CN1372204A - Information processor and real-time distributed process system - Google Patents

Abstract

In order to allocate a plurality of tasks for realizing the protective relay function to calculators by suppressing the possibility of stopping the protective relay function, the present invention is an information processor for processing a plurality of tasks for realizing the protective relay function by a plurality of calculators connected by a network, which has a means for storing combinations of tasks to be processed by a specific calculator among the plurality of tasks and a means for selecting any one of the combinations of tasks to be processed by the specific calculator and executing the tasks of the selected combination by the specific calculator.

Description

Message handler and real-time distributed process system

Background of invention

Invention field

The present invention relates to a kind of structure of distributed processing system(DPS), it has and is connected to a plurality of computing machines that are used to carry out a plurality of programs on the network, relates more specifically to be provided by the protection controller of power-supply system a kind of disposal system of real-time control.

[prior art]

In the common distributed processing system(DPS) of tradition, designed various method for allocating tasks and dropped into practical capacity and the reliability of improving processing.For example, in Japanese publication patent disclosure publication No.Hei 05-257908 and Japanese publication patent disclosure publication No.Hei 07-282013, advised a kind of method for allocating tasks according to the characteristic and the condition of each computing machine.By with reference to the capacity of each computing machine such as processing capacity and reliability or the condition such as computational load, set the tasks and distribute and right of priority.The task that importance is different with emergency is handled by the computing machine with suitable reliability, thereby carries out a reliable generally distributed treatment.

Except above-mentioned structure, a kind of mechanism that pulls together to handle can be used a mutual relationship as task.For example, in Japanese publication patent disclosure publication No.Hei 09-231183, provide a kind of control gear, be used to realize the time restriction of task collaborative process, for example begun task and so on synchronously., from viewpoint of reliability, do not solve the mutual relationship of task.

On the other hand, in requiring a real-time processor of reliability, taked a kind of multiplexing countermeasure as usual.For example, in the many controllers such as the Control on Communication of the power control of power generation, power delivery and power division and aviation and railway, though be necessary in real time, trouble-saving is strictly limited.Therefore, in a kind of so real-time processing, the multiplexing high reliability that realizes of multiplexing, the multiplexing and algorithm that calculates by controlled target.

As a result, when above-mentioned target is mounted as a kind of highly reliable real-time distributed processing,, between the each task that will be distributed, produce a kind of mutual relationship from the viewpoint of reliability.For example, when by a plurality of multiplexing tasks being installed when realizing reliability for same target, very desirable from the viewpoint of reliability, task is not handled by identical computation process.Therefore, in highly reliable real-time distributed processing, not only must consider that the mutual relationship of task also will consider the mutual relationship of task from the viewpoint of reliability from the viewpoint that totalizes.

As usual, when will be in a distributed treatment during real-time processor of installation requirement reliability, so relative restriction must be just in time be satisfied in front the design.For example, in Japanese publication patent disclosure publication No.Sho 57-3516, advised being used for installing a kind of method of the protection controller of power-supply system a distributed processing system(DPS)., the adjustment that totalizes when design and the adjustment of reliability will manually be carried out.Therefore, must be optimal configuration of each processor structure design.In addition, be used for the configuration that sets the tasks in advance, so this effect has not just had when being changed when processor is constructed owing to only make.

[problem that the present invention solves]

As described in " DESCRIPTION OF THE PRIOR ART ", the common distributed processing system(DPS) of tradition can be applied in the controller that requires reliability.

Previous distributed proccessing can not be stipulated each computing machine or mission reliability and emergency and realize high-performance and high reliability.

Here be not provided for device into the mutual relationship of reliability assignment of mission.

Therefore, in order to realize a kind of highly reliable real-time distributed process system,, be used for reflecting that from multiplexing viewpoint the mechanism of the mutual relationship of reliability also is essential except as the mutual relationship of the task of the number of computations of tradition.

Summary of the invention

An object of the present invention is to stipulate reliability and emergency and the realization high-performance and the high reliability of each computing machine or task.

Another object of the present invention is allocating task so that forbids stopping the possibility of protective relay function and by few computer realization protective relay function.[means of dealing with problems]

In requiring a real-time processor such as the Control on Communication of power control and aviation and railway of reliability, run well in order to keep for a mistake in the computing machine, must be by the multiplexing carrying out safety backup of failing of realizing.When realizing that by a distributed processing system(DPS) this handles target, only make the fault tolerant of this system own, then can't avoid fault.According to calculating the multiplexing and multiplexing notion of algorithm, must realize the carrying out safety backup of failing by equipping a plurality of tasks.

In this case, the essential relation of considering a plurality of tasks.As the mutual relationship of task, the calculating of same computing machine can be limited or it may not be suitable.Therefore, in highly reliable real-time distributed process system of the present invention, an object computer control gear that is used to control this mutual relationship is mounted.

Object computer control gear (being used for judging the computing machine of each task of execution) is controlled the emergency of each task executions time and the exclusive degree of importance and reliability, and the exclusive degree of this reliability is the mutual relationship of the task in the database.

This object computer control gear can be for optimal task configuration of condition judgment of distributed processing system(DPS) by the rule in the reference database.

Consider emergency and the importance of each task executions time, can carry out with the maximization of the service efficiency of CPU, to the given right of priority of most important task and do not reduce the processing that the exclusive degree of reliability (it is the mutual relationship of task) is selected the task that will be performed, and can carry out the processing that selection is assigned to related task in the computing machine on it.

In addition, when on above-mentioned distributed processing system(DPS), executing the task scheduling, can realize high reliability all the time.In the moment of computer failure or maintenance, obtain to be used to realize that the function of CPU service efficiency continuity and reliability keeps.

In addition, the present invention is a message handler that is used to handle by a plurality of tasks of a plurality of computer realization protective relay functions that connect by network, and it has device, is used to store among a plurality of tasks the task combination that will be handled by certain computer; And device, being used to select any one that will be made up and carrying out selected combination by certain computer of task by the task that certain computer is handled.

In addition, the present invention is a message handler that is used to handle by a plurality of tasks of a plurality of computer realization protective relay functions that connect by network, and it has device, is used to store among a plurality of tasks the task combination that will be handled by certain computer; And device, be used to carry out not by the combined task beyond the combination of the task of specific Computer Processing.

Description of drawings

Fig. 1 represents to use the structure of a kind of highly reliable real-time distributed process system of the present invention.

Fig. 2 represents another structure No.1 of highly reliable real-time distributed process system.

Fig. 3 represents another structure No.2 of highly reliable real-time distributed process system.

Fig. 4 represents the operation of No.1 when equipment is failed.

Fig. 5 represents the operation of No.2 when equipment is failed.

Fig. 6 represents the task example by this highly reliable real-time distributed process system scheduling.

Fig. 7 represents the example of schematic views of the database of an object computer control gear.

Fig. 8 represents the time portion of No.1 task scheduling.

Fig. 9 represents the time portion of No.2 task scheduling.

Figure 10 represents the time portion of No.3 task scheduling.

Figure 11 represents from a kind of design tool in the external communication network.

Figure 12 represents to have the structure of the distributed processing system(DPS) of communication channel monitor task.

Figure 13 represents to use the structure of a kind of digital protective relay equipment of the present invention.

Figure 14 represents the example of schematic views of database of an object computer control gear of this digital protective relay.

Figure 15 represents the example of schematic views of database of an input and output control gear of this digital protective relay.

Figure 16 represents the example of textual form of the database of this object computer control gear.

Figure 17 represents the example of textual form of the database of this input and output control gear.

Figure 18 represents to use the example (system configuration) of this digital protective relay equipment of the present invention.

Figure 19 is the conceptual view of the task in the example of this digital protective relay equipment.

Figure 20 is illustrated in the operating process at object computer control gear place's zero hour in the example of this digital protective relay equipment.

Figure 21 represent in the example of this digital protective relay equipment the zero hour place the initiating task configuration.

Figure 22 is illustrated in the normal running flow process of the object computer control gear in the example of this digital protective relay equipment.

