Distributed elevator shuttle dispatching

Abstract

 A plurality of elevator shuttles (1-10) are dispatched in a sequence at regular intervals which extend across the average round trip run time for all of the shuttles, by enabling a shuttle to run (101, in response to the shuttle being ready to run (96) following expiration of a dispatching interval (97) which equals the average run time of all the shuttles (91).

Description

[0001] This invention relates to dispatching elevator shuttles, such as in extremely tall buildings, in a dispersed manner so that elevators leave lower and upper lobbies served by the shuttles at regular intervals.

[0002] The sheer weight of the rope in the hoisting system of a conventional elevator limits their practical length of travel. To reach portions of tall buildings which exceed that limitation, it has been common to deliver passengers to sky lobbies, where the passengers walk on foot to other elevators which will take them higher in the building. However, the milling around of passengers is typically disorderly, and disrupts the steady flow of passengers upwardly or downwardly in the building. All of the passengers for upper floors of a building must travel upwardly through the lower floors of the building. Therefore, as buildings become higher, more arid more passengers must travel through the lower floors, requiring that more and more of the building be devoted to elevator hoistways (referred to as the "core" herein).
Reduction of the amount of core required to move adequate passengers to the upper reaches of a building requires increases in the effective usage of each elevator hoistway. For instance, the known double deck car doubled the number of passengers which could be moved during peak traffic thereby reducing the number of required hoistways by nearly half.
Suggestions for having multiple cabs moving in hoistways have included double slung systems in which a higher cab moves twice the distance of a lower cab due to a roping ratio, and elevators powered by linear induction motors (LIMs) on the sidewalls of the hoistways, thereby eliminating the need for roping. However, the double slung systems are not practical for shuttling passengers to sky lobbies in very tall buildings, and the LIMs are not yet practical, principally because, without a counterweight, motor components and power consumption would be prohibitively large.

[0003] In order to reach longer distances, an elevator cab may be moved in a first car frame in a first hoistway, from the ground floor up to a transfer floor, moved horizontally into a second elevator car frame to a second hoistway, and moved therein upwardly in the building, and so forth. Since the loading and unloading of passengers takes considerable time, in contrast with high speed express runs of elevators, another way to increase hoistway utilization, thereby decreasing core requirements, includes moving the elevator cab out of the hoistway for unloading and loading, as is described in our European patent applications claiming priority of U.S. patent applications Serial Nos. 08/564,534 and 08/565,606 and filed contemporaneously herewith.

[0004] In extremely tall buildings, which may be several hundred stories or floors high, elevator shuttles of any of the foregoing types consume significant time in making a round trip run, such as from an entry lobby back to the entry lobby.
In conventional elevators, timers typically determine when elevator car doors begin to close, thereby beginning the process of commencing a run. However, passengers interrupt by means of pressing a "door open" switch on the car operating panel, or by activating the door safety switch. In a large bank of elevators, with long round trip run times, controlled by door closings, the commencement of runs from one car to the next is often quite irregular.

[0005] Objects of the present invention include controlling the dispatching of elevator shuttles, such as in very large buildings, in a manner to distribute the departure times in a regular fashion, evenly, so that a shuttle will leave a lobby on a regular recurring basis.

[0006] According to the present invention, each elevator car of a bank of shuttle elevators is allowed to run (that is, to undertake response to a motion controller) only when the car is ready to run and a distributed dispatching controller (dispatcher) signifies that the time for departure, under a scheme of distributed departures, has arrived. According to the invention, if the car is ready to run before the dispatcher indicates that the car should commence a run, the car will wait until the dispatcher signifies the run can commence. If the dispatcher signifies that the run can commence prior to the car being ready, the car will commence its run as soon thereafter as it is ready.

[0007] According further to the invention, distributed dispatching of a plurality of shuttle elevators is accomplished by means of elapsed time intervals, the beginning of which is established each time that a car commences a run; any delay in commencement of the run of any elevator car causes a corresponding subsequent delay in the running of all cars, but not in the frequency of dispatching.

