JP2003140706A - Multiplexed parallel processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain restoration in a state that control is continued without stopping a system. SOLUTION: An operation control part 61 is provided with a data transfer requesting function part 611, a request accepting function part 612 and a data transferring function part 613 and a transfer data generating function part 614 and a difference data generating function part 615 and a transmitted data update monitoring function part 616 and a data receiving function part 617. After restart, a data base stored in a main storage part 122 in another controller are requested, and fetched in its own main storage part 122.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multiplex parallel processing device having a redundant configuration and controlling processes and the like.

[0002]

2. Description of the Related Art Conventionally, a process control device used for process control of a plant or the like has been used which has a redundant structure in which a plurality of controllers for performing control calculations are provided when high reliability is required. A conventional process control device having such a redundant structure generally has two controllers, a master system and a slave system, and when an abnormality occurs in the master controller, a redundant standby redundancy system is generally used in which the standby controller switches to the standby controller. is there.

However, in such a redundant standby redundancy system process control device, when the controller switches from the master system to the slave system, the slave system, which has been waiting, starts processing after detecting the stop of the master system. There is a problem in that it cannot be used for process control with a short control cycle of the process because of a control blank time. For this reason, there has been a demand for a process control device in which a control blank time does not occur even when an abnormality occurs in the operating controller.

[0004]

On the other hand, there is a method in which three or more controllers having the same configuration are operated in parallel, a majority decision is made on the control outputs of these controllers, and the process control is performed using the decision result. Conceivable. In such a triple redundant parallel control system, it is assumed that an abnormality has occurred in the control processing unit that has output a control result different from that of the other two, and the control processing unit is disconnected from the system such as by stopping the control processing unit and investigated. I am trying to make adjustments. However, in a semiconductor device or the like in which elements such as a transistor forming the control processing unit are highly integrated, it is not possible to avoid memory corruption due to cosmic rays or the like. In addition, it is practically impossible for a program for performing more advanced control to completely eliminate the bug.

Due to such memory corruption and bugs,
There may be a case where a control result different from that of the other redundant section is temporarily output at a certain timing. In such a case, stop the target control processing unit and disconnect it from the system
Adjustment is made, and after this, the multiple parallel parallel state is restored. At the time of this restoration, it is necessary to synchronize the control processing unit to be restored with another control processing unit in normal operation such as matching the database contents. Conventionally, in this restoration, other control processing units that are operating normally were temporarily stopped, and the contents of the database were duplicated. Therefore, the system was temporarily stopped.

As described above, in the conventional multiplexed parallel redundant system, when a control processing unit that has once stopped due to an abnormality or the like is to be restored, it is necessary to stop other control processing units that have been operating normally. It was necessary to temporarily stop the operation of the system. The present invention has been made to solve the above problems, and an object of the present invention is to enable recovery in a state where control is continued without stopping.

[0007]

According to one aspect of the present invention, a multiplex parallel processing apparatus generates and outputs predetermined input information and control data corresponding thereto according to preset setting conditions. A multiplexing parallel processing apparatus comprising N controllers (N is a natural number of 3 or more) for performing a control operation, wherein the controller includes a receiving unit that receives an operation stop signal sent via a network, and a predetermined unit. Re-starting means for restarting the controller to start the control operation according to conditions, data transfer requesting means for requesting transfer of specific data to another controller after restarting by the restarting means, and other specific controller Request reception means for receiving the transfer request for the specific data sent by the request, and the specific data targeted by the transfer request received by the request reception means for the specific data. After restarting by the data transfer means for transferring to the roller, the data receiving means for receiving the specific data transferred from another controller, and the restarting means, the control operation is restarted after the data receiving means receives the specific data. And an operation control section.