Figure 23 represents a kind of hypothesis of the computing machine mistake in the example of this digital protective relay equipment.

Figure 24 represents the fault tolerant operation (1) in the example of this digital protective relay equipment.

Figure 25 represents the fault tolerant operation (2) in the example of this digital protective relay equipment.

The detailed description of preferred forms

To at length explain embodiments of the invention with reference to the accompanying drawings at this.First embodiment as shown in Figure 1.

A plurality of computing machines are three computing machines 102 to 104 in this example, are connected to a network by communication network 101.Here Biao Shi each computing machine is to be used for the computing machine that combine digital is calculated.A computing machine has a basic structure independently: for example, power supply, CPU and storer, it is not to be subjected to stopping or a equipment that fault influenced of any other computing machine on operation capacity.In addition, each computing machine has identical program execution enviroment, so independently program module can be carried out by each computing machine.In the accompanying drawings, the shaped as frame numeral 105 to 109 of computing machine 102 to 104 internal descriptions is represented the program module that each computer-internal is carried out abstractively.Each computing machine can be carried out a plurality of independently program modules simultaneously.

Utilize above mentioned a plurality of computing machine to ask to carry out a plurality of tasks.For each task, the execution time is limited.In example as shown in Figure 1, execute the task all five tasks 105 to 109 of 1 to 5 of request.

About above mentioned a plurality of tasks, as the mutual relationship of task, the calculating of same computing machine can be limited or it may be unsuitable.For example, when a function of equipment so that the result of calculation of multiplexing task " with " so that avoid fault, do not wish by these tasks of same COMPUTER CALCULATION for the reason of reliability.Hereinafter, it is known as reliability interactive task restrictive condition.

The object computer control gear 110 that is equipped in one or more processors is carried out a processing: the distribution of executing the task of the rule of considering the interactive task restrictive condition of reliability.Therefore as a kind of realization means, some available method is arranged here.

At first, as shown in Figure 1, a kind of like this structure of the object computer control gear 110 to 112 by preserving a database can be realized it.The object computer control gear obtain the interactive task restrictive condition of the reliability of task separately in advance and with their storages as data-base content 113.

Data-base content has been described the interactive task restrictive condition of task executions time restriction and reliability separately.In Fig. 1, represent to be used to determine estimate index 115 and estimate index 116 execution time about the execution time restriction of real time capacity about an object computer of the relation between task 1 and the task 4 about the description 114 of task 1.

In view of the above, in task 1, the emergency of execution time and importance are estimated index 115 regulations by the execution time and are stipulated by object computer estimation index 116 from the exclusive degree of the reliability in the task 4.Similarly, other each tasks and mutual task are described.That is, constructing this processor makes and is used for judging emergency, importance and the security of the contents processing of being made up of data-base content 113 a plurality of tasks by acquired information.

Object computer control gear 110 to 112 obtains the information of data-base content 113 and judges to be used to carry out the computing machine of task separately.Consider task executions time restriction separately, object computer control gear 110 to 112 is carried out with the service efficiency maximization of CPU, to the given right of priority of most important task and do not reduce processings (task choosing processings) that the exclusive degree of reliability (it is the mutual relationship of task) is selected the task that will be performed, and carries out a processing (computing machine selection processing) of selecting related task is assigned to the computing machine on it.

In example as shown in Figure 1, the strict interactive task restrictive condition of reliability forced at task 1 and 4, task 1 and 2, task 2 and 3 and task 3 and 5 on.Therefore, work out a task equipment pattern be not these combined equipments on same computing machine.One of solution is a Task Distribution pattern as shown in Figure 1.Task 1 and 3 is assigned to computing machine X, and task 2 and 4 is distributed to computing machine Y, and task 5 is distributed to computing machine Z.The interactive task restrictive condition of above-mentioned reliability all is satisfied.

Next will explain the another one means, by it, the object computer control gear is considered the rule of the interactive task restrictive condition of reliability.The object computer control gear does not have database can obtain above-mentioned effect yet.

In example as shown in Figure 2, execution time that each task all has this task estimate index and in program module an object computer of each mutual relationship between other task estimate index.Communication network 101 and a plurality of computing machines 102 to 104 in textural and Fig. 1, explain those be identical.With the difference of embodiment shown in Figure 1 be exactly: task 201 to 205 comprises above-mentioned index.Object computer control gear 206 to 208 for has as shown in Figure 2 increased a device, is used for being included in by the communication acquisition index of task 201 to 205.Therefore, when the index that this object computer control gear in case of necessity transmits from task and points out to launch, above-mentioned task choosing is handled and computing machine is selected processing thereby can carry out.

The same with device (by it, the object computer control gear is considered the rule of the interactive task restrictive condition of reliability) next will explain another kind of method.Even all other index is not included in the program module in advance, also can obtain above-mentioned effect.In example as shown in Figure 3, show a kind of method, wherein, each task has a task names determining grammer according to pre-.Communication network 101 and a plurality of computing machines 102 to 104 in textural and Fig. 1, explain those be identical.With the difference of embodiment shown in Figure 1 be exactly: task 301 to 305 has the task names that presets.Object computer control gear 306 to 308 is as shown in Figure 3 pointed out the task names of task 301 to 305 and is calculated object computer and estimate index.In order to carry out this operation, this object computer control gear 306 to 308 is preserved a pre-definite syntax rule 309.The exclusive restrictive condition 313 of same computing machine described in grammer 310 to 312, as the relation of two task names patterns.

Object computer control gear 306 to 308 is discerned this task names in the device level of program module.Object computer control gear 306 to 308 is pointed out syntax rule 309, extracts the grammer of this task names coupling from grammer 310 to 312, and can obtain the interactive task restrictive condition of reliability.With with as illustrated in fig. 1 and 2 the identical mode of embodiment, object computer control gear 306 to 308 can be executed the task and be selected to handle and computing machine is selected processing.

Shown in Fig. 1 to 3, do not need computing machine to carry out this object computer control gear.As the function of OS, this mechanism can be installed in all computing machines.Can use principal and subordinate's structure.In addition, they can be installed in the computing machine, and this computing machine for example is connected on the remote operation platform off and on via an external communication channel (rather than communication network 101).

By introducing the object computer control gear of explaining above, the continuity that can realize the restrictive condition of real time capacity and reliability according to the structure and the capacity of computing machine.

Next, a structure of the application function of the task of realizing keeping high reliability will be used the present invention to explain to be used to.Except the function that task choosing of the present invention is handled and computing machine is selected to handle, have that task is regenerated and a distributed processing system(DPS) of the function that computing machine moves is assumed to be.

By explaining this structure with reference to figure 4.In example as shown in Figure 4, supposed the wrong a kind of situation that appears among the computing machine X according to example as shown in Figure 1.Mistake in computing machine X occurred, communication channel 401, a plurality of computing machine 402 to 404, task 405 to 409 and object computer control gear 410 to 412 were all identical with shown in Figure 1 those.

In the example of representing in this figure, task 1 to 5 is tasks of requirement identical calculations total value., about the emergency and the importance of task, suppose that task 5 is in the state of highest priority.In addition, in computing machine Y and Z, having only four in the task 1 to 5 can be performed.

When by being installed in the object computer control gear 410 among the computing machine X or being installed in object computer control gear 411 among computing machine Y and the Z and 412 routine observation detects of computing machine X when wrong in this state, object computer control gear 411 and 412 is executed the task again corresponding to the mistake of X and is selected to handle and computing machine is selected processing.

The situation of the distributed processing system(DPS) after carrying out again as shown in Figure 5.Out of order computing machine 502 and normal computing machine 503 and 504 are connected to communication channel 501.Task choosing is handled and computing machine selects processing to be performed computer control mechanism 505 and 506 execution.Before mistake took place, the distribution of three tasks 507 to 509 (task 2,4,5) was not by its change.A difference before occurring with mistake is: the task 1 that substituting for computer 502 calculates, task 510 is reallocated to computing machine 504.

