A cross substation line massive telemetry peer-end distributed processing method and system
By modeling and segmenting substation lines, and combining distributed architecture and asynchronous processing, the problem of insufficient telemetry data processing capability in traditional power grid control systems has been solved, achieving high throughput and low latency telemetry data processing and improving data qualification rate.
Patent Information
- Application Number
- CN202311362980.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-20
- Publication Date
- 2025-12-26
- Estimated Expiration
- 2043-10-20
AI Technical Summary
Traditional SCADA systems, which employ a primary/backup architecture, cannot meet the high-throughput telemetry data processing requirements of large power grids, resulting in data delays and reduced accuracy, thus affecting the safe operation of the power grid.
A distributed processing method for massive telemetry data across substation lines is adopted. This method involves modeling substations, AC lines, and endpoints, processing them in segments, and performing distributed peer-to-peer processing based on endpoint type. It uses asynchronous processing mechanisms and a distributed architecture to reduce cross-segment access.
It improves data processing speed and pass rate, reduces latency, provides reliable data assurance for advanced applications, and supports massive telemetry processing for large power grids.
Smart Images
Figure CN117439260B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to a power grid dispatching automation management system distributed SCADA real-time data processing method and system, in particular to a cross substation line massive telemetry opposite end distributed processing method and system. BACKGROUND
[0002] In the power system, substation is connected to each other through AC line, and the AC line has two line ends, which belong to different substations. Each substation collects real-time data of the equipment in the station and sends it to the master station front-end system. According to the telemetry quality code of the two ends of the line, when the telemetry of one end is abnormal and the telemetry of the other end is normal, SCADA automatically replaces the data when receiving the front-end data, and uses the telemetry of the normal end to replace the telemetry of the abnormal end. Among them, the active value and the reactive value are replaced in reverse, and the current value is directly replaced. In addition, if the dispatcher considers that the data of a certain line end is incorrect, the measurement point can be set to manual opposite end replacement by manual setting, and the telemetry of the opposite end of the line is forced to be replaced. The automatic opposite end replacement and the manual opposite end replacement operation complements the system data in a large area, effectively improves the data qualification rate of the whole system, and provides reliable basic data guarantee for AGC, AVC and other advanced applications.
[0003] The traditional power grid control system SCADA adopts a master-slave architecture, and the SCADA master and the backup machine process all telemetry data at the same time, and write the processed data into the local real-time database. The master machine is responsible for providing data access outside. The real-time telemetry processing performance of this system is limited within several ten thousand TPS, and cannot support horizontal expansion. In a large power grid control system, the real-time throughput rate of telemetry transmitted from the front-end to the SCADA background is as high as several hundred thousand TPS. The traditional master-slave control system cannot meet such a huge throughput rate. When the throughput rate increases beyond the processing capacity of the system, data accumulation occurs, which in turn causes a large delay in data, a significant decrease in data qualification rate, and a serious impact on the control decision of the advanced application of the power grid AGC, AVC, and even a threat to the safe operation of the power grid. In order to ensure the low delay and high throughput rate of telemetry processing, the new generation of power grid control system adopts a distributed architecture for telemetry processing. The endpoints of the two ends of the line are inevitably distributed in different shards, and mutual access to the model and real-time data of the opposite side shard is required in the opposite end processing process. This cross-shard data access seriously affects the data processing performance. SUMMARY
[0004] The purpose of the present application is to provide a cross-substation line massive telemetry opposite end distributed processing method and system, which solves the performance problem of cross-shard telemetry opposite end processing of AC lines in a large power grid.
[0005] Technical scheme: The cross-substation line massive telemetry opposite end distributed processing method provided by the present application comprises the following steps:
[0006] (1)Modeling substation, AC line, AC line end point and peer information;
[0007] (2)Fragmenting the modeled model;
[0008] (3)Identifying end point type according to the result of fragmenting;
[0009] (4)Distributed peer replacement according to different end point types.
[0010] Further, the fragmenting of step (2) includes fragmenting the substation according to the dispatching mechanism of the region to which it belongs, attributing the substation of one region to the same fragment, reducing the cross-fragment connection between substations; each AC line belongs to two different substations and is not fragmented; the fragment number of the AC line end point is consistent with the fragment number of the substation to which it belongs; and the fragment number of the peer information table is consistent with the fragment number of the substation of the replaced end point.
[0011] Further, the specific steps of the end point type identification of step (3) are as follows:
[0012] (31)Reading all fragment information of the node;
[0013] (32)Reading all AC line end point information by network interface, including end point ID and end point belonging line ID;
[0014] (33)Constructing a hash table with line ID as key and end point pair as value, and pairing the end point pairs to form an end point mapping relationship;
[0015] (34)Iterating the hash table to determine whether the two end points in each line end point pair belong to the same fragment, if the end point pair belongs to the same fragment, it is an intra-fragment end point, and if the end point pair does not belong to the same fragment, it is an inter-fragment end point; for the inter-fragment end point, determine whether the fragments to which the two end points belong are distributed in the same node, if the fragments are distributed in the same node, the end point pair is a node-intra-fragment end point, and if the fragments are not distributed in the same node, it is a cross-node inter-fragment end point.