According to this multiplex parallel processing apparatus, after restarting, the controller itself requests transfer of specific data, and after receiving the specific data transferred from another controller by this request, its own control operation To resume. In the above multiplexed parallel processing device, the specific data is
For example, it includes setting conditions, and also includes setting conditions, predetermined information, and control data at the time when the request by the data transfer requesting means occurs.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram showing a configuration example of a multiplexing parallel processing device according to an embodiment of the present invention. The process control device 100 includes a network interface 110, a controller A 120a, a controller B 120b, a controller C 120c, and a process equipment interface 130. The network interface 110 is a monitoring device 1 via the network 103.
02, and the process equipment interface 130
It is connected to a plurality of process devices 104 via the I / O bus 105. Further, the controller A 120 a, the controller B 120 b, and the controller C 120 c are connected to each other by the internal network 140, and are also connected to the network interface 110.

In the process control device 100, first, the network interface 110 is, for example, the monitoring device 10.
2 receives a request from the network 103 and receives the received request from the internal network 14
Controller A120a, controller B1
20b, given to the controller C120c. Each controller given a request generates control data in response to the request, and the process device interface 13
Output to 0.

The process equipment interface 130 that has received the control data output from each controller transfers one control data determined by the majority decision to the I / O bus 1.
It outputs to the corresponding process equipment 104 via 05.
Further, the process data from the process equipment 104 is first transmitted to the process equipment interface 130 via the I / O bus 105. The process device interface 130 that has received this process data outputs the received process data to the controller A 120a, controller B 120b, and controller C 120c.

Each controller given the process data from the process equipment 104 generates a response corresponding to the given process data and outputs it to the network interface 110. The response output from each controller is accepted by the network interface 110 via the internal network 140. Upon receiving these responses, the network interface sends one response determined by the majority vote to the monitoring apparatus 102 via the network 103.

As described above, in the process equipment interface 110, even if any one of the controller A 120a, the controller B 120b, and the controller C 120c has an abnormality and outputs different control data and response from the other two. Therefore, normal control data and response can be output to the process equipment 104 and the monitoring device 102 without interruption.

Next, the network interface 11
0 will be described in detail with reference to FIG. The network interface 110 includes a response / request transmission / reception unit 111, a load check unit 112, and a request distribution unit 1.
13, a response receiving unit 114, a response comparing unit 115, and an abnormality notifying unit 116. In such a configuration, first, the request received via the network 103 is received by the response / request transmission / reception unit 111 and sent to the load check unit 112. The load check unit 112 to which the request is sent is the controller A 120
a, controller B120b, controller C120c
Check each processing load.

The load check unit 112 is a controller A.
120a, controller B120b, controller C1
When the number of processes executed by 20c is less than or equal to a predetermined number, it is determined that a request for the next process can be accepted. The number of processes being executed can be obtained by subtracting the number of responses from the number of processes such as transferred requests. The number of responses can be obtained from the response receiving unit 114 described later.

As described above, the processing load of each controller is confirmed, and when all the controllers are ready to accept the request, the sent request is transferred to the request delivery unit 113. on the other hand,
If there is a controller that cannot be accepted, the response / request transmission / reception unit 111 is notified that the transmitted request cannot be accepted, and the transmitted request is deleted. When it is notified that the request cannot be accepted, the response / request transmission / reception unit 111 notifies the request source that the request cannot be executed.

The request delivery unit 1 to which the request is sent
The controller 13 adds processing timing information indicating the timing to be processed to the transmitted request, and the controller A
120a, controller B120b, controller C1
Broadcast to 20c. The processing timing information to be added includes the timing of executing the request and the timing of transmitting the response generated as a result of processing the request. Here, the processing timing information is a 3-bit code indicating the processing timing in one control cycle, and the request delivery unit 113 outputs a 3-bit processing timing signal output to a data line (not shown) by the process equipment interface 130 described later. With reference to the current value of, the value of the processing timing signal generated in the future is set as the timing of executing the request.

Next, the network interface 11
The processing of the response output from each controller at 0 will be described. When the responses are transmitted from the controller A 120a, the controller B 120b, and the controller C 120c, these are received by the response receiving unit 114 and sent to the response comparing unit 115. In addition, the response receiving unit 114
The load check unit 112 is notified of the controller that has returned the response.

