222 results about "Alarm trigger" patented technology

The invention relates to the technical field of public safety and intelligent fire protection, in particular to a real-time processing method and system based on fire alarm data, and the method comprises the steps: environment steady state reconstruction: accessing non-fire environment dynamic parameters; presetting thermotechnical static parameters of the building space; constructing a space thermal inertia differential equation; generating an ideal reference baseline; knowledge-driven simulation: receiving the ideal reference baseline; calling a synthesis operator in a preset disaster interference knowledge base; executing dynamic superposition injection; generating a virtual sensor state flow which has physical characteristics and contains environmental background characteristics; homomorphic judgment: executing double difference calculation; obtaining a real residual feature vector; geometric homomorphism verification is executed; outputting fire alarm triggering, interference filtering or fault prompting instructions; according to the method, the problem of false alarm caused by non-stable fluctuation of the environment background in the background technology is solved, and dynamic fusion and scene adaptation of the standard signal model and the real environment background are realized.

The invention discloses a monitoring equipment fault rapid positioning method based on causal diagram reasoning, and the method comprises the following steps: building a thermal diffusion model composed of equipment nodes and links through collecting the topological structure and operation data of a monitoring system; normalizing and fusing the operation deviation information into a node abnormal energy value, and establishing an abnormal energy propagation equation; when an alarm is triggered, collecting an energy change sequence of each node to generate a space-time energy distribution matrix, and reversely solving a thermal diffusion equation to calculate a disturbance source item; performing clustering backtracking on disturbance source items, determining fault root cause nodes, and comprehensively evaluating confidence; and when multi-source abnormity exists, thermal field orthogonal decomposition is executed, and model parameters are adaptively updated according to field results. According to the method, abnormal propagation of monitoring equipment is modeled into a thermal diffusion process, and reverse solution and adaptive calibration are combined, so that rapid, explainable and high-precision positioning of fault root causes in a complex monitoring network is realized.

The invention discloses a single-axis mechanical accelerometer and a measuring method, and relates to the technical field of accelerometer measuring equipment.The single-axis mechanical accelerometer comprises a base and a dome shell installed on the base, a first cavity is formed in the base, and an I-shaped balancing weight capable of axially sliding is installed in the first cavity; a first spring is mounted between the I-shaped balancing weight and the cavity bottom of the first cavity; a conical sleeve capable of sliding in the radial direction is installed in the dome shell, a transverse shaft is inserted into the conical sleeve in a penetrating mode, the two ends of the transverse shaft are fixedly connected to the inner wall of the dome shell, the transverse shaft is sleeved with a second spring, one end of the second spring is fixedly connected to the transverse shaft, and the other end of the second spring is fixedly connected with the conical sleeve. An alarm triggering module is installed on the dome shell, and through the synergistic effect of an I-shaped balancing weight-spring inertia system and a conical sleeve triggering mechanism, the acceleration detection and alarm functions of a pure mechanical structure are achieved.

The invention discloses a contact type alarm output device for a mechanical accelerometer, and relates to the technical field of accelerometers, the contact type alarm output device comprises a dome shell, and the dome shell is provided with a first cavity; the first contact, the second contact, the first alarm output lead and the second alarm output lead are inserted into the dome shell in a penetrating mode, the first alarm output lead is connected to the end, inserted into the outside of the dome shell, of the first contact, and the second alarm output lead is connected to the end, inserted into the outside of the dome shell, of the second contact. The second alarm output lead is connected to the end, penetrating into the outer portion of the dome shell, of the second contact, one end of the metal spring piece is fixedly connected to the first contact, the other end of the metal spring piece is in point contact with the second contact, and the metal spring piece and the second contact are in point contact design. And in combination with mechanical pushing of the conical sleeve, tight contact and reliable conduction during alarm triggering are ensured, and mistaken contact or poor contact caused by vibration or impact is avoided.

The invention belongs to the technical field of foreign matter detection and alarm, and discloses a vehicle-mounted multi-sensor fusion runway FOD detection and alarm device. Comprising a data acquisition module for acquiring radar data, visual data and infrared data, and performing space-time synchronization and coordinate calibration to obtain synchronous data; the feature analysis module is used for analyzing the synchronous data to obtain radar features, visual features and infrared features; the analysis and complementation module is used for detecting the radar features, the visual features and the infrared features to obtain combined features, and performing shielding detection and feature complementation on the combined features to obtain complemented features; constructing an FOD dynamic feature tree; the positioning alarm module is used for analyzing the FOD dynamic feature tree and making an FOD alarm decision according to an alarm triggering condition; according to the method, accurate detection and reliable alarm of the runway FOD are realized, the accuracy, robustness and practicability of FOD detection in a complex scene are remarkably improved, and the runway FOD safety management and control requirements in a vehicle-mounted scene are effectively met.

