A method for torque detection in a single-cylinder rotary aero engine

Abstract

This invention provides a torque detection method for a single-cylinder rotary aero-engine, relating to the field of aero-engine technology. The method includes: Step 2, acquiring the airflow flow rate and pressure coupling signal flowing through the valve body in real time during engine operation based on a dynamic pressure sensing orifice and a one-way pressure valve to obtain raw flow time-series data; Step 3, performing multi-dimensional feature analysis on the raw flow time-series data to obtain multiple process parameters with inherent correlation; quantifying the synergistic relationship of different parameter change directions by calculating the angle between the tangents of the process parameter change trajectory curves at feature points, and analyzing the steady-state and transient response characteristics based on the synergistic relationship to establish parameterized correction rules and generate corresponding correction parameter sets. This invention improves the accuracy and reliability of real-time torque monitoring of a single-cylinder rotary engine under complex operating conditions by acquiring and processing flow signals reflecting the working state of the engine's internal compression chamber.

Description

Technical Field

[0001] This invention relates to the field of aero-engine technology, and in particular to a method for detecting torque in a single-cylinder rotor aero-engine. Background Technology

[0002] In the field of aero-engine monitoring and control, torque is one of the key parameters for evaluating engine output performance and operating status. For single-cylinder rotary engines, accurate and real-time monitoring of their output torque is crucial for optimizing control strategies, ensuring operational safety, and improving maintenance efficiency. Currently, engine torque is mostly measured through indirect calculation or direct sensing. A common approach is to calculate the torque value by measuring the torsional deformation of the crankshaft. These methods usually require installing strain gauges on the drive shaft or setting up corresponding phase detection devices.

[0003] However, in the specific application of a single-cylinder rotary engine, the above methods may face some limitations. For example, due to the compact structure and limited internal space of the rotary engine, the installation of additional sensing devices is sometimes limited by the installation location and space, which may affect the original structure and balance of the engine. At the same time, factors such as high-frequency vibration, temperature changes and electromagnetic interference during engine operation may also interfere with the measurement system based on strain or phase difference signals, resulting in a decrease in the signal-to-noise ratio. Especially under transient conditions, the real-time performance and accuracy of the measurement results are sometimes difficult to guarantee. In addition, most of these methods rely on monitoring the overall transmission components of the engine, and their ability to directly reflect the internal working process of a single-cylinder rotor is relatively limited, which may make it difficult to accurately capture the torque fluctuation characteristics generated by the dynamic changes of the internal compression chamber. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a torque detection method for a single-cylinder rotary aero-engine, which improves the accuracy and reliability of real-time torque monitoring of a single-cylinder rotary engine under complex operating conditions by acquiring and processing flow signals that reflect the working state of the compression chamber inside the engine.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] In a first aspect, a method for detecting torque in a single-cylinder rotary aero engine, the method comprising:

[0007] Step 1: Open a dynamic pressure sensing hole on the axial end face of the single-cylinder rotor of the engine, which communicates with the internal compression chamber, and integrate and install a one-way pressure valve at the dynamic pressure sensing hole.

[0008] Step 2: Based on the dynamic pressure sensing orifice and the one-way pressure valve, the airflow flow rate and pressure coupling signal flowing through the valve body are collected in real time during engine operation to obtain the raw flow time series data.

[0009] Step 3: Perform multi-dimensional feature analysis on the original flow time series data to obtain multiple process parameters with intrinsic correlation; quantify the cooperative relationship of different parameter change directions by calculating the angle between the tangents of the process parameter change trajectory curves at feature points, and analyze the steady-state and transient response characteristics based on the cooperative relationship, and establish parameterized correction rules to generate the corresponding correction parameter set.

[0010] Step 4: Using the correction parameter set, synchronous filtering and phase compensation are performed on the flow signal in the original flow time series data and the real-time speed and angular position signal of the single-cylinder rotor; by identifying and fitting the key extreme points of the flow signal waveform within a cycle, a feature polygon is constructed and the geometric symmetry axis is obtained, and the flow characteristic parameters corresponding to the rotor periodic motion are determined according to the position of the symmetry axis.

[0011] Step 5: Based on the flow characteristic parameters, combined with the intake manifold absolute pressure and intake temperature under the current engine operating conditions, the real-time output torque value of a single-cylinder rotor is calculated using a torque mapping model to achieve online torque monitoring of the engine operating state.

[0012] In a second aspect, a computing device includes:

[0013] One or more processors;

[0014] A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the method.

[0015] Thirdly, a computer-readable storage medium storing a program that, when executed by a processor, implements the method.

