77 results about "Detection early" patented technology

The invention discloses an inter-satellite interconnection routing component radiation hardening method and system, and the method comprises the following steps: obtaining preset radiation hardening configuration parameters in response to a single event upset detection instruction started by a user, the configuration parameters comprising a detection time interval, a detection table item range and a single event upset exception processing strategy; in a system initialization stage, register table entries of a switch chip are backed up through a memory error correction mechanism, a reference table entry backup library is constructed, write operation on the table entries of the switch chip is intercepted during system operation, and the reference table entry backup library is synchronously updated in real time. According to the method, through regular scanning and real-time synchronous backup, efficient detection of single event upset errors is realized, and it is ensured that the system can be found and recovered as soon as possible when upset occurs; a user can set a detection interval, a table item range and a self-defined recovery behavior according to requirements, and extremely high flexibility is provided.

The invention, which relates to the technical field of medical information management, discloses an obstetrical diagnosis and treatment information integrated management and leak detection early warning system comprising a data acquisition module, a diagnosis and treatment trajectory modeling module, a rule knowledge base module, a leak detection identification module, a leak detection risk calculation module, a high-risk cross analysis module and an early warning push module. A gestational week time axis and diagnosis and treatment trajectory model is constructed by collecting basic information of pregnant and lying-in women and previous antenatal care data; matching the trajectory model with a standard obstetrical diagnosis and treatment path rule set, identifying missing or overdue examination, calculating a missing detection weight value, and forming an individualized risk map; further performing cross analysis in combination with high-risk pregnancy indexes, determining high-risk omission items, generating an early warning task queue, and pushing the early warning task queue to a doctor-patient terminal to realize closed-loop management and real-time intervention; the accuracy of missed detection identification and the intelligent level of obstetrical risk management can be effectively improved.

The utility model discloses a fault detection and early warning device for a fan in a tunnel. The fault detection and early warning device comprises an assembly shell, the side wall of the assembly shell is fixedly connected with a connecting block, an installation mechanism is fixedly assembled in the assembly shell, a vibration sensor is fixedly assembled on the assembly shell through the installation mechanism, and an alarm is fixedly assembled on the top of the assembly shell. The state of the fan in the operation process can be detected through the vibration sensor assembled in the assembly shell, when the fan operates abnormally, vibration of different frequencies can occur, and when the vibration frequency exceeds the range set by the vibration sensor, the alarm gives an alarm, so that the fan is prevented from being damaged. According to the utility model, the operation state of the fan is monitored in the tunnel, the structure is simpler, the installation process is convenient, the debugging process of the vibration sensor is simpler and more convenient through the arranged installation mechanism, and the time for erecting a ventilation system is saved.

This invention discloses a high-temperature detection, early warning, and fire prevention and extinguishing system for coal mine goaf areas, relating to the field of coal mine safety monitoring technology. It includes an environmental sensing module, a data processing module, a fire prevention module, and a fire extinguishing module. The environmental sensing module is used to detect and collect environmental data within the coal mine in real time. The data processing module is used to observe CO distribution, analyze the CO concentration distribution patterns in different areas, and, combined with temperature, locate the high-temperature zone of spontaneous combustion point in the goaf area, determining the early warning level. When the early warning level is primary, the fire prevention module is activated; when the early warning level is not primary, fire extinguishing slurry or inert gas is injected into the high-temperature area. This invention can improve the fire prevention and control level in coal mine goaf areas and ensure coal mine production safety.

The tunnel pavement foreign matter intelligent detection early warning system comprises a processing terminal, the processing terminal is electrically connected with a clock circuit, the input end of the processing terminal is connected with a plurality of distance measuring terminals, and the output end of the processing terminal is connected with a plurality of alarm terminals; the distance measuring terminal comprises a distance measuring instrument and a reflecting plate which are oppositely arranged on the road edge in the tunnel, and a plurality of distance measuring terminals are arranged on the road edge at intervals in the length direction; the alarm terminal comprises an alarm lamp arranged at the top of the tunnel, and the alarm lamp is located on the side, close to the tunnel entrance, of the adjacent distance measuring terminal. According to the device, the distance measuring terminal is used for detection, when the reading of the distance measuring terminal is separated from a threshold value and the timing of the clock circuit exceeds a certain time period, the existence of a foreign matter can be judged, a driver can be reminded in an alarm mode, and the situation that the foreign matter does not exist in time and an alarm is caused by vehicle parking is avoided; and positive feedback can be generated, namely, a subsequent driver is reminded to slow down, and a negative effect cannot be caused.