According to the intelligent operation monitoring system based on multi-source sensing, through cooperative work of the sensing terminal, the positioning base station, the intelligent video monitoring device and the back-end control system, combined monitoring of the operation environment, the personnel state and behavior violation is achieved. The system realizes indoor and outdoor seamless positioning through cooperation of UWB positioning and GPS positioning, and constructs a multi-dimensional security monitoring network in combination with environment sensor data and a video behavior recognition result. Through automatic processing of the alarm module and the video linkage module, rapid response of alarm triggering and video linkage is realized. The system supports historical data backtracking analysis to generate an operation risk thermodynamic diagram, and provides data support for safety management. The system improves the real-time performance and accuracy of operation safety monitoring, reduces the manual inspection cost, improves the emergency response efficiency, achieves the precise early warning of safety risks through multi-source data fusion, and improves the scientificity and refinement level of operation safety management.

The invention relates to a broadcast audio quality discrimination method and device, and belongs to the technical field of broadcast audio monitoring, and the discrimination method comprises the steps: receiving and preprocessing a broadcast audio signal stream, and outputting a framed signal sequence; extracting multi-dimensional features of the framed signal sequence in parallel, and generating a multi-dimensional feature set comprising a signal strength feature, an interference feature, a noise feature and a propagation disturbance feature; performing dynamic weighted fusion calculation based on the multi-dimensional feature set to generate a comprehensive audibility score; performing quality discrimination and root cause positioning according to the comprehensive audibility score and the multi-dimensional feature set, and outputting a quality grade and a fault type identifier; and when the quality grade is lower than a preset grade threshold value or the multi-dimensional feature set satisfies an alarm triggering condition, generating real-time alarm information and visual data. According to the invention, accurate evaluation and accurate traceability of the broadcast audio quality are realized.

The invention provides a cable fault accurate monitoring method, and belongs to the technical field of cable fault monitoring, and the method comprises the following steps: 1, arranging a sensor on the surface of a cable, and carrying out the data collection, 2, carrying out the preprocessing and feature extraction of a collected signal, 3, carrying out the recognition and positioning of a cable fault, and 4, carrying out the recognition and positioning of the cable fault. And step 4, realizing dynamic updating by using edge computing and cloud collaboration. Through multi-modal sensor fusion and intelligent algorithm optimization, a fault positioning error is controlled to be smaller than or equal to 1.8 m in an onshore scene and smaller than or equal to 10 m in a seabed scene, millisecond-level fault recognition and positioning are achieved, the time from fault occurrence to alarm triggering is smaller than or equal to 5 seconds, and the time is more than 100 times faster than that of a traditional method; the false alarm rate is reduced from 87% of a traditional method to less than or equal to 2%, through accurate monitoring and positioning, the manual inspection and troubleshooting time is reduced, and the operation and maintenance cost is reduced by about 40%.

The utility model relates to an electric control cabinet intelligent safety alarm device, which comprises an electric control cabinet main body, an ultrasonic ranging module, an alarm module and a communication module, and is characterized in that the ultrasonic ranging module measures the time difference of transmitting and receiving ultrasonic waves to determine the actual distance between an obstacle and the electric control cabinet main body, so that non-contact ranging is realized; the timeliness of alarm triggering is ensured, and accurate measurement can be carried out even in a severe environment full of dust, smog or corrosive substances and the like; the multi-mode alarm module and the ultrasonic ranging module are combined for use, multi-mode alarm is carried out when the ultrasonic ranging module triggers the alarm function, a multi-level verification and confirmation mechanism is realized, the possibility of false alarm is reduced, diversified alarm signals are provided, and the alarm can be carried out even in a noisy or poor-sight environment. Normal receiving of alarm information can also be ensured; by means of the communication module, the intelligent safety alarm device of the electric control cabinet has remote monitoring and management functions, and the quick response capability is remarkably enhanced.