[0016] The above-described solution of the present invention has at least the following beneficial effects:

[0017] By setting dynamic pressure sensing holes and one-way pressure valves on the axial end face of the single-cylinder rotor of the engine, the stability of the airflow and pressure coupling signal acquisition during engine operation is ensured. The synergistic relationship of process parameters is quantified by multi-dimensional feature analysis and a set of correction parameters is generated to reduce interference factors and deviations in the original data. Through synchronous filtering, phase compensation and precise extraction of flow characteristic parameters, the flow signal is accurately matched with the real-time rotor speed and angular position. Combined with operating parameters such as intake manifold absolute pressure and intake air temperature and torque mapping model, online monitoring of the real-time output torque of the single-cylinder rotor is realized, improving the response speed of torque monitoring, distinguishing the steady-state and transient response characteristics of the engine, and reducing problems such as insufficient data reliability, unclear feature correlation and poor real-time monitoring. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of a torque detection method for a single-cylinder rotor aero-engine provided by an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of an embodiment of the present invention, which uses a dynamic pressure sensing orifice and a one-way pressure valve to collect the airflow and pressure coupling signal of the airflow passing through the valve body in real time during engine operation to obtain the original flow time sequence data. Detailed Implementation

[0020] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0021] like Figure 1 As shown, an embodiment of the present invention proposes a torque detection method for a single-cylinder rotary aero engine, the method comprising the following steps:

[0022] Step 1: Open a dynamic pressure sensing hole on the axial end face of the single-cylinder rotor of the engine, which communicates with the internal compression chamber, and integrate and install a one-way pressure valve at the dynamic pressure sensing hole.

[0023] Step 2: Based on the dynamic pressure sensing orifice and the one-way pressure valve, the airflow flow rate and pressure coupling signal flowing through the valve body are collected in real time during engine operation to obtain the raw flow time series data.

[0024] Step 3: Perform multi-dimensional feature analysis on the original flow time series data to obtain multiple process parameters with intrinsic correlation; quantify the cooperative relationship of different parameter change directions by calculating the angle between the tangents of the process parameter change trajectory curves at feature points, and analyze the steady-state and transient response characteristics based on the cooperative relationship, and establish parameterized correction rules to generate the corresponding correction parameter set.

[0025] Step 4: Using the correction parameter set, synchronous filtering and phase compensation are performed on the flow signal in the original flow time series data and the real-time speed and angular position signal of the single-cylinder rotor; by identifying and fitting the key extreme points of the flow signal waveform within a cycle, a feature polygon is constructed and the geometric symmetry axis is obtained, and the flow characteristic parameters corresponding to the rotor periodic motion are determined according to the position of the symmetry axis.

[0026] Step 5: Based on the flow characteristic parameters, combined with the intake manifold absolute pressure and intake temperature under the current engine operating conditions, the real-time output torque value of a single-cylinder rotor is calculated using a torque mapping model to achieve online torque monitoring of the engine operating state.

[0027] In this embodiment of the invention, by setting a dynamic pressure sensing hole and a one-way pressure valve on the axial end face of the single-cylinder rotor of the engine, the stability of the airflow and pressure coupling signal acquisition during engine operation is ensured. The synergistic relationship of process parameters is quantified by multi-dimensional feature analysis and a correction parameter set is generated to reduce interference factors and deviation problems in the original data. Through synchronous filtering, phase compensation and accurate extraction of flow characteristic parameters, the flow signal is accurately matched with the real-time speed and angular position of the rotor. Combined with operating condition parameters such as intake manifold absolute pressure and intake temperature and torque mapping model, online monitoring of the real-time output torque of the single-cylinder rotor is realized, improving the response speed of torque monitoring, distinguishing the steady-state and transient response characteristics of the engine, and reducing problems such as insufficient data reliability, unclear feature correlation and poor real-time monitoring.

[0028] In a preferred embodiment of the present invention, step 1 above, which involves opening a dynamic pressure sensing hole communicating with the internal compression chamber on the axial end face of the single-cylinder rotor of the engine, and integrating and installing a one-way pressure valve at the dynamic pressure sensing hole, may include:

[0029] In this embodiment of the invention, step 110 involves analyzing the internal structure of the single-cylinder rotor of the engine to determine the dynamic pressure region where the compression chamber of the single-cylinder rotor generates regular high pressure during the rotation cycle, and obtaining the three-dimensional spatial coordinates of the dynamic pressure region. Specifically, this includes: firstly, according to the assembly manual of the single-cylinder rotor engine, using specialized disassembly tools to sequentially disassemble the engine's external housing, intake components, and other auxiliary structures, gradually exposing the main body of the single-cylinder rotor, ensuring that the core structures such as the internal compression chamber and rotor shaft are not damaged during disassembly; then, using a high-definition industrial endoscope inserted into the rotor, combined with strong light illumination equipment, to comprehensively observe the chamber shape, inner wall contour, and connection position with the rotor shaft of the compression chamber, while simultaneously using several... Precision measuring tools such as calipers and dial indicators are used to measure key dimensions of the compression chamber, such as depth and inner diameter, and these basic structural parameters are recorded. Next, the disassembled single-cylinder rotor is stably installed on a dedicated test bench, ensuring that the rotor shaft is coaxially connected to the test bench drive mechanism to avoid unstable rotor rotation due to installation deviations. Then, miniaturized temporary pressure sensing patches are attached to different areas of the inner wall of the compression chamber, including the area near the air inlet, the middle area of ​​the chamber, the area near the exhaust outlet, and multiple points evenly distributed around the circumference. Each patch is connected to the data acquisition instrument of the test bench through a shielded wire. The shielded wire is arranged to avoid the rotor rotation trajectory to prevent the wire from getting tangled or broken during rotation.

