Electric power centralized control intelligent alarm system and method based on multi-mode communication fusion

Abstract

The invention discloses an electric power centralized control intelligent alarm system and method based on multi-mode communication fusion, the system is composed of a system hardware platform deployed in four areas of electric power and software on the hardware platform, and the hardware platform comprises an application server, a database server, a communication hardware unit and a network security device. The communication hardware unit comprises a short message modem module, an automatic dialing module and a text-to-voice module; the software is deployed on a hardware platform and comprises a strategy configuration and management module used for providing a visual interface and configuring a multi-mode communication strategy. The system dynamically selects and fuses various communication modes such as short messages, voice communication and voice broadcast according to alarm event characteristics, so that related notifications are more targeted, information noise is reduced, and accuracy and individuation of alarm notifications are realized; and meanwhile, the whole communication process meets the requirements of high safety and high reliability of the power system, the cryptographic algorithm encryption is supported, the data is not out of the local, and the communication state can be monitored and traced in the whole process.

Description

Technical Field

[0001] This invention relates to the field of power grid equipment fault monitoring and alarm technology, and in particular to a power centralized control intelligent alarm system and method based on multimodal communication fusion. Background Technology

[0002] Currently, the new generation of centralized control systems is the core technical support platform for power grid equipment monitoring and fault handling. In the network security partitioning architecture of the power system, the centralized control system located in the fourth zone (management information zone) is mainly responsible for non-real-time monitoring, data analysis, and information dissemination. Currently, this system generally uses a single communication mode for interaction with maintenance personnel, especially in the delivery of emergency alarm information, primarily relying on manual text messaging via WeChat and SMS. When the monitoring system detects equipment defects, faults, or abnormal events, the typical process is as follows: the system generates an alarm work order, which is then compiled and summarized by monitoring and maintenance personnel. A text message containing fields such as equipment name, alarm content, and time is then manually sent to the designated personnel.

[0003] The existing technology can be summarized as follows: After detecting an alarm event, the centralized control system application server calls the communication gateway's application programming interface (API), submitting the alarm text and a list of receiving numbers as parameters; the communication gateway then sends the information out via a mobile communication network in the form of point-to-point SMS and voice messages. The entire process is a one-way, static, and text-based information transmission. Current monitoring systems have the following problems:

[0004] (1) Inefficient and unreliable information transmission: Manual notification by monitoring personnel is a non-real-time, best-effort service, which is prone to delays and false alarms, especially during large-scale accidents or anomalies. In case of emergency faults (such as power outages or fires), it cannot be guaranteed that information will reach key responsible persons in the first instance, thus delaying the opportunity for emergency response.

[0005] (2) Lack of interactive capabilities: SMS is a one-way notification. Monitoring personnel can only passively receive information through system accidents and then send it back. They cannot quickly confirm, provide feedback or obtain further handling guidance through the same channel, making it difficult to form closed-loop management.

[0006] (3) Poor information reception experience and easy to overlook: Among numerous work text messages and information, plain text notifications are not intuitive and key information is easily buried. The lack of communication methods with stronger reminder effects, such as voice and telephone, may cause recipients to ignore important alarms in noisy environments or during non-working hours.

[0007] (4) Low level of intelligence and rigid configuration: Existing systems usually adopt a fixed "one-to-many" group sending mode, which cannot dynamically match the optimal communication method based on information such as alarm event type, level, and recipient role (e.g., sending text messages for general warnings and automatically making phone calls for major accidents). The configuration and management of personnel, groups, and notification strategies are complex and lack flexibility.

[0008] (5) Lack of compliance guarantee: Simple SMS sending is difficult to meet the power industry's requirements for traceability and strong confirmation of key operation instructions. Summary of the Invention

[0009] To overcome the above problems, the purpose of this invention is to provide a power centralized control intelligent alarm system and method based on multimodal communication fusion. This system dynamically selects and integrates multiple communication modes such as SMS, voice calls, and voice broadcasts according to the characteristics of alarm events, making relevant notifications more targeted, reducing information noise, increasing the attention of recipients, and achieving accurate and personalized alarm notifications. At the same time, the entire communication process meets the high security and high reliability requirements of the power system, supports national cryptographic algorithms for encryption, ensures that data does not leave the local machine, and allows for full monitoring and traceability of the communication status.