The invention discloses a thermal management system fault processing method, a thermal management system fault processing device and a related product. And under the condition that the fault alarm signal is monitored, fault information corresponding to the fault alarm signal is input to the influence evaluation model. The influence evaluation model can carry out objective, automatic and standardized influence grade evaluation on a fault from three dimensions of a refrigeration / heating function of the thermal management system, acquisition of parameters of a preset type and a preset safety alarm triggering condition, and finally obtains an influence grade and executes a control strategy corresponding to the influence grade. And a closed-loop automatic process from fault monitoring to grade judgment to strategy execution is formed. And through refined grading and differentiated fault-tolerant strategies, the functions of the thermal management system can be maintained to the maximum extent on the premise of safe operation of the thermal management system so as to improve the availability of the thermal management system. And meanwhile, the method can be popularized and applied only by defining the influence of each fault on three dimensions according to requirements, has high universality and is suitable for various thermal management systems.

The invention relates to the technical field of network security, in particular to a network security threat real-time analysis method and system based on a multi-source log, and the method comprises the following steps: based on a firewall log, completing the verification of a source IP and a time reference, unifying a field format to form an alignment sequence, dividing a time window according to the source IP, and generating an abnormal behavior track according to the cross-window change. And combining the port path and the authentication result to identify a conflict record to form an alarm triggering record. According to the method, a cross-log alignment frame is established through a unified field structure and a time reference, scattered records are converted into a data sequence capable of being continuously associated, a behavior track link is established around continuous window changes, abnormal judgment is changed from fragment recognition to process recognition, and a conflict evidence set is formed by relying on consistency verification of a port path and authentication information; and alarm generation has a centralized basis and a hierarchical judgment basis, so that the continuity and judgment definition of threat identification are improved.

The invention discloses a virtual technology medical linear accelerator maintenance process method, and relates to the technical field of linear accelerator maintenance, and the method comprises field equipment control: the field equipment control is connected with linear accelerator hardware, and the method is provided with a field maintenance terminal, a remote maintenance platform, an expert cooperation module and a fault knowledge graph database; a perfect maintenance process system is constructed, the maintenance process is managed through process and safety control, the operation safety is ensured, remote expert interaction authentication, key parameter transmission and abnormal alarm triggering are realized through expert access, the overall maintenance decision speed is increased, multi-brand equipment access is supported, the use range and scene are expanded, and the maintenance efficiency is improved. And meanwhile, the annual operation and maintenance cost of the equipment is reduced, the cost is effectively controlled, maintenance process data and corresponding analysis results are managed, visualized and stored, maintenance process management can be realized through cloud management and control of the Internet of Things, and the intelligent level of maintenance process management is improved.

The invention discloses an engine oil consumption calculation and early warning method for a vehicle engine. The engine oil consumption calculation and early warning method comprises the following steps that S1, engine operation parameters are obtained, wherein engine rotating speed and fuel consumption parameters in engine operation data are obtained; s2, extracting engine oil consumption by interpolation: extracting the engine oil consumption under a corresponding working condition from a pre-constructed engine oil consumption model through an interpolation method according to the engine rotating speed and the fuel consumption; s3, accumulative calculation and unit conversion, wherein accumulative calculation and unit conversion are conducted on the multiple sets of engine oil consumption, and accumulative engine oil consumption is obtained; and S4, comparison and alarm triggering: triggering a corresponding alarm prompt according to a comparison result of the accumulated engine oil consumption and a preset engine oil consumption limit value. By means of the method, real-time and accurate calculation of the engine oil consumption under the full working condition of the engine is achieved, the hysteresis quality and estimation errors of manual inspection are broken through, and engine oil consumption data are upgraded to a real-time continuous value from a regular discrete value.

The invention provides a vehicle self-diagnosis method and a vehicle-mounted TBOX. The vehicle self-diagnosis method comprises the steps that a vehicle-mounted bus signal is collected in real time, the confidence coefficient is obtained based on the vehicle-mounted bus signal through a time sequence prediction model, and meanwhile a reconstruction error is obtained based on the vehicle-mounted bus signal through an anomaly detection model; judging whether an alarm triggering condition is met or not through a diagnosis decision engine in combination with the confidence coefficient and the reconstruction error, and if yes, triggering alarm work; otherwise, continuously monitoring the vehicle-mounted bus signal. According to the method, the self-diagnosis model is embedded in the vehicle-mounted TBOX, potential faults or abnormal states can be independently found in the vehicle driving process without depending on the cloud, alarm work can be triggered at the moment when the vehicle is abnormal, the delay problem caused by depending on cloud operation is avoided, and the method has remarkable significance in guaranteeing safe operation of the vehicle; meanwhile, the whole self-diagnosis scheme is automatically operated, manual intervention is not needed, a local scene can be actively adapted for early warning, and the safety performance of the vehicle is further guaranteed.