[0030] The test bench drive mechanism is activated, driving the single-cylinder rotor to rotate uniformly at three typical speeds (low, medium, and high) within the rated speed range. At each speed, the rotor rotates continuously for at least 10 complete cycles. The data acquisition unit simultaneously collects real-time pressure data from all pressure sensor patches and records the time point corresponding to each data point. After the rotation ends, the collected pressure data is exported. Through point-by-point comparative analysis, the regions where the pressure values ​​consistently exceed a preset high-pressure threshold (calculated based on the engine's compression ratio, i.e., 70% of the maximum design pressure of the compression chamber) and where the pressure values ​​exhibit regular, periodic peak changes with rotor rotation are identified. When the rotor rotates to this region, forming a closed chamber with the intake and exhaust ports, the pressure rapidly increases. The pressure drops rapidly when the closed chamber is opened, and the area that conforms to this pattern is the dynamic pressure region. Finally, a three-dimensional laser scanning measuring instrument is used to scan the determined dynamic pressure region. A three-dimensional rectangular coordinate system is established with the center of the axial end face of the single-cylinder rotor as the origin. The rotor axis is the Z-axis, pointing inwards from the rotor as the positive direction. The horizontal direction passing through the origin in the axial end face is the X-axis, and the vertical direction passing through the origin in the axial end face is the Y-axis. The coordinate data of at least 20 feature points evenly distributed on the boundary of the dynamic pressure region are obtained by scanning. The coordinates of each feature point are directly read and recorded by the three-dimensional laser scanning measuring instrument. The coordinate values ​​of these feature points are integrated to form the complete three-dimensional spatial coordinate range of the dynamic pressure region.

[0031] Step 111: Project the three-dimensional spatial coordinates onto the axial end face of the single-cylinder rotor. Based on the mapped position, machine a dynamic pressure sensing hole on the axial end face of the single-cylinder rotor to connect the dynamic pressure sensing hole with the inside of the compression chamber, converting the periodic pressure changes in the compression chamber into periodic gas pressure signals exported through the dynamic pressure sensing hole. Specifically, this includes: importing the three-dimensional spatial coordinate data of the dynamic pressure region obtained in step 110 into professional coordinate projection software, setting the projection direction in the software to the negative Z-axis direction, i.e., from the inside of the rotor to the axial end face, and projecting the three-dimensional coordinates onto the two-dimensional plane of the axial end face of the single-cylinder rotor. On the surface, the software automatically generates a two-dimensional mapped area after projection, marking the center point and boundary range of this mapped area in the software, which serves as the positioning reference for the sensor hole opening; then, the single-cylinder rotor is fixed on the fixture of the CNC machining tool, the fixture position is adjusted, and the axial end face of the rotor is calibrated using the positioning device built into the machine tool to ensure that the axial end face is perpendicular to the machining spindle of the machine tool; according to the center point marked by the projection, the machining spindle of the machine tool is moved so that the drill bit mounted on the spindle is aligned with the center point. The drill bit material is a high-speed steel drill bit that matches the rotor material, and the initial diameter of the drill bit is smaller than the preset preliminary hole diameter; start The machine tool is used for drilling. A slow feed rate is set during the process, pausing every 0.5 mm. An industrial endoscope is inserted into the borehole to observe whether the drilling direction deviates from the dynamic pressure zone. If a deviation occurs, it is corrected by fine-tuning the spindle position on the machine tool. When the drilling depth reaches the preset value, which is determined based on the compression chamber depth measured in step 110, ensuring the borehole penetrates the axial end face and communicates with the inside of the compression chamber, drilling is stopped. A grinding drill bit is then used to grind the hole wall until it is smooth and burr-free. After grinding, the rotor is removed and reinstalled on the test bench. The dynamic pressure zone is then tested. The external port of the pressure sensing hole is connected to a temporary pressure acquisition device via a sealed joint. The test bench is started to drive the rotor to rotate, and the exported gas pressure signal is acquired. The signal curve displayed by the pressure acquisition device is observed. If the curve shows periodic fluctuations consistent with the rotor rotation period, and the peak and trough values ​​of the fluctuations conform to the pressure change law of the compression chamber, it indicates that the sensing hole and the inside of the compression chamber are well connected, and the periodic pressure changes in the compression chamber can be converted into periodic gas pressure signals exported through the sensing hole. If the signal has no obvious periodic fluctuations or the fluctuation amplitude is too small, the drilling position and depth need to be rechecked, and re-machining may be necessary.

