JP2766362B2 - Error information file generation method for power control digital controller - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial application field) The present invention relates to an error information file generation method for improving a RAS (reliability / maintenance) function of a power control digital controller.

(Prior Art) The RAS function is one of important functions in a power control digital controller applied to a microprocessor. In operation status, watchdog timer, memory diagnosis, response check between connected hardware modules or bus transfer check Component-based diagnosis, verification of plant operation (eg, reasonable check of analog input, check of response to output) Various diagnostics for safe driving are performed according to the configuration. Recently, a function of recording a result of the diagnosis on a memory in the apparatus and utilizing the result for failure analysis has also been implemented.

FIG. 4 shows a conventional example of such an error information file generation method for recording a diagnosis result in a memory and utilizing it for failure analysis. FIG. 4 shows a file in which error information is set in a certain memory space. This method has a space of n words, and assigns an error type to each bit in each word of each address in advance. For example, when a memory parity error occurs in D15 of ADRSa, this bit is set to "1".

However, in this method, it is not clear when monitoring how many memory parity errors have occurred. Also, if D1 of ADRSa is set to “1” when the unmounted memory space is accessed for some reason, if a memory parity error occurs due to this factor, D1
Also set with 5, D14. For this reason, when monitoring later, it is unclear which is the true cause. That is, the memory is broken, the microprocessor executes based on the incorrect memory read data, accesses the space that is not normally accessed, and detects a memory parity error, or the microprocessor erroneously fetches due to noise or poor timing. It is not clear whether a parity error has occurred due to access to the unmounted space, and it is difficult to analyze.

In order to eliminate such a problem, an improved system shown in FIG. 5 is known. In this method, information is logged in one block every time an error occurs. In the example of this figure, one block is composed of four words, and an error code, an occurrence location, and related information (for example, an address counter value of a microprocessor at the time of occurrence of a memory parity error) occurrence time are set for each word.

In this method, the state of occurrence can be monitored later in a time series, and failure cause analysis becomes easier. However, the area is limited, and when the device is left unattended, a mechanical large-capacity external memory such as a hard disk cannot be attached, and this file is created in a limited space of the main memory. Therefore, although there are n blocks in FIG. 5, as a method of logging a failure that occurs after the n blocks are logged, (1) no more logs (2) a circular list overwriting from 0 block Format There are two ways.

(Problems to be Solved by the Invention) However, in the method (1), a trigger that fails can be left, but it is not clear what state the device is in before being logged and monitored in n blocks. If the logged content is not necessarily sent to the outside (logging but successful in retry, no alarm will be issued to the outside, for example, transmission line disconnection is detected for a moment, If there is no effect on the plant during operation), the subsequent analysis may be insufficient data.

On the other hand, in the above method (2), the latest error information before monitoring remains, but if overwritten, it becomes unknown what the first error was, and the causal relationship with the obtained error information may not be understood. There is.

Therefore, the present invention is a power supply that detects errors including over-abnormal errors due to minor failures that do not cause a system down in the device, external factors such as lightning strikes and ground faults, etc. An object of the present invention is to provide an error information file generation method for a control digital controller, which is useful for analyzing a cause of a failure and improves maintainability.

[Structure of the Invention] (Means for Solving the Problems) In the present invention, when writing error information to an error information file having a structure composed of n blocks, m (m
<N) pieces of error information are written and stored in order, and error information that occurs more than that is written cyclically in the remaining (mn) blocks to store the latest (mn) pieces of information. It was made.

(Operation) Since the initial error information group has been saved since the start of plant operation, the trigger of the error factor can be identified. Even if the latest error information group is saved, even if intermediate error information is missing. A certain degree of guess can be added, the transition of the entire error information can be grasped, and the analysis of the error factor becomes easy.

(Embodiment) FIG. 1 shows a configuration diagram of an information file according to an embodiment of the present invention.

This error information file is provided in the main memory (writable RAM area) of the power control digital controller, and is composed of n blocks. The blocks from 0 to m are used as blocks for logging “initial error information” generated after the initialization of the error information file, and (nm) blocks from m + 1 to n are referred to as the “latest error information” in a circular list format. "Information" as the log part.
Each time an error is detected, the diagnostic process performs a block 0, 1, 2, ...,
Error information is logged in the order of m, ... If it is necessary to log the error information even after the block n is reached, overwriting is performed from the block m + 1 which is the start of the portion in the circular list format. Hereinafter, this is repeated.

According to the above-described method, information of m errors that occurred first after the initialization of the error information file is logged in the “initial error information recording portion (1)”, and the latest (nm) pieces of error information are recorded. In the “Latest error information recording part (2)”
Will be logged.

FIG. 2 is a diagram showing a configuration of one block of the error information file shown in FIG.

 One block consists of the error date (3) Error hour / minute / second (4) Unit No. (5) Slot No. (6) Card type (7) Error No. (8) Related information (9) Continuous occurrence It consists of the number of times (10).

The date of occurrence of the error (3) and the hour, minute, and second (4) of the occurrence of the error are portions for logging the value of the “device clock” at the time of the occurrence of the error.

The unit No. (5), the slot No. (6) and the card type (7) are the unit No. to which the board judged as an error as a result of the diagnosis belongs. If it is composed of one or a plurality of chassis, it logs the slot number indicating the position where the board is set in the chassis and the card type indicating the type of the board.