[0010] The technical solution adopted in this invention is:

[0011] The power centralized control intelligent alarm system based on multimodal communication fusion consists of a system hardware platform and software deployed in four power zones. The hardware platform includes an application server, a database server, communication hardware units, and network security equipment. The communication hardware units include an SMS modem module, an automatic dialing module, and a text-to-speech (TTS) module.

[0012] The software is deployed on a hardware platform and includes a policy configuration and management module, which provides a visual interface for configuring multimodal communication policies.

[0013] The event receiving and intelligent decision-making module is used to receive alarm events from the power centralized control platform and intelligently decide the target recipient, communication method, and notification content based on the strategy and event content.

[0014] The multimodal communication execution module is used to call the SMS modem, automatic dialing and text-to-speech services in parallel or serially according to the decision results, and to perform SMS sending, voice calling and voice synthesis broadcasting tasks.

[0015] The closed-loop feedback and record traceability module is used to record end-to-end notification logs and supports query auditing.

[0016] As a further description of the present invention, the SMS modem module adopts an industrial-grade 4G / 5G multi-band device, the device integrates an AT command interface, and supports national cryptographic SM4 encrypted transmission.

[0017] The automatic dialing module is based on the SIP protocol to build a switching platform, integrates a voice gateway algorithm, triggers outbound calls in batches and monitors the call status in real time, and automatically matches the contact level according to the fault level.

[0018] The text-to-speech module uses a lightweight neural network model and achieves localized deployment through domain fine-tuning and model compression technology. It accurately broadcasts power industry terms, synthesizes alarm text into an audio stream, and plays it through a voice gateway after the call is connected, with a synthesis delay of ≤1 second.

[0019] As a further description of the present invention, the strategy configuration and management module is a low-code visual strategy orchestration platform, which provides a strategy canvas and supports users to build communication processes by dragging and dropping predefined strategy nodes. The strategy nodes include receiving events, condition judgment, sending SMS, initiating voice calls, and executing TTS broadcasts. This module compiles the drawn flowcharts into backend executable rule chains for storage and execution.

[0020] As a further description of the present invention, the event receiving and intelligent decision-making module includes a secure data synchronization subsystem and an intelligent decision-making engine.

[0021] The security data synchronization subsystem, through the read-only data mirroring service and forward-type physical unidirectional optical gate deployed in Power Zone 1, uses change data capture technology based on database logs to incrementally synchronize alarm event data from Zone 1 to the database of Zone 4.

[0022] The intelligent decision engine integrates a domain-specific large language model fine-tuned from power industry corpus. It is used to perform semantic understanding and structured information extraction on synchronous alarm events, generate summary text containing event summary, current status, preliminary judgment and suggested measures, and perform policy matching based on the semantics of the summary text and event tags to determine the final notification policy.

[0023] As a further description of the present invention, the multimodal communication execution module is managed by a unified communication scheduling middleware, which monitors the SMS sending status and voice call status in real time and supports a fault tolerance mechanism: when the preferred communication method fails to reach the target within a set time, the backup communication method or the upgrade notification object is automatically triggered.

[0024] The intelligent alarm method for power centralized control based on multimodal communication fusion includes the following steps:

[0025] S1: Configure alarm notification policies in a visual manner through the policy configuration and management module.

[0026] S2: Through the event receiving and intelligent decision-making module, alarm events are received from the centralized control platform, and the target recipient, communication method combination, and notification content are intelligently determined.

[0027] S3: Through the multimodal communication execution module, based on the decision result of S2, SMS sending, automatic voice calling and voice synthesis broadcasting are executed in parallel or serially.

[0028] S4: Record and store end-to-end notification logs through the closed-loop feedback and record traceability module.