[0016] Further, the replacement types of the peer replacement include manual peer replacement and automatic peer replacement, the manual peer replacement does not judge the telemetry quality code and forcibly replaces the opposite side with the current side; the automatic peer replacement automatically determines whether to replace according to the telemetry quality code of the measurement end point, the priority of the manual peer replacement is higher than that of the automatic peer replacement; the telemetry quality code of the replaced opposite side is stacked in the state of being replaced by the opposite side; when the peer replacement occurs for the first time or is canceled, an alarm notification is sent.
[0017] Further, the peer-end processing of step (4) includes in-slice peer-end processing, inter-slice peer-end processing within a node, and inter-slice peer-end asynchronous processing across nodes; through an asynchronous processing mechanism, the peer-end processing all adopts a real-time library local interface; the peer-end processing adopts a distributed processing architecture, and a plurality of slices are distributed on each SCADA node, and a telemetry processing process is started on each slice, and the telemetry processing process of each slice only receives telemetry of a channel where the slice is located and can send inter-process interaction information to other channels.
[0018] Further, specific steps of the in-slice peer-end processing are as follows:
[0019] (41a) The telemetry processing process receives telemetry data of an end point;
[0020] (41b) Through the peer-end information table, whether there is a peer-end processing record of the end point and the peer end point and a replacement type and a replacement coefficient of the peer-end processing are queried;
[0021] (41c) If the replacement type of the end point is manual peer-end processing and the peer end point is replaced, the telemetry value of the peer end point is read, multiplied by the replacement coefficient, and then the telemetry value of the current end point is replaced to perform database storage processing; if the peer end point is replaced, the telemetry value of the end point is replaced to perform database storage processing;
[0022] (41d) According to whether there is an automatic replacement record in the peer-end information table and whether the quality codes of the two end points of the line are normal, processing is performed according to the in-slice end point peer-end processing logic.
[0023] Further, the inter-slice peer-end processing within a node is consistent with the in-slice processing, but when accessing data of the peer end point, the slice number of the real-time library access interface of the peer end point needs to be set.
[0024] Further, specific steps of the inter-slice peer-end asynchronous processing across nodes are as follows:
[0025] The node where the current end point a is located is provided with a slice x, and a corresponding telemetry processing process Pm receives telemetry information of a front channel x; the node where the peer end point b is located is provided with a slice y, and a corresponding telemetry processing process Pn receives telemetry information of a front channel y; first, the current peer-end pre-processing is performed:
[0026] (43a-1) The telemetry processing process Pm receives the telemetry value and the quality code of the end point a sent by the front channel x;
[0027] (43a-2) The ID and the slice number y of the peer end point b are acquired, and the channel number of information interaction is determined according to the peer slice number;
[0028] (43a-3) Query the opposite end proxy information table to determine whether the opposite end proxy record of end point a exists, and obtain the proxy type and proxy coefficient;
[0029] (43a-4) Send the telemetry of end point a and the queried relevant information to the telemetry processing process Pn on the node where the opposite end point is located, and the message type is MT_YC_EXCHANGE_GO;
[0030] Secondly, after the telemetry processing process Pn of the opposite end point b receives the MT_YC_EXCHANGE_GO message, the opposite processing is entered:
[0031] (43b-1) Analyze the MT_YC_EXCHANGE_GO message to obtain the telemetry value and quality code of end point a, and determine whether the b end needs to be automatically proxied by the opposite side, whether the b end can replace the opposite side, and whether the b end needs to be manually proxied by the a end;
[0032] (43b-2) Query the local real-time database to determine whether the b end needs to be proxied by the opposite side according to the quality code of the b end telemetry and the record of the b end proxy information, whether the b end can replace the opposite side, and whether the b end needs to be manually proxied by the a end;
[0033] (43b-3) If the a end needs to be manually proxied by the b end, regardless of whether the b end telemetry quality code is normal, replace the telemetry value and quality code of end point a with the product of the telemetry value and the proxy coefficient of the b end and the quality code thereof, and send back to the opposite process Pm, and the message type is MT_YC_EXCHANGE_BACK; if the b end needs to be manually proxied by the a end, regardless of whether the a end quality code is normal, multiply the value of end point a by the proxy coefficient and the quality code thereof, update the value and quality code of end point b in the real-time database, and the value and quality code of the a end remain unchanged, send back to the opposite telemetry processing process Pm, and the message type is MT_YC_EXCHANGE_BACK;
[0034] (43b-4) If there is no manual proxy, execute the opposite processing automatic proxy logic of end point b;
[0035] Finally, after the process Pm on the node where the local end point a is located receives the MT_YC_EXCHANGE_BACK message sent back by the process Pn on the node where the opposite end point b is located, the local proxy subsequent processing is performed:
[0036] (43c-1) According to whether there is a proxy mark in the sent back message and whether there is a proxy record of the telemetry aYc of end point a in the opposite end proxy information table, execute the local proxy subsequent processing logic of end point a;
[0037] Similarly, when the telemetry processing process Pn receives the telemetry of the front-end b, the processing flow is the same as that of the process Pm, except that the process Pn transmits and receives messages and processes message operations in reverse, the process Pn sends the MT_YC_EXCHANGE_GO message, and processes the MT_YC_EXCHANGE_BACK message; and the process Pm sends the MT_YC_EXCHANGE_BACK message and processes the MT_YC_EXCHANGE_GO message.