[0032] Step 112 involves extracting the frequency and pressure amplitude characteristics of the periodic gas pressure signal, calculating the minimum gas flow capacity required to transmit the pressure signal based on these characteristics, and designing and determining the final aperture size of the dynamic pressure sensing orifice based on this minimum flow capacity. Specifically, this includes importing the periodic gas pressure signal acquired in step 111 into a signal analysis device, and using the device's signal feature extraction function to extract the frequency and pressure amplitude characteristics of the signal. The frequency feature extraction method involves counting the number of pressure peaks that occur in the signal curve within one minute, dividing this number by 60, and obtaining the pressure signal variation per unit time. Frequency is defined as the number of times a pressure peak occurs per second. The pressure amplitude feature is extracted by reading the maximum and minimum pressure values ​​from the signal curve and subtracting the minimum pressure value from the maximum to obtain the pressure amplitude. Based on the extracted frequency and pressure amplitude features, the minimum gas flow capacity required to transmit the pressure signal is calculated. First, the number of cycles of pressure change in the compression chamber per second is determined based on the frequency features. Then, combined with the dynamic pressure region volume of the compression chamber measured in step 110, the gas volume that needs to be discharged through the sensing orifice in each pressure change cycle is calculated. The gas volume discharged in each cycle is equal to the dynamic pressure region volume multiplied by the pressure amplitude. The corresponding volume expansion ratio is calculated by dividing the difference between the gas volume corresponding to the maximum pressure and the gas volume corresponding to the minimum pressure by the gas volume corresponding to the minimum pressure. Then, the required gas volume for each cycle is multiplied by the number of cycles per second to obtain the minimum gas volume that needs to pass through the sensing orifice per second. This volume represents the minimum gas flow capacity required to transmit the current pressure signal. Based on the calculated minimum gas flow capacity, 3-5 different candidate orifice sizes are initially determined, ranging from smaller than the orifice size corresponding to the minimum flow capacity to larger than that orifice size. Subsequently, each candidate orifice size is tested on a test bench. Testing was conducted by mounting drill bits of the corresponding aperture on a CNC machine tool and machining the sensing hole by either enlarging or shrinking it. During enlargement, the drill bit diameter was gradually increased, while during shrinkage, a precision boring tool was used. After machining, a temporary pressure acquisition device was connected to collect the exported pressure signals. The frequency characteristics and pressure amplitude characteristics of the signals under different apertures were compared and analyzed to ensure consistency with the original signals. The aperture size with the highest signal frequency deviation and pressure amplitude retention rate was selected and determined as the final aperture size of the dynamic pressure sensing hole. This size can meet the minimum gas flow capacity, ensure the complete transmission of pressure signals, and avoid excessive pressure loss in the compression chamber due to an excessively large aperture.

[0033] Step 113: Match a one-way pressure valve of the appropriate specification according to the final orifice size, and integrate the one-way pressure valve into the external port of the dynamic pressure sensing orifice; specifically, this includes: based on the final orifice size of the dynamic pressure sensing orifice determined in Step 112, consulting the product specification manual of the one-way pressure valve, selecting one-way pressure valves whose orifice size is completely consistent with the final orifice size of the sensing orifice, whose rated working pressure range covers the maximum design pressure of the compression chamber, and whose installation interface type is compatible with the axial end face of the single-cylinder rotor; at the same time, check the valve core flexibility and sealing performance of the candidate one-way pressure valves, confirm that the valve core can smoothly reset by manually pressing the valve core, connect the valve to a simple pneumatic circuit, introduce preset pressure gas, and observe whether there is gas leakage, confirming... Ensure the valve itself meets performance standards. Then, pre-treat the mounting surface of the one-way pressure valve and the external port of the sensing hole by wiping the surfaces with anhydrous ethanol to remove oil, iron filings, and other impurities. Apply a suitable sealing gasket to the valve's mounting flange or threaded interface. The gasket material should be fluororubber, resistant to high and low temperatures and oil, to ensure it can withstand the harsh environment of engine operation. Align the input end of the one-way pressure valve with the external port of the dynamic pressure sensing hole, and adjust the valve direction so that the output end faces a direction conducive to subsequent sensor installation. Ensure the valve and sensing hole are coaxially aligned. Measure the perpendicularity of the valve end face and the rotor axial end face using a right-angle ruler, ensuring the perpendicularity error is within 0.02 mm.

[0034] If a threaded connection is used, tighten the valve clockwise with a torque wrench. The tightening torque should be set according to the recommended value in the valve specification manual to avoid excessive torque causing thread stripping or insufficient torque causing poor sealing. If welding is used, select laser welding equipment, adjust the welding power and welding speed, and perform circumferential welding along the joint between the valve mounting flange and the axial end face of the rotor. During the welding process, continuously observe the weld pool to ensure that the weld is continuous and free of pores and cracks. After installation, perform an airtightness test. Connect the sealing joint to the output end of the one-way pressure valve, and fill the channel formed by the sensor hole and the valve with gas equal to 80% of the maximum design pressure of the compression chamber. Maintain the pressure for 30 minutes and observe the change in the reading of the pressure monitoring instrument. If the pressure drop does not exceed 2% of the filling pressure, the installation and sealing are good. If leakage is found, the valve needs to be disassembled to re-inspect the sealing gasket or weld, reinstall and test until the airtightness is qualified.

[0035] The dynamic pressure zone is determined by multi-speed testing, the sensing holes are machined and the final aperture is matched to ensure that the derived periodic gas pressure signal can truly reflect the pressure changes inside the compression chamber; the matching and sealing installation of the one-way pressure valve prevents external airflow from flowing back into the compression chamber and avoids interference from reverse airflow on the original signal.

