Multi-parameter oil state on-line monitoring system and method for engine test

Abstract

The present application relates to the technical field of engine test run lubricating oil monitoring, and discloses a multi-parameter oil state online monitoring system and method for engine test run.The system comprises four units of multi-parameter sensing, signal conditioning and acquisition, data health assessment and early warning display; five types of sensors are integrated to collect lubricating oil signals in situ, and the signals are sent to an embedded processor after processing; an OHI weighted model is constructed to perform hierarchical dynamic sampling, random forest is used to identify four types of faults, the threshold values of high-cold and highland working conditions can be adaptively corrected, and a three-stage sound-light remote early warning is matched.The present application is equipped with an anti-interference circuit, a shock-absorbing support and a compensation degassing cavity, is suitable for complex test run environment, has an accurate fault identification rate of 92%, has a recall rate of various faults of over 85%, solves the problems of offline monitoring lag, single-parameter blind area and false alarm in extreme environment, and realizes automatic online monitoring of aero-engine in all working conditions.

Description

Technical Field

[0001] This invention relates to the field of engine test lubricating oil monitoring technology, specifically to a multi-parameter online oil condition monitoring system and method for engine test. Background Technology

[0002] Engine testing is a crucial step in verifying engine performance, reliability, and durability. The condition of the lubricating oil system directly affects the lubrication of friction pairs, component wear, and test safety. Traditional oil monitoring relies primarily on offline sampling and laboratory analysis, which suffers from poor timeliness and inability to capture transient anomalies. Existing online monitoring systems mostly use single-parameter monitoring, lacking multi-parameter collaborative analysis capabilities, and generally employ fixed sampling frequencies, making it difficult to simultaneously meet the requirements of high-precision monitoring and low-power operation. In extreme environments such as high altitudes and frigid zones, oil viscosity, pressure, and gas solubility exhibit strong nonlinear changes, easily leading to monitoring threshold failure, sensor drift, and even false alarms.

[0003] Existing technologies lack a closed-loop control logic of "multi-parameter fusion → health index quantification → dynamic sampling → extreme environment adaptation," and specifically suffer from four core defects:

[0004] 1. Single-parameter monitoring cannot identify multi-factor coupled faults such as oil oxidation, water intrusion, and seal leakage, resulting in a large number of blind spots in the monitoring.

[0005] 2. The fixed sampling frequency design causes excessive power consumption due to continuous high-frequency acquisition under normal steady-state conditions, while fixed low-frequency sampling under fault transient conditions cannot capture sudden data changes, and the monitoring accuracy and power consumption cannot be balanced.

[0006] 3. The sensor hardware has not been optimized and the monitoring threshold has not been adaptively corrected for extreme environments such as low temperature and low air pressure, which makes it very easy for sensor drift, threshold failure and a large number of false alarms to occur.

[0007] 4. The lack of a standardized and quantitative oil health assessment system and a hierarchical early warning mechanism means that only parameter values ​​can be displayed individually, making it impossible to intuitively quantify the overall health of the oil. Fault warning response is delayed, making it difficult to ensure the safety of operation throughout the entire test run.

[0008] Existing technologies cannot simultaneously solve all the problems of insufficient real-time performance of oil monitoring, single monitoring dimensions, conflict between power consumption and accuracy, poor adaptability to extreme environments, and lack of standardized assessment and early warning. Therefore, there is an urgent need for an integrated online monitoring solution that integrates multi-parameter synchronous acquisition, quantitative assessment of health index, dynamic adaptive sampling, self-correction of thresholds in extreme environments, and intelligent fault classification and early warning. Summary of the Invention

[0009] The purpose of this invention is to provide a multi-parameter online oil condition monitoring system and method for engine testing, which can solve the above-mentioned problems. Based on the closed-loop technical solution of multi-parameter synchronous acquisition, oil health index quantification algorithm, dynamic graded sampling, adaptive threshold correction for operating conditions and environment, and random forest intelligent fault diagnosis, it completely eliminates all the defects of the prior art and realizes accurate, low-power, adaptive online monitoring and graded fault early warning of oil condition under all operating conditions and environments of engine testing.

[0010] To achieve the above objectives, the present invention provides the following technical solution:

[0011] This invention proposes a multi-parameter online oil condition monitoring system for engine testing, comprising: a multi-parameter sensing module, a signal conditioning and acquisition unit, a data processing and health assessment unit, and an early warning and display unit. The multi-parameter sensing module integrates a temperature sensor, a pressure sensor, a kinematic viscosity sensor, a humidity sensor, and a gas composition sensor, and is installed in the engine's main oil circuit or return oil circuit to achieve in-situ, minimally interference-free data acquisition. Each sensor transmits the raw oil parameter signals to the signal conditioning and acquisition unit. The signal conditioning and acquisition unit includes an amplification and filtering circuit, a 24-bit ADC, and a CAN / LTE communication module. The amplification and filtering circuit receives the raw signals transmitted from the multi-parameter sensing module and performs amplification and anti-aliasing filtering. The processed signal undergoes analog-to-digital conversion via a 24-bit ADC and is then transmitted to the data processing and health assessment unit via a CAN / LTE communication module. The data processing and health assessment unit, employing an embedded processor, receives the digital signal from the signal conditioning and acquisition unit, performs data preprocessing, calculates the Oil Health Index (OHI), implements dynamic sampling control, and establishes a fault diagnosis model. It outputs the health index and fault type and feeds the results back to the multi-parameter sensing module for dynamic sampling adjustment. Simultaneously, it transmits warning-related information to the warning and display unit. The warning and display unit includes a local touchscreen, an audible and visual alarm, and a remote terminal. It receives warning information from the data processing and health assessment unit, enabling three-level warnings and status visualization.

[0012] Preferably, the temperature sensor is a PT100 with a temperature range of -40℃ to 200℃, and is wall-mounted; the pressure sensor is a piezoresistive type with a range of 0~5MPa, connected in series in the main oil circuit; the kinematic viscosity sensor is a vibrating tube type with a range of 0.8~500cSt, and is installed by immersion in the return oil pipe; the humidity sensor is a polymer capacitive type with a range of 0~5000ppm, which draws oil through a capillary tube; and the gas composition sensor is an infrared and electrochemical combination type that detects CO2, H2, and CH4, and is equipped with a 60℃ degassing chamber.

[0013] Preferably, the amplification and filtering circuit is an electromagnetic interference (EMI) filtering module used to suppress high-frequency noise in the engine ignition system; the 24-bit ADC is a Σ-Δ analog-to-digital converter chip to achieve synchronous sampling of multiple parameter signals; the CAN / LTE communication module supports CAN2.0B bus and 4G / LTE wireless transmission, and the data packet includes timestamp, sensor ID, original value and CRC-16 checksum.