The invention discloses a disaster recovery switching method and device for a vehicle management system and a storage medium, and relates to the technical field of digital information transmission, and the disaster recovery switching method for the vehicle management system comprises the steps: responding to a region fault alarm triggered for a fault region, executing a main and standby database switching action based on the databases of the fault area and the to-be-transferred area; in response to operation and maintenance input for the fault area, updating a target protocol address and a target area address, corresponding to a to-be-transferred area, of each vehicle type on the basis of the to-be-transferred area indicated by the operation and maintenance input; and controlling the cluster management platform of the fault area to disconnect the service connection with each vehicle type based on the fault protocol address. The technical effect of improving the fault service recovery efficiency of the server area can be achieved.

The invention discloses a wind turbine generator running state monitoring method based on multi-source heterogeneous data, and the method comprises the steps: collecting SCADA multi-source heterogeneous data, removing abnormal samples, and carrying out the standardization; converting the collected non-normal data into approximate multivariate normal distribution based on a space rank; an elastic network penalty likelihood function is constructed, and mean vector estimation representing SCADA sparse fault signals is solved; calculating a basic monitoring statistic based on mean vector estimation and performing robust optimization to obtain a robust monitoring statistic and an upper control limit; real-time robust monitoring statistics are calculated, and if the real-time robust monitoring statistics exceed the upper control limit, a fault alarm is given out; and after a fault alarm is triggered, reversely deducing mean vector estimation according to the optimal penalty coefficient, and identifying and outputting fault component information. According to the method, the problems of complex SCADA data distribution, variable grouping correlation interference, difficulty in sparse fault identification, insufficient controlled samples and high multi-unit monitoring cost in the prior art are solved.

The invention provides a smoke and fire detection alarm method and a camera, and relates to the technical field of fire safety intelligent monitoring and emergency response, and the method comprises the steps: collecting an image of a monitoring area through the camera; analyzing the image of the monitoring area, extracting the form, texture, color and dynamic change multi-dimensional features of smoke and flame in the image, comparing the features with a preset smoke and flame feature database, analyzing the possibility of existence of smoke or flame features, and obtaining a possibility value; according to the possibility value, if smog or flame characteristics are analyzed and confirmed, an alarm triggering instruction is automatically obtained and sent out; based on an alarm triggering instruction, a remote alarm notification is automatically sent to a specified contact person through a wireless communication network in various modes of making a call, sending a short message or pushing a message by means of a built-in communication function. According to the invention, intelligent and efficient management and control of the whole process from early-stage accurate recognition of the fire behavior to emergency disposal can be realized.

The invention provides a tunnel safety monitoring and early warning management method and system, and relates to the technical field of tunnel safety, and the method comprises the steps: carrying out the multi-mode data collection of a tunnel environment through a plurality of intelligent monitoring terminals connected to an RS485 bus, and carrying out the environment abnormality alarm triggering analysis; when any intelligent monitoring terminal detects an alarm event, the alarm event is broadcasted to other terminals through an RS485 bus, and other terminals receiving the alarm event autonomously analyze a diffusion probability map, calculate an emergency evacuation route and control an escape indicating device to display an evacuation direction; communication authority differentiation scheduling based on the RS485 bus is executed, and alarm event situation awareness based on multi-modal data dynamic updating is completed. The technical problems that in the prior art, tunnel safety monitoring and early warning management often adopts a centralized control mode, information transmission and processing are lagged, the alarm response time is long, and effective emergency disposal cannot be carried out in time are solved.

The invention discloses an alarm linkage system based on a multi-modal large model and a working method of the alarm linkage system. The alarm linkage system comprises a target detection module, an alarm screenshot module, a data preprocessing module, a multi-modal large model re-judgment module and an alarm triggering module. The working method comprises the following steps: S1, the target detection module collects a monitoring video stream and identifies unsafe behaviors in a scene; s2, when an unsafe behavior is recognized, an alarm signal is generated immediately, and an alarm image is intercepted by an alarm screenshot module; s3, the data preprocessing module processes the alarm image; s4, the multi-modal large model re-judgment module performs secondary verification on the alarm behavior in combination with image details; s5, the alarm triggering module receives the re-judgment result and then judges whether real alarm is given or not; s6, the historical data storage module receives the alarm related data and performs classified storage; s7, performing voice query of alarm information through a text and image search module; and S8, the model optimization module periodically scans alarm related data and performs module optimization.