[0036] like Figure 2As shown, in a preferred embodiment of the present invention, step 2 above, based on the dynamic pressure sensing orifice and the one-way pressure valve, to collect the airflow flow rate and pressure coupling signal flowing through the valve body in real time during engine operation and obtain the original flow time series data, may include:

[0037] In this embodiment of the invention, step 220 involves adapting and fixing a miniature flow sensor and a miniature dynamic pressure sensor to the outlet end face based on the outlet specifications of the one-way pressure valve. Specifically, this includes: First, surveying and recording the outlet specifications of the one-way pressure valve. Using a digital caliper, accurately measure the outlet end face of the installed one-way pressure valve, focusing on the outlet inner diameter, outlet end face outer diameter, end face thickness, and the dimensions of the pre-drilled mounting holes on the end face. Simultaneously observe the interface type of the outlet end face. Record all measurement data in the test record table. Each dimensional parameter is measured three times, and the average value is taken to ensure data accuracy. This data will serve as the core basis for sensor adaptation selection, avoiding airflow leakage or sensor malfunction after installation due to size mismatch. Second, selecting and verifying the compatibility of the miniature sensor. Based on the surveyed outlet specifications, select miniature sensors that meet the following conditions: 1) Miniaturization: The overall outer diameter of the sensor is less than 1 / 2 of the outer diameter of the outlet end face, ensuring that the two sensors can be installed simultaneously on the outlet end face without interference, adapting to the compact installation space of a single-cylinder rotary engine; 2) ... First, the measurement range must be matched. The range of the miniature flow sensor needs to cover 1.2 to 2 times the minimum gas flow capacity calculated in step 112, and the range of the miniature dynamic pressure sensor needs to cover 1.2 times the maximum design pressure of the compression chamber to ensure accurate measurement of airflow and pressure changes. Second, the installation interface must be compatible. The size of the sensor's mounting base must match the reserved hole on the outlet end face, or it can be securely connected to the outlet end face via an adapter. Third, the signal output type must be an analog electrical signal with a response frequency of not less than 1kHz to meet real-time acquisition requirements. The selection is now complete. Next, compatibility verification is performed by attaching the mounting bases of the two sensors to the outlet face of the one-way pressure valve. Check whether the mounting holes of the bases are aligned with the reserved holes on the outlet face, and whether the sensor measuring end can completely cover the airflow path. At the same time, check whether the lead-out direction of the sensor wires avoids other moving parts of the engine to prevent the wires from being tangled or worn during engine operation. If there is any misalignment of the holes, select an adapter plate that matches the holes. One end of the adapter plate is adapted to the hole position on the outlet face, and the other end is adapted to the hole position on the sensor base to ensure that the sensor can be accurately positioned and installed.

[0038] The third step involves fixing and installing the sensors and arranging the wiring. Thoroughly wipe the contact surfaces of the one-way pressure valve outlet face and the sensor mounting base with alcohol wipes to remove oil, dust, and other impurities. After the alcohol has completely evaporated, apply a thin, even layer of high-temperature resistant sealant to the contact surfaces to enhance the seal after installation. Place the two sensors symmetrically on the outlet face, ensuring a uniform airflow path through the two sensor measuring ends for consistent measurement data. Adjust the sensor angle so that the measuring end faces the airflow direction. Then, use an Allen wrench to slowly tighten the bolts through the pre-drilled holes in the sensor base and outlet face to the preset torque. During the process, apply force alternately and evenly to avoid uneven stress on the sensor base, which could lead to deformation or poor sealing. After installation, arrange the wires by leading the shielded wires of the two sensors along the edge of the rotor's axial end face. Cover the outer layer of the wires with high-temperature resistant insulating sleeves and secure them to the engine mounting components with cable ties. The spacing between the cable ties should be controlled at about 5cm to ensure that the wires do not loosen or wobble. Connect the aviation plug end of the wire to the corresponding channel of the signal acquisition equipment. When connecting, ensure that the plug is fully inserted and locked to avoid poor contact that could lead to signal loss. Finally, check the firmness of the sensor installation by gently shaking the sensor to ensure that it is not loose and that the sealant has not overflowed from the bonding surface.

[0039] Step 221: Start the engine and run it. The periodic pressure changes generated in the compression chamber drive the airflow through the installed sensor measurement end, simultaneously activating the miniature flow sensor and the miniature dynamic pressure sensor to sense the airflow in real time and generate two continuous analog electrical signals: an analog flow signal and an analog pressure signal. Specifically, this includes: First, pre-start engine checks. Before starting the engine, thoroughly check the status of all components. Firstly, check if the miniature flow sensor and the miniature dynamic pressure sensor are securely installed, if the wiring connections are reliable, and if the aviation connectors are tightened. Secondly, check if the power supply to the signal acquisition equipment is connected, if the channel settings are correct, and if the voltage or current signal channel is selected according to the sensor type. Also check if the equipment is... The engine is in standby mode; thirdly, check whether the engine's fuel and lubricating oil supply is normal and whether the cooling system is intact to ensure that the engine can start and run normally; at the same time, adjust the engine's start control device to manual mode to facilitate control of the start-up process and speed adjustment; the second step is to start the engine and stabilize the speed. Operate the engine start control device to start the engine and gradually increase the engine speed from idle to a target speed within the rated speed range. During the increase, slowly adjust the throttle to avoid sudden speed changes that may cause engine instability. After the speed stabilizes, observe that the fluctuation range of the engine tachometer needle is within the allowable range and remains unchanged for 30 seconds. Maintain this speed. At this time, the single-cylinder rotor of the engine moves with the engine. The compressor rotates at a constant speed, continuously forming a closed chamber and compressing the gas during the rotation cycle, generating periodic pressure changes. In the third step, the airflow drive and sensor are activated synchronously. The periodic pressure changes generated in the compressor chamber push the gas within the chamber through the dynamic pressure sensing hole processed in step 111 and the one-way pressure valve installed in step 113, forming a stable airflow. This airflow exits from the one-way pressure valve outlet and flows directly through the measuring ends of the installed miniature flow sensor and miniature dynamic pressure sensor. Simultaneously, the start button of the signal acquisition device is pressed, triggering the simultaneous activation of the miniature flow sensor and miniature dynamic pressure sensor. The measuring ends of the sensors come into contact with the flowing airflow, beginning real-time sensing of airflow changes and... Pressure changes; the fourth step is the generation and monitoring of analog electrical signals. Under the action of airflow, the sensing element inside the miniature flow sensor undergoes periodic temperature changes. The sensor converts this physical change into a continuous analog electrical signal, i.e., an analog flow signal. Under the action of airflow pressure, the pressure-sensitive element inside the miniature dynamic pressure sensor undergoes a slight deformation. This deformation is also converted into a continuous analog electrical signal, i.e., an analog pressure signal. The waveforms of the two analog electrical signals are observed through the real-time monitoring interface of the signal acquisition device to ensure that the signals are continuous and uninterrupted, without significant noise interference. If an abnormal signal occurs, the engine speed is immediately reduced and the engine is stopped. The sensor installation is checked for looseness and the wires have poor contact. After troubleshooting, the test is restarted.