[0029] As a further description of the present invention, the specific process of intelligent decision-making in S2 is as follows:

[0030] S21: Through the security data synchronization subsystem, alarm event data generated in Power Zone 1 is synchronized to Power Zone 4 in real time and incrementally.

[0031] S22: Input the synchronized event data into the domain-specific large language model that has been fine-tuned for the power sector, and the model generates structured, conversational summary text.

[0032] S23: Combining the semantics of the summary text with the original level and type label of the event, perform matching or vector similarity retrieval in the strategy library to determine the final notification strategy.

[0033] As a further description of the present invention, the specific process of performing speech synthesis and broadcasting in S3 is as follows:

[0034] S31: Extract key text from decision results or event information.

[0035] S32: Input to a locally deployed lightweight TTS model to generate an audio stream, which is then automatically played via a voice gateway after a call is connected.

[0036] S33: The communication scheduling middleware monitors the call status in real time.

[0037] S34: If the call is successful, the call is completed.

[0038] If the call fails or goes unanswered, a backup notification process is triggered.

[0039] As a further description of the present invention, the alarm notification strategy configured in S1 supports combination configuration by personnel group, alarm type, and receiving time period, and supports setting different levels of recipients and differentiated communication method sequences for the same event type.

[0040] As a further description of the present invention, the method further includes:

[0041] Record the event source, decision-making strategy, communication execution status, recipient, and timestamp of each alarm notification task as a full-link log.

[0042] It provides a query interface that supports querying and tracing historical notification records by at least one of the following criteria: recipient, sending type, and time range.

[0043] The beneficial effects of this invention are:

[0044] This invention is based on a power centralized control intelligent alarm system that integrates multimodal communication. The system uses multimodal communication methods such as "telephone + SMS" to ensure that information is perceived as soon as possible in case of emergency faults, taking advantage of the strong reminder characteristics of telephone. Especially when staff are in noisy environments, telephone reminders can be responded to in a timely manner, solving the problem of delayed response caused by SMS delays or omissions, and significantly improving the speed and reliability of emergency response.

[0045] This invention relates to a power centralized control intelligent alarm system based on multimodal communication fusion. The software module uses a flexible subscription strategy based on event type, level, and personnel role to change the "one-size-fits-all" mass notification mode, making notifications more targeted, reducing information noise, increasing the attention of recipients, and achieving accurate and personalized alarm notifications.

[0046] This invention relates to a power centralized control intelligent alarm system based on multimodal communication fusion. The system converts complex text procedures into voice broadcasts through a text-to-speech module, enabling on-site maintenance personnel to quickly understand the nature of the fault and key operating steps by listening to the audio in emergency situations without having to read lengthy texts. This reduces the risk of misoperation and enhances the effectiveness of information understanding and operational guidance.

[0047] This invention is a power centralized control intelligent alarm system based on multimodal communication fusion. The software module provides a visual configuration interface, which makes the adjustment of personnel grouping and communication strategies convenient and efficient, and adapts to the dynamic changes in the power grid organizational structure and operation and maintenance mode. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the overall architecture of the power centralized control intelligent alarm system based on multimodal communication fusion proposed in this invention.

[0049] Figure 2 This is a flowchart of the intelligent alarm method for power centralized control based on multimodal communication fusion proposed in this invention. Detailed Implementation

[0050] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0051] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0052] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0053] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0054] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0055] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0056] like Figures 1-2 As shown, it illustrates a specific embodiment of the present invention:

[0057] Example 1:

[0058] The power centralized control intelligent alarm system based on multimodal communication fusion consists of a system hardware platform and software deployed in four power zones. The hardware platform includes an application server, a database server, communication hardware units, and network security equipment. The communication hardware units include an SMS modem module, an automatic dialing module, and a text-to-speech module.

[0059] The software is deployed on a hardware platform and includes:

[0060] The strategy configuration and management module provides a visual interface for configuring multimodal communication strategies.

