 190.

In the case of bad MCD parameters, after responding to the alarm, the operator makes the determination as to whether or not he or she can correct the error. If not, then a determination is made as to whether or not the part can be skipped, and the above-described procedure is followed. If the error can be corrected, then the operator corrects it and returns the work station to normal operation as shown at 192, 194.

in the case of tool outage, as indicated at 172, the exception processor responds to the alarm and instructs the operator to replenish tools, as indicated at 196, 198. Thereafter, the exception processor returns the cell to normal operation.

Referring again to FIG. 6, after nesting has been completed (12.8), the sheet is unclamped (12.10) and transferred to a buffer cart (12.11).

FIGS. 7-17 illustrate the remaining manufacturing procedures at each work station in FIG. 3, and where further possible recovery decisions or procedures should be implemented. All of the possible recovery decisions are not disclosed here as they are not necessary in order to understand the invention, and therefore add nothing to the above description.

With the above description, it should be clear how the exception processor 10 is implemented in the manufacturing cell 12 shown in FIG. 3. The exception processor 10 is designed to evolve and be improved upon over time. That is to say, the manner of implementing automated recovery procedures will change as further improvements are made. Therefore, in order to implement the invention, it is not necessary to describe each and every recovery procedure that Was, or ever will be, developed for use in the FIG. 3 cell.

III. Hardware Requirements

As mentioned above, the cell controller 16 of the manufacturing cell 12 shown in FIG. 3 is a VAX 3800 computer. It is connected to individual cell Work stations via an Ethernet communications network, Which utilizes TCP/IP protocols. Both Ethernet and TCP/IP protocols would be familiar to the skilled person.

FIG. 21 illustrates the interface protocols between the exception processor and the manufacturing cell. The cell controller interfaces with cell work stations via an Ethernet type of network. As mentioned above, the exception processor 10 utilizes object-oriented code that interfaces with the TCP/IP protocols of the Ethernet. The user interface or operator station 22a interfaces with the exception processor via RS/232 protocols. An advantage of the exception processor 10 is that it has the capability for off-line statistical analysis which is indicated schematically at 200 in FIG. 21. This is designated by the legend "SPC Server" where the acronym "SPC" stands for statistical process control.

Suitable hardware required to implement all aspects of the invention as described above are listed in Table I below.

TABLE I______________________________________1. Exception Processor:1 IMB Compatible 486-25 Mhz with 8 Mb 32-bit RAM, tower case, 200 Mb IDE hard drive and controller, 256 kb CPU cache, 1.2 Mb and 1.44 Mb floppy drives, ultra keyboard, (MS) DOS 4.01, Weitek 4167 co- processor socket, 14" amber monitor, Hercules compatible mono graphics card, clock/calendar, 5-year battery backup, 1 parallel port and 2 serial ports, and ground transportation.1 Upgrade hard drive and controller to ESDI 150 Mb.1 Specify 1.44 Mb floppy as drive A:1 Specify 1.2 Mb floppy as drive b:1 Specify Omnikey Ultra keyboard1 Specify tower case1 Specify (MS) DOS 4.01 Specify NO monitor1 32-bit memory expansion card with 8 Mb RAM (8 @ 1 Mb .times. 9 .times. 80 ns)1 HardCache/ESDI Controller with 4 Mb memory1 Ungermann-Bass PC-NIU/ex (ISA bus network adaptor card) with Ungermann-Bass Net/One work station software (TCP9191) and Ungermann-Bass TCP software (TCP1010)1 Microsoft Serial Mouse1 NEC Multisync Graphics Engine with 1 Mb memory (256 color)1 NEC 4D 16" Color Monitor2. Operator Station:1 IBM Compatible 486-25 Mhz with 8 Mb 32-bit RAM, tower case, 200 Mb IDE hard drive and controller, 256 kb CPU cache, 1.2 Mb and 1.44 Mb floppy drives, ultra keyboard, (MS) DOS 4.01, Weitek 4167 co- processor socket, 14" amber monitor, Hercules compatible mono graphics card, clock/calendar, 5-year battery backup, parallel port and 2 serial ports, and ground transportation.1 Upgrade hard drive and controller to ESDI 600 Mb.1 Add additional ESDI 600 Mb hard drive only.1 Specify 1.44 Mb floppy as drive A:1 Specify 1.2 Mb floppy as drive b:1 Specify Omnikey Ultra keyboard1 Specify tower case1 Specify (MS) DOS 4.011 Specify NO monitor1 32-bit memory expansion card with 8 Mb RAM (8 @ 1 Mb .times. 9 .times. 80 ns)1 HardCache/ESDI Controller with 4 Mb memory1 Ungermann-Bass PC-NIU/ex (ISA bus network adaptor card) with Ungermann-Bass Net/One work station software (TCP9191) and Ungermann-Bass TCP software (TCP1010)1 Microsoft Serial Mouse1 NEC Multisync Graphics Engine with 1 Mb memory (256 color)1 NEC 5D 19" Color Monitor3. SPC Server:1 IBM Compatible 486-25 Mhz with 8 Mb 32-bit RAM, tower case, 200 Mb IDE hard drive and controller, 256 kb CPU cache, 1.2 Mb and 1.44 Mb floppy drives, ultra keyboard, (MS) DOS 4.01, Weitek 4l67 co- processor socket, 14" amber monitor, Hercules compatible mono graphics card, clock/calendar, 5-yr. battery backup, 1 parallel port and 2 serial ports, and ground transportation.1 Upgrade hard drive and controller to ESDI 600 Mb.1 Add additional ESDI 600 Mb hard drive only.1 Specify 1.44 Mb floppy as drive A:1 Specify 1.2 Mb floppy as drive b:1 Specify Omnikey Ultra keyboard1 Specify tower case1 Specify (MS) DOS 4.011 Specify 14" amber monochrome monitor with hercules compatible graphics card1 32-bit memory expansion card with 8 Mb RAM (8 @ 1 Mb .times. 9 .times. 80 ns)1 HardCache/ESDI Controller with 4 Mb memory1 Ungermann-Bass PC-NIU/ex (ISA bus network adaptor card) with Ungermann-Bass Net/One server software (hardware from the TCP9191 bundle and software from the TCP9193 bundle) and Ungermann-Bass TCP software (TCP1010)4. Auxiliary Equipment:1 Hewlett Packard LaserJet III laser printer1 Pacific Data Products MB460 4 Mb memory board for HP LaserJet III1 Mountain FileSafe Model 7500 (525 Mb external tape drive with controller, cable, FileSafe software and FileTalk software)10 3M DC-6525 Data Cartridge Tape (525 Mb, .about.$30@)______________________________________

