MEMS inertial sensor reliability analysis method

Abstract

The invention discloses an MEMS inertial sensor reliability analysis method, which comprises the following steps: step 1, failure mode and mechanism analysis: carding environmental factors influencing the work of an MEMS inertial sensor, identifying failure modes caused by the environmental factors, and determining internal action mechanisms corresponding to the failure modes; 2, selecting a representative sample, and formulating a sample selection rule according to an application scene, a production batch, a structure type and performance parameter distribution of the sensor; the method has the beneficial effects that the system carding the association relationship between various environmental factors and failure modes, defining the internal mechanism of failure, breaking through the limitation that the existing method only pays attention to a single factor, enabling the analysis result to be more fit with the actual working scene of the sensor, and remarkably improving the comprehensiveness of reliability analysis; representative samples are selected based on scientific sample selection rules, key variables such as production batches, structure types and performance parameters are covered, and the problem of insufficient sample representativeness of an existing method is solved.

Description

Technical Field

[0001] This invention belongs to the field of MEMS inertial sensor technology, and specifically relates to a reliability analysis method for MEMS inertial sensors. Background Technology

[0002] MEMS inertial sensors are high-precision sensing devices manufactured based on microelectromechanical systems technology. They can accurately measure physical quantities such as acceleration, angular velocity, and angular acceleration of objects. With significant advantages such as small size, light weight, low power consumption, low cost, and high integration, they have become one of the core components in modern technology fields. In the aerospace field, MEMS inertial sensors are widely used in critical scenarios such as satellite attitude control, aircraft navigation and positioning, and rocket launch process monitoring. Their reliability is directly related to the success or failure of aerospace missions and personnel safety. In the consumer electronics field, this sensor is a core component for products such as smartphones, smartwatches, and virtual reality devices to achieve functions such as motion detection, image stabilization, and human-computer interaction. Its performance stability directly affects the user experience. In addition, MEMS inertial sensors also play an irreplaceable role in scenarios such as autonomous driving assistance systems in automotive electronics, equipment status monitoring in industrial control, and motion attitude perception in medical equipment.

[0003] With the continuous expansion of application scenarios, MEMS inertial sensors face increasingly complex and diverse operating environments, making their reliability issues more prominent. Extreme temperature cycling, strong vibration and shock, and high-altitude low-pressure environments in the aerospace field; frequent temperature changes, humidity corrosion, and electromagnetic interference in the consumer electronics field; and high temperature and humidity, mechanical vibration, and chemical corrosion in the automotive electronics field—all these environmental factors significantly impact the microstructure and performance of MEMS inertial sensors. In practical applications, MEMS inertial sensors often exhibit various failure phenomena: for example, temperature changes cause thermal expansion mismatch in the internal microstructure materials of the sensor, leading to adhesion failure between the microbeams and the substrate; long-term vibration causes stress accumulation in the microstructure, resulting in fatigue fracture; humidity and environmental chemicals work together to cause corrosion failure of the sensor's packaging materials or internal structure; the continuous influence of environmental factors can also cause the sensor's output accuracy to deviate from the standard value, resulting in parameter drift, and in severe cases, even complete sensor failure. These failure modes not only cause the sensor itself to lose its function but may also trigger the failure of the entire equipment system, leading to huge economic losses, and in critical fields such as aerospace, they may even cause catastrophic consequences.