[0008] The invention results in a small phase delay in the departure of all cars subsequent to a delay in the departure of one car, rather than disrupting the distributed departure of all of the cars. By means of the invention, the bunching up of departures, interleaved with sparse departures, cannot occur.

[0009] The invention greatly improves the effective handling of traffic of a bank of shuttle elevators, particularly in very tall buildings.

[0010] Other objects, features and advantages of the present invention will become more apparent in the light of the following detailed description of exemplary embodiments thereof, as illustrated in the accompanying drawings, in which:

[0011] Fig. 1 is a simplified, stylized block diagram of a bank of elevators with which the present invention may be used.

[0012] Fig. 2 is a logic flow diagram of a portion of a car control routine, of a conventional elevator system known to the art, adapted for use with the present invention.

[0013] Fig. 3 is a logic flow diagram of a portion of a car control routine for an elevator shuttle having off-hoistway loading and unloading, adapted for use with the present invention.

[0014] Fig. 4 is a logic flow diagram of a low lobby dispatch routine according to the present invention.

[0015] Fig. 5 is a timing diagram illustrating the invention.

[0016] Referring now to Fig. 1, a bank of elevator shuttles may comprise elevators numbered 1-10, arranged in two tiers facing each other, so that each time an elevator of a given number if dispatched, the next elevator to leave will be an adjacent elevator (either in the same tier or across the lobby). The elevators depicted in Fig. 1 may be conventional elevators known to the prior art; conventional double deck elevators known to the prior art; elevators in which a cab is transferred from a lower shaft to an adjacent upper shaft so that passengers can travel more than the aforementioned rope-limited distance without having to leave the elevator cab, or elevators having cabs that are removed to landings for off-hoistway loading and unloading thereof, as in the aforementioned co-pending patent applications. Or, the elevators depicted in the bank of elevators in Fig. 1 might be other sorts of elevators, such being irrelevant to the present invention. Therefore, as used herein, the term "elevator car" may mean a conventional elevator car with a cab fixed thereon, or an elevator car frame carrying a horizontally moveable cab. The term includes the motion control, hoisting motor, ropes, brake, and so forth.

[0017] The manner in which a conventional car establishes its running condition, and submits itself to control by a motion controller, is modified to take advantage of the present invention. Various functions such as responding to car calls or to a hall call dispatcher, or the like might normally be performed. However, in a shuttle elevator, it is anticipated in the example herein that all passengers travel only between a lower lobby and a middle lobby or an upper lobby, with no stops in between. In such a case, a typical routine will simply sense the expiration of a suitable time for loading the elevator, close the doors, recognize that the elevator must go up or down to reach the opposite lobby floor, and set direction. Each shuttle has one or more car controllers as in Figs. 2 and 3; an exemplary shuttle is shuttle S. The fragment of a car control program illustrated in Fig. 2 is reached through an entry point 21 and a test 23 determines if direction has been set for the car or not. If not, a negative result of test 23 causes other programming to be reached through a return point 24. In subsequent passes through the routine of Fig. 2, eventually direction will have been established for the car of shuttle S, so an affirmative result of test 23 will reach a step 25 to issue a door close command to close the doors of the car. Then a test 26 determines if the doors are fully closed or not. Initially, they will not be, so a negative result of test 26 reaches other programming through the return point 24. In a subsequent pass through the routine of Fig. 2, test 23 will be affirmative and eventually the doors will be fully closed so that test 26 will be affirmative. This reaches a step 27 to set a "run ready for shuttle S" flag, which indicates to the dispatcher that shuttle S is ready to run. Then a test 32 determines if the dispatcher has sent back a "enable run for shuttle S" flag, or not. Initially, it will not unless the shuttle has been delayed for some reason, such as passengers holding the doors open for a prolonged period of time. If shuttle S is not yet enabled, a negative result of test 32 reaches other programming through the return point 24. Eventually, test 32 will be affirmative reaching a step 33 which sets run for this car (which is the car of shuttle S), a step 34 which resets the "run ready for shuttle S" flag, and a step 35 which resets the "enable run shuttle S" flag.