[0038] Further, the distributed processing architecture supports horizontal expansion, and in the case of insufficient system processing capacity, the processing capacity of the peer generation can be further improved by increasing the number of shards.
[0039] The cross-substation line mass telemetry peer generation distributed processing system provided by the application comprises:
[0040] A modeling module is configured to model the substation, the AC line, the AC line endpoint and the peer generation information, the main attributes of the substation modeling including the substation ID, the substation name and the dispatching organization to which the substation belongs, the main attributes of the AC line modeling including the line ID, the line name, the starting substation and the ending substation, the main attributes of the AC line endpoint modeling including the endpoint ID, the line to which the endpoint belongs, the substation to which the endpoint belongs, the active power, the reactive power and the current measurement, and the peer generation information modeling, that is, the peer generation information table is established to record the endpoint information being processed by the peer generation.
[0041] A model shard processing module is configured to perform shard processing on the modeled model.
[0042] An endpoint type identification module is configured to identify the type of the AC line endpoint, including the in-chip endpoint, the node-in-chip inter-chip endpoint and the cross-node chip inter-chip endpoint.
[0043] A distributed peer generation processing module is configured to adopt the distributed processing architecture for the peer generation processing according to different endpoint types.
[0044] Beneficial effects: compared with the prior art, the present application has the following remarkable advantages: 1, the rate of data processing is improved, distributed architecture, asynchronous processing mechanism and multi-thread technology are adopted, network interface access model and real-time data are avoided, the mass of end-to-end processing performance of large-scale regulation and control system is greatly improved, the telemetry processing throughput is improved, and the processing delay is reduced; 2, the qualified rate of data is improved, through the application of automatic end-to-end generation and manual end-to-end generation, the telemetry data of large-scale power grid regulation and control system is greatly completed and corrected, the qualified rate of data of the whole system is effectively improved, and reliable basic data guarantee is provided for AGC, AVC and other advanced applications; 3, by adopting a reasonable fragmentation method, the end point model of alternating current line is divided into three types of in-slice end point, node-in-slice inter-slice end point and cross-node inter-slice end point, and the cross-slice cross-node calculation of end-to-end generation is reduced as much as possible. BRIEF DESCRIPTION OF DRAWINGS
[0045] Figure 1 The substation liaison and end-to-end generation of the present application are shown in the figure;
[0046] Figure 2 The end-to-end generation processing diagram of the present application is shown in the figure;
[0047] Figure 3 The end-to-end generation equipment fragmentation and end point classification diagram of the present application is shown in the figure;
[0048] Figure 4 The alternating current line end point type identification flow chart of the present application is shown in the figure;
[0049] Figure 5 The distributed processing architecture diagram of the present application is shown in the figure;
[0050] Figure 6 The cross-process end-to-end generation asynchronous processing diagram of the present application is shown in the figure;
[0051] Figure 7 The cross-node inter-slice end point end-to-end generation processing flow chart of the present application is shown in the figure. DETAILED DESCRIPTION
[0052] The technical scheme of the present application will be further described below in combination with the drawings.
[0053] As shown in Figure 1 and Figure 2 , the embodiment of the present application provides a kind of mass telemetry end-to-end generation distributed processing method and system across substation line, and its specific steps are as follows:
[0054] Step 1: modeling of substation, alternating current line, alternating current line end point and end-to-end generation information.
[0055] Step 1.1: substation modeling, main attributes include substation ID, substation name, belonging dispatching agency.
[0056] Step 1.2: AC line modeling, main attributes include line ID, line name, start substation, end substation.
[0057] Step 1.3: AC line endpoint modeling, main attributes include endpoint ID, belonging line, belonging substation, active, reactive, current measurement, etc.