IV. Programming Requirements

The software architecture of the exception processor 10 is based on a technology called "object-oriented" programming. There are three features that distinguish object-oriented programming from traditional programming techniques.

The first feature is encapsulation. Each "object" has its own self-contained data processing capability and/or unique behavior that is independent from other objects. An object responds to a particular message according to its own internal behavior and internal data, and responds back with an answer.

The second feature that distinguishes object-oriented programming is its inheritance behavior. Hierarchies of objects can be easily developed where additional objects are defined by making successive refinements of properties of their parents. This capability makes it easy to generate a large number of objects with very specific behavior in a very fast manner.

The third feature that distinguishes object-oriented programming is its polymorphism. Different objects can respond to the same message uniquely. This makes it very easy to modify objects without doing a tremendous amount of self-test.

Attached hereto is an appendix of object-oriented code which implements the exception processor 10 as described above. The particular language package is implemented by a system called "SmallTalk " and a particular dialect identified by the trademark SmallTalk)IV 286. These packages are provided by Digitalk, Inc. of Los Angeles, Calif. The particular objects provided in the appendix are as follows: (1) exception processor handler object; (2) operator screen object; (3) recovery log object; (4) communication interface object; (5) protocol object; (6) message object; and (7) statistical processing object.

The skilled person, having the above description, and the hardware and software identified above, including the code appendix appended hereto, would be able to implement the exception processor 10 in conjunction with the automated cell shown in FIG. 3, or in other cells, as the case may be. It is to be understood that the preceding description of the exception processor is not intended to be limited to the FIG. 3 cell in any way, shape or form. It is to be further appreciated that the preceding description sets forth the best mode for carrying out the invention as it is presently known. It is a certainty that further developments and improvements to the invention will be made over time. For this reason, the preceding description is not to be taken in a limiting sense. Instead, the spirit and scope of the invention is to be limited only by the patent claim or claims which follow, the interpretation of which is to be made in accordance with the well-established doctrines of patent claim interpretation. ##SPC1##

Claims

1. For use in a manufacturing facility involving a plurality of preplanned steps and procedures for making parts, such steps and procedures being implemented by a plurality of individual work stations whose respective operations are coordinated by a cell controller, an exception processing system for correcting unplanned events during the operation of said facility, comprising:

a computer connected to said cell controller to send and receive information to and from the same, said computer being programmed to receive an alarm from said controller indicating an unplanned event in the operation of said automated manufacturing facility, and to acknowledge to said controller receipt of said alarm, and being further programmed to selectively request certain data from said controller concerning said alarm, and to receive such data;

a set of recovery procedures, each procedure providing a preplanned pattern of potential corrective actions for returning said manufacturing facility to normal operation after a given unplanned event, wherein said computer is further programmed to access and implement, by determining a cause of an unplanned event and implementing corrective actions, at least one of said recovery procedures in response to said cell controller alarm; and

a user interface connected to said computer, for enabling communications between said computer and a human operator, wherein said computer is still further programmed to prompt the input of certain information from said operator via said user interface, to assist said computer in implementing a recovery procedure.

