A power dispatch accident backtracking method based on dispatch telephone identification

Abstract

The application discloses a power dispatching accident backtracking method based on dispatching telephone identification, which comprises the following steps: collecting power dispatching accident information and dispatcher voice information in a certain area, and establishing a first DIKW model based on the power dispatching accident and a second DIKW model based on the dispatcher voice information; performing content and logic matching analysis based on the first DIKW model and the second DIKW model to obtain specific dispatcher voice information corresponding to a specific power dispatching accident; performing semantic recognition on the specific dispatcher voice information, and displaying an event chain of the specific power dispatching accident based on the semantic recognition result; and performing content and logic matching analysis on the first DIKW model and the second DIKW model to obtain specific dispatcher voice information corresponding to a specific power dispatching accident, so that an analyst can quickly establish a topological relationship between dispatcher recording and the power dispatching accident, and the analyst can conveniently and quickly analyze the power dispatching accident.

Description

Technical Field

[0001] This invention belongs to the field of power dispatching technology, and in particular relates to a power dispatching accident backtracking method based on dispatching telephone identification. Background Technology

[0002] Currently, during the telephone dispatching process in the power system, it is necessary to record the dispatching process. However, the existing dispatching recording system can only complete the voice recording function and store the voice in the existing recording system. The handling of power dispatching accidents and the retrospective of accidents rely solely on manual listening to historical audio information. It is impossible to extract and centralize contact information for individuals and power dispatching accidents, nor can it complete the statistical analysis of information. This important data cannot currently play its role in promoting the continuous improvement of the standardization and safety of production operations. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a power dispatching accident backtracking method based on dispatching telephone identification, so as to help analysts quickly establish the topological relationship between dispatcher recordings and power dispatching accidents, and facilitate analysts to quickly analyze power dispatching accidents.

[0004] Technical solution of the present invention:

[0005] A method for retrospectively tracing power dispatching incidents based on dispatching telephone identification, the method comprising:

[0006] S1. Collect power dispatching accident information and dispatcher voice information in a certain area, and establish a first DIKW model based on power dispatching accidents and a second DIKW model based on dispatcher voice information.

[0007] S2. Based on the first DIKW model and the second DIKW model, perform content and logic matching analysis to obtain specific dispatcher voice information corresponding to specific power dispatching accidents;

[0008] S3. Perform semantic recognition on the voice information of specific dispatchers, and display the event chain of specific power dispatching accidents based on the semantic recognition results.

[0009] Methods for establishing the first DIKW model based on power dispatching accidents include:

[0010] Based on the known power dispatching accident information, each type of power dispatching accident information is taken as a dispatching information node, and each dispatching information node constitutes a dispatching information resource. A model is then created based on the dispatching information resource to obtain the dispatching information model.

[0011] Each power equipment data is used as a scheduling data node, and each scheduling data node constitutes a scheduling data resource. A scheduling data model is obtained by modeling based on the scheduling data resource. The power equipment data includes basic data and operational data. The basic data includes at least the power equipment model, number and name.

[0012] Various recognition algorithms are used as scheduling intelligent nodes, and each scheduling intelligent node constitutes a scheduling intelligent resource. Based on the scheduling intelligent resource, a scheduling intelligent model is obtained.

[0013] By matching scheduling data nodes and scheduling information nodes with each other through the scheduling intelligence model, the logical relationship between scheduling data nodes and scheduling information nodes is constructed. Each logical relationship constitutes a scheduling knowledge node, and each scheduling knowledge node constitutes a scheduling knowledge resource. Based on the scheduling knowledge resource, a scheduling knowledge model is obtained.

[0014] The first DIKW model is constructed based on the scheduling information model, scheduling data model, scheduling intelligence model, and scheduling knowledge model.

[0015] The scheduling intelligent model matches scheduling data nodes and scheduling information nodes. Specifically, each intelligent node in the scheduling intelligent model analyzes and identifies scheduling information nodes, determines faulty power equipment, establishes a logical relationship between faulty power equipment and scheduling data nodes, and constructs scheduling knowledge nodes based on the above logical relationship.