[0014] Preferably, the embedded processor is a TI AM62x series with a main frequency of 1GHz and 2GB of memory. The fault diagnosis model it runs is a random forest model, which takes five-parameter deviation and OHI as input and outputs moisture intrusion, oxidation degradation, metal wear, seal leakage and corresponding confidence levels.

[0015] Preferably, the local touchscreen is used to display real-time parameter tables, OHI trend curves, fault diagnosis reports, and historical comparison data; the audible and visual alarm uses a buzzer and LED flashing alarm method, supporting differentiated alarms for different warning levels; the remote terminal supports warning information push, historical data query, and remote parameter configuration.

[0016] This invention also proposes a method for online monitoring of multi-parameter oil condition for engine testing, based on the aforementioned online monitoring system for multi-parameter oil condition for engine testing, comprising the following steps:

[0017] S1. Real-time acquisition of multiple parameters: The original signals of oil temperature, pressure, kinematic viscosity, humidity and gas composition are acquired synchronously through the multi-parameter sensing module.

[0018] S2, Signal Conditioning and Calibration Conversion: The original signal is amplified, filtered, and converted from analog to digital by the signal conditioning and acquisition unit, and then converted into a physical quantity in engineering units by the calibration curve;

[0019] S3. Oil Health Index (OHI) Calculation: The OHI is calculated using a weighted algorithm through data processing and health assessment units. The formula is as follows:

[0020] OHI=100- ),

[0021] in, This is the real-time value of the i-th parameter. The normal baseline value for the i-th parameter. Let i be the allowable fluctuation range for the i-th parameter. The weight of the i-th parameter;

[0022] S4. Dynamic Sampling Control: The sampling frequency is adjusted in stages according to the real-time OHI value. The sampling frequency is updated in real time every 5 seconds based on the OHI. When OHI≥80, a low-frequency sampling of 1Hz is used to collect only temperature and pressure. When 60≤OHI<80, a medium-frequency sampling of 10Hz is used to collect all parameters. When OHI<60, a high-frequency sampling of 100Hz is used to collect all parameters transiently.

[0023] S5. Adaptive threshold correction: Automatically corrects the baseline values ​​and allowable fluctuation ranges of each parameter based on engine test conditions and extreme environments such as high altitude and cold.

[0024] S6. Fault Diagnosis and Graded Early Warning: The fault type and confidence level are diagnosed by random forest model, and a three-level early warning is triggered based on the OHI value and pushed to the early warning and display unit.

[0025] Preferably, in step S3, the weights of each parameter are as follows: temperature 20%, kinematic viscosity 30%, gas composition 25%, humidity 15%, and pressure 10%; the sampling control OHI classification threshold is OHI≥80 for normal state, 60≤OHI<80 for warning state, and OHI<60 for fault state.

[0026] Preferably, in step S5, in a cold environment of ≤-30℃, the sensor heating module is activated, the upper limit of temperature is relaxed to 120℃, and the kinematic viscosity weight is reduced from 30% to 20%; in a high-altitude environment of ≥3000m, atmospheric pressure compensation is performed on the pressure sensor, the gas threshold is relaxed, and the humidity threshold is tightened; under test conditions, the corresponding threshold table is automatically called during cold start, rated speed, and maximum thrust stages.

[0027] Preferably, in step S6, the OHI grading thresholds of the three-level early warning mechanism are as follows: when OHI is between 70 and 80, it is a mild anomaly, and the local touchscreen displays a yellow warning and records the parameter trend; when OHI is between 50 and 70, it is a moderate risk, and the audible and visual alarm is activated and a warning message is pushed to the remote terminal; when OHI < 50, it is a serious fault, and the audible and visual alarm sounds urgently and pushes fault details and shutdown suggestions.

[0028] Preferably, in step S2, the original signal is amplified, filtered, and converted to digital by the signal conditioning and acquisition unit. Specifically, the original temperature signal output by the temperature sensor is sent to the amplification and filtering circuit, converted to a voltage signal by the built-in Wheatstone bridge using a three-wire connection, and then amplified by the built-in instrumentation amplifier before being sent to the 24-bit ADC to complete the analog-to-digital conversion. The original pressure signal output by the pressure sensor is sent to the amplification and filtering circuit, converted to a voltage signal by the built-in 250Ω resistor, and then subjected to low-pass filtering before being sent to the 24-bit ADC to complete the analog-to-digital conversion. The original signals output by the kinematic viscosity sensor, humidity sensor, and gas composition sensor are sent to the amplification and filtering circuit, isolated and anti-aliasing filtered by the built-in voltage follower, and then sent to the 24-bit ADC to complete the analog-to-digital conversion.

[0029] The beneficial effects of this invention are as follows:

[0030] (1) This invention integrates five dimensions of parameters—temperature, pressure, kinematic viscosity, humidity, and gas—for synchronous in-situ acquisition, covering all types of faults, including oil lubrication attenuation, oxidation, moisture contamination, seal leakage, and metal wear. Based on a unified OHI quantification model, the multi-dimensional data is standardized to a score of 0 to 100. The parameter weights and baseline thresholds can be adaptively adjusted. Combined with automatic correction logic for high-altitude and cold environments, it solves the defect of fault blind spots in single-parameter monitoring, eliminates false alarms caused by sensor drift and threshold failure in extreme environments, and greatly improves the reliability of full-condition monitoring, making it accurate and comprehensive.

[0031] (2) This invention incorporates OHI-linked dynamic sampling closed-loop logic. Under normal operating conditions, it collects only core temperature and pressure parameters at low frequency to reduce power consumption. Under warning / fault conditions, the sampling frequency is gradually increased and full parameter acquisition is enabled, balancing monitoring accuracy and system energy consumption. This solves the problem that fixed sampling frequency cannot balance accuracy and power consumption, resulting in efficient resource utilization. The entire system uses industrially standardized and mature sensor hardware, resulting in low cost per unit. The modular hardware design allows for rapid expansion to other industrial oil monitoring scenarios such as gas turbines and marine power systems, making it highly adaptable for engineering implementation.

[0032] (3) This invention integrates an embedded processor to realize real-time quantitative evaluation of OHI and is equipped with a three-level standardized early warning mechanism. It can push early warning prompts, audible and visual alarms and shutdown suggestions according to the different levels of oil health, solve the defects of lack of standardized health assessment and delayed early warning, and make fault diagnosis and early warning intelligent. It has a built-in random forest fault diagnosis model, which judges the fault type based on the five-parameter deviation and OHI. The average accuracy of fault identification is 92%, and the recall rate of various faults is higher than 85%. At the same time, it supports online incremental learning, which can continuously identify new test faults, capture transient anomalies in advance, and ensure the safety of engine test operation throughout the entire process.