[0018] According to the embodiment, dispatching is dependent upon the length of time it takes for the shuttles to make a round trip, and the number of shuttles in the group to which the traffic may be distributed. In order to determine the average run time for the shuttles, each shuttle determines its own run time by initiating a run timer whenever the run command is set in step 33, and then recording the elapsed time therein the next time that the run command is set. However, the run command is set when a car leaves the lower lobby as well as when a car leaves an upper lobby, and therefore a test 33 determines when the direction is up so as to have a step 34 record the setting of the run timer and a step 35 reinitiate the run timer only when the car is leaving the lobby to make a shuttle run, indicated by direction being up. Then, other programming is reverted to through the return point 24.

[0019] If the invention is to be used in an embodiment wherein passengers are loaded and unloaded off-hoistway, the ability to start a run is based upon a cab being on a car frame ready to go, rather than upon the closure of the doors which occurs earlier in the process. In the present example, it is considered that the beginning of a run is conditioned upon a horizontally moveable cab having been moved into the car frame of a car and locked during a period of time when that car could be designated to start a run by the dispatcher of the invention.

[0020] In Fig. 3, routines for a single elevator shuttle may include door control routines 38, 39 for each of the cabs in a shuttle. For instance, in the aforementioned application claiming priority of U.S. Serial No. 08/564,534, there are two elevator cars, each with a car frame in adjacent overlapping hoistways, and five horizontally moveable cabs, serving two landings on each of three different levels; ten cabs are used in a double deck embodiment. A transfer routine 40 may be performed to exchange a recently arrived cab on the car frame with one which has been loaded on a landing. In the aforementioned application, the transfer routine 40 will be causing the transfer of three or four cabs at a time. In the application claiming priority of US Serial No. 08/565,606, the transfer routine 40 will be controlling the horizontal movement of two cabs at a time in a single deck embodiment, and four cabs at a time in a double deck embodiment, between a single car frame and the landings.

[0021] A car control routine for shuttle S 42 is then reached through an entry point 43. The final steps of that routine, before allowing the car or cars of the shuttle to begin a run, include the portion of the routine which is reached only after a test 44 determines that there is a cab (or cabs) locked in the car (or cars) of shuttle S. Before all cabs are firmly locked to respective cars, a negative result of test 44, in sequential passes through the routine of Fig. 3, will reach other programming through a return point 45. But once all cabs are locked in the cars of the shuttle, then direction for the car will be set to cause it to leave one lobby and run toward another in a portion 46 of the routine, after which a step 47 will set a "run ready shuttle S" flag which is sent to the dispatcher to inform the dispatcher that this shuttle is ready to go. Then a test 50 determines if the dispatcher has sent back an "enable run shuttle S" flag or not. In a typical case, it will not so other programming is reached through the return point 45. Eventually, an affirmative result of test 50 reaches a step 51 to set the run command for shuttle S, a step 52 to reset the "run ready shuttle S flag" and a step 53 to rest the "enable run shuttle S flag". Now the cars of shuttle S are under control of their respective motion controllers. A test 54 determines if the direction set for the low car is up. If it is, this is taken as the beginning of a run so a step 55 registers the run time as whatever is set in the run timer and a step 56 reinitiates the run timer to time the run which is about to begin. If a shuttle that moves cars onto landings for off-hoistway loading and unloading has only a single car in it, test 54 would relate to that car alone. In a case where there is an upper elevator and a lower elevator dispatched synchronously with each other, test 54 need only sense the direction of one of the cars. Then other programming is reverted to through the return point 45.

[0022] In Fig. 4, the upper half of a low lobby dispatch routine simply determines how many cars are in the group and divides a recent average round trip time for the shuttles by the number of shuttles, to determine how often the shuttles should be dispatched to have shuttles dispatched on a regular basis. The first part of the routine of Fig. 4 simply determines how many shuttles, T, are currently in the dispatching group, and a new dispatching interval is determined. Then, the dispatcher waits until both (a) the shuttle is ready to run and (b) the dispatching interval has expired, after which the dispatching interval is reinitiated for the next shuttle in turn.