This distribution is by the combination of the continuity of the character of character that is used for realizing each task and task mutual relationship.Task choosing is handled the importance of whole task and emergency maximization and computing machine selects to handle the reliability maximization that makes the task combination that is installed in the same computing machine.In example as shown in Figure 5, the calculating total value of each task is identical, so in two computing machines, a plurality of equipment patterns of four tasks can be considered., it can obtain uniquely, satisfies that equipment pattern that the execution time shown in Figure 1 is estimated and object computer is estimated most to be among each equipment pattern: computing machine Y (task 2, task 4), computing machine Z (task 1, task 5).

When the object computer control gear is handled these task choosing and computing machine selection, as shown in Figure 5, utilize fault detect as a triggering, can realize such function: automatically distribute by the replacement task 510 of computing machine 502 to the task 1 of computing machine 504 calculating.

Next, will use the present invention to explain a structure, this is configured to realize carrying out the function of the task scheduling that keeps high reliability.Be used to the to execute the task object computer control gear of scheduling except the function of previous explanation, is confirmed will to be existed by the task that current calculating section produces in the future.For each incident part of wanting processed task to be produced, the object computer control gear is carried out above-mentioned task choosing processing and computing machine is selected to handle, and obtains an optimal solution for these a plurality of time portion, and realizes task scheduling.

By explaining embodiment with reference to figure 6 to 10.Fig. 6 has conceptually represented by the timetable of the task generation of object computer control gear affirmation and the result who these Task Distribution is given the task scheduling of computing machine.In the accompanying drawings, transverse axis 601 express time axles are used to represent about the generation of the task of task 1 and processing for this reason.The appearance of the execution request of the arrow 602 expression tasks 1 under transverse axis 601.The execution of the rectangle 603 expression tasks 1 on the transverse axis 601.Computer name in the rectangle 603 represents to be responsible for the computing machine of execution.This situation represents that the execution request 602 of task 1 is assigned to the execution 603 of computing machine X.In addition, when carrying out during 603, in the time of will distributing execution corresponding to a new execution request 604, it is represented as rectangle 605.

In very first time part 606, can realize this processing by the object computer control gear Task Distribution of previous explanation.The task requests that occurs in the time portion 604 has the content identical with first embodiment.That is, use Task Distribution structure shown in Figure 1.That is, the computing machine of task 1 to 5 as shown in Figure 6 and the content of Fig. 1 are harmonious.

In second time portion 607, task 1 request 604 reappears.In response to this request, the task scheduling of object computer control gear is planned Task Distribution in second time portion again by the interactive task restrictive condition that keeps reliability as shown in Figure 7.Except Task Distribution condition as shown in Figure 1, task 1 is redistributed to computing machine Z, just as carrying out 605.The result becomes situation as shown in Figure 8.In Fig. 8, among the computing machine Z 801, between new allocating task 802 (representing by thick line in the accompanying drawing of task) and executory task 803, there is not the interactive task restrictive condition of reliability.Generally, the interactive task restrictive condition of reliability as shown in Figure 7 satisfies.

At the time point place of the 3rd time portion 608, the task 1 that computing machine X carries out carry out handle 603 and the task 2 carried out of computing machine Y carry out and handle 610 and be done and on the contrary, task 5 is carried out request 611 and is received.The task scheduling of object computer control gear is carried out Task Distribution as shown in Figure 9.From the viewpoint of Computer Processing capacity, 5 the computing machine of can executing the task is computing machine X 901 or computing machine Y 902., when task 5 is distributed to computing machine X, just consider that in executory task 903 (task 3), then the interactive task restrictive condition of reliability is not followed.As the allocation result that obtains the highest estimation, select execution by the new task 904 (task 5) of computing machine Y.Reason is because in commission task 905 (task 4), no problem here forcing in reliability.

At the time point place of the 4th time portion 609, carry out the task execution processing of handling 612 except the task of carrying out by computing machine Y 5 and all finished.Here, the execution request 613 of task 1 to 5 is received.By with above mentioned same treatment, the task scheduling of object computer control gear is carried out Task Distribution as shown in figure 10.Except computing machine Y 1001 executory tasks 1002 (task 5), five tasks 1003 are reallocated to 1007.In the interactive task restrictive condition of reliability, there is not anything to run counter to.

As above mentioned, in each time portion that the request carried out takes place, when task choosing and computing machine are selected can realize task scheduling with at present just when executory task is processed.

About in a plurality of time portion by Task Distribution be used to the to execute the task method of scheduling, can rely on the difference in the method for providing of optimization apparatus to consider a plurality of methods.In above-mentioned example, along the scheduling of executing the task of the flow process adjoining land of time portion.The problem of the combination that is used for making simultaneously the Task Distribution optimization of all time portion by solution discretely thus and the difference in the Task Distribution condition between the adjacent time portion is minimized, the scheduling of can executing the task.

By this task scheduling function, this distributed processing system(DPS) can realize high reliability all the time intentionally.For example, even when a new task except confirming of task by the object computer control gear begin to ask in system, to take place the time or the initiation command of a new task when externally being published, then task scheduling is carried out along with the variation that is reflected again, thereby still can keep high reliability after changing.In addition, wrong or maintenance for computing machine, even when ordering for one of computer application condition issue by an outer computer, then suppose this computer application condition and then the scheduling of executing the task, thereby can keep having the successional function that is used to realize the CPU service efficiency of reliability according to this command context.

Next, as second embodiment, use a kind of developing instrument of the present invention as shown in figure 11.

The disposal system that is developed is identical with embodiment shown in Figure 1.By the communication channel 1101 that is used to control, a plurality of computing machines, that is, three computing machines 1102 to 1104 in this example are connected on the network.In these computing machines, a plurality of tasks 1105 to 1109 are carried out in request, and these a plurality of task execution times are limited.

In order to design this disposal system, used an outside development computer.Development computer 1111 is not by being that an external communication network 1110 that is used to control is connected.By the use of this development computer, carry out the basic engineering and the various setting that numerical value is set of above-mentioned task.

Development computer 1111 has an object computer control gear 1112, its have with embodiment shown in Figure 1 in the identical function explained.Therefore, the equipment of the disposal system 1101 to 1104 that is developed by order structure, can carry out have reliability be used to realize that the successional initiating task of CPU service efficiency distributes.

By the use of such developing instrument, can finish an initial design--be used to carry out the selection processing that the computing machine of this selections task was handled and was used to calculate in the selection of the task that will be performed; And can design a kind of highly reliable real-time distributed process system.In addition, by an external communication network increase, deletion or correction task, thereby can be implemented and needn't increase the object computer control gear by most of functions that this highly reliable real-time distributed process system keeps to the computing machine that constitutes this disposal system.

As the 3rd embodiment, will explain a kind of distributed processing system(DPS) of utilizing the present invention to have to be used for the function of monitors communication channels below.

In order to constitute this disposal system, a plurality of computing machines, that is, three computing machines 1202 to 1204 in this example are connected to a network by communication channel 1201.

Here, increased monitoring and be used to examine a communication channel of communication channel task, so make this disposal system more reliable.As a kind of communication channel method for supervising, a plurality of tasks that are used for executive communication by the communication channel that is monitored are formed as a group and the validity of mutual Content of Communication is verified, thus the normality of monitor communication channel.

In this case, many-sided problem appears in the distribution of these monitor tasks and task scheduling.At first, in order to realize function for monitoring, preferably just do not carry out needed monitor task when its minimum value but also will carry out the monitor task that will be performed as much as possible., the task execution as the initial treatment target must not be hindered really.Secondly, the individual task of a group of formation must be by carrying out with the different computing machine of same group task.Use the present invention can realize having the distribution and the scheduling of these monitor tasks of many restrictions.

With the mode identical with embodiment shown in Figure 1, a plurality of computing machines in the disposal system have object computer control gear 1205 to 1207.The implication of the task that these object computer control gear controls will be performed is as data-base content 1208.Data-base content has been described the interactive task restrictive condition of each task executions time restriction and reliability.In Figure 12, represent to be used to determine estimate index 1210 and estimate index 1211 execution time about the execution time restriction of real time capacity about an object computer of the relation between task 1 and 4 about the description 1209 of task 12.