[0058] Step 1.4: modeling of opposite end information, establishing an opposite end information table to record the endpoint information in the system that is being processed by the opposite end, the main fields are as follows: 1.4.1: replaced endpoint telemetry ID, set as the primary key; 1.4.2: replacement endpoint telemetry ID; 1.4.3: substation ID of the replaced endpoint, used to determine the shard; 1.4.4: replacement type, divided into automatic opposite end replacement and manual opposite end replacement; 1.4.5: replacement coefficient, such as the replacement coefficient of active and reactive values is -1.
[0059] The opposite end information table is also set as a sharded table, and the shard number recorded in the table is consistent with the shard number of the replaced endpoint telemetry ID.
[0060] Step 2: Sharding processing of the modeled model.
[0061] Step 2.1: substation sharding, according to its belonging dispatching agency, because a dispatching agency in an area governs multiple substations, so the substations in an area all belong to the same shard, which maximally reduces the cross-shard connection between substations.
[0062] Step 2.2: AC line, a line belongs to two different substations, so it is not sharded.
[0063] Step 2.3: AC line endpoint, the shard number is consistent with the belonging substation.
[0064] Step 2.4: opposite end information table, consistent with the substation shard number of the replaced telemetry.
[0065] It should be noted that there are several shards distributed on each SCADA node, and all SCADA node shards form a complete system model.
[0066] The record in the fragmentation table is only downloaded to the SCADA node containing the fragmentation, and the non-fragmentation table is downloaded, that is, cached to all SCADA nodes; for example, assuming that two fragments 1 and 2 are deployed on a scada1 node, and two fragments 3 and 4 are deployed on a scada2 node, the AC line terminal table has 5 records, the AC line terminal 1 belongs to the 1st fragment, the AC line terminal 2 belongs to the 4th fragment, the AC line terminal 3 belongs to the 2nd fragment, the AC line terminal 4 belongs to the 3rd fragment, and the AC line terminal 5 belongs to the 3rd fragment, then the scada1 node finally has 2 records, that is, the line terminals 1 and 3, and the scada2 node has 3 records, that is, the line terminals 2, 4 and 5; therefore, the process on the scada1 can use the local real-time library interface to obtain the information of the line terminals 1 and 3 with high performance, and if the process on the scada1 wants to obtain the information of the line terminals 2, 4 and 5, then the real-time library network interface must be used, which greatly reduces the data access performance; the non-fragmented table is downloaded to all SCADA servers, and the high-performance local interface can be used, but the model involving these tables is not suitable for distributed processing.
[0067] The AC line terminal in the embodiment of the application is divided into in-chip terminal, intra-node inter-chip terminal and cross-node inter-chip terminal, three types, as shown in Figure 3
[0068] Step 3: The telemetry processing process on each node needs to identify the AC terminal type before processing the opposite terminal, as shown in Figure 4
[0069] Step 3.1: Read all the fragmentation information of the node, assuming that the node has deployed fragments 1 and 2.
[0070] Step 3.2: Read all the AC line terminal information of the system using the network interface, wherein the main information includes the terminal ID and the line ID to which the terminal belongs.
[0071] Table 1 is a fragmentation table, but the model changes frequently, and the entire data needs to be obtained using the network interface, only when the static model of the table changes.
[0072] Table 1 terminal model key information
[0073] End point ID Line ID Substation End point 1 1 Line 1 Substation 1 End point 1 2 Line 1 Substation 2 End point n 1 Line n Substation x End point n 1 Line n Substation y
[0074] Step 3.3: Construct a hash table as shown in Table 2 with the line ID as the key and the terminal pair as the value, and pair the terminal pairs to form the terminal mapping relationship.
[0075] Table 2 hash table structure
[0076] Line ID Terminal Pair Line 1 (End point 1 1, End point 1 2) Line 2 (End point 2 1, End point 2 2) Line n (End point n 1, End point n 2)
[0077] Step 3.4: Traverse the hash table, determine whether the two endpoints in each line endpoint pair belong to the same shard. Assuming that the endpoints 11 and 12 of line 1, call the shard obtaining interface respectively, and obtain the shard numbers as 1; the endpoints 21 and 22 of line 2, call the shard obtaining interface respectively, and obtain the shard numbers as 1 and 2; the endpoints n1 and n2 of line n, call the shard obtaining interface respectively, and obtain the shard numbers as 2 and 4.
[0078] Step 3.5: The endpoint pairs belonging to the same shard are identified as intra-shard endpoints, such as the endpoints 11 and 12 of line 1 above, which belong to shard 1; the endpoint pairs not belonging to the same shard are identified as inter-shard endpoints, such as the endpoints 21 and 22 of line 2 above, and the endpoints n1 and n2 of line n.
[0079] Step 3.6: For inter-shard endpoints, determine whether the shards to which the two endpoints belong are deployed on the same node. The endpoint pairs whose shards are deployed on the same node are identified as intra-node inter-shard endpoints, such as the endpoints 21 and 22 above; the endpoint pairs not deployed on the same node are identified as cross-node inter-shard endpoints, such as the endpoints n1 and n2 above.