[0004] Currently, various reliability analysis methods for MEMS inertial sensors have been developed in the industry, but they still have many significant limitations: First, existing analysis methods mostly focus on testing single environmental factors, such as high-temperature testing, vibration testing, or humidity testing alone, ignoring the coupling effect of multiple environmental factors such as temperature, vibration, and humidity in actual applications. This results in the analysis results failing to accurately reflect the actual working state of the sensor and deviating significantly from actual failure scenarios. Second, the selection of samples lacks scientific and systematic rules, often selecting only a small number of conventional samples, failing to cover key variables such as process fluctuations, structural differences, and batch differences during production. This leads to insufficient representativeness of the analysis results, making it difficult to apply them to sensors in mass production. Third, the testing conditions... The existing methods suffer from several shortcomings. First, they lack specificity and fail to adequately consider the specific application scenarios and failure mode characteristics of the sensors. This results in unreasonable test parameter ranges, excessively long or short test cycles, and an inability to effectively trigger failure modes, leading to low analysis efficiency and excessive costs. Second, the data analysis methods are relatively simplistic, often relying on basic statistical analysis or trend observation. This makes it difficult to accurately identify abnormal data characteristics, quantify the influence weight of environmental factors, and determine failure thresholds, resulting in insufficient scientific rigor and accuracy of the analysis results. Third, existing methods are largely limited to the identification of failure phenomena and the determination of reliability levels. They fail to establish the correlation between failure modes and sensor design, materials, and processes, and cannot provide targeted optimization suggestions to guide sensor performance improvement and design enhancement. Summary of the Invention

[0005] The purpose of this invention is to provide a reliability analysis method for MEMS inertial sensors, enabling systematic identification, clear mechanism understanding, accurate analysis, and targeted optimization of sensor failure modes.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a reliability analysis method for MEMS inertial sensors, comprising the following steps: Step 1: Failure Mode and Mechanism Analysis. This step involves identifying the environmental factors that affect the operation of MEMS inertial sensors, determining the failure modes caused by these environmental factors, and clarifying the underlying mechanisms of each failure mode. Step 2: Representative sample selection. Based on the sensor's application scenario, production batch, structural type, and performance parameter distribution, formulate sample selection rules and select a representative sample group that meets the coverage requirements. Step 3: Test condition setting. Simulate the actual use environment of the sensor. Based on the failure mode analysis results, set test conditions that include single environmental factors and coupling of multiple environmental factors, and clarify the test parameter range, duration and cycle rules. Step 4: Performance data monitoring. Build a data monitoring system to collect and store the core performance parameters of the sensor and the test environment parameters in real time during the testing process. Step 5: Performance data analysis. Multi-dimensional data analysis methods are used to process the monitoring data, identify abnormal data characteristics, fit performance change trends, and determine the failure threshold and the influence weight of environmental factors. Step Six: Output Analysis Results and Generate Optimization Suggestions. Based on the data analysis results, output a sensor reliability assessment report and propose targeted design, material, or process optimization suggestions for different failure modes.

[0007] As a preferred technical solution of the present invention, in step one, the environmental factors include one or more of temperature, vibration, humidity, air pressure, and electromagnetic interference; the failure modes include at least two of adhesion failure, fatigue failure, corrosion failure, parameter drift, and structural fracture; the intrinsic mechanism of action is determined by microscopic characterization technology combined with simulation analysis, specifically including the mechanisms of material thermal expansion mismatch, structural stress concentration, interfacial electrochemical reaction, and signal transmission interference.

[0008] As a preferred technical solution of the present invention, in step two, the sample selection rules include: the samples cover at least three consecutive production batches, covering more than two types of structural types, including basic, integrated, and high-precision types, and the performance parameter distribution covers more than 80% of the design tolerance range; the number of samples is determined according to the orthogonal experimental design principle, and the number of samples corresponding to each set of test conditions is not less than five.

[0009] As a preferred technical solution of the present invention, in step three, the parameter range of the single environmental factor test conditions is: temperature -40℃~125℃, vibration frequency 10Hz~2000Hz, humidity 10%RH~95%RH, air pressure 5kPa~101kPa, and electromagnetic interference intensity 1V / m~30V / m; the multi-environmental factor coupled test conditions include combinations of temperature-vibration, humidity-air pressure, and temperature-humidity-electromagnetic interference, with a single test duration of 1h~100h and a cycle count of 5~100 times.