[0040] Step 222: Receive two analog electrical signals. Based on the rated speed range of the engine rotor and the expected change period of the pressure signal, calculate the minimum sampling frequency required to satisfy the signal, and set fixed values ​​higher than the minimum frequency as the unified sampling frequency of the signal acquisition unit. Specifically, this includes: First, analog electrical signal reception and preprocessing. Through the signal receiving channel of the signal acquisition device, simultaneously receive the analog flow signal output by the miniature flow sensor and the analog pressure signal output by the miniature dynamic pressure sensor. During reception, activate the device's pre-amplification function to amplify the weak analog electrical signals to a processable amplitude. Simultaneously, activate the anti-interference filtering function to filter the received two analog electrical signals. The signal is temporarily stored in the buffer of the signal acquisition device to ensure that the data is not lost. The second step is to confirm the rated speed range and expected change period. By consulting the design manual and technical parameter table of the single-cylinder rotor aero-engine, the rated speed range of the engine's single-cylinder rotor is determined. The maximum rated speed value is extracted from this range, as the maximum speed corresponds to the shortest pressure signal change period, requiring the highest sampling frequency. The minimum sampling frequency is calculated based on the maximum speed, ensuring accurate signal acquisition under all rated speed conditions. Simultaneously, according to the engine's working principle, each revolution of the single-cylinder rotor completes a full compression and expansion process in the compression chamber, corresponding to one complete pressure change. Therefore, the expected change period of the pressure signal is consistent with the rotation period of the rotor, that is, one rotation period corresponds to one pressure signal change period; the third step is to calculate the minimum sampling frequency. The calculation process is as follows: First, convert the maximum rated speed value into the maximum number of rotations per second of the rotor. The calculation method is to divide the maximum rated speed value by 60. For example, if the maximum rated speed is 15000 r / min, then the maximum number of rotations per second is 15000 divided by 60, and the result is 250 rotations / second; second, since each rotation corresponds to one pressure signal change period, the maximum number of pressure signal changes per second is equal to the maximum number of rotations per second, that is, 250 times / second; finally; According to the basic requirements of signal acquisition, in order to fully preserve all the characteristics of the signal and avoid signal distortion, the minimum sampling frequency needs to be greater than or equal to twice the maximum number of signal changes per second. Therefore, multiplying the maximum number of pressure signal changes per second by 2 gives the minimum sampling frequency. For example, 250 times / second multiplied by 2 results in 500Hz, which is the minimum sampling frequency. The fourth step is to set a unified sampling frequency. Taking into account the performance of the signal acquisition equipment, the needs of subsequent data processing, and the signal's anti-interference capability, a fixed value higher than the minimum sampling frequency is selected as the unified sampling frequency of the signal acquisition unit. This fixed value should be 1 times the minimum sampling frequency.A sampling frequency 5 to 2 times higher ensures complete capture of signal characteristics while avoiding excessive data volume and increased processing burden due to excessively high sampling frequencies. For example, if the minimum sampling frequency is 500Hz, 1000Hz can be selected as the uniform sampling frequency. Through the parameter setting interface of the signal acquisition device, the sampling frequency of both signals can be set to 1000Hz to ensure that both signals are acquired using the same sampling frequency.