[0061] The event receiving and intelligent decision-making module is used to receive alarm events from the power centralized control platform and intelligently decide the target recipient, communication method, and notification content based on the strategy and event content.

[0062] The multimodal communication execution module is used to call the SMS modem, automatic dialing and text-to-speech services in parallel or serially according to the decision results, and to perform SMS sending, voice calling and voice synthesis broadcasting tasks.

[0063] The closed-loop feedback and record traceability module is used to record end-to-end notification logs and supports query auditing.

[0064] Specifically, the SMS modem module adopts an industrial-grade 4G / 5G multi-band device, which integrates an AT command interface and supports SM4 encryption transmission.

[0065] The automatic dialing module is based on the SIP protocol to build a switching platform, integrates a voice gateway algorithm, triggers outbound calls in batches and monitors the call status in real time, and automatically matches the contact level according to the fault level.

[0066] The text-to-speech module uses a lightweight neural network model and achieves localized deployment through domain fine-tuning and model compression technology. It accurately broadcasts power industry terms, synthesizes alarm text into an audio stream, and plays it through a voice gateway after the call is connected, with a synthesis delay of ≤1 second.

[0067] In this embodiment, as Figure 1 As shown, this system utilizes multimodal communication methods such as "telephone + SMS" to ensure that information is perceived immediately in case of emergency faults, leveraging the strong alerting characteristics of telephone. Especially when staff are in noisy environments, telephone alerts can be responded to promptly, solving the problem of delayed or missed responses caused by SMS, and significantly improving the speed and reliability of emergency response. The software module, based on a flexible subscription strategy according to event type, level, and personnel role, changes the "one-size-fits-all" mass sending mode, making notifications more targeted, reducing information noise, increasing the attention of recipients, and achieving precise and personalized alarm notifications.

[0068] In this embodiment, the system is based on multimodal communication technology. The SMS modem module uses industrial-grade 4G / 5G multi-band equipment, which achieves highly reliable SMS sending and receiving through the AT command interface. It is adapted to the complex electromagnetic environment of the four power zones and supports the national cryptographic standard SM4 encryption transmission to ensure the security of the communication link. The automatic dialing module is based on the SIP protocol to build a switching platform, integrates a voice gateway algorithm, can trigger outbound calls in batches and monitor the call status in real time, automatically match the contact level according to the fault level, and combine a heartbeat detection mechanism to ensure service continuity. The text-to-speech (TTS) module adopts a lightweight neural network model and achieves localized deployment through domain fine-tuning and model compression technology. It supports accurate broadcasting of power professional terms ("differential protection", "switching operation") with a synthesis delay of ≤1 second.

[0069] In this embodiment, at the centralized control system level of power equipment, the system supports defining SMS and voice messages for data sources such as alarms, defects, and events; provides personnel grouping information input function; provides personnel information input function, including name, telephone number, personnel group, etc.; provides broadcast type classification function, supporting the classification of subscribed messages, including: accident, abnormal, application condition, etc.; provides subscriber configuration function, supporting different types of subscriptions, and supporting push to different personnel for the same type; supports batch push of SMS messages and making phone calls by personnel and groups; supports configuring the date and time period for receiving SMS messages, and supports publishing through telephone platforms, traditional SMS gateways, etc.; provides a sent history query module, supporting queries based on receiving personnel, sending type, SMS and voice push time, etc.; supports manual definition and modification of SMS information, and SMS publishing function. All of the above are implemented through relevant modules in the software.

[0070] In this embodiment, collaboration is achieved through a unified communication middleware: after the system analyzes the device alarm data in real time, it dynamically selects the optimal communication method (such as emergency fault-triggered voice outbound call + SMS synchronous push), and at the same time converts the handling suggestion text into voice commands, forming a "perception-decision-reach" closed loop, comprehensively improving the intelligence and reliability of power emergency response.

[0071] Example 2:

[0072] Specifically, the strategy configuration and management module is a low-code visual strategy orchestration platform that provides a strategy canvas, allowing users to build communication processes by dragging and dropping predefined strategy nodes. The strategy nodes include receiving events, condition judgments, sending SMS messages, initiating voice calls, and executing TTS broadcasts. This module compiles the drawn flowcharts into a backend executable rule chain for storage and execution.