[0023] In Fig. 4, the routine is reached through an entry point 80 and a first step 81 resets the total number of cars in the group, T, to zero, and a step 82 sets a shuttle pointer, n, to point to the first shuttle in the group. In Fig. 1, the total number of shuttles, N, which may be operating in the group is ten. Then a total run time accumulator, TOTL, is reset to zero in a step 83. A test 85 determines if the shuttle currently being pointed to by the n pointer is in the dispatching group or not. If it is, the T counter is incremented in a step 86 and the total run time is incremented in step 87 by the run time of shuttle n (from steps 34 and 35). If shuttle n is not in the group, steps 86 and 87 are bypassed. Then, the n pointer is advanced in a step 87 and a test 88 determines if the n pointer is pointing to one more than the total number of cars, N, that could be in the group, indicating that all cars have been tested. If not, a negative result of test 88 reverts the routine to test 85 to see if the next car in turn is in the group and increment T and the total run time if it is. When all of the possible cars have been tested, an affirmative result of test 88 reaches a step 91 to generate an average run time by taking the total count of step 87 and dividing it by the number of shuttles, T, that contributed thereto. And then the average run time is divided by T to establish a dispatch interval which evenly divides the average run time. Then a test 92 is reached to see if a shuttle currently being pointed to by an S pointer, which identifies the shuttles in sequence for dispatching as described hereinbefore, is pointing to a shuttle which is in the group or not, as described more fully below. If the shuttle being pointed to by the S pointer is not in the group, a negative result of test 92 reaches a step 93 to advance the S pointer to the next shuttle in turn. Thus, if a shuttle, such as shuttle 8 in Fig. 1 is not within the group at this time, after dispatching shuttle 7, the S pointer would point to shuttle 8. But when the time comes for dispatching the next shuttle in turn, shuttle 8 would not be counted in the total number of cars, T, its round trip run time would not be incorporated into the dispatching interval, and it would not be dispatched. When the S pointer points to a shuttle in the group, an affirmative result of test 92 reaches a pair of tests 96, 97 to determine if shuttle S is ready to run and the interval since the last shuttle was dispatched has timed out. If either of these have not occurred, negative results cause other programming to be reverted to through a return point 98. But when shuttle S is ready to run and the dispatching interval has expired, affirmative results of tests 96 and 97 reach a step 101 to set the "enable run S" flag, for use in either of Figs. 2 and 3. A step 102 initializes the interval timer with the interval determined in step 91, and the S pointer is advanced in a step 103, after which other programming is reverted to through the return point 98.

[0024] In an embodiment in which a single elevator comprises a shuttle, or where a shuttle comprises more than two elevators, and the lobby elevators are dispatched separately, tests and steps 96-103 of Fig. 4 may be used with the low lobby dispatch routine as in Fig. 2 or Fig. 3 as appropriate, to dispatch shuttles from the upper lobby, in an obvious fashion. When used in a two elevator system wherein the elevators are synchronously dispatched, only the low lobby routine of Fig. 5 need be used, since the upper elevator in the shuttle is always dispatched simultaneously with the lower elevator.

[0025] In the embodiment of the invention shown, shuttles can be added to and removed from the dispatching group, at will without disrupting the dispatching process. A shuttle being added to the dispatching group can have its run time initiated as a non-zero run time of any other shuttle in the system, or it may retain the run time last determined for it in either of steps 34 or 35, retained from the last time it operated in the dispatching group.

[0026] The invention is basically self-leveling: if the door times allotted for the cars to permit passengers to unload and load are too short, the run times will be increased until adequate time is provided. Although not treated herein, the time allocated for doors to remain open is readily adjustable, as is known in the art.