[0041] Step 223: Using a unified sampling frequency, the signal acquisition unit synchronously and periodically samples and converts the two analog electrical signals to digital, generating digital flow and pressure value sequences with strictly aligned timestamps, forming the original flow time series data. Specifically, this includes: First, confirming sampling parameters and debugging the equipment. The sampling parameters of the signal acquisition unit are checked again to confirm that the unified sampling frequency is accurately set, the sampling channels correctly correspond to the miniature flow sensor and the miniature dynamic pressure sensor, and that the signal amplification factor and filtering parameters are set appropriately. Then, the equipment is debugged. The trial sampling function of the signal acquisition unit is started, acquiring 10 seconds of analog electrical signals to check if the sampling process is stable and if the signal can be received normally. If there are problems such as sampling interruption or signal distortion, the equipment parameters are adjusted or the sensor connection is checked until the trial sampling is normal. Second, synchronous periodic sampling. After debugging, the formal sampling function of the signal acquisition unit is started. The sampling unit synchronously and periodically samples the two analog electrical signals according to the set unified sampling frequency. Synchronous sampling means that at each sampling moment, the sampling unit simultaneously acquires the instantaneous values ​​of the analog flow signal and the analog pressure signal, ensuring that the two signals are accurately matched. The sampling points of the road signal correspond perfectly in time; periodic sampling means that the sampling unit continuously acquires signals at fixed time intervals, the sampling period, which is 1 divided by the sampling frequency, such as 1000Hz corresponding to a sampling period of 0.001 seconds, until the preset acquisition duration is completed or acquisition is manually stopped; the third step is analog-to-digital conversion and digital signal generation. After acquiring each instantaneous value of an analog signal, the sampling unit immediately starts the internal analog-to-digital conversion mode to convert the analog electrical signal into a discrete digital signal. For analog flow signals, the converted value is the digital flow value; for analog pressure signals, the converted value is the digital flow value. Digital pressure value; During analog-to-digital conversion, the sampling unit automatically performs precision calibration on the converted digital signal to ensure that the digital signal accurately reflects the actual changes in the analog signal; Fourth step, timestamp addition and sequence generation: While generating each digital flow value and digital pressure value, the sampling unit automatically adds a corresponding timestamp to it. The timestamp starts at the sampling start time and is recorded as time 0. The timestamp of each sampling point is the sampling period multiplied by the number of samples. For example, the timestamp of the first sampling point is 0.001 seconds × 1 = 0.001 seconds, and the timestamp of the second sampling point is 0.001 seconds × 2 = 0.001 seconds.002 seconds, and so on. Because the two signals are sampled synchronously, the digital flow rate value and digital pressure value at the same sampling time will be given the exact same timestamp, ensuring strict alignment in the time dimension. Subsequently, the sampling unit arranges all timestamped digital flow rate values ​​in chronological order of sampling time, forming a digital flow rate value sequence. Similarly, all timestamped digital pressure values ​​are arranged in chronological order, forming a digital pressure value sequence. The fifth step involves the formation and storage of raw flow time-series data. The generated digital flow rate value sequence and digital pressure value sequence are integrated, establishing a one-to-one correspondence between them through the same timestamp. The integrated dataset is the raw flow time-series data. Using the export function of the signal acquisition unit, the raw flow time-series data is exported and stored in a dedicated data storage device. The storage format is a common text or table format for easy retrieval during data processing. Simultaneously, the stored data undergoes an integrity check to confirm the absence of missing data, incorrect timestamps, or other issues. If any problems are found, sampling is performed again.

[0042] By employing miniaturized sensors and rationally planning the installation location and wiring layout, the limited space at the outlet end of the one-way pressure valve is fully utilized without occupying the core transmission area of ​​the engine. At the same time, no additional modifications are required to the original engine structure, ensuring the structural integrity and operational stability of the engine.

[0043] In a preferred embodiment of the present invention, step 3 above involves performing multi-dimensional feature analysis on the original flow time-series data to obtain multiple process parameters with intrinsic correlation; quantifying the collaborative relationship between different parameter change directions by calculating the angle between the tangents to the trajectory curves of the process parameters at feature points; and analyzing the steady-state and transient response characteristics based on the collaborative relationship to establish parameterized correction rules and generate a corresponding set of correction parameters, which may include:

[0044] In this embodiment of the invention, step 330 involves calculating the mean and variance of the original traffic time series data, identifying the peak value of the time series data, and the corresponding rise and fall times. The mean, variance, peak value, rise time, and fall time are then used to form a first process parameter set. Specifically, this includes: First, filtering and organizing the original traffic time series data. The original traffic time series data stored in step 223, including digital traffic values, digital pressure values, and corresponding timestamps, is exported to a data processing computer. A new Excel spreadsheet is created, and only the timestamp and digital traffic value columns are extracted to form a separate digital stream. A value sequence table is created; then, outlier removal is performed. Each numerical flow value is examined row by row and compared with the average of the three adjacent data points. If the difference between the value and the average exceeds twice the average, it is considered an outlier. For example, if the average of the three adjacent data points is 50 L / s, and the current value is 120 L / s, the difference of 70 L / s exceeds twice that of 50 L / s, thus it is considered an outlier. The row containing the outlier is then marked and deleted. After removal, the table data is reordered in ascending order of timestamp to ensure the temporal continuity of the data and avoid errors caused by sampling processes. Minor time deviations cause data misalignment; the second step is mean calculation. First, count the total number of rows in the digital flow value sequence in an Excel spreadsheet, i.e., the total number of digital flow values, let's say N. Then, use the summation formula in a blank column of the spreadsheet to add up all the digital flow values ​​sequentially, obtaining the sum of all digital flow values, denoted as S. Finally, divide the sum S by the total number N to calculate the mean of the original flow time series data, denoted as μ. That is, mean μ = sum S ÷ total number N. For example, if the total number N = 1000 and the sum S = 50000 L / s, then the mean μ = 50000 ÷ 1000. =50L / s, this mean reflects the average airflow rate during the entire collection period; the third step is variance calculation. Based on the mean calculation, a new column for difference is added to the table. Each digital flow rate value is calculated row by row, denoted as the difference between xᵢ and the mean μ, i.e., difference = xᵢ - μ. Then, a new column for squared difference is added, and each difference is squared, i.e., squared difference = difference × difference. Subsequently, all squared differences are summed to obtain the total sum of squared differences, denoted as Q. Finally, the total sum of squared differences Q is divided by the total number of digital flow rate values ​​N to obtain the variance of the original flow time series data, denoted as σ. 2 That is, variance σ 2 = Sum of squared differences Q ÷ Total number of differences N, for example, Sum of squared differences Q = 10000 (L / s) 2 If the total number of items N = 1000, then the variance σ 2 =10000÷1000=10 (L / s) 2 The variance reflects the dispersion of the flow data; the larger the variance, the more drastic the flow fluctuations.