[0010] As a preferred technical solution of the present invention, in step four, the core performance parameters include output accuracy, response time, stability, repeatability, and zero drift; the sampling frequency of the monitoring system is dynamically adjusted according to the testing stage. During the normal testing stage, the sampling frequency is 1 time / 5min to 1 time / 10min. When abnormal fluctuations occur in the data, the sampling frequency is automatically increased to 1 time / 1min to 1 time / 30s; the monitoring system synchronously records the real-time parameters of the testing environment, and the data storage accuracy is not less than 0.01 level.

[0011] As a preferred technical solution of the present invention, in step five, the multi-dimensional data analysis method includes trend fitting analysis, outlier detection, correlation analysis, and failure threshold calculation; wherein, the trend fitting analysis adopts a multinomial regression or exponential fitting algorithm, the outlier detection adopts the 3σ criterion or Grubbs criterion, the correlation analysis is used to determine the correlation strength between environmental factors and performance parameters, and the failure threshold is determined by combining the statistical confidence interval method with industry standards.

[0012] As a preferred technical solution of the present invention, step five further includes establishing a failure prediction model, which is trained by machine learning algorithm based on historical monitoring data and failure cases to predict the remaining service life of the sensor.

[0013] As a preferred technical solution of the present invention, it also includes full-process quality verification, which cross-verifies the rationality of sample selection, the effectiveness of test conditions, and the accuracy of data analysis; the rationality of sample selection is evaluated by statistical sampling inspection, the effectiveness of test conditions is verified by pre-testing, and the accuracy of data analysis is confirmed by comparing the results of at least two independent algorithms, so as to ensure the repeatability and credibility of the reliability analysis results.

[0014] Compared with the prior art, the beneficial effects of the present invention are: The system systematically analyzed the correlation between various environmental factors and failure modes, clarified the underlying failure mechanism, and broke through the limitation of existing methods that only focus on a single factor. This made the analysis results more consistent with the actual working scenarios of sensors and significantly improved the comprehensiveness of reliability analysis. Representative samples were selected based on scientific sample selection rules, covering key variables such as production batch, structural type, and performance parameters, thus solving the problem of insufficient sample representativeness in existing methods. Based on practical application scenarios and failure mode characteristics, single-factor, multi-factor coupling, and accelerated aging test conditions were set, which not only ensured the relevance and effectiveness of the test, but also took into account the analysis efficiency and reduced the test cost. By employing multi-dimensional data analysis methods, we have achieved data anomaly identification, trend fitting, failure threshold determination, and remaining life prediction, which has improved the depth and accuracy of data analysis and provided a scientific basis for reliability assessment. By implementing a full-process quality verification process and a data traceability system, the repeatability and credibility of the reliability analysis results are ensured, thereby enhancing the scientific rigor and authority of the methodology. Attached Figure Description