[0016] Methods for establishing a second DIKW model based on dispatcher voice information include:

[0017] Each dispatcher's voice feature data is used as a dispatcher data node, and each voice data node constitutes a voice data resource. The dispatcher data resource is modeled to obtain a dispatcher data model.

[0018] Each dispatcher's personal information and the power equipment information they are responsible for are used as dispatcher information nodes. The information of each dispatcher constitutes the dispatcher information resource. The dispatcher information resource is modeled to obtain the dispatcher information model.

[0019] Each speech recognition algorithm is used as a dispatcher's intelligent node, and each speech recognition algorithm constitutes a dispatcher's intelligent resource. The dispatcher's intelligent resource is modeled to obtain the dispatcher's intelligent model.

[0020] The logical relationships between dispatcher information nodes and dispatcher data nodes are constructed. Each logical relationship constitutes a dispatcher knowledge node, and all dispatcher knowledge nodes constitute dispatcher knowledge resources. Based on the dispatcher knowledge resources, a model is obtained to acquire the dispatcher knowledge model.

[0021] The second DIKW model is constructed based on the dispatcher information model, dispatcher data model, dispatcher intelligence model, and dispatcher knowledge model.

[0022] The logical relationship between the dispatcher information model and the dispatcher data model is constructed as follows: the dispatcher intelligent model is used to identify and analyze the voice feature data in the dispatcher data node, determine the dispatcher identity corresponding to the voice feature, and compare and match it with the dispatcher information node. Based on the matching result, the logical relationship between the dispatcher information node and the dispatcher data node is constructed.

[0023] Speech data resources include speech time data, speech pitch data resources, speech timbre data resources, speech volume data resources, and speech voiceprint feature data resources.

[0024] Based on the first DIKW model and the second DIKW model, content and logic matching analysis is performed to obtain specific dispatcher voice information corresponding to a specific power dispatching accident. Specifically, this includes: putting the node content in the dispatching information model, dispatching data model, and dispatching knowledge model of the first DIKW model into the dispatcher information model, dispatcher data model, and dispatcher knowledge model of the second DIKW model for matching, and obtaining specific dispatcher personal information and specific dispatcher voice information corresponding to the faulty power equipment.

[0025] The method for semantic recognition and display of specific dispatcher voice information is as follows: Natural language processing algorithms are used to parse the dispatcher's voice information to obtain relevant semantic results, and a logical association chain is constructed between the dispatcher, the dispatcher's semantic results, the faulty power equipment, and the dispatching accident, which is then displayed in a graphical form.

[0026] Beneficial effects of this invention:

[0027] This invention provides a method for backtracking power dispatching accidents based on dispatching telephone identification. By establishing a first DIKW model based on power dispatching accidents and a second DIKW model based on dispatcher voice information, and performing content and logic matching analysis on the first and second DIKW models, specific dispatcher voice information corresponding to specific power dispatching accidents can be obtained. This helps analysts quickly establish the topological relationship between dispatcher recordings and power dispatching accidents, facilitating rapid analysis of power dispatching accidents. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0029] See Figure 1 A method for retrospectively tracing power dispatching incidents based on dispatching telephone identification includes the following steps:

[0030] S1. Collect power dispatching accident information and dispatcher voice information in a certain area, and establish a first DIKW model based on power dispatching accidents and a second DIKW model based on dispatcher voice information.

[0031] In step S1, a first DIKW model based on power dispatching accidents is established. DIKW stands for Data, Information, Knowledge, and Wisdom. Each of these elements is composed of corresponding type resources. For example, data type resources are discrete objects collected without binding intent through direct observation of existing semantics. Data resources are categorized into numerical data, probabilistic data, logical data, and range data based on their attributes and structures. Numerical data consists of specific numerical values ​​obtained through observation, measurement, or simple calculation, including Arabic numerals, time, date, or numerical data associated with character data. Probabilistic data represents an inference or prediction of the development trend of other things based on personal perception; it is an uncertain description. Logical data uses (yes / no) to represent data resources, allowing for filtering and classification of things and using numerical codes to represent each category. Range data describes a single value of a certain data type, expressing multiple data points of a single data object. The higher the overlap of range data, the more similar they are. Range data is further divided into bounded ranges and unbounded ranges based on whether they have boundaries.