[0073] In this embodiment, this module adopts a front-end and back-end separation architecture. A React-based front-end interface is deployed in four zones, while the back-end uses a Spring Cloud micro-distributed service architecture, together forming a low-code visual strategy orchestration platform. Its core function is to transform complex multimodal communication logic into intuitive graphical configuration, significantly reducing the operational threshold for operations and maintenance personnel.

[0074] In this embodiment, the policy configuration and management module mainly consists of the following components:

[0075] 1. Strategy Canvas and Component-Based Design: A visual orchestration interface is adopted, with the front end providing a "Strategy Canvas" similar to a flowchart editor. Users can build complete communication processes by dragging and dropping predefined strategy nodes ("Receive Event", "Condition Judgment", "Send SMS", "Initiate Voice Call", "Execute TTS Broadcast", "Wait for Feedback", "Elevation Notification").

[0076] 2. Dynamic Node Parameter Form: After each strategy node is dragged into the canvas, a corresponding parameter configuration form will be dynamically generated on the right side of the interface. The "Conditional Judgment" node can be configured with rules; the "Send SMS" node can be associated with a preset SMS template and the source of variables to fill the template can be selected.

[0077] 3. Event-Response Rule Engine: The system compiles the flowcharts drawn by the user on the canvas into an executable rule chain. In this system, the rule chain is stored in the database in JSON format. The rule engine supports complex logic, including: parallel branches (execute SMS and phone calls simultaneously) and sequential steps (make a phone call first, and send an SMS if there is no answer).

[0078] In this embodiment, the module is not only about updating a single interface, but also about deeply integrating low-code visual orchestration with a domain model for power alarms. This allows non-R&D personnel to flexibly customize intelligent communication processes that meet the requirements of complex on-site management, achieving "what you see is what you get" and "rapid iteration" of operation and maintenance strategies.

[0079] Specifically, the event receiving and intelligent decision-making module includes a secure data synchronization subsystem and an intelligent decision-making engine.

[0080] The security data synchronization subsystem, through the read-only data mirroring service and forward-type physical unidirectional optical gate deployed in Power Zone 1, uses change data capture technology based on database logs to incrementally synchronize alarm event data from Zone 1 to the database of Zone 4.

[0081] The intelligent decision engine integrates a domain-specific large language model fine-tuned from power industry corpus. It is used to perform semantic understanding and structured information extraction on synchronous alarm events, generate summary text containing event summary, current status, preliminary judgment and suggested measures, and perform policy matching based on the semantics of the summary text and event tags to determine the final notification policy.

[0082] In this embodiment, the event reception and intelligent decision-making module consists of two parts: a secure data synchronization subsystem and an intelligent decision engine. It combines reduced data synchronization with the semantic understanding capabilities of a domain large language model (LLM). This module specifically consists of the following components:

[0083] 1. Secure Data Synchronization Subsystem: Employing a unidirectional synchronization architecture, a read-only data mirroring service is deployed in Zone 1 (production control zone). Through a forward-looking physical unidirectional optical gateway, critical alarm event tables and real-time device status tables are incrementally synchronized to the database in Zone 4. The optical gateway ensures data can only flow from Zone 1 to Zone 4, fundamentally preventing reverse engineering attacks. Based on an SQL change data capture synchronization mechanism, using CDC technology in database logs, INSERT (add alarm) and UPDATE (status update) operations in the source tables of Zone 1 are captured in real time, converted into standard SQL statements or messages, and transmitted to Zone 4 through the unidirectional optical gateway. The receiving service in Zone 4 executes these SQL statements or parses the messages to update the local mirror database. Compared to full table polling, this method is more efficient, has lower latency (second-level), and places minimal pressure on the source database.