In this example, suppose that with embodiment shown in Figure 1 two differences are arranged.At first, suppose about execution time of communication channel monitor task A 1212 and estimate index 1213, estimate that about the execution time of communication channel monitor task B 1214 object computer between index 1215 and the task estimates that index 1216 is increased.

Secondly, estimating that about the execution time of task 5 index 1217 is set to estimate than the execution time of task 1 to 4 the lower execution priority of execution time estimation index of index and monitor task.That is, task 5 is set to not be the task an of necessity.

Here, two tasks form a group and have increased the function that is used for monitors communication channels.Minimum value place monitoring function at it is essential, so that the higher execution priority than task 5 is set up.

When provide such one when being provided with to object computer control gear 1205 to 1207, three 1202 to 1204 Task Distribution is performed to 1225 just as 1220.The execution of task 5 is omitted.In addition, about other six tasks, utilize the mutual exclusive relation that is satisfied to decide the configuration of task separately.

When the task of being used for monitors communication channels is set to controlled target a part of of object computer control gear like this, can realize being used for a kind of highly reliable real-time distributed process system of binding communication channel normality.

In addition, the use of the task scheduling by utilizing the object computer control gear, monitor communication channel termly.By the monitor task of one group of low execution priority of Request Processing termly, within the especially calculated capacity of disposal system, can realize high reliability as much as possible.

The monitoring of communication channel has been described in the above-described embodiment.By identical method, can equip a monitoring function for a computing machine.At first, as the computer monitoring task, carry out one and be used to check pre-a plurality of other computing machines and a plurality of tasks that are used for verification result through communicating by letter of determining the test of computing formula and preparing to connect by network.Then, concerning this object computer control gear, the restriction or the condition of each computer monitoring task of being carried out by the various computing machine are given.Thus, the computer monitoring task is attached to automatically that appropriate tasks is distributed and scheduling and can realize the sign of unusual computing machine or the detection of network error.

As the 4th embodiment, use and use a kind of disposal system of the present invention, will consider such example: wherein, be used to protect a kind of digital protective relay equipment of one or more power apparatus to be implemented.It is indicated among Figure 13.

At first; equipment as input; with respect to each power apparatus 1301 protected and control; for example bus, converter and transmission line, be used in the needed fault detection capability of equipment protection with the needed observing function of various types of controls in voltage and current information 1302 be transmitted to the input equipment 1303 that is connected with them.Similarly, as the equipment of output, with respect to controlled power apparatus 1304, isolating switch for example, a control signal 1305 such as ON-and OFF-command is transmitted to the output device 1306 on the equipment separately of being connected to.

These input equipments 1303 and output device 1306 are connected to a plurality of computing machines 1308 to 1311 by communication channel 1307.In this case, computing machine is to be used for the computing machine that combine digital is calculated.A computing machine has a basic structure independently: for example, power supply, CPU and storer, it is not to be subjected to stopping or a equipment that fault influenced of any other computing machine on operation capacity.In addition, each computing machine has identical program execution enviroment, so independently program module can be carried out by each computing machine.In the accompanying drawings, the shaped as frame numeral 1311 of a plurality of computing machine 1308 to 1310 internal descriptions is represented the program module that each computer-internal is carried out abstractively.Each computing machine can be carried out one or more independently program modules simultaneously.

The execution of computing machine program module separately distributes to be controlled by an object computer control gear 1312.This function with the object computer control gear of above-mentioned first embodiment shown in Figure 1 is identical.Therefore, in the database of object computer control gear, the interactive task restrictive condition of each task executions time restriction and reliability has been described.

In Figure 13, object computer control gear control data-base content as shown in figure 14.The interactive task restrictive condition of each task executions time restriction and reliability is described.In description as shown in the figure, comprise about an execution time of the execution time restriction that is used to determine real time capacity about the description 1401 of each task and to estimate index 1402 and to estimate index 1403 about an object computer of other tasks.

For example; be used to protect among the computing machine of device A; be used for carrying out all the time main protection Program for Calculation module " device A; main frame; normal system " and when this program module become make can not carry out the time switch immediately and back up standby " device A; lead, back-up system " must be carried out by different computing machines.Therefore in, the interactive task restrictive condition of reliability the is imposed to task relation of " device A, main, normal " and task " device A is led back-up system ".

Similarly, for fear of fault, " device A, main, normal system " and " device A, failure safety, normal system " must be carried out by different computing machines.Therefore, the interactive task restrictive condition of reliability is imposed in the relation of these two tasks.When having forced such computing machine restriction, unless wherein the meaning is meant that two computing machines that the program module of two tasks is performed enter inoperable state in identical time domain, otherwise the fault of the defencive function of device A can not produce.

In addition, even the single error of equipment takes place, keeping more, the Task Distribution of high reliability can be applied in.For example, even any task of normal system is being switched in the situation of back-up system owing to a Device Errors, if main task is designed to be carried out by different counters with the failure safe task, then exist " device A, main, normal system " and " device A; failure safety; back-up system " the task relation in, as the interactive task condition of reliability, a task object computer can be set estimate index.In this case, can lower right of priority than the restriction of the interactive task of reliability 1403 be set to an exclusive condition 1404.

As above mentioned, for the design of carrying out about whole tasks of device A as shown in figure 14.In addition, according to identical notion, design is about the task of equipment B.By having the object computer control gear of this database, the computing machine that is used to carry out each program module is normally determined.As a result, obtain Task Distribution shown in Figure 13.Can't realize satisfying the interactive task restriction of all reliabilities and the condition of forcing from the quantity of computing machine.The configuration of satisfying more high priority restriction 1403 and satisfying exclusive condition 1404 simultaneously as far as possible is set up.Utilize the processing capacity of total system at an optional time point place, then can realize a highly reliable system.

In addition, in the above-described embodiment, a control gear also has been installed so that control above-mentioned input and output relation.As shown in figure 13, input-output device J is fixedly attached to device A to D to M.The structure control gear makes by the input/output information of emission of multi-point net or receiving equipment.A kind of like this situation of input/output relation control gear 1313 controls: the task in the computing machine is used the V and the I information of any input equipment and is launched a control signal to any output device in the database.

As an example, will consider the operation of relevant system shown in Figure 13, wherein, input equipment J1314 and input equipment J ＇ 1315 prepare as the input equipment that is connected to device A.The input/output relation control gear receives indication 1503: an input equipment J 1501 and the input equipment J ＇ 1502 a kind of fallback relationship for the data-base content that schematically shows as Figure 15.Therefore, after the detection of the mistake of input equipment J, each computing machine of input/output relation control gear order makes the input data that the input data replacement input equipment J that will launch with input equipment J ＇ will launch.Thus, can avoid the erroneous effects of input equipment J.

In the superincumbent explanation, by during Design of processing system stage or maintenance, on rule-based approach, specifying the data-base content of equipping shown in Figure 14 and 15 at it.This method will be by explaining with reference to an example.

Figure 16 represents to the given a kind of data layout of object computer control gear.Can be according to a kind of pre-content of grammer of determining by schematically showing among the next given Figure 14 of text data.

In Figure 16, comment statement of first row, 1601 expressions.On second row 1602 and subsequent rows, be set up about the defencive function of device A and B.

At first, the task about the protection of device A is declared.By the explanation on the third line 1603, the functional programs module that is used for treatment facility A, master, normal system is declared and is called " module_A1 ", as a task that is used to carry out normal process." " (" device A, main, normal system ") is a comment statement to normal system to character string after mark # for Device A, main.In addition, by describing on the fourth line 1604, the functional programs module that is used for treatment facility A, primary, spare system is declared and is called " module_A2 " (modules A 2), as a task that is used to carry out standby processing.Similarly, for device A, four tasks have been stated altogether.About the explanation of each task real time capacity, for example, task processing starting, concluding time and right of priority etc. are provided with respectively.

Next, be declared about the normal system of the protection of device A and the relation between the alt hierarchy.On the 7th row 1605, defined by module_A2 and backed up module_A1.Hereinafter, similarly, will the backup of failure security function be described.