[0033] (4) The system architecture of this invention has a built-in dual adaptive correction link for environment and working condition, which can automatically identify the switching threshold parameters of cold start, rated speed, maximum thrust and other test stages; automatically start sensor heating and adjust viscosity weight in high cold and low temperature environment; automatically compensate pressure reference and correct gas / humidity threshold in high plateau and low air pressure environment, fully covering the full test environment of normal temperature, high cold and high plateau, without the need for manual on-site adjustment of monitoring parameters, solving the defects of poor adaptability to extreme environment and lack of adaptive logic for working condition, and enabling full-scenario adaptive operation. Attached Figure Description

[0034] Figure 1 Hardware architecture diagram of the system of this invention;

[0035] Figure 2 This is a schematic diagram of the adaptive threshold correction for test conditions according to the present invention;

[0036] Figure 3 This is a schematic diagram of the adaptive threshold correction for high-altitude / cold environments according to the present invention;

[0037] Figure 4 This is a flowchart of the oil health index calculation method of the present invention;

[0038] Figure 5 This is the dynamic sampling frequency control logic diagram of the present invention. Detailed Implementation

[0039] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The scope of protection of the present invention is not limited to the following embodiments. Equivalent modifications and substitutions made by those skilled in the art based on the inventive concept all fall within the scope of protection defined by the claims of the present invention.

[0040] Example 1:

[0041] Figure 1 For system hardware architecture diagrams, such as Figure 1 As shown in the figure, this embodiment discloses a multi-parameter oil condition online monitoring system for engine testing, which includes four hierarchical modular hardware units: a multi-parameter sensing module, a signal conditioning and acquisition unit, a data processing and health assessment unit, and an early warning and display unit.

[0042] The multi-parameter sensing module integrates a temperature sensor, a pressure sensor, a kinematic viscosity sensor, a humidity sensor, and a gas composition sensor. The entire module is fixed to the engine's main oil circuit or return oil line using an in-situ, minimally disruptive installation method. The temperature sensor collects the original analog signal of lubricating oil temperature, the pressure sensor collects the original analog signal of lubrication line oil pressure, the kinematic viscosity sensor collects the original analog signal of lubricating oil viscosity, the humidity sensor collects the original analog signal of lubricating oil dissolved water, and the gas composition sensor collects the original analog signal of lubricating oil decomposition gases. The shielded wiring harnesses of the five types of sensors are then connected to the signal input terminal of the signal conditioning and acquisition unit.

[0043] The signal conditioning and acquisition unit, acting as the signal conversion and transmission layer, integrates three processing sub-modules: an amplification and filtering circuit, a 24-bit ADC, and a CAN / LTE communication module. The front end receives five weak raw analog signals from the temperature sensor, pressure sensor, kinematic viscosity sensor, humidity sensor, and gas composition sensor in the multi-parameter sensing module. The amplification and filtering circuit simultaneously amplifies, anti-aliasing, and anti-interference filtering of the five raw signals. The five processed analog signals are input in parallel to the 24-bit ADC, which simultaneously performs analog-to-digital conversion, transforming the five analog physical signals into digital signals. The digitized five parameter data are packaged together, with timestamps, sensor IDs, and checksums added, and then transmitted synchronously to the data processing and health assessment unit via the CAN / LTE communication module through an industrial CAN bus or a 4G wireless link.

[0044] The data processing and health assessment unit serves as the core computing and control layer. Its core hardware is an embedded processor, which continuously receives five digital parameters—temperature, pressure, kinematic viscosity, humidity, and gas—from the signal conditioning and acquisition unit. Internally, it has a complete set of monitoring and control programs that sequentially execute four core operations: raw data preprocessing, weighted calculation of the Oil Health Index (OHI), dynamic sampling frequency closed-loop control, and random forest fault diagnosis. The generated OHI values, fault determination results, and graded early warning commands form a two-way control link: one control command is fed back to the multi-parameter sensing module, controlling the sampling start / stop and acquisition frequency of the temperature, pressure, kinematic viscosity, humidity, and gas composition sensors to achieve dynamic acquisition adjustment; the other early warning information is transmitted forward to the early warning and display unit, outputting visualized human-machine interactive information.

[0045] The early warning and display unit, serving as the human-machine interface output layer, integrates three types of output devices: a local touchscreen, an audible and visual alarm, and a remote terminal. It receives five channels of oil parameters, OHI values, fault diagnosis reports, and graded early warning commands from the data processing and health assessment unit. The local touchscreen displays real-time parameters (temperature, pressure, kinematic viscosity, humidity, and gas), continuous OHI curves, and historical fault records. The audible and visual alarm outputs differentiated light and buzzer alarms based on different OHI warning levels. The remote terminal synchronously receives all monitoring data via a wireless network, enabling remote engineers to view the five parameters, configure sensor acquisition rules, and receive fault warning push notifications.

[0046] The entire system consists of four layers forming a complete signal link: temperature sensor, pressure sensor, kinematic viscosity sensor, humidity sensor, and gas composition sensor simultaneously collect five raw oil signals, which are then converted by the signal conditioning and acquisition unit, processed by the data processing and health assessment unit, and finally output by the early warning and display unit. At the same time, the core computing unit controls the acquisition strategy of the five types of front-end sensors in reverse, forming a complete closed-loop monitoring link of "five-channel synchronous acquisition - signal conversion - intelligent computing - dynamic adjustment - graded early warning". No manual offline sampling is required, and the entire process is automated online to complete the full-dimensional status monitoring of engine lubricating oil.

[0047] Example 2:

[0048] Figure 1 This is a system hardware architecture diagram. Figure 2 This is a schematic diagram of adaptive threshold correction under test conditions. Figure 3 This is a schematic diagram of adaptive threshold correction for high-altitude / cold environments, as shown below. Figure 1 , Figure 2 , Figure 3 As shown in the figure, this embodiment discloses a multi-parameter oil condition online monitoring system for engine testing, comprising four hierarchical modular hardware units: a multi-parameter sensing module, a signal conditioning and acquisition unit, a data processing and health assessment unit, and an early warning and display unit. The specific structure is as follows:

[0049] 1. Multi-parameter sensing module.

[0050] The multi-parameter sensing module integrates a temperature sensor, a pressure sensor, a kinematic viscosity sensor, a humidity sensor, and a gas composition sensor. The entire module is fixed to the engine's main oil circuit or return oil line using an in-situ, minimally disruptive installation method.

[0051] The temperature sensor is a Class A precision PT100 armored platinum resistance thermometer with a temperature range of -40℃ to 200℃ and an accuracy of ±0.5℃. It is installed on the outer wall of the engine oil return pipe (contacting the oil through thermal grease, with a response time of <3s). The thermal grease fills the gap between the sensor and the pipe wall to improve heat conduction efficiency. The sensor housing is equipped with a ceramic heat insulation layer and a low-temperature heating resistance wire for thawing in cold-weather test environments. The sensor uses a three-wire resistor output signal, and the shielded cable is connected to the input terminal of the amplification and filtering circuit.