The application discloses a kind of energy storage controller performance monitoring method and device, belong to energy storage system technical field, this method includes: in the embedded operating environment of energy storage controller, multiple-source performance index is collected;According to preset sampling period and time window, the multiple-source performance index collected is aggregated, and feature vector is generated;Feature vector is input in online self-calibration abnormal score model, for each feature in the feature vector, the normalized deviation of feature is calculated using the baseline parameter corresponding to the feature;According to the normalized deviation of each feature and corresponding weight, the abnormal score result corresponding to feature vector is calculated;According to abnormal score result, it is judged whether to meet alarm triggering condition, and alarm mechanism is triggered in the case where alarm triggering condition is met;The multiple-source performance index collected and the abnormal score result are displayed in real time through visual panel.

This invention discloses an intelligent alarm and control method for personnel safety in power plants, relating to the field of power industry safety control technology. The method includes: acquiring multimodal sensing data related to personnel based on the existing sensing system of the power plant; using a multimodal analysis model to fuse and analyze the collected multimodal sensing data to identify personnel safety risks; generating an alarm trigger signal when the risk level reaches a preset threshold; combining intelligent agent technology to generate a personalized alarm strategy after receiving the alarm trigger signal, simultaneously finding the nearest alarm device based on the real-time location of the personnel for continuous alarm, and establishing a continuous monitoring mechanism to track personnel status in real time; continuously monitoring the behavior and location changes of the alarmed personnel until the risk is confirmed to be completely eliminated. This invention enables cross-system correlation analysis, improves the accuracy of risk identification, effectively shortens response time, and improves the safety level of power plants.

This invention provides a method and system for alarm event reporting and remote response linkage based on wireless communication, relating to the field of intelligent security monitoring technology. The method includes: collecting raw sensor data at three-dimensional spatial state sensing anchor points deployed on the top of the equipment casing (acoustic resonant cavity), the bottom inertial reference plane, and the lateral electromagnetic shield); performing trajectory decoupling and phase synchronization processing on the raw sensor data to obtain processed data; converting the processed data into a unified alarm trigger event; performing multi-source feature analysis on the alarm trigger event to obtain event analysis results; extracting the spatial vector coordinates and dynamic stress distribution gradient of the three-dimensional spatial state sensing anchor points based on the event analysis results, introducing the attenuation characteristics of intermolecular van der Waals forces to map the signal field, and constructing a three-dimensional spatial situation fitting ellipsoid. This invention achieves efficient end-to-end coordination between on-site alarms and remote dispatching.

This invention belongs to the field of industrial automation and quality inspection technology, and is used to solve the problem that existing technologies cannot logically distinguish between "physical mixing transition state" and "actual production failure". Specifically, it is an integrated monitoring system for flexible production and quality inspection of alcohol-based fuels, including: a basic parameter acquisition module, a parameter preprocessing module, a dynamic monitoring and analysis module, and a fault alarm triggering module. This invention achieves accurate differentiation between physical mixing transition state and actual production failure by constructing a dynamic logical alignment mechanism between the upstream formula change event flow and the downstream sensor response flow. Specifically, this invention uses an active event pool to store historical formula switching events within a preset survival time window, and calculates the theoretical prediction deviation based on the component change gradient vector of each event and its time decay weight, and then uses residual analysis to remove the mixing transition effect that inevitably exists in the pipeline.

This application is directed to systems and methods for delivering notification for a user identification. An exemplary system may include memory configured to store instructions and at least one processor in electronic communication with the memory. The at least one processor is configured to execute the instructions to obtain a notification preference profile associated with the user identification, receive an alert trigger message associated with the user identification, determine whether to send the notification based on the alert trigger message and the notification preference profile, upon determining to send the notification, determine whether to send the notification in real time or at a scheduled time, and in response to determining to send the notification in real time: generate a real-time alert based on the notification preference profile and send a real-time notification message based on the real-time alert to a user device associated with the user identification.