The response comparing section 115, to which the response is sent, compares the contents of the three received responses, and when the three are all the same or when any one is different and the other two are the same. Any one of the matched responses is sent to the response / request transmission / reception unit 111. The response / request transmission / reception unit 111 that has received the response thus transmitted transmits the received response to the request source. In addition, the response comparison unit 115 notifies the abnormality notification unit 116 of the controller that has returned a response different from the other responses. Upon receipt of the notification, the abnormality notification unit 116 sends an operation stop signal to the target controller.

The network interface 110
Is, for example, an interface for the network 103, an interface for the internal network 140, a digital input interface for processing timing signal input, a storage unit that stores a program that realizes the functions of the network interface 110 described above, and this program. Can be composed of a CPU for executing the program, a main memory used for executing the program, and a bus connecting these.

Next, the process equipment interface 13
0 will be described. As shown in FIG. 3, the process device interface 130 includes a controller communication processing unit 13
1, data comparison unit 132, I / O bus input / output processing unit 13
3, abnormality notification unit 134, processing timing signal generation unit 15
It has 0. The controller communication processing unit 131 receives commands and control data transmitted from the controller A 120a, the controller B 120b, and the controller C 120c almost at the same time, and transfers them to the data comparison unit 132. The data comparison unit 132 compares the transferred three signals such as command and control data, and when the three match and the two match, the matched data is input to the I / O bus input / output processing unit. Transfer to 133.

The I / O bus input / output processing unit 133 sends the transferred data to the corresponding process device 104 via the I / O bus 105. In addition, the data comparison unit 132
Notifies the abnormality notifying unit 134 of the controller that is the transmission source of data different from the other two. Abnormality notification section 1
At 34, in response to the notification from the data comparison unit 132, an operation stop signal is sent to the corresponding controller via the controller communication processing unit 131.

Further, the I / O bus input / output processing unit 133 is
Data from the process equipment 104 is received via the I / O bus 105, and the received data is transferred to the controller communication processing unit 131. The controller communication processing unit 131 to which the data from the process equipment 104 is transferred receives the transferred data from the controller A 120a and the controller B.
120b, controller C120c.

The processing timing signal generator 150 generates a processing timing signal and sends it to the network interface 110 and each controller. The processing timing signal to be transmitted is, for example, a 3-bit parallel signal. For example, eight processings are performed by sequentially taking eight states obtained from a combination of three bits in one control cycle by the process control device 100. Timing is provided. Note that the processing timing signal is not limited to 3 bits, and another number of bits may be used.

Next, the controller A 120a, controller B 120b, and controller C 120c will be described. These have the same configuration, and hereinafter, the controller A 120a will be described as a representative with reference to FIG. As shown in FIG. 4, the controller A 120 includes a program storage unit 121, a set value 122a and a current value 12 which will be described later.
2b and a control data 122c, a main storage unit 122 storing a database, a network control unit 123,
An arithmetic processing unit (CPU) 124, an I / O control unit 125, and an abnormality processing unit 126 are provided, and these are connected by a bus 127.

The program storage unit 121 stores a startup program and a control program for the arithmetic processing unit 124 to perform a startup operation and various control operations. In the controller A 120a, each program stored in the program storage unit 121 is stored in the main storage unit 122.
The arithmetic processing unit 124 is operated by the program loaded into the main storage unit 122 and each control operation is performed. For example, the arithmetic processing unit 124 uses the program loaded in the main storage unit 122 to execute the request processing function unit 1
24a, a control data generation function unit 124b, and a control processing function unit 124c are realized.

For example, in the controller A 120a, the set value initially sent from the monitoring device 102 (FIG. 1) is
Network 103, network interface 11
0, when the network control unit 123 receives it via the internal network 140, the arithmetic processing unit 124 analyzes the content by the request processing function unit 124a and stores it as a set value 122a in a predetermined area of the main storage unit 122. On the other hand, the arithmetic processing unit 124 collects predetermined control values and state values from the process equipment 104 by the control processing function unit 124c, and stores them in the main storage unit 122 as the current value 122b.