[0084] 2. Intelligent Decision Engine: A domain model fine-tuning and tool invocation are embedded on the four-zone server. A private model, fine-tuned using extensive power operation and maintenance procedures, fault cases, and dispatch reporting data, is deployed on the four-zone server. This model not only possesses general semantic understanding capabilities but has also been trained to master power grid jargon. This leverages the innovative aspects of current AI development and application, providing structured output and framework filling for event information. The core change in the engine lies in its working mode. When a new event arrives synchronously, information extraction and summarization are performed. The engine inputs the original event data (e.g., equipment: 330kV #1 main transformer, telemetry: oil temperature 85℃, alarm: high oil temperature, time: 10:05) into the model, along with a prompt, requiring the model to organize the information according to the preset power grid reporting framework.

[0085] A framework template is as follows: "[Event Summary] A problem occurred on [device] at [time]. [Current Status] The key parameter is [parameter]. [Preliminary Assessment] Possible causes are [Cause 1] and [Cause 2]. [Recommended Actions] We recommend immediate [Action 1] and checking [Action 2]."

[0086] 3. Multi-round Policy Matching: The structured, conversational summary text generated by the model, along with the original event level, type, and other tags, is sent to the policy matching module. This module does not perform simple matching but utilizes the semantics of the summary text to perform vector similarity retrieval in the policy library to find the most suitable handling policy template for the current situation, ultimately determining the contact person and communication method combination. 4. Execution and Feedback Learning: The decision results drive the 103 module to execute communication actions. The logs of the entire process (events, model summaries, matched policies, execution results) form a high-quality data pair, used to continuously strengthen the model training and form a self-evolving closed loop of decision-making ability.

[0087] Specifically, the multimodal communication execution module is managed through a unified communication scheduling middleware. This middleware monitors the SMS sending status and voice call status in real time and supports a fault tolerance mechanism: when the preferred communication method fails to reach the target within a set time, the backup communication method or the upgrade notification object is automatically triggered.

[0088] In this embodiment, the multimodal communication execution module performs the following operations in parallel through a unified communication scheduling middleware:

[0089] SMS channel:

[0090] The SMS modem driver is invoked, and AT commands are used to send the formatted alarm text SMS to all target numbers.

[0091] Voice channel:

[0092] (1) Automatic dialing: The voice gateway initiates automatic calls to high-priority targets according to the SIP protocol.

[0093] (2) Speech Synthesis and Broadcasting: For alarms requiring broadcasting, the decision engine extracts key text from the knowledge base or event information and feeds it into a locally deployed lightweight TTS neural network model, which is then fine-tuned using power industry terminology in the Tacotron2 model. This model runs on the AI ​​server, synthesizing the text into an audio stream, which is then automatically played via a voice gateway after the call is connected. The entire process delay is controlled within 1 second to ensure real-time information delivery. It also supports TCP communication with the server IP and port number, and in cases where the AI ​​cannot recognize or support the alarm, it sends customized templates for accident alarms.

[0094] (3) Status monitoring and fault tolerance: The communication middleware monitors the status of each SMS message and each call in real time (successful / failed, rejected / busy / connected). If the preferred method (telephone) fails to reach the recipient within the set time, an upgrade notification for the backup method will be automatically triggered, such as dialing a second contact or sending a more detailed SMS message.

[0095] Example 3:

[0096] The intelligent alarm method for power centralized control based on multimodal communication fusion includes the following steps:

[0097] S1: Configure alarm notification policies in a visual manner through the policy configuration and management module.

[0098] S2: Through the event receiving and intelligent decision-making module, alarm events are received from the centralized control platform, and the target recipient, communication method combination, and notification content are intelligently determined.

[0099] S3: Through the multimodal communication execution module, based on the decision result of S2, SMS sending, automatic voice calling and voice synthesis broadcasting are executed in parallel or serially.

[0100] S4: Record and store end-to-end notification logs through the closed-loop feedback and record traceability module.

[0101] Specifically, the intelligent decision-making process in S2 is as follows:

[0102] S21: Through the security data synchronization subsystem, alarm event data generated in Power Zone 1 is synchronized to Power Zone 4 in real time and incrementally.