[0032] The information resources mentioned are formed by data directly or indirectly related to one or more intentions and are time-sensitive. Information can be partially ordered, purpose-oriented, or change-oriented. Partially ordered information compares data objects of the same or different types based on the same standard and under the premise of identical attributes. Purpose-oriented information analyzes the functional attributes of data to achieve a final purpose based on usage intentions. Change-oriented information describes the process of modifying or processing selected objects, resulting in changes.

[0033] The knowledge resources are structured and formalized categorical reasoning or empirical rules derived from probability and statistics of data and / or information resources. The wisdom resources describe value orientations, i.e., what things have higher value. The intent resources reflect what the developers or system users want to do and what goals they want to achieve.

[0034] Specifically, in this embodiment, based on known power dispatching accident information, each type of power dispatching accident information is used as a dispatching information node, and multiple dispatching information nodes constitute a dispatching information resource. Based on the dispatching information resource, a model is created to obtain a dispatching information model, which contains multiple types of power dispatching accidents.

[0035] Each power equipment data is used as a scheduling data node, and multiple scheduling data nodes constitute a scheduling data resource. A scheduling data model is obtained by modeling based on the scheduling data resource. The power equipment data includes basic data and operational data. The basic data includes at least the power equipment model, number, and name.

[0036] Each recognition algorithm is used as a scheduling intelligent node, and multiple scheduling intelligent nodes constitute a scheduling intelligent resource. A scheduling intelligent model is obtained by modeling based on the scheduling intelligent resource.

[0037] The scheduling intelligence model matches scheduling data nodes and scheduling information nodes. Specifically, each intelligent node in the model analyzes and identifies scheduling information nodes to determine faulty power equipment and establishes a logical relationship between the faulty power equipment and the scheduling data nodes. Scheduling knowledge nodes are then constructed based on this logical relationship. For example, analyzing scheduling information node A through the scheduling intelligence model reveals faulty power equipment B. Faulty power equipment B is then compared and matched with the power equipment in scheduling data node C. If a match is found, a logical relationship is generated: the name of faulty power equipment B is XX, its model number is XXX, and its serial number is XXX. This process continues, with each formed logical relationship constituting a scheduling knowledge node. Multiple scheduling knowledge nodes constitute a scheduling knowledge resource. Modeling is then performed based on this scheduling knowledge resource to obtain a scheduling knowledge model.

[0038] Finally, based on the aforementioned scheduling information model, scheduling data model, scheduling intelligence model, and scheduling knowledge model, the first DIKW model is constructed.

[0039] Furthermore, in one embodiment of the present invention, different intelligent nodes, i.e. different identification algorithms, are used to analyze and identify the scheduling information nodes. If each analysis identifies a unique faulty power device, then the faulty power device is the final analysis and identification result of the scheduling intelligent model on the scheduling information node.

[0040] If multiple faulty power devices are identified, the faulty power device that is identified the most times is selected as the final analysis and identification result of the scheduling intelligent model for the scheduling information node.

[0041] For the second DIKW model, its data, information, knowledge, and wisdom are all composed of corresponding types of resources. For example, the voice feature data of each dispatcher is used as a dispatcher data node, and multiple voice data nodes constitute voice data resources. The dispatcher data resources are modeled to obtain the dispatcher data model. The voice data resources include: voice time data, voice pitch data resources, voice timbre data resources, voice volume data resources, and voiceprint feature data resources.

[0042] It should be noted that the dispatcher's voice data is obtained from the dispatcher's telephone recordings, and the voice data is preprocessed to obtain voice feature data.

