8 results about "Abnormal test result" patented technology

The invention provides a substation simulation test method based on self-supervised learning, and relates to the field of substation simulation test. The method comprises the following steps: acquiring data of a transformer substation in a normal operation state to form an unmarked training data set; forming positive and negative sample pairs based on the data set, and constructing a time sequence context comparison prediction task; based on a time sequence context comparison prediction task, training to obtain a causal expansion convolution feature extraction network; a data set depth feature vector set is extracted, and a multivariate Gaussian normal feature reference is calculated; constructing a transformer substation digital twinborn simulation model; generating a simulation fault data set based on the transformer substation digital twin simulation model; and generating an abnormal test result of the simulation fault data set through the causal expansion convolution feature extraction network and the multivariate Gaussian normal feature reference. According to the invention, self-supervised learning and a digital twinborn simulation model are combined, the transformer substation fault detection method is optimized, and the method has the advantages of high efficiency, accuracy and self-adaption.

The application discloses a kind of test operation management systems suitable for automobile temperature sensor, belong to automobile temperature sensor test evaluation technical field, the application is by static test module, dynamic test module respectively to temperature sensor by static to dynamic many aspects are tested, rely on static test obtains temperature sensor in stable condition under temperature difference float value, voltage difference float value, and comprehensive analysis will test object be divided into qualified object and object to be detected, that is, before carrying out dynamic test, screening is carried out, then rely on dynamic test obtains temperature sensor in unstable condition under normal thermal response coefficient, normal linear coefficient, and dynamic performance coefficient is generated from the two, whether the performance detection of qualified object meets the requirement according to dynamic performance coefficient is judged and analyzed, in addition, in the process of testing, abnormal test results are checked and analyzed by static investigation module, dynamic analysis module, effectively improve test accuracy.

The invention discloses a test feedback method, and relates to the technical field of test feedback, and the method comprises the steps: determining a corresponding first test task and a first test server after a first voice instruction is received, sending a first test instruction carrying a test script to the first test server, and enabling the first test server to execute the first test task, monitoring the execution state of the first test task in real time, determining a first feedback strategy according to a target exception type when the target exception type is detected, and sending the corresponding feedback content to a specified first feedback device; according to the method, by establishing an automatic feedback mechanism of voice instruction triggering, state monitoring, anomaly classification, strategy matching and active feedback, the technical problem that in the prior art, due to the fact that a test result is manually inquired, anomaly information cannot reach a tester in time, and anomaly processing is delayed is solved; the technical effect of improving the abnormal response speed and the test feedback efficiency is achieved.

The application discloses an AI Agent-based storage chip mass production test method and device and a medium, relates to the technical field of storage chip testing, and comprises the following steps: based on mass production test double images, utilizing an AI Agent to comprehensively test coverage requirements, life consumption constraints and misjudgment risks, forming a risk constraint mass production test configuration, determining a main test channel for executing a current mass production test task from test channels, when a channel health score of the main test channel reaches a preset execution threshold and a historical test result fluctuation value exceeds a preset fluctuation threshold, selecting a backup test channel with the highest channel health score and a jig contact deviation value lower than a preset contact deviation threshold as a shadow review channel to form a main shadow channel binding test schedule; and the application forms the main shadow channel binding test schedule, calls the shadow review channel to review the same storage chip to be tested, and realizes the re-verification of abnormal test results in an independent review path.

The present application relates to the field of medical artificial intelligence and clinical auxiliary decision-making technology, in particular to an abnormal test result combination pattern recognition method and system based on deep learning, comprising: receiving test results output by a hospital test information system to construct a 48-dimensional test index vector; calculating a multi-dimensional joint deviation degree based on a group health joint distribution reference model; constructing a patient individual baseline with stable period test values for baseline drift correction, and inferring an individual baseline and giving a baseline disturbance double-channel representation when the stable period is missing with a meta-learning baseline inference subnetwork output; inputting the individualized joint deviation degree and the baseline disturbance double-channel representation into a multilayer perceptron network classifier to output five types of severe clinical event prediction probabilities of sepsis, acute kidney injury, disseminated intravascular coagulation, acute liver failure and acute exacerbation of chronic diseases, and the training loss contains a pathogenic causal diagram prior constraint term; when the prediction probability exceeds 40%, an orange reminder is pushed to the mobile terminal of the responsible nurse.

The application discloses a method and device for monitoring product testing, electronic equipment and a storage medium. The method comprises the following steps: collecting a test result picture and a test environment picture of a product through a camera; inputting the test result picture and the test environment picture into a trained neural network model to output a test condition corresponding to the product; and if the test condition is abnormal, uploading abnormal test result information to a user terminal according to the test condition. The application fully considers the environmental factors during testing, avoids the problem of test failure caused by external environmental factors, improves the reliability of test monitoring, uses a neural network model for monitoring, saves manpower and time for manual monitoring, and automatically uploads the abnormal condition to the user terminal, so that the test personnel or the research and development personnel can understand the test condition, analyze and process the abnormal condition in a timely manner, thereby improving the test efficiency and accelerating the product research and development process.

This invention discloses a sintering material sampling device using a sagger as a sintering carrier, relating to the field of sagger sampling technology. The device includes a cover plate with a fastening assembly on its lower surface. Three circular slots are provided on the cover plate, each containing a sampling component. Each sampling component includes an outer cylinder. By using a fixed cover plate relative to the sagger and with three circular slots on the cover plate, this invention allows for sampling sintering material at three different points. The fixed sampling location significantly improves the reliability of material compaction after firing, reducing the likelihood of abnormal test results due to different sampling points.

Embodiments of the present application provide a kind of intelligent teaching data processing method and related device, applied to intelligent teaching system, the intelligent teaching system includes first teaching equipment and second teaching equipment, the method includes: first teaching equipment receives the first application program code generated by target user;First teaching equipment sends first application program code to second teaching equipment;Second teaching equipment tests first application program code to obtain first test result;If first test result is abnormal test result, then second teaching equipment obtains the abnormal information in first test result;Second teaching equipment determines abnormal correction information according to abnormal information;Second teaching equipment sends abnormal correction information to first teaching equipment;First teaching equipment shows abnormal correction information, can when the test result of application program code is abnormal, corresponding abnormal correction information is determined and shows, improve the reliability when student carries out application program code programming.