[0027] Fig. 5 is a simplified timing diagram illustrating distributed dispatching in accordance with the invention. Therein, a plurality of time periods 1-31 represent just more than a complete round trip for any one shuttle. As an example, shuttle 1 is shown being dispatched from an upper landing at the beginning of the first time period, reaching the lower landing at the beginning of the 11th time period, and being dispatched to return to the upper landing at the beginning of the 12th time period. Shuttle 2 is simply two time periods behind shuttle 1. The time periods 12 and 13 represent a dispatching interval, shown by arrow 110. In the example shown in Fig. 5, shuttle 3 has difficulty closing its doors and preparing to run at the beginning of the 16th period. Thus, although an arrow 111 shows that the dispatching interval has not changed, shuttle 3 does not in fact run until nearly half-way through the 16th period. Since that is when the dispatching interval is initiated in step 102 of Fig. 4, the dispatching interval illustrated by the arrow 112 is the same as that illustrated by the arrows 110 and 111. Shuttle 4 has no difficulty, so it will be-dispatched as soon as the dispatching interval expires. Notice that shuttle 4 will have a car at a lower landing for a longer period of time, since it was dispatched toward the lower landing prior to any delay in the system, at the start of time period 7. On the other hand, notice that in time period 27, at the upper landing, car 3 is ready at a time indicated by an arrow 113, which, even though it arrived nearly half a cycle late, is still after the as-yet unadjusted dispatching interval indicated by the arrow 114. Therefore, when car 3 is dispatched at the point of the arrow 113, the dispatching interval indicated by the arrow 115 is established.

[0028] Notice that car 4 left in response to the end of the now adjusted timing interval indicated by the arrow 112 nearly in the middle of time period 18. However, at the upper landing, it had sufficient time within time period 28 and a small portion of time period 29 so as to be ready to leave when the dispatching interval indicated by the arrow 115 expired. This means that from now on, car 4 is back on a totally regular cycle, having fully assimilated the delay incurred by car 3 in time period 16. In Fig. 5, the remaining shuttles 5-10 respond essentially the same as shuttle 4 as indicated by the arrows 116 and 117, and quickly get back into normal operation.

[0029] The illustration of Fig. 5 is for a system having a separate low lobby dispatcher and high lobby dispatcher. In the case where upper and lower lobbies are dispatched in synchronism, and only one lobby dispatch routine such as that shown in Fig. 4 is used, the delay of the 16th period caused by car 3 would immediately have affected, say, shuttle 10 since the shuttle cars in such case are dispatched together. In such a system, a departure from an upper lobby being interleaved with a departure from a lower lobby as in Fig. 5 cannot, of course, occur.

[0030] Thus, although the invention has been shown and described with respect to exemplary embodiments thereof, it should be understood by those skilled in the art that the foregoing and various other changes, omissions and additions may be made therein and thereto, without departing from the scope of the invention, which is defined by the claims.

Claims

1. A system for distributed dispatching of a group of elevator shuttles that carry passengers between a first lobby at a low level in a building and a second lobby at a high level in said building, comprising:

a plurality of elevator shuttles, each including an elevator car vertically moveable in a hoistway in response to a run command and each determining when it is ready to run and providing a run ready signal indicative thereof;

a controller for determining the number of shuttles operating in said group and providing a number signal indicative thereof, for calculating a dispatching interval that will result in the shuttles in said group being dispatched in a repetitive sequence of substantially equal intervals, and for sequentially dispatching said shuttles in a repetitive sequence, each shuttle being dispatched in response to a corresponding run ready signal after the expiration of said dispatching interval from the time at which a next prior one of said shuttles in said sequence has been dispatched.

2. A dispatching system according to claim 1 wherein said controller comprises means for determining the time period required for each shuttle to make a round trip run between said lobbies and providing a plurality of run time signals respectively indicative thereof, and for calculating said dispatching interval by dividing said average run time for all of the shuttles in said group into as many intervals as there are shuttles in said group.

3. A method of dispatching elevator shuttles in a group, comprising the steps of:

(a) providing a dispatching interval related to the number of shuttles operating in said group; and

(b) dispatching said shuttles in a sequence, each shuttle being dispatched when it is ready to commence a run after the expiration of said dispatching interval from the time at which a next prior shuttle in said sequence was dispatched, said interval being selected to cause said shuttles to be dispatched in a repetitive sequence of substantially equal intervals.

4. A method of dispatching according to claim 3 wherein said step (a) comprises:

(c) determining the average round trip run time for the shuttles in said group; and

(d) providing said dispatching interval having a duration equal to said average round trip run time divided by the number of shuttles in said group.

Drawing