[0080] Step 4: According to different endpoint types, perform distributed peer processing.
[0081] The distributed processing architecture in this embodiment is shown in Figure 5 Assuming that shard 1 and shard 2 are deployed on the SCADA-1 node, and shard n is deployed on the SCADA-n server, then the telemetry processing processes P1 and P2 are started on SCADA1, P1 subscribes to message channel 1, and P2 subscribes to message channel 2, and the telemetry processing process Pn is started on SCADA-n, and subscribes to message channel n.
[0082] The front-end system first locates which shard the telemetry ID belongs to according to the telemetry ID, and then sends it to the corresponding channel. If the process P1 of shard 1 wants to send a message to the process Pn of shard n, it can send the message to channel n.
[0083] Step 4.1: The intra-shard endpoint pair processing flow is as follows:
[0084] Step 4.1.1: The telemetry processing process Pn receives telemetry data of an endpoint: aYc, indicating telemetry of endpoint a.
[0085] Step 4.1.2: Query the opposite side information table, whether the opposite side record exists for the end point and the opposite end point, and whether the replacement type is manual opposite side replacement or automatic opposite side replacement. As shown in Table 3, "Line 1-aYc" is forcibly manually replaced by "Line 1-bYc", and it is noted that bYc represents the telemetry of the b end. "Line 2-aYc" has been automatically replaced by "Line 2-bYc", and for automatic replacement, whether it needs to continue to be replaced needs to be seen from whether the quality code of "Line 2-aYc" is abnormal and whether the quality code of "Line 2-bYc" is normal. Only when the quality code of "Line 2-aYc" is abnormal and the quality code of "Line 2-bYc" is normal, the automatic opposite side replacement can be continued, otherwise, the record needs to be deleted and the automatic opposite side replacement needs to be removed. Typical abnormal quality codes include "working condition exit", "telemetry does not change", and "collection problem".
[0086] Table 3: Structure of opposite side replacement information table
[0087] Substitute telemetry ID Substitute telemetry Substation Substitute type Substitute coefficient Line 1 -a Yc Line 1 -b Yc Substation 1 Manual pair 1 Line 2 -a Yc Line 2 -b Yc Substation 2 Automatic pair -1
[0088] Step 4.1.3: If it is manual opposite side replacement and the opposite side is replaced, read the telemetry value of the opposite end point, multiply the replacement coefficient to replace the telemetry value of the end point on this side and update the real-time database; if it is replaced on the opposite side, multiply the telemetry value by the replacement coefficient at the same time, and replace the telemetry value of the opposite side device into the database; it is noted that since the in-chip end point is processed here, the records of whether aYc and bYc are replaced can be found at the same time.
[0089] Step 4.1.4: According to whether there is an automatic replacement record in the opposite side replacement information table and whether the quality codes of aYc and bYc are normal, the logic processing is as shown in Table 4.
[0090] Table 4: In-chip end point opposite side replacement processing logic
[0091]
[0092]
[0093] Step 4.2: Intra-node inter-chip end point opposite side replacement processing:
[0094] The intra-node inter-chip end point opposite side replacement processing flow is consistent with the in-chip end point processing flow, but when accessing the opposite end point data, the slice of the real-time database access interface needs to be set to the slice number of the opposite side, and the data access is still the local access interface.
[0095] The opposite side replacement processing adopts an asynchronous processing mechanism, and the processing flow is as follows Figure 6As shown, the telemetry endpoint sends the information of endpoint a to process Pn after process Pm completes the preliminary processing of endpoint a, combines the information of endpoints a and b, performs comprehensive processing, and finally sends endpoint a back to Pm for processing.
[0096] The single-process peer-end synchronous processing method provided in the embodiment has a processing time of t1+t2+t3 for the peer-end consumption of a measurement point, wherein t1 and t3 are read and write operations on a local real-time library, and the time performance is about tens of microseconds; t2 is a call of a network interface to access remote data, and the time performance is about tens of milliseconds, t2 is about 1000 times of t1, and the ratio is larger in a poor network environment.
[0097] The time of peer-end asynchronous processing across shards is about tm1+tm2+tn, and the main time consumption of the three items is the consumption of local real-time library operations. Since the asynchronous mechanism is used, the message transmission delay can be ignored, and the message transmission only slightly affects the latency of telemetry and has little effect on the throughput. Therefore, the performance of the asynchronous processing mechanism is much better than that of the synchronous processing mechanism.
[0098] Step 4.3: Peer-end asynchronous processing across node shards
[0099] Suppose that shard x is distributed on node SCADA-m, and the telemetry processing process Pm on the shard subscribes to receive telemetry messages of front channel x. The flow is as shown in Figure 7 .