At the end of the definition of data of device A, the exclusive relation between the task is declared.At the 9th row 1606, with restriction of highest priority definition: " device A, main, normal system " function and " device A, failure safety, " function is assigned to different computing machines to normal system.In the tenth delegation 1607, with than low priority definition as a condition: ideally " device A, main, normal system " function and " device A, failure safety, " function is distributed to different computing machines to normal system.Hereinafter, similarly, five relations that are present between four tasks are defined.

As shown in figure 16 in the given text data of object computer control gear, also use with about the identical mode of the protection of device A, the setting about equipment B has been described.The rule of the interactive task restrictive condition of the reliability that schematically shows in this way, can given Figure 14.

Next, will explain to the given data layout of input/output relation control gear.Can be according to a kind of pre-content of grammer of determining by schematically showing among the next given Figure 15 of text data as shown in figure 17.

In Figure 17, comment statement of first row, 1701 expressions.On second row 1702 and subsequent rows, carry out setting, about the setting of output device and about the setting of input and output about input equipment.

At first, the input total value by input equipment J emission is declared.By the explanation on the third line 1703, be set to the title of " a_bus_v1 " in the first input total value of input equipment J.Hereinafter, be declared by all input total values of input equipment J emission and title is set up.

Similarly, the output total value that is received by output device also is declared.The 17 row 1704 and and subsequent rows on, output device L is defined.With with the identical mode of explanation about the input total value, on the 18 row 1705, the output total value is declared and title is set up.

At last, for each task, the input and output total value that this task is used is set.On the 22 row 1706 and subsequent rows, the input and output total value of task module_A1 is set up.The argument of being occupied by the program module that is used to realize above-mentioned task is provided with in proper order.On the 20 triplex row 1707, definition uses input a_bus_v1 as first argument.Similarly, the needed input and output total value of said procedure module all is defined.Then, can be by being set up that the employed input and output total value of task substitutes.On the 28 row 1708, change definition into when in using, mistake taking place and use a2_bus_v1 as the input a_bus_v1 of first argument.

By as shown in figure 17 to the given text data of input/output relation control gear, the relation of the input and output total value that schematically shows in can given Figure 15 and therefore rule about substituting.

When being introduced in shown in Figure 16 and 17 a kind of method for expressing on the grammer basis, deviser and operator can directly write rule in text.In addition, a kind of mechanism of the text by from design with GUI and operation tool, being used for exporting above-mentioned expression, deviser and operator obtain the environment of easier design and running.

As above mentioned; when a kind of highly reliable real-time distributed process system control with object computer control gear of being illustrated among the embodiment 1 is used to realize the task of digital protective relay of power apparatus, can realize a kind of highly reliable protective relay system.

Here there are a plurality of mutual relationships of task to control.For example, as what explain among the above mentioned embodiment 3, as the normal system task of the first calculating target and a kind of mutual relationship between the back-up system task, the execution restriction or the condition of various computing machine can be cited.In view of the above, when failing, computing machine is used to avoid a kind of digital protective relay of protective relay function stop to be implemented.

In addition; in a kind of protective relay (this protective relay has a main detection task of the main measuring ability that is used to realize protective relay and is used to realize a fault detect task of fault detection capability as the failure safety element); as the main detection task of same protective relay and the mutual relationship between the fault detect task, the execution restriction or the situation of various computing machine can be cited.In view of the above, be used to avoid a kind of digital protective relay of fault to be implemented.

Except it; in a kind of protective relay (this protective relay has an assembling bus protection task of the assembling bus protection function that is used to realize the bus protection relay and is used to realize a separate bus protection task of separate bus defencive function); as the assembling bus protection task of same bus and the mutual relationship between the separate bus protection task, the execution restriction or the situation of various computing machine can be cited.In view of the above, be used to avoid a kind of digital protective relay of bus protection relay fault to be implemented.

In addition; about each transmission line that constitutes by parallel circuit; at a protective relay, can be cited for the execution restriction or the condition of a plurality of transmission line task various computing machines of the transmission line protection that belongs to same parallel circuit with the transmission line that is used for realizing transmission line protective relay function.In view of the above, being used for a kind of digital protective relay that the computation error influence with parallel transmission protection minimizes is implemented.

By with reference to the accompanying drawings,, a kind of highly reliable protective relay system that controls by as above mentioned highly reliable real-time distributed process system will be explained as the 5th embodiment.

At first, the total system image as shown in figure 18.By an optical communication channel 1801 that is used to realize that optical medium is communicated by letter, a plurality of computing machines, for example, three computing machines 1802 to 1804 in this example are connected to a network.Here Biao Shi each computing machine is to be used for the computing machine that combine digital is calculated.A computing machine has a basic structure independently: for example, power supply, CPU and storer, it is not to be subjected to stopping or a equipment that fault influenced of any other computing machine on operation capacity.In addition, each computing machine has identical program execution enviroment, so independently program module can be carried out by each computing machine.In addition, each computing machine obtains and a synchronous signal of industrial frequency by optical communication net 1801.Therefore, can carry out the protective relay calculating synchronous with the electric angle of industrial frequency.

Concerning optical communication net 1801, not only computing machine but also input-output device all are connected.Consider a phase place of line of electric force, this configuration as shown in drawings.At the line of electric force 1805 that is used to provide industrial power, isolating switch 1806 and converter 1807 and equipment separately have been installed have been had input-output device 1808 to 1811.This system has these input-output device is connected to a structure on the optical communication net.

Input-output device 1808 to 1811 have analog digital translation function, communication function and electric light translation function and be installed in isolating switch 1806 and converter 1807 on.Input-output device 1808 to 1809 is realized the current information of voltage from the isolating switch to the computing machine and the communication function of the blocking-up control information from the computing machine to the isolating switch as dual communication terminal of isolating switch 1806.Similarly, input-output device 1810 and 1811 dual communication terminals as converter 1807 realize communication function.From the viewpoint and the viewpoint from insulating of power supply, input-output device separately and computing machine are irrelevant each other.

With the mode identical with the embodiment of previous explanation, in this system, object computer control gear 1812 and input/output relation control gear 1813 are installed in one or more computing machines arbitrarily.

In a kind of highly reliable protective relay system by the highly reliable real-time distributed process system control with above-mentioned structure, the task handling with real time capacity is requested.Those tasks are illustrated among Figure 19 abstractively.The task that shaped as frame numeral 1901 to 1906 expressions are as shown in figure 19 handled by computing machine 1802 to 1804.

Task by numeral 1901 expressions will be known as " Task1 " or " TrPrMN " hereinafter.Task1 is used to realize that main converter differential relay element 87T calculates and be used to stipulate the task of order computation of the trigger condition of this element.

Task by numeral 1902 expressions will be known as " Task2 " or " TrPrFD " hereinafter.Task2 is a task, be used to realize to finish above mentioned Task1 the failure security function another algorithm fault detection capability and be used for the order computation of the trigger condition of regulation element.

Task by numeral 1903 expressions will be known as " Task3 " or " TrPrMN-FT " hereinafter.Task3 is a task, is used to realize finish the standby functions of the fault tolerant function of above mentioned Task1.When Task1 becomes can not carry out the time, back up this function Task3 at once, continue standby by the calculating of carrying out with the minimum value part of Task1 identical algorithms.Task3 carries out calculating, is used for continue preserving the necessary processing of aforementioned numerical value, for example import total value read or digital filtering handles and so on.

Task by numeral 1904 expressions will be known as " Task4 " or " TrPrFD-FT " hereinafter.Task4 is the same with Task3 to be a task, is used to realize finish the standby functions of the fault tolerant function of above mentioned Task2.

Task by numeral 1905 expressions will be known as " Task5 " or " SS-P/P ＇ " hereinafter.Task5 finishes the function of the normality of the input-output device P that is used for monitoring Figure 18 and explains (as shown in figure 18 1808) and input-output device P ＇ (as shown in figure 18 1809).For example, concerning input-output device, increase a hardware capability, be used for the 12nd harmonic component of industrial frequency imported quantitatively as an electricity (analogue value) of checking that signal is superimposed upon from line of electric force.Thus, when any one computer monitoring be included in input-output device emission or the information that receives in harmonic component the time, the normality of input signal can be monitored.