[0052] The pressure sensor is a piezoresistive pressure transmitter with a range of 0~5MPa and an accuracy of ±0.1%FS. It is installed in series in the lubrication branch pipeline of the engine's main oil circuit, near the outlet of the lubrication bearing. It is used to monitor the pressure of the main oil circuit / return oil circuit, determine the oil supply capacity of the lubricating oil pump and the risk of pipeline blockage, and output a 4~20mA standard current signal, which is connected to the amplification and filtering circuit to convert the voltage.

[0053] The kinematic viscosity sensor is a vibrating tube kinematic viscosity sensor based on the Coriolis principle, with a range of 0.8~500 cSt, an accuracy of ±2%, and a response time of <10s. It is immersed and fixed in the straight and stable flow channel section of the return oil pipe to measure the kinematic viscosity of the oil in real time (reflecting the oil film carrying capacity). It outputs a 0~10V viscosity voltage signal in real time, which is directly connected to the amplification and filtering circuit for isolation processing.

[0054] The humidity sensor is a polymer capacitive water content sensor with a range of 0~5000ppm and an accuracy of ±50ppm. It draws a small amount of oil from the return oil pipe through a capillary tube (to avoid direct immersion corrosion) to detect the dissolved water content in the oil (excessive water content will emulsify the oil and accelerate corrosion), and outputs a 0~10V voltage signal.

[0055] The gas composition sensor employs a miniature infrared sensor (detecting CO2, range 0~1000ppm, accuracy ±1ppm) + an electrochemical sensor (detecting H2 and CH4, range 0~100ppm, accuracy ±0.5ppm), integrated into a sealed gas sampling chamber (dissolved gas is extracted through an oil degassing module at a degassing temperature of 60℃). It is used to determine the oxidation and corrosion of the oil (e.g., the degree of oxidation due to CO2 reaction, and H2 indicating the reaction between metal and water), and outputs a gas concentration voltage signal.

[0056] The gas sampling chamber of the gas composition sensor is equipped with a two-stage desiccant, with silica gel and molecular sieve as the drying media, which can completely eliminate the interference of condensation in the chamber on gas concentration detection. The sampling chamber also integrates an independent temperature compensation module, which can automatically correct the drift problem of gas sensor output signal in high and low temperature environments.

[0057] All sensors in the multi-parameter sensing module are equipped with a dedicated sensor bracket. The bracket is made of titanium alloy, which has better high and low temperature resistance than aluminum alloy. The bottom of the bracket is equipped with shock-absorbing rubber pads with a natural frequency of less than 50Hz, which can completely avoid the main vibration frequency of the engine test. The surface of the bracket is coated with an antifreeze and lubricating coating to prevent mechanical jamming of the moving parts of the bracket under cold conditions.

[0058] The independent signal output terminals of the five types of sensors are combined into a shielded cable bundle, which is then connected to the multi-channel signal input terminal of the back-end signal conditioning and acquisition unit to achieve synchronous transmission of the five parameters.

[0059] 2. Signal conditioning and acquisition unit.

[0060] The signal conditioning and acquisition unit, acting as the signal conversion and transmission layer, integrates three processing sub-modules: an amplification and filtering circuit, a 24-bit ADC, and a CAN / LTE communication module. The front end receives five weak raw analog signals from the temperature sensor, pressure sensor, kinematic viscosity sensor, humidity sensor, and gas composition sensor in the multi-parameter sensing module. The amplification and filtering circuit simultaneously amplifies, anti-aliasing, and anti-interference filtering of the five raw signals. The five processed analog signals are input in parallel to the 24-bit ADC, which simultaneously performs analog-to-digital conversion, transforming the five analog physical signals into digital signals. The digitized five parameter data are packaged together, with timestamps, sensor IDs, and checksums added, and then transmitted synchronously to the data processing and health assessment unit via the CAN / LTE communication module through an industrial CAN bus or a 4G wireless link.

[0061] The amplification and filtering circuit is an integrated EMI electromagnetic interference suppression filter module, which can suppress high-frequency noise in the engine ignition system, ensure the stability of CAN bus and LTE wireless transmission signals, and adapt to the complex electromagnetic environment of the test bench.

[0062] The circuit processes five sensor signals through independent signal channels: the PT100 temperature signal uses a three-wire connection and is converted into a voltage signal (0~50mV, corresponding to -40℃~200℃) using a Wheatstone bridge resistor → amplified to 0~3.3V by an AD623 instrumentation amplifier → 24-bit Σ-Δ ADC (ADS1248) analog-to-digital conversion. The pressure current signal (4~20mA) is converted to a 1~5V voltage signal via a 250Ω resistor → low-pass filter (cutoff frequency 10Hz) → ADC. Kinematic viscosity, humidity, and gas voltage signals (0~10V or 0~5V) are isolated by a voltage follower → anti-aliasing filter (cutoff frequency 50Hz) → ADC.

[0063] The 24-bit ADC uses a Σ-Δ type multi-channel synchronous analog-to-digital converter chip, supports synchronous sampling of five sensors, has a maximum hardware sampling frequency of 100Hz, and converts the received and filtered analog signals into digital quantities.

[0064] The CAN / LTE communication module is compatible with the CAN2.0B bus, with a transmission rate of 500kbps, and also integrates a 4G / LTE wireless module. Each transmitted data packet includes a timestamp, sensor ID, original value, and CRC-16 checksum to avoid data transmission distortion and packet loss.

[0065] 3. Data processing and health assessment unit.

[0066] The core component of the data processing and health assessment unit is an embedded processor. The embedded processor receives digital oil parameters and internally runs four types of programs: data preprocessing, oil health index (OHI) calculation, dynamic sampling control, and fault diagnosis. It generates bidirectional instructions: control instructions are transmitted in reverse to the multi-parameter sensing module to adjust the sampling start / stop and acquisition frequency of five types of sensors; and early warning information is sent in forward to the early warning and display unit.

[0067] The embedded processor uses a TI AM62x industrial-grade processor with a main frequency of 1GHz and 2GB of RAM. It is equipped with a large-capacity local storage medium to store the monitoring data of the entire test cycle. The processor's built-in software algorithm modules include: a moving average filtering preprocessing module (10-point window smoothing), a 3σ outlier removal module, an OHI weighted calculation module, a dynamic sampling frequency control module, and a random forest fault diagnosis model. The random forest model training dataset contains more than 1,000 sets of aero-engine test data, covering cold start, rated, and maximum thrust full operating conditions. It can output four types of faults: moisture intrusion, oxidation degradation, metal wear, and seal leakage, as well as their corresponding confidence levels. The model supports online incremental learning and updating of decision tree weights every 24 hours.

[0068] 4. Early warning and display unit.

[0069] The early warning and display unit, as the human-machine interaction output layer, integrates three types of output devices: local touch screen, sound and light alarm, and remote terminal. It receives five channels of oil parameters, OHI values, fault diagnosis reports, and graded early warning instructions from the data processing and health assessment unit.