[0100] Step 4.3.1: The preprocessing steps of the local-side endpoint a are as follows:
[0101] Step 4.3.1.1: The telemetry processing process Pm of shard x receives endpoint aYc.
[0102] Step 4.3.1.2: The ID of the opposite-side endpoint bYc and the shard number y are obtained.
[0103] Step 4.3.1.3: The peer-end information table is queried to determine whether there is a record of the peer-end of endpoint a, and whether it is a manual peer-end or an automatic peer-end.
[0104] Step 4.3.1.4: aYc and the related information of endpoint a obtained by querying are sent to the telemetry processing process Pn of node SCADA-n, and the message type is MT_YC_EXCHANGE_GO.
[0105] In the embodiment, the telemetry processing process Pn receives the MT_YC_EXCHANGE_GO message of endpoint a sent by the telemetry processing process Pm, and starts the endpoint opposite-side processing flow.
[0106] Step 4.3.2: End-to-end processing on the opposite side, the specific operation steps are as follows:
[0107] Step 4.3.2.1: Analyze the MT_YC_EXCHANGE_GO message to obtain the telemetry value and quality code of endpoint A, and determine whether it needs to be automatically replaced by b, whether it can be replaced by the opposite side, and whether it needs to be manually replaced by b.
[0108] Step 4.3.2.2: Query the local real-time library to obtain the quality code of b's telemetry and its opposite replacement information table, and calculate whether b needs to be replaced by the opposite side, whether b can replace the opposite side, and whether b needs to be manually replaced by a.
[0109] Step 4.3.2.3: If a needs to be manually replaced by b, regardless of whether the quality code of b's telemetry is normal, replace a's telemetry value and quality code with b's telemetry value and quality code, and send back to the opposite SCADA node telemetry processing process Pm, with the message type being MT_YC_EXCHANGE_BACK; if b needs to be manually replaced by a, regardless of whether the quality code of a is normal, multiply the value of endpoint a by the replacement coefficient and its quality code, update the value and quality code of endpoint b in the real-time library, and keep the value and quality code of a unchanged, and send back to the opposite SCADA node telemetry processing process Pm, with the message type being MT_YC_EXCHANGE_BACK.
[0110] Step 4.3.2.4: If there is no manual opposite replacement, execute the automatic opposite replacement process as shown in Table 5.
[0111] Table 5 Automatic opposite replacement processing logic of the opposite side process B
[0112]
[0113]
[0114] Again, after the process Pm on the node SCADA-m receives the MT_YC_EXCHANGE_BACK message sent back by the process Pn on the node SCADA-n at B, the process Pm performs the subsequent processing of the opposite side replacement on the local side, and according to whether there is a "replacement" mark in the sent-back message and whether there is a "replacement" record of aYc in the opposite replacement information table, executes the logic as shown in Table 6.
[0115] Table 6 Subsequent processing logic of the opposite replacement on the local side A
[0116]
[0117]
[0118] Step 4.3.3: The specific operation steps of the local side to the end of the subsequent processing are as follows:
[0119] Step 4.3.3.1: Update the real-time library, write the telemetry value and quality code of the end point a sent back from the b end, and the telemetry value and quality code sent back may be data replaced by the end point b.
[0120] Step 4.3.3.2: If there is a "substitute" mark in the message and there is a "substitute" record in the table, end the processing; if there is a "substitute" mark in the message and there is no "substitute" record in the table, insert a record of a being automatically substituted by b in the opposite end of the information table, and send an alarm "Telemetry a is automatically replaced by telemetry b"; if there is no "substitute" mark in the message and there is a "substitute" record in the table, delete the record of a being substituted by b in the table, and send an alarm "Telemetry a cancels automatic opposite end substitution"; if there is no "substitute" mark in the message and there is no "substitute" record in the table, end the processing.
[0121] Further, for the above cross-node inter-slice end point opposite end substitution asynchronous processing, when the telemetry processing process Pn receives the telemetry of the front-end sent end point b, its processing flow is similar to that of Pm, except that its message sending and processing message operation are opposite, Pn sends MT_YC_EXCHANGE_GO message, and processes MT_YC_EXCHANGE_BACK message; and Pm sends MT_YC_EXCHANGE_BACK message, and processes MT_YC_EXCHANGE_GO message.
[0122] The cross-substation line massive telemetry opposite end substitution distributed processing system of the application comprises:
[0123] The modeling module is used for modeling processing of the substation, the AC line, the AC line end point and the opposite end substitution information, the main attributes of the substation modeling include the substation ID, the substation name and the belonging dispatching organization; the main attributes of the AC line modeling include the line ID, the line name, the starting substation and the ending substation; the main attributes of the AC line end point modeling include the end point ID, the belonging line, the belonging substation, the active power, the reactive power and the current measurement; the opposite end substitution information modeling, that is, establishing the opposite end substitution information table, records the end point information being processed in the system.