Task by numeral 1906 expressions will be known as " Task6 " or " SS-Q/Q ＇ " hereinafter.Task6 is the same with Task5, finishes the function of the normality of the input-output device Q that is used for monitoring Figure 18 and explains (as shown in figure 18 1810) and input-output device Q ＇ (as shown in figure 18 1811).

For these a plurality of tasks, real time capacity is requested.In example as shown in figure 19, request determines to utilize within the execution time three computing machines 1802 to 1804 as shown in figure 18 to carry out six task task 1 to Task6 pre-.

In Figure 19, be that each execution time of determining in advance restriction of six task task 1 to Task6 is represented as 1907 to 1912.With with the foregoing description, in advance to the given data of database of object computer control gear control about these execution time restrictions.To explain the content of each execution time restriction below.

The content that is illustrated in the execution time restriction 1907 is: Task1 has the execution priority of level 1, and the request of this task takes place with the temporal frequency of 30deg (number of degrees), then from the cycle that occurs to processing be 30deg.With respect to higher-priority of 1, one littler numeric representation of level, so level 1 means a priority task the highest.Handle periodic unit deg and be a round-robin electric angle with commercial frequency and be unit representation, as institute's explanations among Figure 18, it only specifies in carries out the time of locating with the time of the synchronous protective relay calculating of industrial frequency electric angle.

Similarly, execution time restriction 1908 specifies Task2 to have the execution priority of level 1, and occurrence frequency is 30deg, so the processing cycle is 30deg.That is, being given for the Task1 and the Task2 that realize the main protection function is the highest priority task.

In addition, execution time restriction 1909 and 1910 is given for and realizes that Task3 and Task4 that defencive function is fault-tolerant are the right of priority inferior to highest priority, and similarly, occurrence frequency is 30deg, and the processing cycle is 30deg.

On the other hand, execution time the restriction 1911 and 1912 minimum execution priorities that have in the design.Stipulate that this frequency of occurrences is 10deg and handles and can finish within 2 second cycle.In addition,, also mentioned aperiodic task generation rule, do not produced task so so that after system start-up in 1 minute for Task6.

In this embodiment; except execution time restriction for each term of reference of explaining above; with with the identical mode of previous embodiment of highly reliable real-time distributed process system, it is essential being used for the interactive task restrictive condition of reliability of assignment of mission mutual relationship the same with highly reliable protective relay system.

The interactive task restrictive condition of Task1 to the Task4 relation depended in exclusive restriction 1913 to 1916 expressions.With mode same as the previously described embodiments, in advance to the data of the given interactive task restrictive condition about these reliabilities of the database of object computer control gear control.To explain the content of the interactive task restriction of each reliability below.

The content representation Task1 of expression and Task3 have the exclusive right of priority of level 1 in exclusive restriction 1913.With respect to higher-priority of 1, one littler numeric representation of level, so level 1 is meant the exclusive configuration of the computing machine that is considered to highest priority.Reason is that then it is insignificant fully because if be used to realize that the Task3 of the fault tolerant function of Task1 can't help another computing machine and handles.

Similarly, exclusive restriction 1914 regulations are for Task2 and Task4, and the exclusive configuration of computing machine is considered to highest priority.

On the other hand, exclusive restriction 1915 expression Task2 and Task4 are in exclusive right of priority 2.The task of being used to realize the failure security function of Task1 must be handled by another computing machine in principle., Task2 is made of as a fault detection system different algorithms, so even it is handled by the computing machine identical with main detection Task1, it also is inessential.Therefore, the feasible exclusive configuration of considering computing machine as much as possible of the interactive task restrictive condition of regulation reliability.

Similarly, the feasible exclusive configuration of considering computing machine as much as possible of the interactive task of the reliability of exclusive restriction 1916 regulation Task3 and Task4 restriction.

According to the specified content of the interactive task restriction of execution time of explaining above restriction and reliability, configuration that the object computer control gear sets the tasks and control are executed instruction.It the processing stage will make an explanation in order.At first, with the operation of explaining particularly when the system start-up, that is, and this

Initial operation among the embodiment.

Figure 20 is the operational flowchart of object computer control gear when being illustrated in system start-up.At first, handling in 2001, the object computer control gear is with all computer initializations.Then, it is confirmed the processing capacity conditions of each computing machine and carries out of the calculated capacity be used to confirm each computing machine and handle 2002.Next, handling in 2003, the object computer control gear is introduced the specified content about all tasks from database.Particularly, they are that the execution time restriction explained above and the interactive task of reliability limit.It is carried out and handles 2004, is used for selecting to have the startup tasks requested from them at the beginning from system.In example as shown in figure 19, Task1 is an initial start request task to Task5, but Task6 is not.

By handling 2004, finish about the affirmation of the interactive task restriction of initial start request task executions time restriction and reliability.According to data about restrictive condition, handling in 2005, the object computer control gear will hinder total value to minimize, promptly, follow each task executions time restriction, with the maximization of CPU service efficiency, to the given right of priority of important task, avoid the interactive task restrictive condition loss of reliability, thereby select to give being performed of task and distributed the computing machine of this task.For this purpose, it carries out a computing, is used to solve combinatorial problem and result's configuration that sets the tasks from then on.After the allocation position of each task was determined, the object computer control gear was carried out a processing 2006 and is handled 2007, handled for 2006 (hereinafter, being called task module) that are used for those tasks of form predistribution of carrying out; Processing 2007 is used to export an order and carries out this predistribution task module.

In addition, the object computer control gear not only starts initiating task but also calculates task configuration and predistribution task module in next time portion in advance.Next time portion is to represent that next a task produces a time portion of request.The 30deg synchronous with electric angle after start-up time of next time portion in this embodiment, because as shown in figure 19, it is 30deg that minimum task produces frequency.

Handling in 2008, the task requests in the task real time capacity of object computer control gear regulation from the execution time restriction of previous explanation in next time portion of affirmation is also carried out next task for it with mode same as described above and is disposed optimal treatment 2009.It is compared predistribution initiating task modules configured and carries out one and handle 2010 with the configuration of next task module of determining now, be used for when needed next task module being pre-assigned to object computer.By, in next time portion, be to produce and task resembles and takes place predetermined owing to external cause if there is not task, then this task can be performed with optimal configuration.

The state of the initiating task that operation the distributed configuration when Figure 21 represents to be started by the top object computer control gear of mentioning.In computing machine X (2101 in the accompanying drawing), Task1 (2102 in the accompanying drawing) and Task5 (2103 in the accompanying drawing) have been arranged, and in computing machine Y (2104 in the accompanying drawing), Task2 (2105 in the accompanying drawing) and Task3 (2106 in the accompanying drawing) have been arranged, in computing machine Z (2107 in the accompanying drawing), arranged Task4 (2108 in the accompanying drawing).In this configuration, be illustrated in as shown in figure 19 1913 to 1916 in the interactive task restriction (here being the exclusive restriction of two tasks) of reliability all be satisfied.All task handling of asking in the current generation are performed and total system is under the sound condition.For example, even past and Task6 request constantly takes place, the object computer control gear also can be given the configuration section of Task5 the processing plan of distribution T ask6 automatically and after finishing it be returned to task, so that can keep the sound status of total system.

When the condition except plan during as one of system whole the generation, for example, when any computing machine all becomes when unavailable, the then reallocation that must execute the task of object computer control gear.

Its treatment scheme as shown in figure 22.

At first, the object computer control gear, a supposition time (being 30deg this embodiment) past tense begins this processing after the last time from object computer control gear itself starts.Handle 2201 as an of down trigger synchronously and install by being used to be provided with a signal and electric angle.

The object computer control gear is at first carried out one and is handled 2202 after starting, judge the calculated capacity table by gathering various monitored results, that is, whether the condition of computing machine is changed.When the discovering device mistake, it starts a processing of execution error tolerance function.This will describe in detail subsequently.To explain a processing when not having the equipment mistake below.