[0103] S22: Input the synchronized event data into the domain-specific large language model that has been fine-tuned for the power sector, and the model generates structured, conversational summary text.

[0104] S23: Combining the semantics of the summary text with the original level and type label of the event, perform matching or vector similarity retrieval in the strategy library to determine the final notification strategy.

[0105] Specifically, the process of performing speech synthesis and broadcasting in S3 is as follows:

[0106] S31: Extract key text from decision results or event information.

[0107] S32: Input to a locally deployed lightweight TTS model to generate an audio stream, which is then automatically played via a voice gateway after a call is connected.

[0108] S33: The communication scheduling middleware monitors the call status in real time.

[0109] S34: If the call is successful, the call is completed; if the call fails or no one answers, the backup notification process is triggered.

[0110] Specifically, the alarm notification strategy configured in S1 supports combination configuration by personnel group, alarm type, and receiving time period, and supports setting different levels of recipients and differentiated communication method sequences for the same event type.

[0111] Specifically, the method further includes:

[0112] Record the event source, decision-making strategy, communication execution status, recipient, and timestamp of each alarm notification task as a full-link log.

[0113] It provides a query interface that supports querying and tracing historical notification records by at least one of the following criteria: recipient, sending type, and time range.

[0114] Based on the above description, in this embodiment, as follows: Figure 2 As shown, this alarm method can also be described as:

[0115] Step 1: Configure personnel, groups, event types, and communication strategies.

[0116] Step 2: Receive alarm events from the centralized control system.

[0117] Step 3: Determine the target personnel, communication methods, and content templates.

[0118] Step 4: Call the communication middleware to execute tasks in parallel, including calling the SMS modem to send formatted SMS messages; calling the voice gateway to initiate automatic dialing calls; and calling the TTS engine to synthesize voice content and play it when the call is connected.

[0119] Step 5: Monitor the execution status of each channel.

[0120] Step 6: Check whether all preset communications have been completed or timed out. If they have been completed or timed out, proceed to the next step. If they have not been completed, return to step 5 for monitoring.

[0121] Step 7: If completed, record the complete communication log;

[0122] If the timeout occurs, an upgrade notification will be sent to the backup contact person, and a complete communication log will be recorded.

[0123] The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

[0124] Many other changes and modifications can be made without departing from the concept and scope of this invention. It should be understood that this invention is not limited to the specific embodiments, and the scope of this invention is defined by the appended claims.

Claims

1. A power centralized control intelligent alarm system based on multimodal communication fusion, comprising a system hardware platform deployed in four power zones and software on the hardware platform, characterized in that, The hardware platform includes an application server, a database server, a communication hardware unit, and a network security device. The communication hardware unit includes an SMS modem module, an automatic dialing module, and a text-to-speech module. The software is deployed on a hardware platform and includes: The strategy configuration and management module provides a visual interface for configuring multimodal communication strategies. The event receiving and intelligent decision-making module is used to receive alarm events from the power centralized control platform and intelligently decide the target recipient, communication method and notification content according to the strategy and event content. The multimodal communication execution module is used to call the SMS modem, automatic dialing and text-to-speech services in parallel or serially according to the decision results, and to perform SMS sending, voice calling and voice synthesis broadcasting tasks. The closed-loop feedback and record traceability module is used to record end-to-end notification logs and supports query auditing.

2. The power centralized control intelligent alarm system based on multimodal communication fusion according to claim 1, characterized in that, The SMS modem module adopts an industrial-grade 4G / 5G multi-band device, which integrates an AT command interface and supports national cryptographic SM4 encrypted transmission. The automatic dialing module is based on the SIP protocol to build a switching platform, integrates a voice gateway algorithm, triggers outbound calls in batches and monitors the call status in real time, and automatically matches the contact level according to the fault level. The text-to-speech module uses a lightweight neural network model and achieves localized deployment through domain fine-tuning and model compression technology. It accurately broadcasts power industry terms, synthesizes alarm text into an audio stream, and plays it through a voice gateway after the call is connected, with a synthesis delay of ≤1 second.