[0124] The model slice processing module is used for slice processing of the modeled model;
[0125] The end point type identification module is used for identifying the type of the AC line end point, including the in-slice end point, the node-in-slice inter-slice end point and the cross-node slice inter-slice end point.
[0126] The distributed opposite end substitution processing module is used for processing the opposite end substitution through the distributed architecture according to different end point types.
[0127] In summary, the application divides the AC line end point model into three types of in-chip end point, intra-node inter-chip end point and cross-node inter-chip end point by using a distributed peer processing method, reduces the cross-chip cross-node calculation of peer as much as possible, greatly improves the mass peer processing performance of large-scale control system, improves the telemetry processing throughput, and reduces the processing delay; through the application of automatic peer and manual peer, the telemetry data of the large power grid control system is completed and corrected in a large area, the data qualification rate of the whole system is effectively improved, and reliable basic data guarantee is provided for AGC, AVC and other advanced applications.
Claims
1. A method for distributed processing of cross-substation line massive telemetry peer, characterized in that, The method comprises the following steps: (1) modeling the transformer substation, AC line, AC line end point and opposite end information; (2) performing fragmentation processing on the modeled model; (3) identifying the end point type according to the fragmentation processing result; specifically comprising: (31) reading all the fragmentation information of the node; (32) reading all the AC line end point information through the network interface, and the information comprises the end point ID and the line ID to which the end point belongs; (33) constructing a hash table with the line ID as the key and the end point pair as the value, and pairing the end point pair to form an end point mapping relationship; (34) traversing the hash table to determine whether the two end points of each line end point pair belong to the same fragment, if the end point pair belongs to the same fragment, it is an intra-fragment end point, and if the end point pair does not belong to the same fragment, it is an inter-fragment end point; for the inter-fragment end point, it is determined whether the fragments to which the two end points belong are distributed in the same node, if the fragments are distributed in the same node, the end point pair is a node-intra-fragment end point, and if the fragments are not distributed in the same node, the end point pair is a cross-node inter-fragment end point; (4) performing distributed opposite end substitution processing according to different end point types.
2. The method of claim 1, wherein the method is a method for distributed processing of massive telemetry of a cross-substation line pair, and the method comprises: The fragmentation processing in step (2) comprises fragmenting the transformer substation according to the dispatching mechanism of the region to which the transformer substation belongs, and attributing the transformer substations in one region to the same fragment; each AC line belongs to two different transformer substations, and no fragmentation processing is performed on the AC line; The fragmentation number of the AC line end point is consistent with the fragmentation number of the transformer substation to which the AC line end point belongs; and the fragmentation number of the opposite end substitution information table is consistent with the fragmentation number of the transformer substation to which the substituted end point belongs.
3. The method of claim 1, wherein the method further comprises: The substitution types of the opposite end substitution processing comprise manual opposite end substitution and automatic opposite end substitution; the manual opposite end substitution does not judge the telemetry quality code, and forcibly substitutes the opposite side for the current side; the automatic opposite end substitution automatically determines whether to substitute according to the telemetry quality code of the end point measurement; the priority of the manual opposite end substitution is higher than that of the automatic opposite end substitution; the telemetry quality code of the substituted opposite side is stacked in the state of being substituted by the opposite side; when the opposite end substitution occurs for the first time or is cancelled, an alarm notification is sent.
4. The method of claim 3, wherein the method further comprises: The opposite end substitution processing in step (4) comprises intra-fragment end point opposite end substitution processing, node-intra-fragment end point opposite end substitution processing, and cross-node inter-fragment end point opposite end substitution asynchronous processing; through the asynchronous processing mechanism, the opposite end substitution processing all adopts a real-time library local interface; the opposite end substitution processing adopts a distributed processing architecture, and there are fragments distributed on each SCADA node; each fragment starts a telemetry processing process; and the telemetry processing process of each fragment only receives the telemetry of the channel in which the fragment is located, and can send inter-process interaction information to other channels.
5. The method of claim 4, wherein the method further comprises: The specific steps of the intra-fragment end point opposite end substitution processing are as follows: (41a) the telemetry processing process receives the telemetry data of an end point; (41b) through the opposite end substitution information table, it is determined whether the end point and the opposite end point exist opposite end substitution records and the substitution type and substitution coefficient of the opposite end substitution processing; (41c) if the substitution type of the end point is manual opposite end substitution and the opposite side is substituted, the telemetry value of the opposite end point is read and multiplied by the substitution coefficient to substitute the telemetry value of the current side for storage processing; if the opposite side is substituted, the telemetry value of the end point is multiplied by the substitution coefficient to substitute the telemetry value of the opposite end point for storage processing; (41d) According to whether there is an automatic substitution record in the opposite end substitution information table and whether the quality code of the opposite end point is normal, the in-slice end point opposite end substitution processing logic is processed.