The utility model discloses a synthetic electric field intensity detection and early warning device, relates to the technical field of electric field detection, and aims to solve the problems that a conventional electric field intensity detection and early warning device is mounted on a stand column, birds are easy to nest on the detection device when the device is used outdoors, and related power supply equipment is required to supply power to the detection device. According to the scheme, the device comprises a stand column, an electric field detection probe is installed at the top of the stand column, an instrument box is arranged on one side of the stand column, a lifting plate is arranged at the top of the instrument box, a lifting motor is installed at the bottom of the lifting plate, a telescopic block is arranged in the lifting plate, a sliding block is installed at the end of the telescopic block, and a telescopic rod is arranged on one side of the telescopic block; the outer sides of the telescopic rods are sleeved with limiting springs, and a solar panel is arranged at the top of the box body door. The device is novel in structure, solar energy is adopted to supply power to the detection device, meanwhile, the solar panel can be adjusted according to the irradiation angle of sunlight, and the situation that flying birds nest is effectively prevented.

The application discloses a thunder and lightning comprehensive detection early warning system and a thunder and lightning detection method thereof, which comprises an atmospheric electric field detection module, a meteorological radar module and a system module. The atmospheric electric field detection module is used for detecting characteristic value A of a thundercloud in a specific area, which at least contains electric field data. The meteorological radar module is used for detecting characteristic value B of the thundercloud in the specific area, which at least contains geographic position data. The system module is used for connecting the atmospheric electric field detection module and the meteorological radar module, and combining the received characteristic value A with the characteristic value B to form thunder and lightning early warning information. The application combines the meteorological radar and the atmospheric electric field equipment, embeds the meteorological radar device into the atmospheric electric field instrument, detects echo data by the radar, converts the echo data into an echo intensity color map, and superimposes the color map on a map to form a thunder and lightning detection early warning radar system.

The invention relates to the field of power equipment monitoring, and discloses an oiled paper insulation capacitive bushing full fault detection early warning system, which comprises a multi-parameter sensor module, a data acquisition and transmission module, a data processing and analysis module and an early warning and decision module, the multi-parameter sensor module comprises an ultrahigh frequency (UHF) sensor which is installed near a sleeve flange plate and is used for monitoring a partial discharge signal in a frequency range of 300MHz-3GHz in the sleeve in real time; a plurality of high-precision optical fiber temperature sensors are arranged on a bushing capacitor core, an insulating oil layer and the surface of a bushing, and various types of sensors such as a partial discharge sensor, a temperature sensor, a pressure sensor, a humidity sensor and the like are integrated; according to the invention, parameter changes of electrical, thermal, mechanical, humidity and the like of the oil-paper insulation capacitive bushing in the operation process can be comprehensively monitored, so that comprehensive detection of various faults is realized, and the limitation of a single detection method is effectively avoided.

The application discloses a kind of railway construction area train proximity automatic detection early warning method, belong to railway construction safety monitoring technical field;Through the train perception equipment of operation line both ends, real-time acquisition track image and train movement data, after data pre-processing, feature extraction and multi-source data fusion, accurate identification train target;Record train time, travel speed and other parameters, calculate the real-time distance of train and construction area;When real-time distance reaches preset early warning threshold, trigger early warning, control the sound and light alarm of fixed and vehicle-mounted horn in construction area, call HTTP interface to push SMS or voice alarm notice to management personnel;After train leaves, close alarm device and archive detection data, alarm log traceability;The method is through sensor collaborative identification, accurate distance calculation and double linkage alarm, reduce manual intervention, significantly improve the timeliness and accuracy of railway construction area train proximity early warning, comprehensively guarantee construction safety.

The application discloses a geological loosening detection early warning device applied to the geological detection field, which is installed in a crack of the geological body to be monitored and extends up and down along the geological surface, and longitudinally spans multiple cracks. When the geological body deforms and moves, the sliding part drives the corresponding position of the embedded nails or the embedded cylinder to jointly move downward, and the cross-crack rope is pulled. When the geological deformation reaches a moving degree, the cross-crack rope is separated from the embedded cylinder, the upper missing disc or the lower missing disc which loses the limitation automatically rotates under the action of the magnetic force until the upper missing disc and the lower missing disc are aligned at the gap, the light of the laser lamp smoothly passes through the pair of gaps and is emitted to the air, a timely and effective long-distance light warning effect is given to the monitoring personnel, the monitoring personnel can timely prevent the geological body with disaster precursors, and personnel casualty is effectively reduced.