[0043] Each dispatcher's personal information and the power equipment information they are responsible for are used as dispatcher information nodes. Multiple dispatchers with personnel information constitute dispatcher information resources. The dispatcher information resources are modeled to obtain the dispatcher information model.

[0044] Each speech recognition algorithm is used as a dispatcher smart node, and multiple speech recognition algorithms constitute dispatcher smart resources. The dispatcher smart resources are modeled to obtain the dispatcher smart model.

[0045] The logical relationship between dispatcher information nodes and dispatcher data nodes is constructed based on a dispatcher intelligence model. Specifically, the dispatcher intelligence model identifies and analyzes voice feature data in dispatcher data nodes to determine the dispatcher identity corresponding to the voice features. This identity is then compared and matched with dispatcher information nodes. Based on the matching results, the logical relationship between the dispatcher information nodes and dispatcher data nodes is constructed. For example, the dispatcher intelligence model identifies and analyzes voice feature data A in a dispatcher data node to determine the corresponding dispatcher identity B. Dispatcher identity B is then matched with dispatcher information node C. If the match is successful, the logical relationship is generated: Dispatcher identity B's voice feature data A is A, and the power equipment they are responsible for is XXX. This logical relationship constitutes dispatcher knowledge nodes. Multiple dispatcher knowledge nodes constitute dispatcher knowledge resources. A model is then built based on these dispatcher knowledge resources to obtain the dispatcher knowledge model.

[0046] The second DIKW model is constructed based on the dispatcher information model, dispatcher data model, dispatcher intelligence model, and dispatcher knowledge model.

[0047] Furthermore, in one embodiment of the present invention, different dispatcher smart nodes, i.e., different voice algorithms, are used to analyze and identify dispatcher data nodes. If each analysis identifies a unique dispatcher identity, then the dispatcher identity is the final analysis and identification result of the dispatching smart model on the dispatcher data node.

[0048] If multiple dispatcher identities are identified, the dispatcher identity that is identified most frequently is selected as the final analysis and identification result of the dispatcher intelligent model for the dispatcher data node.

[0049] S2. Based on the first DIKW model and the second DIKW model, perform content and logic matching analysis to obtain specific dispatcher voice information corresponding to a specific power dispatching accident. Specifically, this includes: matching the node content from the dispatching information model, dispatching data model, and dispatching knowledge model of the first DIKW model with the dispatcher information model, dispatcher data model, and dispatcher knowledge model of the second DIKW model to obtain specific dispatcher personal information and voice information corresponding to the faulty power equipment. For example, based on the faulty power equipment B's name being XX, model being XXX, and number being XXX, determine the specific dispatcher's identity and corresponding voice feature data corresponding to power equipment B named XX.

[0050] S3. Perform semantic recognition on the specific dispatcher's voice information, and display the event chain of the specific power dispatching accident based on the semantic recognition results.

[0051] In step S3, a natural language processing algorithm is used to parse the dispatcher's voice information to obtain relevant semantic results, and a logical association chain is constructed between the dispatcher, the dispatcher's semantic results, the faulty power equipment, and the dispatching accident, which is then displayed in a graphical form.