[0070] The local touchscreen displays five-channel fluid parameter tables, OHI trend curves, fault diagnosis reports, and historical comparison data of the same engine model in real time; the audible and visual alarm is equipped with a three-color LED and a segmented buzzer, which outputs differentiated light and buzzer prompts for different warning levels; the remote terminal supports warning information push, five-year historical data query, and remote sensor parameter configuration.

[0071] The three-level early warning mechanism is as follows:

[0072] Mild abnormality (OHI 70~80): The local touch screen HMI displays a yellow warning and records parameter trends (such as temperature approaching the upper limit).

[0073] Medium risk (OHI50~70): Local audible and visual alarm (buzzer + LED flashing) + remote terminal push warning (e.g., "Kinematic viscosity decreases, oxidation should be monitored");

[0074] Critical fault (OHI<50): Local urgent alarm (continuous beeping + red indicator light) + remote push fault details (such as "moisture content is seriously excessive, stop the machine immediately"), and generate a test run interruption suggestion.

[0075] Figure 2 This is a schematic diagram of adaptive threshold correction under test conditions, as shown below. Figure 2 As shown, the system identifies six test phases—cold start, idle, rated speed, maximum thrust, deceleration, and stop—by engine speed and torque signals, and automatically retrieves the corresponding baselines and allowable fluctuation thresholds. During the cold start phase, the temperature baseline is relaxed, while during the maximum thrust condition, the range of humidity and gas concentration fluctuations is tightened.

[0076] Figure 3 This is a schematic diagram of adaptive threshold correction for high-altitude / cold environments, as shown below. Figure 3 As shown, in cold-weather conditions with an ambient temperature ≤ -30℃, the temperature sensor heating module is automatically activated, the upper limit of the allowable temperature is increased to 120℃, and the kinematic viscosity weight is reduced from 30% to 20%. In high-altitude conditions with an altitude ≥ 3000m, the pressure sensor measurement value is compensated for atmospheric pressure, the CO2 gas threshold is relaxed, the dissolved water humidity threshold is tightened, and the gas composition calculation weight is reduced accordingly to adapt to the characteristic that the oxidation rate of sliding oil slows down in low-pressure conditions at high altitudes.

[0077] Example 3:

[0078] Figure 4 Flowchart for calculating oil health index Figure 5 The dynamic sampling frequency control logic diagram is as follows: Figure 4 , Figure 5 As shown, the present invention proposes a method for online monitoring of multi-parameter oil condition for engine testing, which relies on the complete set of hardware equipment of the online monitoring system for multi-parameter oil condition for engine testing disclosed in Example 2. The specific steps are as follows:

[0079] S1. Real-time acquisition of multiple parameters:

[0080] After the engine test equipment is powered on, the temperature sensor, pressure sensor, kinematic viscosity sensor, humidity sensor, and gas composition sensor inside the multi-parameter sensing module simultaneously enter the standby acquisition state. During the cold start phase, the default 1Hz low-frequency acquisition mode is used, with only the temperature sensor and pressure sensor working continuously, while the kinematic viscosity sensor, humidity sensor, and gas composition sensor go into sleep mode to reduce the overall power consumption. The five raw analog signals are transmitted to the signal conditioning and acquisition unit in real time.

[0081] S2, Signal Conditioning and Calibration Conversion:

[0082] The original signal is amplified, filtered, and converted from analog to digital by the signal conditioning and acquisition unit, and then converted into a physical quantity in engineering units by the calibration curve.

[0083] Specifically, the signal conditioning and acquisition unit amplifies, filters, and performs analog-to-digital conversion on the original signal, which involves:

[0084] The raw temperature signal output by the temperature sensor is sent to the amplification and filtering circuit. It is converted into a voltage signal by the built-in Wheatstone bridge using a three-wire connection. After being amplified by the built-in instrumentation amplifier, it is sent to the 24-bit ADC to complete the analog-to-digital conversion.

[0085] The raw pressure signal output by the pressure sensor is sent to the amplification and filtering circuit. After being converted into a voltage signal by the built-in 250Ω resistor, it is subjected to low-pass filtering. The instantaneous noise of oil pressure is eliminated by the 10Hz low-pass filtering, and then sent to the 24-bit ADC to complete the analog-to-digital conversion.

[0086] The raw signals output from the kinematic viscosity sensor, humidity sensor, and gas composition sensor are sent to the amplification and filtering circuit. After isolation and anti-aliasing filtering by the built-in voltage follower, all standardized voltage signals are sent in parallel to the 24-bit ADC to complete the analog-to-digital conversion.

[0087] The embedded processor retrieves the sensor's factory calibration curve, converts the digital quantity into engineering physical quantities such as ℃, MPa, cSt, and ppm, and simultaneously reads the current test phase, ambient temperature, and altitude data for later use.

[0088] S3, Oil Health Index (OHI) Calculation:

[0089] OHI is calculated using a weighted algorithm through the data processing and health assessment unit, with the following formula:

[0090] OHI=100- ),

[0091] in, This represents the real-time value of the i-th parameter (e.g., temperature, pressure, kinematic viscosity, humidity, gas concentration). The normal baseline value for the i-th parameter (set based on the test outline and historical data, such as temperature baseline 80℃, kinematic viscosity baseline 50cSt). The allowable fluctuation range for the i-th parameter (e.g., temperature ±20℃, kinematic viscosity ±20cSt); The weights for the i-th parameter are as follows (the specific weights for each parameter are: temperature 20%, kinematic viscosity 30%, gas composition 25%, humidity 15%, and pressure 10%).

[0092] Example as follows: If the monitored values ​​at a certain moment are: temperature 95℃ (deviation 0.75), kinematic viscosity 40cSt (deviation 0.5), humidity 800ppm (deviation 1), CO2 150ppm (deviation 1.5), and pressure normal (deviation 0), then:

[0093] OHI = 100 - (0.2 × 0.75 + 0.3 × 0.5 + 0.15 × 1 + 0.25 × 1 + 0 × 0) = 99.3. If humidity and gas levels are severely exceeded (e.g., humidity 2000 ppm, CO 2500 ppm), OHI may drop below 60, triggering a fault warning.

[0094] Note: Baseline values ​​and allowable fluctuation ranges can be dynamically adjusted according to operating conditions (such as cold start, high load) to adapt to different monitoring needs.

[0095] S4. Dynamic sampling control:

[0096] The sampling frequency is adjusted according to the real-time OHI value (balancing accuracy and power consumption). Figure 5 The dynamic sampling frequency control logic is executed, and the sampling frequency is updated in real time every 5 seconds according to OHI.

[0097] The OHI classification thresholds are as follows: OHI ≥ 80 indicates a normal state, using 1Hz low-frequency sampling and collecting only temperature and pressure, while pausing other parameters; when 60 ≤ OHI < 80, it is a warning state, and kinematic viscosity, humidity, and gas composition are increased, using 10Hz medium-frequency sampling and collecting all parameters; when OHI < 60, it is a fault state, using 100Hz high-frequency sampling and collecting all parameters transiently.