The error No. (8) is a part for logging the No. indicating the type of error as a diagnosis result.

The related information (9) is a part for logging related information useful for failure cause analysis for each error number. For example,
This is an address counter value of the microprocessor when a memory parity error occurs.

The number of continuous occurrences (10) includes a part for logging the number of continuous occurrences when the same error occurs continuously for the same substrate. This is provided to avoid logging the exact same error information to a plurality of blocks continuously. In addition, since the number of blocks in the error information file is actually limited, this is an effective means for particularly small error information files.

The RAM area secured as the error information file is desirably provided in a battery-backed RAM area in order to prevent the error information file from being lost due to a power failure or momentary power failure of the device.

FIG. 3 is a flowchart showing a process in which the control unit logs the error information to the error information file shown in FIGS. 1 and 2 by the power control digital controller (not shown).

When the control unit starts the process in response to the error information file clear request (11), first, it clears all blocks of the error information file (12). Next, the block number for which the error information is to be logged next is represented by a "block pointer" and is cleared to 0 (13). If an error occurs, the following steps are executed (14).
Check whether an error has already occurred (1
Five). If an error has already occurred, that is, if “block pointer ≠ 0”, the flow branches to processing (16) to check whether or not the error is the same as the current error by referring to the information of the error that occurred last time. If no error has occurred, that is, if “block pointer = 0”, the flow branches to processing (20). If an error has already occurred, the previous error is compared with the current error (16). That is, it is checked whether the error number is the same as the previous time. If the error number is the same as the previous one, the process branches to the process (18) for the next comparison. If the error numbers are different, the process branches to processing (20), and error information is logged in the block indicated by the block pointer as an error different from the previous time. Also, the unit N of the board that caused the error
Check if o. and slot number are the same as before. In the case of the same substrate as the previous time, the process branches to the process (19) to perform the same error process as the previous time. If the board is different from the previous one, the process branches to processing (20), and error information is logged in the block indicated by the block pointer as an error different from the previous one. When it is diagnosed that the error that occurred last time and the error that occurred this time are the same, the value of the “number of consecutive occurrences” of the previous block, that is, the block of (block pointer 1), is incremented by one. Then, it waits until the next error occurs (19). As the first error or an error different from the previous error, new error information (ie, date, hour, minute, second, unit No., slot No., card type, error No., related information, Log the number of consecutive occurrences = 0) (20 to 24).

As a result, since the error information has been logged in the block indicated by the current block pointer, the process 25 increments the block pointer by 1 so as to indicate the block in which the error information is to be logged next. It is determined whether or not the block pointer has become larger than n as a result of the increment (26). When the block pointer becomes larger than n, that is, when the error information is logged in all the blocks (0 to n) of the error information file, as shown in FIG. In order to make it possible, the value of the block pointer is reset to "m + 1". If the block pointer is equal to or less than n, nothing is performed. After that, it waits until the next error occurs.

In order to refer to the error information logged by the diagnostic processing and to issue a request to clear the error information file to the main unit, a terminal or the like is connected to the power control digital controller main unit, and this is operated. Is done. The clearing of the error information file can also be realized by providing a switch on the substrate for clearing the error information file provided on the RAM and operating the switch.

As described above, the error information file is all cleared by a human operation, and the error information file is cleared by the operation of the maintenance staff at the start of the operation of the plant (or at the time of periodic inspection). For example, when the content of the error is a memory parity error, the operation is restarted after the self-reset. In this case, the address where the memory parity error has occurred is logged in the error information file as related information together with the error number indicating the memory parity error. Based on this related information,
The cause of failure is analyzed, but if the error information file is cleared at the time of restart, this analysis cannot be performed. For this reason, in this embodiment, the error information file is cleared by a human operation.

As described above, after the error information file is cleared (that is, after the plant operation is started), a plurality of pieces of error information generated in a relatively small file area are left, and
The latest error information up to the monitor can be obtained. As a result, a trigger for an error factor can be obtained, and the relation with the error information generated during this time can be assumed to some extent, which becomes useful data for analysis.

As another embodiment, an “initial error information recording portion” capable of logging m errors from the first error detected after clearing the error information file, and a circular list format file is generated after detecting m + 1 or more errors. It is also conceivable to set separately from the “latest error information recording portion”.

[Effects of the Invention] As described above, according to the present invention, failure analysis is facilitated by the error file structure and generation means, and power control digital control with a more complete RAS function can be achieved. This can be useful for improving maintainability and extracting improvements.

[Brief description of the drawings]

FIG. 1 is a diagram showing the configuration of an error information file showing one embodiment of the present invention, FIG. 2 is a diagram showing the configuration of one block of the error information file shown in FIG. 1, and FIG. FIG. 4 is a flowchart showing a method of logging error information to the error information file of FIG. 2, and FIGS. 4 and 5 are block diagrams of a conventional error information file.

Claims (1)

(57) [Claims] In an error information file generation method of a power control digital controller having a function of logging an accident diagnosis and a diagnosis result in a memory in addition to a plant control function, a log storing n error information in a memory area. A file is set, and m pieces of error information generated after the initialization of the log file are sequentially stored and saved, and error information generated thereafter is cyclically stored in (nm) storage locations. Error information generation method for power control digital controller.