Further, the arithmetic processing section 124 uses the control data generation function section 124b to generate control data and commands by referring to the set value 122a and the current value 122b stored in the main storage section 122, and to generate this. The current value 122b is stored in a predetermined location in the main storage unit 122. Arithmetic processing unit 1
24, when the control data 122c is stored in the main storage unit 122 as described above, the control processing function unit 124c
The control data 122c stored by is sent to the target process device 104. Control data 1
22 c is sent to the process equipment 104 via the process equipment interface 130 by the I / O control unit 125.

When the network control unit 123 receives a request from the monitoring device 102, the arithmetic processing unit 124 responds to the received request by the request processing function unit 124a. For example, when a request for the current control value of the process equipment 104 is received, the request processing function unit 124 a in the arithmetic processing unit 124 analyzes the content and stores the current value 1 stored in the main storage unit 122.
22b, the corresponding control value is searched and retrieved,
The retrieved control value is sent to the monitoring device 102 as a response.

The abnormality processing unit 126 includes an operation control unit 61, a watchdog timer (WDT: restarting means) 62, and a timer initialization unit 63. The WDT 62 outputs a reset signal, and by this reset signal, the controller A 120a itself is restarted and initialized, and the control operation is restarted. The WDT 62 measures time and outputs a reset signal when a predetermined time T has elapsed. The timer initialization unit 63 measures the time and measures the time of the WDT 62 when a time t shorter than the time T has elapsed. Is to be initialized.

The operation control unit 61 controls the timer initialization unit 63 to perform an operation during normal operation. For example, the abnormality notification unit 116 of the network interface 110 or the abnormality notification unit of the process equipment interface 130. When the operation stop signal is sent from 134,
The operation of the timer initialization unit 63 is stopped.

In addition, in the present embodiment, the operation control unit 6
1 is a data transfer request function unit 611 and a request reception function unit 6
12, a data transfer function unit 613, a transfer data generation function unit 614, a difference data generation function unit 615, a transmitted data update monitoring function unit 616, and a data reception function unit 617, and after restarting, the main memory in another controller. The database stored in the unit 122 is requested and loaded in the main storage unit 122 of itself. The operation control unit 61 including these functional units is realized by an embedded processor that includes the above-described program and operates by itself.

The operation after the restart will be described below with reference to the flow chart of FIG. In addition, the case where the controller A 120a (FIG. 1) is restarted will be described below. First, when restarting (step S501), in the operation control unit 61 of the controller A 120a, the data transfer request function unit 611 requests the other controller via the internal network 140 to transfer the database in the main storage unit 122. Yes (step S502). In another controller that has received this request by the request reception function unit 612 (step S503), the transfer data generation function unit 614 in the operation control unit 61 has its own main storage unit 1
Check the status of the database in 22 and
Transfer data for 120a is generated (step S5).
04).

In another controller which has generated the transfer data, the data transfer function unit 613 is used in the operation control unit 61.
Transfers the transfer data generated via the internal network 140 to the controller A 120a (step S5).
05). Controller A120 to which the transfer data is transferred
At a, in the operation control unit 61, the data reception function unit 617 receives the transfer data, confirms that the plurality of transfer data transferred from other controllers are the same, and stores this in the main storage unit 122. (Step S
506).

In the other controllers, the transmitted data update monitoring function unit 616 in the operation control unit 61 has its own main storage unit 1 until the data transfer is completed.
When a change is detected in the database in step 22 and the change is detected (step S507), the difference data generation function unit 615 generates a difference from the state before the change of the changed data (step S508). Thereafter, in the other controller, the data transfer function unit 613 transmits the generated difference to the controller A1 via the internal network 140.
20a (step S509).