[0015] Figure 1 This is a flowchart of the MEMS inertial sensor reliability analysis method of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1 Please see Figure 1 This is the first embodiment of the present invention, which provides a reliability analysis method for MEMS inertial sensors, including the following steps: Step 1: Failure Mode and Mechanism Analysis. This step comprehensively reviews the environmental factors affecting the operation of MEMS inertial sensors, including single and combined factors such as temperature, vibration, humidity, air pressure, and electromagnetic interference. Through literature review, failure case statistics, and preliminary experiments, it identifies potential failure modes of the sensor under different environmental factors, covering at least two of the following: adhesion failure, fatigue failure, corrosion failure, parameter drift, and structural fracture. A correlation matrix of "environmental factors - structural response - failure phenomena" is constructed to clarify the correspondence between each environmental factor and the failure mode. The correlation matrix includes parameters such as environmental factor type, intensity, duration, failure mode type, and probability of occurrence. Microscopic characterization techniques such as scanning electron microscopy and atomic force microscopy are used to observe the microstructure of the failure site. Combined with simulation analysis methods such as finite element simulation and thermodynamic simulation, the intrinsic mechanisms of each failure mode are clarified. Examples include stress accumulation due to material thermal expansion mismatch caused by temperature, fatigue fracture caused by resonance effect from vibration, interfacial electrochemical corrosion caused by the combined effects of humidity and chemical substances, and signal transmission interference caused by electromagnetic interference. Step Two: Representative Sample Selection. Based on the sensor's application scenario, production batch, structural type, and performance parameter distribution, scientific sample selection rules are established. Samples must cover at least three consecutive production batches to ensure they reflect the stability of the production process; they must encompass at least two structural types to accommodate sensors with different structural designs; the performance parameter distribution must cover more than 80% of the design tolerance range, including upper limit, lower limit, and intermediate value samples; the number of samples is determined according to the orthogonal experimental design principle, with no fewer than five samples corresponding to each test condition to ensure the statistical significance of the test results; the sample groups selected through these rules can comprehensively reflect sensor production fluctuations and product differences, improving the representativeness and versatility of the analysis results. Step 3 involves setting test conditions to simulate the actual operating environment of the sensor. Based on the failure mode analysis results from Step 1, test conditions are set, including both single and coupled environmental factors. The parameter range for single environmental factor test conditions is set as follows: temperature -40℃, vibration frequency 10Hz, humidity 10%RH, air pressure 5kPa, and electromagnetic interference intensity 1V / m, covering the environmental parameter range for most application scenarios. Coupled environmental factor test conditions include common combinations such as temperature-vibration, humidity-air pressure, and temperature-humidity-electromagnetic interference, reflecting scenarios where multiple factors interact synergistically in practical applications. The duration and number of cycles for each test group are clearly defined, and the test time is appropriately extended for modes prone to slow failure (such as corrosion failure and long-term parameter drift). Simultaneously, accelerated aging test conditions are set, with the acceleration factor calculated based on the Arrhenius equation (for temperature factors) or the inverse power law model (for vibration, humidity, etc.). This shortens the test cycle and improves analysis efficiency while ensuring the equivalence of accelerated aging test results with natural aging results. Step four involves performance data monitoring, establishing a real-time data monitoring system comprised of a data acquisition module, an environmental monitoring module, a data transmission module, and a storage module. The data acquisition module collects core performance parameters of the sensor, including output accuracy, response time, stability, repeatability, and zero drift, ensuring a comprehensive reflection of the sensor's operating status. The environmental monitoring module synchronously records real-time environmental parameters such as temperature, vibration, and humidity during the testing process, providing environmental background information for subsequent data analysis. The sampling frequency of the monitoring system is dynamically adjusted according to the testing phase. During routine testing, the sampling frequency is set to once every 5 minutes. When abnormal fluctuations in performance parameters are detected, the sampling frequency automatically increases to once every 1 minute to ensure the capture of critical failure data. The data storage accuracy is no less than 0.01%, ensuring data accuracy and the reliability of subsequent analysis. Step five involves performance data analysis, employing multi-dimensional data analysis methods to comprehensively process the monitoring data. First, trend fitting analysis is used to obtain the curves showing the changes in sensor performance parameters with testing time and environmental factors. Then, the 3σ criterion or Grubbs criterion is used for outlier detection, eliminating invalid data caused by equipment errors, environmental interference, etc. Correlation analysis is used to quantify the correlation strength between various environmental factors and performance parameters, identifying key environmental factors affecting sensor reliability and their influence weights. Combining industry standards and statistical confidence interval methods, the failure thresholds corresponding to each failure mode are calculated. Furthermore, based on historical monitoring data and failure cases, a failure prediction model is trained using machine learning algorithms to predict the remaining service life of the sensor, with the prediction error controlled within 10%, providing a reference for sensor replacement and maintenance. Step Six: Output of Analysis Results and Generation of Optimization Suggestions. Based on the data analysis results from Step Five, a sensor reliability assessment report is output. The report includes sample information, test conditions, monitoring data, failure mode identification results, reliability level determination, and remaining life prediction. For different failure modes, and in conjunction with the failure mechanisms identified in Step One, targeted optimization suggestions are proposed: For adhesion failure, it is recommended to optimize the microstructure gap design, use hydrophobic coating materials, or improve the encapsulation process; for fatigue failure, it is recommended to optimize the structural fillet parameters, select higher-strength structural materials, or add vibration damping design; for corrosion failure, it is recommended to upgrade the encapsulation material, optimize the sealing process, or use an anti-corrosion coating; for parameter drift, it provides solutions for improving the circuit compensation algorithm, calibrating the temperature coefficient, or optimizing the structure of the sensor's sensitive element, thus achieving a closed loop from analysis results to design improvement. This embodiment also includes full-process quality verification, which cross-verifies the rationality of sample selection, the effectiveness of test conditions, and the accuracy of data analysis. The rationality of sample selection is evaluated through statistical sampling inspection, the effectiveness of test conditions is verified through pre-testing, and the accuracy of data analysis is confirmed by comparing the results of at least two independent algorithms. This ensures the repeatability and credibility of the reliability analysis results and establishes a full-process data traceability system to record all key information from sample selection, test condition setting, data collection to data analysis, which is convenient for subsequent query and verification.