The invention relates to the technical field of enterprise management, in particular to an enterprise management-oriented enterprise operation supervision method based on a micro-service architecture, and the method comprises the steps: collecting observability data from a plurality of micro-services deployed in the micro-service architecture, carrying out the standardization processing of the observability data, and injecting a unified label set; calculating a real-time supervision index value through a supervision algorithm model by using the observability data after standardization processing, and generating corresponding interpretability information; and obtaining a real-time supervision index value, calculating to obtain a supervision index risk score at the current moment, comparing the supervision index risk score with a dynamic threshold value, generating alarm events when a preset alarm triggering condition is met, and aggregating the alarm events.

The application discloses a kind of digital risk research management methods and systems for high-risk scene, it is related to industrial manufacturing technical field, the method includes: in each target equipment in high-risk place, real-time monitoring device is installed, and based on the functional attribute of each target equipment, the corresponding data type label is configured to the real-time monitoring data obtained by the real-time monitoring device corresponding to each target equipment;When the data type label of real-time monitoring data is state class signal, then based on the preset determination rule and the state corresponding to real-time monitoring data, fault determination is carried out;When the data type label of real-time monitoring data is numerical value class signal, then based on the preset multiple different alarm levels and corresponding alarm trigger threshold, fault determination is carried out.The application can realize the collection of the key parameters of high-risk equipment, improve equipment fault, environmental anomaly prediction accuracy, convert sudden failure repair to planned maintenance, realize safety management front.

Systems and methods for monitoring dispatch notification are disclosed. One method of providing notifications to a user includes, in response to receiving an alarm trigger from an alarm system from amongst multiple alarm systems being used to monitor respective locations, initiating a sequence of actions to be performed to protect the user of the alarm system. In response to a first action of the sequence of actions being initiated by a monitoring service, a first message indicating that the first action has been initiated may be generated and communicated via a communications network to a network address associated with the user. In response to at least one second action being initiated, at least one second message indicating that the second action has been initiated may be generated and communicated via the communications network to the network address associated with the user.

The invention discloses an accurate positioning system based on an air alarm event of an unmanned aerial vehicle group. The system comprises an alarm triggering module used for carrying out anomaly detection on multi-modal data collected by an unmanned aerial vehicle group, and generating a joint alarm vector set containing a timestamp, an alarm direction and confidence when a plurality of unmanned aerial vehicles recognize the same potential alarm area; the spatial analysis module is used for receiving the joint alarm vector set, constructing a three-dimensional alarm pointing vector group in combination with the position and attitude information of the unmanned aerial vehicle, and analyzing and generating a spatial intersection candidate set through a spatial consistency function; and the positioning optimization module is used for receiving the spatial intersection candidate set, performing weighted iterative fitting based on the confidence coefficient, the timestamp deviation and the spatial residual error, and finally outputting an accurate position coordinate of the alarm event. According to the method, high-precision and high-robustness positioning of the air abnormal event is realized through multi-machine collaborative sensing, space intersection analysis and confidence coefficient optimization.

The utility model belongs to the technical field of preventing water leakage, especially relates to a heating pipeline exhaust valve water leakage prevention device, including original heating pipeline exhaust valve. The utility model can ensure that surrounding decoration materials, instruments and equipment etc. are protected at the root when the automatic exhaust valve appears water spraying and leakage phenomenon when exhaust work is not completed, and when the sound and light alarm triggers, the staff can find that the heating pipeline automatic exhaust valve malfunctions at the first time, can carry out targeted maintenance at the first time, after maintenance, open the manual water drain valve under the water collector, drain the water in the water collector, adjust the signal switch of liquid level sensor, make the electromagnetic valve and sound and light alarm closed in linkage, and return to the initial working state. The utility model is simple in operation, convenient, fast, cost-saving, safe, practical in use and easy to disassemble and assemble.

The utility model relates to the related technical field of detection of broken ends of knitting yarns, and discloses an automatic detection and alarm device for broken ends of yarns of a knitting machine, which comprises a multi-mode sensing module arranged on a yarn advancing path, the infrared transmission type photoelectric sensor and the micro tension strain sensor are arranged at intervals in the yarn moving direction. Through cooperation of the multi-mode sensing module, the alarm triggering module and the control unit, the yarn breaking situation can be judged through double-signal combination, so that the yarn breaking misjudgment rate is reduced to the minimum, the situation that the use process is affected by misjudgment is avoided, and through cooperative detection of the photoelectric sensor and the tension bimode sensor, the detection accuracy is improved. The accuracy of the whole detection process is guaranteed, the overall response speed is increased, and the waste condition caused by slow response of the yarn is avoided.