The utility model discloses a kind of electric leakage detection early warning ground nails, comprising: metal rod and detector on the metal rod, the detector includes shell, pulse sensor, main control unit, detection switch and alarm, the shell is insulating shell and is connected with the metal rod, the detection switch is equipped on the shell, the shell inside encapsulation has the pulse sensor, main control unit and alarm, wherein, the pulse sensor is connected with the metal rod, the detection switch, alarm and pulse sensor are connected with the main control unit. The device can be inserted in the ground at the side of street lamp, power distribution cabinet and other equipment needing to detect whether leakage, when these equipment appears leakage, electromagnetic pulse generated by alternating current can be detected by pulse sensor, after pulse sensor transmits signal to main control unit, main control unit can control alarm to alarm, can directly remind passing personnel to pay attention to leakage danger.

The invention relates to the technical field of pipeline internal damage visual detection, and discloses a pipeline internal damage visual detection system and method, and the system comprises an image collection module, an image pre-analysis module, a pipeline damage detection module, and a detection early warning module. The image acquisition module acquires image data in a current pipeline based on image acquisition equipment carried by the pipeline robot and transmits the image data to the image pre-analysis module. According to the system, the current pipeline fault type and the confidence coefficient of the current pipeline fault type are recognized through an image recognition model by reading image data in the current pipeline, and a pipeline damage assessment coefficient is constructed based on the similarity coefficient of the current pipeline fault type and the environment in a historical database and the confidence coefficient of the current pipeline fault type. And the environment with low confidence and the environment of the identified type are comprehensively evaluated, so that mistakes and omissions of small defects can be avoided.

The application discloses a kind of fault early warning method and system of burying injection molding up and down material automatic machine, it is related to equipment management field, wherein the method includes: obtaining the F-grade working condition information of first injection molding up and down material automatic machine;Based on multidimensional fault analysis index and F-grade working condition information, obtain the first injection machine fault record base;Based on sensitive convergence learning function, according to F-grade working condition information and the first injection machine fault record base carries out deep learning, generates first fault detection channel;Based on real-time operation data source, according to first fault detection channel carries out fault analysis, obtains fault detection coefficient;If fault detection coefficient meets fault detection early warning constraint, generates fault early warning signal.The technical problem that the fault detection accuracy of burying injection molding up and down material automatic machine in prior art is low, leading to the poor fault early warning effect of burying injection molding up and down material automatic machine.

The utility model discloses a kind of battery high temperature detection early warning device, it is related to battery detection early warning technical field, including base, battery detection box, temperature sensor and early warning lamp.The battery high temperature detection early warning device disclosed in the utility model, by making multiple battery main bodies simultaneously remove from the deep place of battery detection box, and then the multiple battery main bodies of detection completion are simultaneously taken out battery detection box, to reduce the time consumption of taking and improve work efficiency;Battery detection box moves, and cleaning cotton will rub with battery cavity lateral wall, and then battery cavity lateral wall is cleaned together, so that staff need not arm to be inserted into battery cavity interior to clean, reduce cleaning workload;By making the battery main body removed from battery detection box enter storage box, then staff only need to move storage box to be able to handle battery main body of detection completion to specified position, to facilitate staff to concentrate on collection and handle battery main body of detection completion.

The application discloses a coal tar component online spectrum detection early warning method and system, and belongs to the field of coal chemical process detection and safety management and control. The method comprises the following steps: continuously sampling and pre-treating the to-be-detected sample at constant temperature and pressure; splitting to an atomic emission and near-infrared spectrum unit for synchronous detection of inorganic components and organic components; after pre-processing the double-spectrum data, performing data fusion and two-way cross correction on the double-spectrum data through multivariate correction and two-way matrix compensation models to obtain full-component detection results; matching the results with a threshold value and a risk rule library to generate a graded early warning signal, uploading the signal to a distributed control system and issuing a process adjustment instruction to form a closed loop. The application realizes synchronous online detection of organic and inorganic components, eliminates matrix interference through two-way cross correction, constructs a graded early warning and process linkage system, and improves the safety management and control level of coal tar processing.