[0018] Example 2 Please see Figure 1 This is the second embodiment of the present invention, which is based on the previous embodiment, but differs in that: Simulate the actual use environment of the sensor, and combine the failure mode analysis results from step one to set test conditions that include single environmental factors and coupling of multiple environmental factors. The parameter range of the single environmental factor test conditions is set as follows: temperature 40℃, vibration frequency 1000Hz, humidity 50%RH, air pressure 50kPa, electromagnetic interference intensity 15V / m, covering the environmental parameter range of most application scenarios. The sampling frequency of the monitoring system is dynamically adjusted according to the testing phase. During the normal testing phase, the sampling frequency is set to 1 time / 7min. When abnormal fluctuations in performance parameters are detected, the sampling frequency is automatically increased to 1 time / 45s to ensure the capture of critical failure data.

[0019] Example 3 Please see Figure 1 This is the third embodiment of the present invention, which is based on the previous embodiment, but differs in that: The actual use environment of the sensor is simulated. Based on the failure mode analysis results in step one, test conditions including single environmental factors and multiple environmental factors coupled are set. The parameter range of the single environmental factor test conditions is set as follows: temperature 125℃, vibration frequency 2000Hz, humidity 95%RH, air pressure 101kPa, electromagnetic interference intensity 30V / m, covering the environmental parameter range of most application scenarios. The sampling frequency of the monitoring system is dynamically adjusted according to the testing phase. During the normal testing phase, the sampling frequency is set to 1 time / 10min. When abnormal fluctuations in performance parameters are detected, the sampling frequency is automatically increased to 1 time / 30s to ensure the capture of key failure data.

[0020] Comparative experiments were designed to verify its effectiveness; the experimental data were obtained by simulating the actual test environment, demonstrating the significant advantages of the method of the present invention over traditional methods in reliability analysis.

[0021] Experimental Design and Methods Experimental subjects Test samples: 45 samples of a certain model of MEMS accelerometer were selected from 3 consecutive production batches; Group settings: Experimental group: Using the method of this invention (multi-factor coupling test + scientific sample selection) Control group: using traditional single-factor testing methods; Test conditions

[0022] Comparison table of zero drift change rate:

[0023] Failure prediction model performance metrics:

[0024] These experimental data fully demonstrate the significant advantages and practical value of the method of this invention in the reliability analysis of MEMS inertial sensors.