3. The power centralized control intelligent alarm system based on multimodal communication fusion according to claim 1, characterized in that, The strategy configuration and management module is a low-code visual strategy orchestration platform. It provides a strategy canvas, which allows users to build communication processes by dragging and dropping predefined strategy nodes. The strategy nodes include receiving events, condition judgment, sending SMS, initiating voice calls, and executing TTS broadcasts. This module compiles the drawn flowcharts into backend executable rule chains for storage and execution.

4. The power centralized control intelligent alarm system based on multimodal communication fusion according to claim 1, characterized in that, The event receiving and intelligent decision-making module includes a secure data synchronization subsystem and an intelligent decision-making engine; The security data synchronization subsystem, through the read-only data mirroring service and forward-type physical one-way optical shutter deployed in Power Zone 1, uses database log-based change data capture technology to incrementally synchronize alarm event data from Zone 1 to the database of Zone 4. The intelligent decision engine integrates a domain-specific large language model fine-tuned from power industry corpus. It is used to perform semantic understanding and structured information extraction on synchronous alarm events, generate summary text containing event summary, current status, preliminary judgment and suggested measures, and perform policy matching based on the semantics of the summary text and event tags to determine the final notification policy.

5. The power centralized control intelligent alarm system based on multimodal communication fusion according to claim 1, characterized in that, The multimodal communication execution module is managed through a unified communication scheduling middleware. This middleware monitors the SMS sending status and voice call status in real time and supports a fault tolerance mechanism: when the preferred communication method fails to reach the target within a set time, the backup communication method or the upgrade notification object is automatically triggered.

6. A power centralized control intelligent alarm method based on multimodal communication fusion, applied to the system described in any one of claims 1-5, characterized in that, Includes the following steps: S1: Configure alarm notification policies in a visual manner through the policy configuration and management module; S2: Through the event receiving and intelligent decision-making module, alarm events are received from the centralized control platform, and the target recipient, communication method combination, and notification content are intelligently determined. S3: Through the multimodal communication execution module, based on the decision result of S2, SMS sending, automatic voice calling and voice synthesis broadcasting are executed in parallel or serially; S4: Record and store end-to-end notification logs through the closed-loop feedback and record traceability module.

7. The intelligent alarm method for power centralized control based on multimodal communication fusion according to claim 6, characterized in that, The specific process of intelligent decision-making in S2 is as follows: S21: Through the security data synchronization subsystem, alarm event data generated in Power Zone 1 is synchronized to Zone 4 in real time and incrementally; S22: Input the synchronized event data into the domain-specific large language model that has been fine-tuned for the power sector, and the model will generate structured, conversational summary text; S23: Combining the semantics of the summary text with the original level and type label of the event, perform matching or vector similarity retrieval in the strategy library to determine the final notification strategy.

8. The intelligent alarm method for power centralized control based on multimodal communication fusion according to claim 6, characterized in that, The specific process of performing speech synthesis and broadcasting in S3 is as follows: S31: Extract key text from decision results or event information; S32: Input to a locally deployed lightweight TTS model to generate an audio stream, which is then automatically played via a voice gateway after a call is connected; S33: The communication scheduling middleware monitors call status in real time; S34: If the call is successful, the call is completed. If the call fails or goes unanswered, a backup notification process is triggered.

9. The intelligent alarm method for power centralized control based on multimodal communication fusion according to claim 6, characterized in that, The alarm notification policy configured in S1 supports combination configuration by personnel group, alarm type, and receiving time period, and supports setting different levels of recipients and differentiated communication method sequences for the same event type.

10. The intelligent alarm method for power centralized control based on multimodal communication fusion according to claim 6, characterized in that, The method further includes: Record the event source, decision-making strategy, communication execution status, recipient and timestamp of each alarm notification task as a full-link log; It provides a query interface that supports querying and tracing historical notification records by at least one of the following criteria: recipient, sending type, and time range.

Images