2. For use in a manufacturing facility involving a plurality of preplanned steps and procedures for making parts, such steps and procedures being implemented by a plurality of individual work stations whose respective operations are coordinated by a cell controller, an exception processing system for correcting unplanned events during the operation of said facility, comprising:

a computer connected to said cell controller to send and receive information to and from the same, said computer being programmed to receive an alarm from said controller indicating an unplanned event in the operation of said automated manufacturing facility, and to acknowledge to said controller receipt of said alarm, and being further programmed to selectively request certain data from said controller concerning said alarm, and to receive such data;

a set of recovery procedures, each procedure providing a preplanned pattern of potential corrective actions for returning said manufacturing facility to normal operation after a given unplanned event, wherein said computer is further programmed to access and implement at least one of said recovery procedures in response to said cell controller alarm;

a user interface connected to said computer, for enabling communications between said computer and a human operator, wherein said computer is still further programmed to prompt the input of certain information from said operator via said user interface, to assist said computer in implementing a recovery procedure; and

statistical process control of said manufacturing facility,

wherein said computer is further programmed to monitor said automated manufacturing facility by periodically requesting information from said cell controller concerning part-making steps and procedures under the cell controller's control, said computer storing such information into a statistical knowledge base, for both monitoring operation of said facility and anticipating when a step and procedure in said facility is deviating from normal operation.

3. The exception processing system of claim 1, wherein said computer is further programmed to log into memory any unplanned event in the operation of said manufacturing facility, and the corrective action taken to correct such unplanned event.

4. For use in a manufacturing facility involving a plurality of preplanned steps and procedures for making parts, such steps and procedures being implemented by a plurality of individual work stations whose respective operations are coordinated by a cell controller, an exception processing system having a combination of heuristic, deterministic and statistical procedures for monitoring operation of said facility, and forth both correcting unplanned events during the operation of said facility and anticipating when a step and procedure in said facility is deviating from normal operation, said exception processing system comprising:

a computer connected to said cell controller to send and receive information to and from the same, said computer being programmed to receive an alarm from said controller indicating an unplanned event in the operation of said automated manufacturing facility, and to acknowledge to said controller receipt of said alarm, and being further programmed to periodically request information from said cell controller concerning part-making steps and procedures under the cell controller's control, said computer storing such information into a statistical knowledge base for monitoring operation of said facility;

a set of recovery procedures, each procedure providing a preplanned pattern of potential corrective actions to return said manufacturing facility to normal operation after a given unplanned event, wherein said computer is further programmed to access and implement at least one of said recovery procedures in response to said cell controller alarm; and

a user interface connected to said computer, for enabling communications between said computer and a human operator, wherein said computer is still further programmed to prompt the input of certain information from said operator via said user interface.

5. The exception processing system of claim 4, wherein said computer is programmed by an "object-oriented" program.

6. The exception processing system of claim 5, wherein said object-oriented program includes a communications interface object.

7. The exception processing system of claim 5, wherein said object-oriented program includes a protocol object.

8. The exception processing system of claim 5, wherein said object-oriented program includes a message object.

9. The exception processing system of claim 5, wherein said object-oriented program includes an operator screen object.

10. The exception processing system of claim 5, wherein said object-oriented program includes a recovery log object.

11. The exception processing system of claim 5, wherein said object-oriented program includes an exception processor handler object.

12. The exception processing system of claim 5, wherein said object-oriented program includes a statistical processing object.

13. The exception processing system of claim 6, wherein said communications interface object includes means for formulating and transmitting information in terms of TCP/IP interface protocols.

14. The exception processing system of claim 6, wherein said communications interface object includes means for formulating and transmitting information in terms of Ethernet interface protocols.

15. The exception processing system of claim 5, wherein said manufacturing facility is an automated manufacturing facility having sensors operable to monitor the steps and procedures of said facility, and wherein said computer is connected directly to said sensors, for receiving periodic data updates from them.