[0025] Although embodiments of the invention have been shown and described (see the detailed description above), it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A reliability analysis method for MEMS inertial sensors, characterized in that: Includes the following steps: Step 1: Failure Mode and Mechanism Analysis. This step involves identifying the environmental factors that affect the operation of MEMS inertial sensors, determining the failure modes caused by these environmental factors, and clarifying the underlying mechanisms of each failure mode. Step 2: Representative sample selection. Based on the sensor's application scenario, production batch, structural type, and performance parameter distribution, formulate sample selection rules and select a representative sample group that meets the coverage requirements. Step 3: Test condition setting. Simulate the actual use environment of the sensor. Based on the failure mode analysis results, set test conditions that include single environmental factors and coupling of multiple environmental factors, and clarify the test parameter range, duration and cycle rules. Step 4: Performance data monitoring. Build a data monitoring system to collect and store the core performance parameters of the sensor and the test environment parameters in real time during the testing process. Step 5: Performance data analysis. Multi-dimensional data analysis methods are used to process the monitoring data, identify abnormal data characteristics, fit performance change trends, and determine the failure threshold and the influence weight of environmental factors. Step Six: Output Analysis Results and Generate Optimization Suggestions. Based on the data analysis results, output a sensor reliability assessment report and propose targeted design, material, or process optimization suggestions for different failure modes.

2. The reliability analysis method for a MEMS inertial sensor according to claim 1, characterized in that: In step one, environmental factors include one or more of temperature, vibration, humidity, air pressure, and electromagnetic interference; failure modes include at least two of adhesion failure, fatigue failure, corrosion failure, parameter drift, and structural fracture; and intrinsic mechanisms are determined through microscopic characterization techniques combined with simulation analysis, specifically including the mechanisms of material thermal expansion mismatch, structural stress concentration, interfacial electrochemical reaction, and signal transmission interference.

3. The reliability analysis method for a MEMS inertial sensor according to claim 1, characterized in that: In step two, the sample selection rules include: the samples cover at least three consecutive production batches, covering more than two structural types, including basic, integrated, and high-precision types, and the performance parameter distribution covers more than 80% of the design tolerance range; the number of samples is determined according to the orthogonal experimental design principle, and the number of samples corresponding to each set of test conditions is not less than five.

4. The reliability analysis method for a MEMS inertial sensor according to claim 1, characterized in that: In step three, the parameter ranges for single environmental factor test conditions are: temperature -40℃~125℃, vibration frequency 10Hz~2000Hz, humidity 10%RH~95%RH, air pressure 5kPa~101kPa, and electromagnetic interference intensity 1V / m~30V / m; the multi-environmental factor coupled test conditions include combinations of temperature-vibration, humidity-air pressure, and temperature-humidity-electromagnetic interference, with a single test duration of 1h~100h and a cycle count of 5~100 times.

5. The reliability analysis method for a MEMS inertial sensor according to claim 1, characterized in that: In step four, the core performance parameters include output accuracy, response time, stability, repeatability, and zero drift. The sampling frequency of the monitoring system is dynamically adjusted according to the testing phase. During the normal testing phase, the sampling frequency is 1 time / 5min to 1 time / 10min. When abnormal fluctuations occur in the data, the sampling frequency is automatically increased to 1 time / 1min to 1 time / 30s. The monitoring system synchronously records the real-time parameters of the test environment, and the data storage accuracy is no less than 0.01 level.

6. The reliability analysis method for a MEMS inertial sensor according to claim 1, characterized in that: In step five, the multi-dimensional data analysis methods include trend fitting analysis, outlier detection, correlation analysis, and failure threshold calculation. Among them, trend fitting analysis adopts multinomial regression or exponential fitting algorithm, outlier detection adopts 3σ criterion or Grubbs criterion, correlation analysis is used to determine the correlation strength between environmental factors and performance parameters, and failure threshold is determined by statistical confidence interval method combined with industry standards.

7. The reliability analysis method for a MEMS inertial sensor according to claim 1, characterized in that: Step five also includes establishing a failure prediction model. Based on historical monitoring data and failure cases, the model is trained using machine learning algorithms to predict the remaining service life of the sensor.

8. The reliability analysis method for a MEMS inertial sensor according to claim 1, characterized in that: It also includes full-process quality verification, cross-validating the rationality of sample selection, the effectiveness of test conditions, and the accuracy of data analysis; the rationality of sample selection is evaluated through statistical sampling inspection, the effectiveness of test conditions is verified through pre-testing, and the accuracy of data analysis is confirmed by comparing the results of at least two independent algorithms, ensuring the repeatability and credibility of reliability analysis results.

Images