Claims

1. A power dispatching accident backtracking method based on dispatching telephone identification, characterized in that: The method includes: S1. Collect power dispatching accident information and dispatcher voice information in a certain area, and establish a first DIKW model based on power dispatching accidents and a second DIKW model based on dispatcher voice information. Methods for establishing the first DIKW model based on power dispatching accidents include: Based on the known power dispatching accident information, each type of power dispatching accident information is taken as a dispatching information node, and each dispatching information node constitutes a dispatching information resource. A model is then created based on the dispatching information resource to obtain the dispatching information model. Each power equipment data is used as a scheduling data node, and each scheduling data node constitutes a scheduling data resource. A scheduling data model is obtained by modeling based on the scheduling data resource. The power equipment data includes basic data and operational data. The basic data includes at least the power equipment model, number and name. Various recognition algorithms are used as scheduling intelligent nodes, and each scheduling intelligent node constitutes a scheduling intelligent resource. Based on the scheduling intelligent resource, a scheduling intelligent model is obtained. By matching scheduling data nodes and scheduling information nodes with each other through the scheduling intelligence model, the logical relationship between scheduling data nodes and scheduling information nodes is constructed. Each logical relationship constitutes a scheduling knowledge node, and each scheduling knowledge node constitutes a scheduling knowledge resource. Based on the scheduling knowledge resource, a scheduling knowledge model is obtained. The first DIKW model is constructed based on the scheduling information model, scheduling data model, scheduling intelligence model, and scheduling knowledge model. Methods for establishing a second DIKW model based on dispatcher voice information include: Each dispatcher's voice feature data is used as a dispatcher data node, and each voice data node constitutes a voice data resource. The dispatcher data resource is modeled to obtain a dispatcher data model. Each dispatcher's personal information and the power equipment information they are responsible for are used as dispatcher information nodes. The information of each dispatcher constitutes the dispatcher information resource. The dispatcher information resource is modeled to obtain the dispatcher information model. Each speech recognition algorithm is used as a dispatcher's intelligent node, and each speech recognition algorithm constitutes a dispatcher's intelligent resource. The dispatcher's intelligent resource is modeled to obtain the dispatcher's intelligent model. The logical relationships between dispatcher information nodes and dispatcher data nodes are constructed. Each logical relationship constitutes a dispatcher knowledge node, and all dispatcher knowledge nodes constitute dispatcher knowledge resources. Based on the dispatcher knowledge resources, a model is obtained to acquire the dispatcher knowledge model. The second DIKW model is constructed based on the dispatcher information model, dispatcher data model, dispatcher intelligence model, and dispatcher knowledge model. S2. Based on the first DIKW model and the second DIKW model, perform content and logic matching analysis to obtain specific dispatcher voice information corresponding to specific power dispatching accidents; Based on the first DIKW model and the second DIKW model, content and logic matching analysis is performed to obtain specific dispatcher voice information corresponding to a specific power dispatching accident. Specifically, this includes: putting the node content in the dispatching information model, dispatching data model and dispatching knowledge model of the first DIKW into the dispatcher information model, dispatcher data model and dispatcher knowledge model of the second DIKW model for matching respectively, to obtain specific dispatcher personal information and specific dispatcher voice information corresponding to the faulty power equipment. S3. Perform semantic recognition on the voice information of specific dispatchers, and display the event chain of specific power dispatching accidents based on the semantic recognition results.

2. The power dispatching accident backtracking method based on dispatching telephone identification according to claim 1, characterized in that: The scheduling intelligent model matches scheduling data nodes and scheduling information nodes. Specifically, each intelligent node in the scheduling intelligent model analyzes and identifies scheduling information nodes, determines faulty power equipment, establishes a logical relationship between faulty power equipment and scheduling data nodes, and constructs scheduling knowledge nodes based on the above logical relationship.

3. The power dispatching accident backtracking method based on dispatching telephone identification according to claim 1, characterized in that: The logical relationship between the dispatcher information model and the dispatcher data model is constructed as follows: the dispatcher intelligent model is used to identify and analyze the voice feature data in the dispatcher data node, determine the dispatcher identity corresponding to the voice feature, and compare and match it with the dispatcher information node. Based on the matching result, the logical relationship between the dispatcher information node and the dispatcher data node is constructed.

4. The power dispatching accident backtracking method based on dispatching telephone identification according to claim 1, characterized in that: Speech data resources include speech time data, speech pitch data resources, speech timbre data resources, speech volume data resources, and speech voiceprint feature data resources.

5. The power dispatching accident backtracking method based on dispatching telephone identification according to claim 1, characterized in that: The method for semantic recognition and display of specific dispatcher voice information is as follows: Natural language processing algorithms are used to parse the dispatcher's voice information to obtain relevant semantic results, and a logical association chain is constructed between the dispatcher, the dispatcher's semantic results, the faulty power equipment, and the dispatching accident, which is then displayed in a graphical form.

Images