At first, handling in 2203, the object computer control gear reads a current task request.When starting the another one task owing to its operation is essential, any task of this system must propose a task requests to the object computer control gear.For example, the main detection task of Task1 have function of evicting unusual input total value from and so structure make the request of the definite mistake of monitoring at that time that proposes.That is, under the condition that unusual input and output detect, Task1 sends a Task6 and asks to the object computer control gear.Such request is read out by handling 2203.

In on the object computer control gear, once carrying out, whether have between the preplanned mission in this current task request and the next part not to be both and judge and this processing is branched by a judgment processing 2204 of the existence of not preplanned mission request.

When here not during preplanned mission, carrying out in the part between at this moment task module on the object computer control gear, once carried out in by predistribution.Therefore, a current task execution command is launched immediately by handling 2205.

When preplanned mission is arranged not here, handle 2206 and make the configuration optimization that can execute the task again, that is, and the task module in the computing machine in the preallocated task.When the execution time of preplanned mission was not restricted to high priority, the object computer control gear can be rearranged this task.In fact whether they be rearranged the interactive task restriction of depending on the reliability of other predistribution tasks.

In arbitrary situation, optimal configuration during the object computer control gear is automatically selected to execute the task.Then, handle execution command of 2207 emissions and give optimal configuration in above-mentioned the executing the task.When distributing when comparing free tolerance limit here with task module, handle 2208 predistribution part task as a supplement task and handle 2209 and launch an execution command immediately and give the task of replenishing.

Even when even a not preplanned mission or when here not during preplanned mission is arranged here, handle 2210 to 2212 and carry out a configuration optimal treatment and carry out a processing that next task module is pre-assigned to object computer when needed for the task requests in next time portion.That explains among this program and Figure 20 is 2008 identical to 2010.

Next, a kind of situation of judgment processing 2202 identification equipment mistakes and execution error tolerance function will be explained.

At first, handle 2213 and search for carrying out of task for the existence of the back-up system task (being called the fault tolerant task hereinafter) that is used for finishing the fault tolerant function at the equipment of mistake generation.In example as shown in figure 19, regulation [Task3] TrPrMN-FT is the fault tolerant task of [Task1] TrPrMN.As above mentioned, from database search when mistake occurs in the fault tolerant task of the executory task of equipment.

In this case, when wrong tolerance task here, must switch to main processing mode to corresponding task immediately from standby processing mode.This is transmitted to the fault tolerant tasks of handling 2113 search to a master-slave mode switching command and realizes this point by handling 2214.By switching master-slave mode, this function is held at once.

After master-slave mode was switched, task was rearranged.This notion is handled identical with the generation of the not preplanned mission of having explained basically.In the combination of the task module of having arranged, an optimal execution module is selected by this mechanism., because the computing machine mistake takes place, at first, handle 2215 and prepare a calculated capacity table again and upgrade the computing machine condition of confirming by the object computer control gear.Then, the previous processing of explaining 2206 to 2209 is optimally executed the task and is launched an execution command and give them selecting under precondition.In addition, the task in next time portion configuration optimization is also carried out to 2212 by the processing 2210 of previous explanation.

According to the treatment scheme of explaining above as shown in figure 22, when a computing machine mistake takes place, the tolerance of execution error immediately function, then in next time portion, that is, and after the step of 30deg, self-discipline ground obtains optimal task configuration, and execution is continued.

Use task configuration example as shown in figure 21, will be explained in the operation that the computing machine mistake goes out now by reference Figure 23.Task specification content (the interactive task restriction of execution time restriction and reliability) is identical with shown in Figure 19 those and will consider that a situation of generation computing machine mistake in time portion, its task dispose identical with as shown in figure 21.

In Figure 23, computing machine 2301 to 2303 is identical with as shown in figure 21 computing machine 2101,2104 and 2107 and the task configuration is also identical.When computing machine 2301 becomes when unavailable, [Task1] TrPrMN (2304 in the accompanying drawing) and [Task5] input and output P/P ＇ monitoring (2305 in the accompanying drawing) become and can not carry out.To explain the operation of object computer control gear during this time according to process flow diagram as shown in figure 22 below.

At first, handling object computer control gear judgement mistake in computing machine X (2301 shown in Figure 23) in processing 2202 of interrupting in 2201 and beginning.In this case, program moves on to fault tolerant startup processing immediately.Handling in 2213, mechanism confirms that fault tolerant task [Task3] TrPrMN-FT is carried out by the computing machine Y of [Task1] TrPrMN (2302 shown in the figure).For [Task5] input and output P/P ＇ monitoring, there is not the fault tolerant task.Handling 2214 in this mechanism immediate command [Task3] TrPrMN-FT switching master-slave mode thereafter.As shown in figure 24, [Task3] TrPrMN-FT 2401 that represents in the accompanying drawings provides the Task1 function of replacement [Task1] TrPrMN or equivalent continuously.

Next, handling in 2215, after computing machine did not have available condition reflection, this mechanism carried out and handles 2216, was used for planning a better task by the predistribution task module and carried out configuration, if necessary.Here, particularly, the search of this mechanism is used to cancel the interactive task restriction (exclusive level 2) of the reliability of [Task2] TrPrFD 2402 shown in [Task1] TrPrMN 2401 shown in the accompanying drawing and the accompanying drawing and a combination of the execution time restriction obstruction (execution priority level 2) of [Task3] TrPrMN-FT of not carried out by any computing machine, though there is not its solution in this example.Therefore, task configuration structure as shown in figure 24 is the urgent configuration of a mistake among the computing machine X.Handling in 2207, an execution command is transmitted to task 2401 to 2403.Owing to do not have task to be added and to carry out here, do not execute the task and distribute and the execution command problem so handle 2208 and 2209.

Next, this mechanism promptly, in the next part of 30deg electric angle after the past, plans optimal configuration and task module is carried out predistribution in next time portion.At first handling in 2210, this mechanism checks a task requests.In this case, according to task specification content as shown in figure 19, this mechanism confirms the task requests of Task1 to Task4.Then, handling in 2211, this mechanism carries out next task configuration optimization.As a result, this mechanism can obtain solution (task configuration) as shown in figure 24.

When as shown in figure 25 task configuration during with comparing as shown in figure 24, though the configuration of [Task4] shown in the drawings TrPrFD-FT 2502 and [Task2] shown in the drawings TrPrFD 2503 is exchanged, but the configuration of [Task1] TrPrMN 2501 in the accompanying drawing does not change.Therefore, can cancel the interactive task restriction (exclusive level 2) of the reliability of [Task1] TrPrMN and [Task2] TrPrFD., about the execution time restriction obstruction (execution priority level 2) of [Task3] TrPrFD-FT, there is not to find to be used to cancel its configuration.Therefore, obtain task configuration shown in Figure 25.

At last, this mechanism is handling next task module of predistribution in 2212.Particularly, this mechanism is pre-assigned to the task module of [Task4] TrPrFD-FT computing machine Y and the task module of [Task2] TrPrFD is pre-assigned to computing machine Z.Thus, in next time portion, work out realizing as shown in figure 25 of task and dispose necessary task module.

If computing machine condition in next time portion and task requests condition do not change, promptly, in the next part of 30deg electric angle after the past, only be transmitted to task 2501 to 2503 by an execution command, the object computer control gear just can continue an optimal task to be handled.

[effect of the present invention]

By distributed processing system(DPS) of the present invention, carry out for real according to structure and the capacity of computer The successional task of real time capacity that now has reliability is distributed. In addition, by the scheduling of executing the task, This distributed processing system(DPS) can realize high reliability all the time intentionally. Simultaneously for needing whenever One of the request of Shi Fasheng task start process or in the situation of computer mistake or maintenance, can To obtain being used for realizing that the successional function of CPU service efficiency with reliability keeps.