The plurality of differences transferred from other controllers are received by the controller A 120a and added by the data receiving function unit 617 to the database already stored in the main storage unit 122 (step S5).
10). After that, the controller A 120a starts the control operation like the other controllers. From the above, it is possible to restart the controller A 120a without stopping the other controller, and immediately after the restart, the control state of the other controller is reflected and the control operation is restarted promptly. Will be able to.

In the above embodiment, the transfer request of the database and the transfer of the transfer data are described as being performed via the internal network 140, but other networks (communication means) to which the network interface 110 is not connected may be used. It goes without saying that it may be carried out via.

[0038]

As described above, according to the present invention,
After restarting, the controller itself requests transfer of specific data, and after accepting the specific data transferred from another controller by this request, it restarts its own control operation, so without stopping, The excellent effect that the recovery can be performed while the control is continued is obtained.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a configuration example of a process control device 100 which is an example of a multiplexing parallel processing device according to an embodiment of the present invention.

FIG. 2 is a configuration diagram showing a configuration example of a network interface 110 that constitutes the process control apparatus 100 of FIG.

3 is a configuration diagram showing a configuration example of a process equipment interface 130 which constitutes the process control apparatus 100 of FIG.

4 is a configuration diagram showing a configuration example of a controller A 120a that constitutes the process control apparatus 100 of FIG.

FIG. 5 is a flowchart showing an operation example of a controller.

[Explanation of symbols]

61 ... Operation control unit, 62 ... Watchdog timer (WD
T), 63 ... Timer initialization unit, 100 ... Process control device, 102 ... Monitoring device, 103 ... Network, 104
... process equipment, 105 ... I / O bus, 110 ... network interface, 120a ... controller A, 1
20b ... Controller B, 120c ... Controller C,
121 ... Program storage unit, 122 ... Main storage unit, 122
a ... Set value, 122b ... Current value, 122c ... Control data, 123 ... Network control section, 124 ... Arithmetic processing section, 124a ... Request processing means, 124b ... Control data generating means, 124c ... Control processing means, 125 ... I /
O control unit, 126 ... Abnormality processing unit, 127 ... Bus, 130
... process equipment interface, 611 ... data transfer requesting means, 612 ... request receiving means, 613 ... data transferring means, 614 ... transfer data generating means, 615 ... difference data generating means, 616 ... transmitted data update monitoring means, 61
7 ... Data receiving means.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) G06F 15/177 678 G06F 15/177 678C F term (reference) 5B034 AA05 5B045 AA05 GG11 JJ13 JJ22 JJ26 JJ45 JJ48 5B083 A01 BB03 CC04 CE01 DD13 EE02 GG04 5H209 DD04 EE11 HH22 JJ07 SS02 SS04 SS07 TT04

Claims (3)

[Claims] 1. N (N is a natural number of 3 or more) controllers for performing a control operation of generating and outputting predetermined input information and control data corresponding to the input predetermined information according to preset setting conditions. A multiplexing parallel processing apparatus comprising: the controller, a receiving unit that receives an operation stop signal transmitted via a network, a restarting unit that restarts the controller according to a predetermined condition, and the restarting unit. After restarting by the means, a data transfer requesting means for requesting the transfer of the specific data to the other controller, a request receiving means for receiving the transfer request for the specific data sent from the other specific controller, and the request receiving means. Data transfer means for transferring the specific data that is the target of the transfer request received by the specific controller to the specific controller, and another controller. Data receiving means for receiving the transferred specific data, and an operation control unit for restarting the control operation after the data receiving means receives the specific data after restarting by the restarting means. A characteristic multiplex parallel processing device. 2. The multiplexing parallel processing device according to claim 1, wherein the specific data includes the setting condition. 3. The multiplexed parallel processing apparatus according to claim 1, wherein the specific data includes the setting condition, the predetermined information, and the control data at a time point when the request is made by the data transfer requesting means. A multiplex parallel processing device characterized by being nothing.