6. The method of claim 4, wherein the method further comprises: The inter-slice end point opposite end substitution processing in the node is consistent with the in-slice processing step, but when accessing the opposite side end point data, the slice number of the opposite side end point needs to be set as the slice number of the real-time library access interface.
7. The method of claim 4, wherein the method further comprises: The specific steps of the inter-slice end point opposite end substitution asynchronous processing across nodes are as follows: The node where the local side end point a is located is provided with a slice x, and the corresponding telemetry processing process Pm receives the telemetry information of the front channel x. The node where the opposite side end point b is located is provided with a slice y, and the corresponding telemetry processing process Pn receives the telemetry information of the front channel y. First, the local side opposite end substitution preprocessing of the local side end point a is performed: (43a-1) The telemetry processing process Pm receives the telemetry value and quality code of the end point a sent by the front channel x; (43a-2) The ID and slice number y of the opposite side end point b are obtained, and the inter-process message interaction channel is determined according to the opposite side slice number; (43a-3) The opposite end substitution information table is queried to determine whether there is an opposite end substitution record and a substitution type for the end point a; (43a-4) The telemetry and related information of the end point a obtained by querying are sent to the telemetry processing process Pn on the node where the opposite side end point is located, and the message type is MT_YC_EXCHANGE_GO; Secondly, when the telemetry processing process Pn of the opposite side end point b receives the MT_YC_EXCHANGE_GO message, the opposite side processing is entered: (43b-1) The MT_YC_EXCHANGE_GO message is parsed to obtain the telemetry value and quality code of the end point a, and it is judged whether the end point a needs to be automatically substituted by the end point b, whether the end point a can be substituted by the opposite side, and whether the end point a needs to be manually substituted by the end point b; (43b-2) The local real-time library is queried to determine whether the end point b needs to be substituted by the opposite side according to the quality code of the telemetry of the end point b and the record of the opposite end substitution information of the end point b, whether the end point b can be substituted by the opposite side, and whether the end point b needs to be manually substituted by the end point a; (43b-3) If the end point a needs to be manually substituted by the end point b, the telemetry value and quality code of the end point a are replaced by the product of the telemetry value and the substitution coefficient of the end point b and the quality code of the end point b, and the opposite side process Pm is sent back, and the message type is MT_YC_EXCHANGE_BACK. If the end point b needs to be manually substituted by the end point a, the value and quality code of the end point b in the real-time library are updated by the product of the value and the substitution coefficient of the end point a and the quality code of the end point a, and the value and quality code of the end point a remain unchanged. The opposite side telemetry processing process Pm is sent back, and the message type is MT_YC_EXCHANGE_BACK; (43b-4) If there is no manual substitution, the opposite side process automatic substitution processing logic of the end point b is executed; Finally, when the process Pm of the node where the local side end point a is located receives the MT_YC_EXCHANGE_BACK message sent back by the process Pn of the node where the opposite side end point b is located, the subsequent processing of the local side opposite end substitution is performed: (43c-1) According to whether there is a substitution mark in the sent back message and whether there is a substitution record of the telemetry aYc of the end point a in the opposite end substitution information table, the subsequent processing logic of the local side opposite end substitution of the end point a is executed. Similarly, when the telemetry processing process Pn receives the telemetry of the front-end uploaded endpoint b, the processing flow is the same as that of the process Pm, except that the process Pn transmits and receives messages and the processing message operation is opposite.
8. The method of claim 4, wherein the method further comprises: The distributed processing architecture can support horizontal expansion by increasing the number of shards.
9. A cross-substation line massive telemetry peer-end distributed processing system, characterized in that, Comprise: The modeling module is used for modeling the substation, AC line, AC line endpoint and opposite end information; The model shard processing module is used for sharding the modeled model; The endpoint type identification module is used for identifying the type of AC line endpoint, including intra-shard endpoint, intra-node inter-shard endpoint and cross-node inter-shard endpoint; Specifically comprising: (31) read all shard information of the node; (32) read all AC line endpoint information with network interface, including endpoint ID and line ID to which the endpoint belongs; (33) construct a hash table with line ID as key and endpoint pair as value, and pair the endpoint pairs to form an endpoint mapping relationship; (34) traverse the hash table, judge whether the two endpoints of each line endpoint pair belong to the same shard, if the endpoint pairs belong to the same shard, they are intra-shard endpoints, if they do not belong to the same shard, they are inter-shard endpoints; for inter-shard endpoints, judge whether the shards to which the two endpoints belong are distributed in the same node, if the shards are distributed in the same node, the endpoint pair is intra-node inter-shard endpoint, if not, it is cross-node inter-shard endpoint; The distributed opposite end processing module is used for processing the opposite end according to different endpoint types, and adopting a distributed processing architecture.