The utility model belongs to environmental detection technical field especially relates to a kind of air microorganism self-checking early warning device, including storage bin, flowing mechanism, detection mechanism, control module, limiting mechanism, air pump and bottom plate, and flowing mechanism includes pipeline and control valve;Detection mechanism includes temperature sensor, electrically conductive probe, shortwave ultraviolet emitter and receiver;Control module is located below limiting mechanism;Limiting mechanism and storage bin form detection space;Storage bin is transported to detection space by first pipeline and control valve with nutrient solution;Air pump is transmitted to nutrient solution by second pipeline with microorganism;Detection data is obtained by detection mechanism;Detection mechanism is connected with control module, and detection data is transmitted control module;Control module processes detection data, and obtains the data reflecting the content of microorganism situation. Only need simple implementation, to realize the real-time detection early warning of microorganism in air with lower cost.

The utility model relates to the field of insulation detection of power system equipment, and particularly discloses an insulation on-line detection and maintenance early warning device for substation high-voltage equipment. According to the utility model, through arranging the dynamic abutting assembly, the detection early warning assembly, the sealing sleeve and the end sleeve on the installation cylinder, a user can stably install the installation cylinder on the cable through the dynamic abutting assembly, then real-time on-line detection is carried out on the insulation performance of the cable through the detection early warning equipment, and the end sleeve is matched with the dynamic abutting assembly, so that the insulation performance of the cable can be detected. Therefore, the installation cylinder can be stably connected with cables with different wire diameters, and the application range of the insulation online detection equipment is effectively expanded.

According to the active early-warning anti-explosion box body, the battery and the electric two-wheeled vehicle, under the condition that the air pressure in the battery is high, the air pressure detection unit detects that the air pressure in the shell will rise, and therefore active early warning can be conducted as soon as possible to remind a rider to find danger as soon as possible, stop the rider in time and get away from the rider, personnel injury is avoided, and meanwhile the safety of the rider is improved. When high-pressure air flow in the shell impacts the bottom surface of the cover body, the elastic belt elastically extends under the impact force, the second side of the cover body can move upwards, and when the ventilation groove is formed in the side plate of the second side, the inner cavity of the shell can be communicated with the external environment through the ventilation groove, so that the high-pressure air flow in the shell can be quickly discharged through the ventilation groove; and personnel and property damage caused by explosion is further reduced.

The utility model relates to the technical field of gas detection, and discloses a combustible gas leakage detection early warning device which comprises a detection early warning instrument body, a fixed seat is fixedly installed on the back face of the detection early warning instrument body, the fixed seat is rotationally connected with a rotating head through a pin shaft, and a four-stage telescopic rod is fixedly installed at the bottom of the rotating head; the four-stage telescopic rod comprises a first hollow rod, the top end of the first hollow rod is fixedly connected with the rotating head, the first hollow rod is sleeved with a second hollow rod in a sliding mode, the second hollow rod is sleeved with a third hollow rod in a sliding mode, the four-stage telescopic rod capable of achieving four-stage telescopic is adopted as a holding section, and detection is conducted on the high place or the place where the hand is not prone to stretching out; the four-stage telescopic rod can be lengthened, then the detection early warning instrument body extends to a detection area for detection, the telescopic length of the four-stage telescopic rod is large, the overall size is not too large when the four-stage telescopic rod is not used, and carrying is not greatly affected.

The utility model provides a high-altitude accumulated hydrogen detection device. The high-altitude accumulated hydrogen detection device comprises an outer balloon, an inner balloon and a hydrogen detection device, the device mainly comprises an annular outer balloon and an annular inner balloon arranged in the outer balloon, and an inner balloon filling body provides buoyancy. An I-shaped support frame is arranged in the annular space of the outer balloon and is used for fixedly mounting a hydrogen detection device which is divided into a body and a servo device, and the body and the servo device are connected through a lead. A protruding part filled with gas is arranged on the outer wall of the inner balloon. The suspension height and posture of the bottom of the device are controlled through a pull rope with an iron balance weight, and the device can be recycled through an auxiliary balloon with a magnet on the top. According to the hydrogen detection early-warning device, the light balloon is made of gas with the density lower than that of air, the light balloon serves as a suspension device, the hydrogen detection early-warning device is carried to the space for accumulation, hydrogen accumulated in the space is detected, operation is easy, and safety is high.