[0098] S5, Adaptive Threshold Correction:

[0099] Based on engine test conditions and extreme environments such as high altitude and cold regions, the baseline values ​​and allowable fluctuation ranges of each parameter are automatically corrected.

[0100] (1) Adaptive operating condition:

[0101] Identify the cold start, idle, rated speed, maximum thrust, deceleration, and stopping stages, and retrieve [data / resources]. Figure 2 Corresponding threshold parameter table.

[0102] Cold start phase: Relax the temperature threshold (e.g., allow short-term exceedance of 100℃ (normal) → 120℃ (high cold environment)), reduce the weight of kinematic viscosity monitoring (avoid false alarms due to the naturally high kinematic viscosity of cold oil), and only maintain low-frequency sampling of temperature and pressure at 1Hz;

[0103] Acceleration / deceleration transients: Dynamically increase the sampling frequency to 200Hz (to capture sudden changes in oil pressure / temperature), and restore the base frequency after steady state;

[0104] Long-term steady-state operation: Gradually tighten the threshold as the cumulative operating time increases (e.g., allowable moisture content changes from 500ppm to 300ppm to 200ppm) and strengthen wear monitoring.

[0105] (2) Environmental adaptation:

[0106] Read ambient temperature and altitude, and match. Figure 3 Adaptive logic. For example... Figure 3 As shown, this figure adopts a horizontal layout of "environment-operating condition axis", and labels typical environments (normal temperature and pressure, high-altitude -30℃, and high-altitude 3000 meters) from left to right. Each environment corresponds to a set of "normal baseline values" and "allowable fluctuation ranges" of oil parameters (represented by upper and lower limit dashed lines), and the correction basis is marked (such as relaxing the temperature threshold in high-altitude environments and compensating for the pressure benchmark value in high-altitude environments).

[0107] In frigid environments of ≤-30℃, the sensor heating module is activated, the upper temperature limit is relaxed to 120℃, and the kinematic viscosity weight is reduced from 30% to 20%. In high-altitude environments of ≥3000m, atmospheric pressure compensation is performed on the pressure sensor, the gas threshold is relaxed, and the humidity threshold is tightened. During test runs, the corresponding threshold tables are automatically called during cold starts, rated speed, and maximum thrust stages.

[0108] S6. Fault Diagnosis and Graded Early Warning:

[0109] The fault type and confidence level are diagnosed by using a random forest model, and a three-level early warning is triggered based on the OHI value and pushed to the early warning and display unit.

[0110] Mild abnormality (OHI 70~80): The local touch screen HMI displays a yellow warning and records parameter trends (such as temperature approaching the upper limit).

[0111] Medium risk (OHI50~70): Local audible and visual alarm (buzzer + LED flashing) + remote terminal push warning (e.g., "Kinematic viscosity decreases, oxidation should be monitored");

[0112] Critical fault (OHI<50): Local urgent alarm (continuous beeping + red indicator light) + remote push fault details (such as "moisture content is seriously excessive, stop the machine immediately"), and generate a test run interruption suggestion.

[0113] Display content includes: real-time parameter table, OHI trend curve (24 hours / full test run), fault diagnosis report (timestamp, type, confidence level) and historical comparison data (compared with the same model engine).

[0114] Fault Diagnosis Model: Based on the Random Forest algorithm (training data comes from 1000+ sets of historical test data of aero-engines, covering typical operating conditions such as cold start and rated speed, with normal samples accounting for 60%, warning samples for 30%, and fault samples for 10%; fault types include moisture intrusion (40%), oxidation degradation (30%), metal wear (20%), and seal leakage (10%)). Input features are the real-time deviations of 5 parameters (ΔT, ΔP, ΔV, ΔH, ΔG) and the real-time value of OHI, and the output is the specific fault type and confidence level (average accuracy of 92% on the test set, recall of each fault type >85%). The model supports online incremental learning (optimizing the decision tree weights every 24 hours using new fault data), which can continuously improve the ability to identify new types of faults.

[0115] Example 4:

[0116] This embodiment uses the complete set of standardized online monitoring methods for multi-parameter oil condition for engine testing described in Embodiment 3 as the execution benchmark. The hardware equipment is completely the same as the online monitoring system for multi-parameter oil condition for engine testing disclosed in Embodiment 2. All formulas, sampling strategies, early warning rules, and adaptive logic are not modified.

[0117] The difference lies in the following: A standard test run procedure for a specific type of turbofan engine is performed: cold start → idle speed → rated speed → maximum thrust → deceleration → shutdown, executing all steps S1 to S6 of Example 3 step by step.

[0118] S1. Real-time acquisition of multiple parameters:

[0119] During the cold start phase, the low-frequency acquisition rules described in Example 3 are strictly followed. Only the temperature sensor and pressure sensor acquire data at a frequency of 1Hz, while the kinematic viscosity sensor, humidity sensor, and gas composition sensor remain in sleep mode. Once the rated steady-state condition is reached, all parameters of the oil are stable, and the low-power acquisition mode is maintained continuously.

[0120] S2, Signal Conditioning and Calibration Conversion:

[0121] The five sensor signals were processed in full accordance with the channel filtering, amplification, analog-to-digital conversion, and calibration conversion process explained in Example 3. The calibration curves and conversion standards were completely consistent with the explanation content.

[0122] S3, Oil Health Index (OHI) Calculation:

[0123] Using the weights of the original parameters disclosed and the baseline values, the OHI under steady-state test conditions was consistently maintained above 95, which was determined to be within the normal range.

[0124] S4. Dynamic sampling control:

[0125] During the maximum thrust and high load phase, the lubricating oil temperature rises to 95°C, approaching the baseline allowable upper limit. The system automatically switches to 10Hz intermediate frequency full parameter acquisition according to the logic of Example 3.

[0126] Simulate a water intrusion coupled with oxidation failure scenario:

[0127] (1) The triggering conditions are: humidity rises to 2000ppm (baseline ≤ 500ppm), CO2 concentration rises to 200ppm (baseline ≤ 100ppm), and other parameters are temperature 95℃, kinematic viscosity 45cSt, and pressure normal.

[0128] (2) Calculation of OHI: Humidity deviation (2000 / 500=4, calculated as the maximum of 1, weighted contribution 0.15×1=0.15) + CO2 deviation (200 / 100=2, calculated as the maximum of 1, weighted contribution 0.25×1=0.25) + other parameters with smaller deviations → Overall OHI≈100-(0.15+0.25)=99.6 (If humidity and gas exceed the standard, the actual OHI may drop to 55).