Claims (24)

1. a message handler is used to handle a plurality of tasks by a plurality of computer realization protective relay functions that connect by network, and this message handler comprises:

Device is used to store among described a plurality of task the combination of will being handled by a certain computer of task; And

Device, be used to select by the described task combination of described specific Computer Processing any one and carry out the described task of described selected combination by described specific computing machine.

2. message handler as claimed in claim 1, wherein, described a plurality of tasks comprise: normal system task is used for carrying out all the time the processing that is included in protective relay; With the back-up system task, be used in case of emergency carrying out the described processing that is included in described protective relay; And

The described combination of the described task that can be handled by described certain computer does not comprise the combination of described normal system task and described back-up system task.

3. message handler as claimed in claim 1, wherein, described a plurality of tasks comprise: mainly detect task, be used to realize the main measuring ability of protective relay; With the fault detect task, be used to realize the failure detection operations of protective relay; And

The described combination of the described task that can be handled by described certain computer does not comprise the main detection task and the fault detect task of same protective relay.

4. message handler as claimed in claim 1, wherein, described a plurality of tasks comprise: assemble the bus protection task, be used to realize the assembling bus protection function of bus protection relay; With separate bus protection task, be used to realize the separate bus defencive function of protective relay; And

The described combination of the described task that can be handled by described certain computer does not comprise the assembling bus protection task of same bus and the separate bus protection task of same bus.

5. message handler as claimed in claim 1, wherein, described a plurality of tasks have the transmission line task, are used to realize the transmission line protective relay function of each transmission line of being made of parallel circuit, and

The described combination of the described task that can be handled by described certain computer does not comprise a plurality of transmission line tasks of the transmission line protection that belongs to same parallel circuit.

6. a message handler is used to handle a plurality of tasks by a plurality of computer realization protective relay functions that connect by network, and this message handler comprises:

Device is used to store the task combination of not handled by a certain computer among described a plurality of task; And

Device, be used for by specific computing machine carry out task combination beyond the described task combination of being handled by described certain computer.

7. message handler as claimed in claim 6, wherein, described a plurality of tasks comprise: normal system task is used for carrying out all the time the processing that is included in protective relay; With the back-up system task, be used in case of emergency carrying out the described processing that is included in described protective relay; And

Can't help the described combination of the described task that described certain computer handles comprises the combination of described normal system task and described back-up system task.

8. message handler as claimed in claim 6, wherein, described a plurality of tasks comprise: mainly detect task, be used to realize the main measuring ability of protective relay; With the fault detect task, be used to realize the failure detection operations of described protective relay; And

Can't help the described combination of the described task that described certain computer handles comprises the main detection task and the fault detect task of same protective relay.

9. message handler as claimed in claim 6, wherein, described a plurality of tasks comprise: assemble the bus protection task, be used to realize the assembling bus protection function of bus protection relay; With separate bus protection task, be used to realize the separate bus defencive function of protective relay; And

The described combination of the described task of being handled by described certain computer comprises the assembling bus protection task of same bus and the separate bus protection task of same bus.

10. message handler as claimed in claim 6, wherein, described a plurality of tasks have the transmission line task, are used to realize the transmission line protective relay function of each transmission line of being made of parallel circuit, and

Can't help the described combination of the described task that described certain computer handles comprises a plurality of transmission line tasks of the transmission line protection that belongs to same parallel circuit.

11. highly reliable real-time distributed process system, be used for carrying out a plurality of tasks by a plurality of computing machines that one or more networks connect, wherein, when carrying out a computing-this computing is that each of described a plurality of tasks adds restriction or the condition-time of determining to finish within the time described task executions pre-, judge and be used to carry out between restriction or a plurality of any task or the restriction of each identical mutual relationship of various computing machine or the abundant degree of condition, thereby can carry out in the scope of all described computing machines, obtain maximum or enough reliabilities.

12. highly reliable real-time distributed process system as claimed in claim 11, wherein, be used for estimating that in described pre-described each execution time of described execution of determining to finish within the time described a plurality of tasks index is provided, and an object computer that is used for being carried out by different computing machines each described mutual relationship between described many tasks estimates that index is provided, and

Described system has an object computer control gear, is used for carrying out within all described computing machines can the execution scope that of the task that will be performed selects to handle and of computing machine that is used to calculate described selection task selects to handle and makes described estimation index maximization.

13. highly reliable real-time distributed process system as claimed in claim 12, wherein, described method has described object computer control gear, and the described execution time that is used for storing in advance at a database each described task estimates that the described object computer of each the described mutual relationship between index and the task is estimated index and described task choosing is handled and described computing machine is selected processing by carrying out with reference to described index.

14. highly reliable real-time distributed process system as claimed in claim 12, wherein, the described object computer estimation index of each mutual relationship between the task except described task is set in advance and is being used for realizing that the program module of described task and described system have described object computer control gear, is used for selecting to handle by carry out described task choosing processing and described computing machine with reference to the described index of launching from described task.

15. highly reliable real-time distributed process system as claimed in claim 12, wherein, when a title that is used to control each described task is provided, in a symbol of the described task names of expression, the described execution time that is used to describe described task is estimated a symbol of index and is used to describe between the task except described task the described execution of each described mutual relationship to calculate a symbol estimating index and comprised, therefore described system has described object computer control gear, is used for extracting the described execution time in advance from the described task names that provides and estimates index and described object computer estimation index.

16., wherein, carry out the processing of described object computer control gear by a plurality of computing machines on the described network as the described highly reliable real-time distributed process system of claim 12 to 15.

17., wherein, do not carry out the processing of described object computer control gear by being connected to one or more computing machines on the described network as the described highly reliable real-time distributed process system of claim 12 to 15.

18. as the described highly reliable real-time distributed process system of claim 12 to 16, wherein, described object computer control gear is monitored after the condition changing of the operating conditions of described computing machine and any described computing machine on detecting described network all the time, re-executes that described task choosing is handled and described computing machine is selected processing.

19. highly reliable real-time distributed process system as claimed in claim 18, wherein, the existence of the task that the control of described object computer control gear will produce from current calculating section in the future and about the described estimation index of described task, thus obtain that described task choosing is handled and described computing machine is selected the optimal solution and the scheduling of executing the task of processing by a plurality of time portion.

20. highly reliable real-time distributed process system as claimed in claim 18, wherein, when a new task in the content that is not included in described object computer control gear control started the request generation or issued the startup command of a new task from the outside, described object computer control gear re-executes described task choosing processing and described computing machine is selected to handle.

21. highly reliable real-time distributed process system as claimed in claim 18, wherein, when ordering application conditions of described computing machine by one of an outer computer issue, described object computer control gear is supposed described computer application condition according to the content of described order, and described task choosing is handled and described computing machine is selected to handle thereby re-execute.

22. highly reliable real-time distributed process system as claimed in claim 11, wherein, be used for estimating that in described pre-described each execution time of described execution of determining to finish within the time described a plurality of tasks index is provided, and an object computer that is used for being carried out by different computing machines each described mutual relationship between described many tasks estimates that index is provided, and

Described system has a developing instrument, can be designed at first carry out within all described computing machines can the execution scope that of the task that will be performed selects to handle and of computing machine that is used to calculate described selection task selects to handle and makes described estimation index maximization.

23. as the described highly reliable real-time distributed process system of claim 11 to 22, wherein, for each of a plurality of communication channels that are used for connecting between a plurality of computing machines, described system has a group communication channel monitoring task and is used to examine the normality of content as described task, and

Described system carries out a processing, judges the restriction that is used to carry out each task that the group communication channel monitoring task by the same communication channel that is used to examine the various computing machine constitutes or the abundant degree of condition, thereby confirms the normality of each described communication channel.

24. as the described highly reliable real-time distributed process system of claim 11 to 22, wherein, described system has the described task of a plurality of computer monitoring tasks, be used to carry out and check pre-test determining computing formula and by communicating verification result with a plurality of other computing machines that are connected by network, and

Described system carries out a processing, judge the restriction be used to carry out each task that the group network monitor task by the identical network that is used to examine the various computing machine constitutes or the abundant degree of condition, thereby can discern an out of order computing machine or detect the mistake of described network.

Images