The invention relates to an offshore platform natural gas leakage detection method and a detection early warning system, and the method comprises the steps: collecting a methane leakage infrared image through an infrared gas imaging device, carrying out the pixel probability modeling through Gaussian mixture background modeling, preliminarily recognizing a suspected gas leakage region, and carrying out the early warning of the suspected gas leakage region. And inputting the natural gas into an improved deep learning detection model to detect natural gas leakage, the model takes YOLOv5 as a core structure, background modeling and a neural network are fused, a backbone network is integrated with a background difference image, a detection head comprises a small target sensing module and a context attention mechanism, and the natural gas leakage is detected according to a target detection result output by the model. And key feature parameters of a leakage area in each frame of image are extracted, and graded early warning and system linkage are carried out in combination with time sequence data. According to the method, the detection accuracy and real-time performance in a complex marine environment are remarkably improved, the omission ratio is reduced, the emergency response is accelerated, and the safety monitoring intelligent level of an offshore platform is powerfully improved.

The invention discloses a sea target detection early warning method based on a ViT algorithm, and the method comprises the steps: obtaining a to-be-detected image, and carrying out the preprocessing of the to-be-detected image; establishing a sea target detection model based on a ViT algorithm, and training the sea target detection model; inputting the to-be-detected image into a marine target detection model for target detection to obtain a target detection result; and judging an early warning level by analyzing the target detection result, and executing an early warning operation according to the early warning level. According to the method provided by the invention, the ViT algorithm can be utilized to realize maritime target detection, and then maritime early warning operation is completed.

The utility model provides a hydrogen leakage detection early warning device for a hydrogen pipeline, which relates to the field of pipeline hydrogen transportation and comprises a detector body and a hoop which is arranged on the outer side of the pipeline and used for sealing a flange connecting end, and a gas groove for accommodating a pipeline flange and a valve flange is arranged on the periphery of the inner side of the hoop. The detection end of the detector body is connected to the inner sides of the upper portions of the hoops in a sealed mode, after the detector body is connected to the inner sides of the hoops, the detection end of the detector body extends into the air groove, through the arrangement of the two hoops, the connecting end of the valve flange and the pipeline flange can be sealed, and meanwhile the detection end of the detector body can be isolated in the hoops; not only can a sensor in the detection end of the detector body be prevented from being covered and polluted by dust, but also the phenomenon that trace leakage between the pipeline flange and the valve flange is blown away by wind and cannot be timely detected and alarmed by the detector body can be prevented.

The invention provides an intelligent detection and early warning method and device for fire points, electronic equipment and a medium, and belongs to the technical field of fire early warning, and the method comprises the steps: obtaining image data collected by a plurality of image sensors; the sensor comprises a visible light wave band image sensor, an infrared wave band image sensor and an ultraviolet wave band image sensor; preprocessing the image data and then fusing the image data to obtain a fused image; inputting the fused image into a fire behavior and fire point recognition model to obtain a fire behavior and fire point recognition result; and executing early warning based on the identification result. The fire behavior and fire point identification model is obtained based on the following steps: preprocessing image samples acquired by a plurality of image sensors and then fusing the image samples to obtain a fused image sample; and training a DEV-YOLOv8 model by taking the fire behavior and fire point as a sample label based on the fused image sample to obtain a fire behavior and fire point identification model. According to the invention, the problems of low fire point identification precision and high possibility of false alarm in the existing fire point early warning method can be solved.

The invention belongs to the technical field of computer networks, and particularly relates to a local area network equipment attack monitoring and early warning method based on asynchronous parallel computing, which utilizes the existing attack data and defense data to construct an attack type knowledge graph, is convenient to inquire known attacks, provides background knowledge for monitoring unknown attacks, and improves the security of the equipment. The attack monitoring precision is improved; an asynchronous parallel computing model is designed, equipment in a local area network is used as a computing model, cross model training is carried out, and internal resources are fully utilized to improve monitoring precision and efficiency; the attack early warning scheme is preset, the defense strategy is intelligently matched for the current attack monitoring result, the attack type knowledge graph is updated in real time, the robustness is improved, the method is carried out in the local area network, and internal data privacy leakage caused by third-party management is avoided.