[0129] S5, Adaptive Threshold Correction:

[0130] This embodiment is a normal temperature flat-ground test, which does not require the activation of high-altitude / cold-weather correction logic. It only calls up the logic according to different test stages such as cold start, rated thrust, and maximum thrust. Figure 2 The corresponding operating condition threshold table completes the adjustment of parameter baseline and fluctuation range.

[0131] S6. Fault Diagnosis and Graded Early Warning:

[0132] The random forest model outputs the standard fault conclusion: "Severe moisture intrusion, accompanied by accelerated lubricating oil oxidation, confidence level 90%", triggering an OHI < 60 severe fault level. The system triggers an early warning strategy: the red light on the audible and visual alarm flashes rapidly and sounds continuously; a shutdown and maintenance suggestion is pushed to the remote terminal. Engineers can decide to terminate the test run early and conduct a lubricating oil system sealing check based on the OHI trend and diagnostic results.

[0133] This verification embodiment fully reuses the complete set of multi-parameter online oil monitoring methods described in Embodiment 3. The hardware equipment adopts the complete monitoring system disclosed in Embodiment 2. The system hardware structure, OHI calculation formula, dynamic sampling logic, fault diagnosis model, and three-level early warning rules are all unchanged. Only a standard air-to-ground flight test scenario at room temperature is built to verify the complete operating effect of the present invention under normal test conditions.

[0134] This embodiment verifies that the standard monitoring process can be stably implemented in conventional flat-ground aerial test scenarios, accurately detects oil and water intrusion coupled with oxidation faults, and achieves graded early warning.

[0135] Example 5:

[0136] This embodiment uses the complete set of standardized online monitoring methods for multi-parameter oil condition for engine testing described in Embodiment 3 as the execution benchmark. The hardware equipment is completely the same as the online monitoring system for multi-parameter oil condition for engine testing disclosed in Embodiment 2. All algorithms, sampling strategies, and early warning rules remain unchanged. Two extreme test environments, -30℃ high-altitude cold and 3000m high-altitude plateau, are simulated respectively to verify the actual operation effect of the core function of adaptive correction of working conditions and environment of this invention.

[0137] Scenario 1: -30℃ High-Altitude Cold Test Procedure

[0138] S1. Real-time acquisition of multiple parameters: After the device is powered on, the system identifies the current low temperature environment and automatically starts the heating resistance wire of the temperature sensor to raise the temperature measurement area to above -10℃ within 5 minutes; the cold start stage still follows the 1Hz low frequency acquisition rule specified in Example 3.

[0139] S2. Signal conditioning and calibration conversion: The five-channel signal amplification, filtering, and analog-to-digital conversion process is consistent with that in Example 3; the built-in temperature compensation module in the gas sampling chamber is started synchronously to correct the drift of the output signal generated by the gas sensor under low temperature conditions.

[0140] S3. Calculation of Oil Health Index (OHI): Strictly implement the weighting correction rules for cold environments. The kinematic viscosity weight is reduced from 30% to 20% before being included in the weighted calculation to avoid misjudgment of the health index caused by the naturally high viscosity of cold lubricating oil.

[0141] S4. Dynamic sampling control: The engine enters the rated steady-state operating condition, and all parameters are normal, continuously maintaining the 1Hz low-frequency sampling mode.

[0142] S5, Adaptive Threshold Correction: The system enables high-altitude cold environment correction logic, raising the upper limit of the allowable temperature to 120℃ to adapt to the lubricating oil temperature rise characteristics during low-temperature cold start-up. Simultaneously, it incorporates different parameters based on different stages of the test drive. Figure 2 The parameter adjustment was completed in coordination with the test run condition threshold.

[0143] S6. Fault diagnosis and graded early warning: During the entire test run, no parameters of the oil exceeded the standard, the OHI remained stable above 90, and the system did not output any early warning, effectively eliminating the false alarm defects that traditional monitoring solutions are prone to in cold environments.

[0144] Scenario 2: High-altitude test drive process at 3000m

[0145] S1. Real-time acquisition of multiple parameters: The logic for synchronous acquisition of five types of sensors is the same as that described in Example 3, and the acquisition mode is unchanged.

[0146] S2. Signal conditioning and calibration conversion: The pressure sensor output value is first corrected by the atmospheric pressure compensation algorithm and then converted into the standard engineering pressure value. The rest of the signal processing flow is the same as in Example 3.

[0147] S3. Oil Health Index (OHI) Calculation: Following the plateau environment parameter adjustment rules, the weight of gas components is reduced, the allowable threshold for CO2 is relaxed to 150ppm, and the upper limit of dissolved water humidity is tightened to 200ppm. The weighted calculation of OHI is completed based on the corrected baseline and fluctuation range.

[0148] S4. Dynamic sampling control: The 1Hz low-frequency acquisition strategy is continuously used throughout the steady-state test.

[0149] S5, Adaptive Threshold Correction: System Matching Figure 3 The plateau environment adaptive control logic automatically completes atmospheric pressure compensation and synchronously adjusts the allowable fluctuation range of gas concentration and dissolved humidity to adapt to the changes in the physicochemical properties of gliding oil in the low-pressure environment of the plateau.

[0150] S6. Fault diagnosis and graded early warning: Under fault-free test conditions, the monitoring baseline automatically adapts to the high-altitude and low-pressure environment, and will not experience the threshold failure and frequent false alarms that are common in conventional monitoring systems.

[0151] This embodiment also uses the standardized monitoring method described in Embodiment 3 as the execution basis, and the hardware uses the complete system of Embodiment 2. All algorithms, sampling and early warning rules remain unchanged. It simulates two extreme test environments: -30℃ high cold and 3000m altitude, to verify the actual operation effect of the core function of adaptive correction of working conditions and environment of this invention.

[0152] The above is an exemplary description of the invention. Obviously, the specific implementation of the invention is not limited to the above-described manner. Any non-substantial improvement made using the inventive concept and technical solution of the invention, or the direct application of the inventive concept and technical solution to other situations without modification, is within the protection scope of the invention.

Claims

1. A multi-parameter online oil condition monitoring system for engine testing, characterized in that, include: Multi-parameter sensing module, signal conditioning and acquisition unit, data processing and health assessment unit, and early warning and display unit; The multi-parameter sensing module integrates a temperature sensor, a pressure sensor, a kinematic viscosity sensor, a humidity sensor, and a gas composition sensor. It is installed in the engine's main oil circuit or return oil circuit to achieve in-situ, minimally interfered data acquisition. Each sensor transmits the raw signals of the acquired oil parameters to the signal conditioning and acquisition unit. The signal conditioning and acquisition unit includes an amplification and filtering circuit, a 24-bit ADC, and a CAN / LTE communication module. The amplification and filtering circuit receives the raw signal transmitted by the multi-parameter sensing module and performs amplification and anti-aliasing filtering. The processed signal is converted from analog to digital by the 24-bit ADC and then transmitted to the data processing and health assessment unit through the CAN / LTE communication module. The data processing and health assessment unit uses an embedded processor to receive digital signals transmitted by the signal conditioning and acquisition unit, perform data preprocessing, calculate the oil health index (OHI), perform dynamic sampling control, and implement a fault diagnosis model. It outputs the health index and fault type and feeds the results back to the multi-parameter sensing module to achieve dynamic sampling adjustment. At the same time, it transmits warning-related information to the warning and display unit. The early warning and display unit includes a local touch screen, an audible and visual alarm, and a remote terminal. It receives early warning information transmitted by the data processing and health assessment unit, and realizes three-level early warning and status visualization.