The invention discloses a network security detection early warning method and system and a storage medium, and belongs to the technical field of network security. The method comprises the following steps: controlling a probe arranged at a target network node to obtain real-time network flow data, and determining first deep packet detection data according to an original data packet corresponding to the real-time network flow data and a first deep packet detection mode; determining a real-time association tag corresponding to each user in the target network node according to the real-time network traffic data, the first deep packet inspection data and a preset tag mapping relationship; determining a target user according to a historical tag data set corresponding to each user and the real-time associated tag; the detection result is determined according to the original data packet, the target user and the second deep packet detection mode, the first computing power consumption of the first deep packet detection mode is smaller than the second computing power consumption of the second deep packet detection mode, and the effects that the detection depth can be considered, and it is guaranteed that the operation performance of the detection equipment is sufficient are achieved.

The invention discloses an integrated photovoltaic equipment detection early warning method and system, and relates to the field of photovoltaic equipment monitoring, and the method comprises the steps: arranging a sensor network in combination with assembly deployment information, collecting multi-dimensional data based on the sensor network, preprocessing the multi-dimensional data to obtain a data set, and building an attention-GNN frame. The method comprises the following steps: constructing an attention-GNN framework, training the attention-GNN framework by adopting a data set to obtain an attention-GNN model, inputting real-time data into the attention-GNN model to obtain a real-time associated feature vector, constructing a VAE framework, training the VAE framework by adopting the data set to obtain a VAE model, inputting the real-time feature vector into the VAE model to obtain a reconstruction error, and constructing a VAE model. And performing fault diagnosis and early warning based on the real-time key feature vector and the reconstruction error. The method provided by the invention is adaptive to different types of integrated photovoltaic equipment, double models cooperate to reduce the false detection rate and improve the anomaly recognition accuracy, and the problem of black box in traditional deep learning is solved by associating abnormal nodes to position faults.

The utility model relates to the technical field of pest detection and early warning, in particular to an early warning mechanism for pest detection. According to the technical scheme, the device comprises an insect condition forecasting instrument and a plurality of mounting frames fixedly connected to the bottom of the insect condition forecasting instrument, and further comprises spiral ground inserting rods arranged in the mounting frames and used for being inserted into the ground. Through cooperation of the mounting rack, the spiral ground inserting rod, the control box, the transmission mechanism and other structures, the spiral rod can be rapidly screwed into the ground only by starting all the motors, complex tools and a large number of materials are not needed, mounting is easy and convenient, consumed time is short, and rapid deployment can be achieved under the emergency insect situation. And if the position needs to be adjusted in the later period, the screw rod is easily screwed out for remounting, so that the complexity of dismounting and remounting in a traditional mode is avoided, and the mounting flexibility is greatly improved. In addition, the instrument and the ground are easy to damage during traditional concrete pouring, mounting and dismounting, the risk of damage to the ground and equipment is greatly reduced through screw rod mounting operation, the instrument is protected, and the ground repairing cost is reduced.

The application discloses an anti-radiation reinforcing method and system for an inter-satellite interconnection routing component, and comprises the following steps: in response to a single event upset detection instruction enabled by a user, preset anti-radiation reinforcing configuration parameters are acquired, the configuration parameters comprise a detection time interval, a detection entry range and a single event upset abnormality processing strategy; in a system initialization stage, a memory error correction mechanism is used to back up register entry of a switch chip, a benchmark entry backup library is constructed, and a write operation on the entry of the switch chip is intercepted during system operation, and the benchmark entry backup library is updated in real time. Through periodic scanning and real-time synchronous backup, the application realizes efficient detection of single event upset errors, ensures that the system can be discovered and recovered as soon as possible when the upset occurs, and provides high flexibility because the user can set the detection interval, the entry range and the custom recovery behavior according to requirements.

The invention discloses a refrigerator key device state early warning method, a detection early warning system and equipment, and the method comprises the steps: synchronously collecting the voltage, current, active power and reactive power of a medical refrigerator power supply loop at high frequency, and carrying out the data preprocessing through a power supply fluctuation compensation algorithm; establishing a load curve reference of a medical exclusive working condition, and extracting typical load characteristic frequencies of a compressor, a fan and a defrosting heater under each working condition; based on an improved independent component analysis algorithm, the whole machine load curve is disassembled into a compressor sub-curve, a fan sub-curve and a defrosting heater sub-curve; and the compressor sub-curve, the fan sub-curve and the defrosting heater sub-curve correspondingly give out state early warning according to a grading early warning mechanism. Through power supply side multi-parameter acquisition and load curve analysis, early abnormality identification of devices such as a compressor and a fan is realized, meanwhile, medical cold chain data tracing and supervision linkage requirements are met, and medical material storage safety is guaranteed.