2. The online monitoring system for multi-parameter oil condition for engine testing according to claim 1, characterized in that, The temperature sensor is a PT100 with a temperature range of -40℃ to 200℃, and is wall-mounted. The pressure sensor is a piezoresistive type with a range of 0 to 5 MPa, connected in series in the main oil circuit. The kinematic viscosity sensor is a vibrating tube type with a range of 0.8 to 500 cSt, and is installed by immersion in the return oil pipe. The humidity sensor is a polymer capacitive type with a range of 0 to 5000 ppm, which draws oil through a capillary tube. The gas composition sensor is a combination of infrared and electrochemical sensors, which detects CO2, H2, and CH4, and is equipped with a 60℃ degassing chamber.

3. The online monitoring system for multi-parameter oil condition for engine testing according to claim 1, characterized in that, The amplification and filtering circuit is an electromagnetic interference (EMI) filtering module used to suppress high-frequency noise in the engine ignition system; the 24-bit ADC is a Σ-Δ analog-to-digital converter chip that enables synchronous sampling of multiple parameter signals; the CAN / LTE communication module supports CAN2.0B bus and 4G / LTE wireless transmission, and the data packet includes timestamp, sensor ID, original value, and CRC-16 checksum.

4. The online monitoring system for multi-parameter oil condition for engine testing according to claim 1, characterized in that, The embedded processor is a TI AM62x series with a main frequency of 1GHz and 2GB of memory. The fault diagnosis model it runs is a random forest model, which takes five-parameter deviation and OHI as inputs and outputs moisture intrusion, oxidation degradation, metal wear, seal leakage and corresponding confidence levels.

5. The online monitoring system for multi-parameter oil condition for engine testing according to claim 1, characterized in that, The local touchscreen is used to display real-time parameter tables, OHI trend curves, fault diagnosis reports, and historical comparison data; the audible and visual alarm uses a buzzer and LED flashing alarm method, supporting differentiated alarms for different warning levels; the remote terminal supports warning information push, historical data query, and remote parameter configuration.

6. A method for online monitoring of multi-parameter fluid condition for engine testing, based on the online monitoring system for multi-parameter fluid condition for engine testing as described in any one of claims 1-5, characterized in that, Includes the following steps: S1. Real-time acquisition of multiple parameters: The original signals of oil temperature, pressure, kinematic viscosity, humidity and gas composition are collected synchronously through a multi-parameter sensing module. S2, Signal Conditioning and Calibration Conversion: The original signal is amplified, filtered, and converted from analog to digital by the signal conditioning and acquisition unit, and then converted into a physical quantity in engineering units by the calibration curve. S3, Oil Health Index (OHI) Calculation: OHI is calculated using a weighted algorithm through the data processing and health assessment unit, with the following formula: OHI=100- ), in, This is the real-time value of the i-th parameter. The normal baseline value for the i-th parameter. Let i be the allowable fluctuation range for the i-th parameter. The weight of the i-th parameter; S4. Dynamic sampling control: The sampling frequency is adjusted according to the real-time OHI value, and the sampling frequency is updated every 5 seconds according to the real-time OHI. When OHI≥80, a low-frequency sampling of 1Hz is used to collect only temperature and pressure. When 60≤OHI<80, a medium-frequency sampling of 10Hz is used to collect all parameters. When OHI<60, a high-frequency sampling of 100Hz is used to collect all parameters transiently. S5, Adaptive Threshold Correction: Based on engine test conditions and extreme environments such as high altitude and cold regions, the baseline values ​​and allowable fluctuation ranges of each parameter are automatically corrected. S6. Fault Diagnosis and Graded Early Warning: The fault type and confidence level are diagnosed by random forest model, and a three-level early warning is triggered based on the OHI value and pushed to the early warning and display unit.

7. The online monitoring method for multi-parameter oil condition during engine testing according to claim 6, characterized in that, In step S3, the specific weights of each parameter are: temperature 20%, kinematic viscosity 30%, gas composition 25%, humidity 15%, and pressure 10%; the OHI classification thresholds for sampling control are: OHI≥80 for normal state, 60≤OHI<80 for warning state, and OHI<60 for fault state.

8. The online monitoring method for multi-parameter oil condition during engine testing according to claim 6, characterized in that, In step S5, in a cold environment of ≤-30℃, the sensor heating module is activated, the upper limit of temperature is relaxed to 120℃, and the kinematic viscosity weight is reduced from 30% to 20%; in a high-altitude environment of ≥3000m, atmospheric pressure compensation is performed on the pressure sensor, the gas threshold is relaxed, and the humidity threshold is tightened; under test conditions, the corresponding threshold table is automatically called during cold start, rated speed, and maximum thrust stages.

9. The online monitoring method for multi-parameter oil condition during engine testing according to claim 6, characterized in that, In step S6, the OHI grading thresholds of the three-level early warning mechanism are as follows: when the OHI is between 70 and 80, it is considered a mild anomaly, and the local touchscreen displays a yellow warning and records the parameter trend; when the OHI is between 50 and 70, it is considered a moderate risk, and the audible and visual alarm is activated and a warning message is pushed to the remote terminal. When OHI < 50, it indicates a serious fault. The audible and visual alarm will sound urgently and push out fault details and shutdown suggestions.

10. The online monitoring method for multi-parameter oil condition during engine testing according to claim 6, characterized in that, In step S2, the original signal is amplified, filtered, and converted from analog to digital by the signal conditioning and acquisition unit, specifically as follows: The raw temperature signal output by the temperature sensor is sent to the amplification and filtering circuit. It is converted into a voltage signal by the built-in Wheatstone bridge using a three-wire connection. After being amplified by the built-in instrumentation amplifier, it is sent to the 24-bit ADC to complete the analog-to-digital conversion. The raw pressure signal output by the pressure sensor is sent to the amplification and filtering circuit. After being converted into a voltage signal by the built-in 250Ω resistor, it is low-pass filtered and then sent to the 24-bit ADC to complete the analog-to-digital conversion. The raw signals output from the kinematic viscosity sensor, humidity sensor, and gas composition sensor are sent to the amplification and filtering circuit. After isolation and anti-aliasing filtering by the built-in voltage follower, they are sent to the 24-bit ADC to complete the analog-to-digital conversion.

Images