[0045] The fourth step is peak identification. The processed digital flow rate value sequence table is iterated row by row. Starting from the fourth row, ensuring there are three adjacent data points before and after the current row, the current row's digital flow rate value is compared with the values ​​of its three preceding and three following adjacent data points. If the current value is greater than both the values ​​of its three preceding and three following adjacent data points, it is initially identified as a potential peak. To avoid misjudging small fluctuations caused by noise interference as peaks, a peak threshold is further set: Peak threshold = Mean μ + 3 × Variance σ, i.e., the mean plus three times the variance. For example, if the mean μ = 50 L / s, the variance σ = ... If the peak flow rate is approximately 3.16 L / s, then the peak threshold is 50 + 3 × 3.16 ≈ 59.48 L / s. The potential peak is compared with the peak threshold. If the potential peak is greater than the peak threshold, it is confirmed as a valid peak. The corresponding row in the table is marked as the valid peak, and the corresponding digital flow value and timestamp are recorded. The fifth step is to calculate the rise and fall times. For each confirmed valid peak, the 10% and 90% values ​​of the peak are calculated: 10% peak value = valid peak value × 0.1, 90% peak value = valid peak value × 0.9. Using the timestamp corresponding to the peak value as a baseline, trace the data backward, examining the digital flow values ​​row by row from the row containing the peak value upwards. Find the time point corresponding to the first rise to 90% of the peak value, denoted as the rise start time t1. Continue tracing backwards to find the time point corresponding to the first rise to 10% of the peak value, denoted as the rise initial time t0. Rise time = rise start time t1 - rise initial time t0. Then, trace the data backwards from the row containing the peak value to find the time point corresponding to the first drop to 90% of the peak value, denoted as the drop start time t2. Continue tracing backwards to find the first drop... The time point corresponding to the drop to 10% of the peak value is denoted as the descent termination time t3; descent time = descent termination time t3 - descent start time t2. For example, effective peak value = 60 L / s, 10% of peak value = 6 L / s, 90% of peak value = 54 L / s; initial rise time t0 = 0.05 s, rise start time t1 = 0.06 s, then rise time = 0.06 - 0.05 = 0.01 s; descent start time t2 = 0.07 s, descent termination time t3 = 0.08 s, then descent time = 0.08 - 0.07 s. =0.01s; Step 6: Construct the first process parameter set. Create a new Word document as the parameter set record document. Create a table for the first process parameter set in the document. The table columns include parameter type, specific value, corresponding timestamp, and remarks. Fill in the calculated mean and variance into the table. Fill in the size of all effective peaks, the corresponding timestamp, and the rise time and fall time of each peak into the table in sequence. Mark the number of each peak in the remarks column. After all the data is filled in, confirm that there are no omissions or errors. This table is the first process parameter set.

[0046] Step 331: Based on the peak position in the first process parameter set and combined with the rise time, extract the corresponding key data segments from the original traffic time series data; perform Fourier transform on the key data segments and extract the signal energy distribution characteristics from the transformed spectrum to form the second process parameter set; specifically including: First, determine the key data segment extraction range. Open the first process parameter set table, check the peak position and corresponding rise time of each valid peak, and for each valid peak, determine the extraction range centered on its timestamp. The forward extension time = the rise time of the peak × 2, and the backward extension time = the fall time of the peak × 2; for example, if a peak timestamp = 0.1s, rise time = 0.01s, fall time = 0.01s, then extend forward by 0.02s (extraction start time = 0.1 - 0.02 = 0.08s), and extend backward by 0.02s (extraction end time = 0.1 + 0.02 = 0.08s). The peak value is 0.02 = 0.12s, and the corresponding cutoff range is 0.08s-0.12s. If the cutoff ranges of two adjacent peaks overlap, such as the cutoff end time of the previous peak = 0.12s and the cutoff start time of the next peak = 0.11s, the two cutoff ranges are merged into a continuous range, which is the cutoff start time of the previous peak to the cutoff end time of the next peak, to avoid duplicate data cutoff. The second step is to extract key data segments. Return to the sorted digital flow value sequence Excel table, and according to each determined cutoff range, filter out all data rows with timestamps within the range, including the corresponding timestamps and digital flow values. Copy the data in each cutoff range to a new Excel worksheet, and name each worksheet with the corresponding peak number. Mark the start and end times of the cutoff range in the header of each worksheet to ensure that the source of each key data segment is clear and traceable.

[0047] The third step is the preparation and operation of the Fourier transform. Export the Excel worksheet containing each key data segment as a CSV file for easy reading by the signal analysis equipment. Open the dedicated signal analysis equipment and import the CSV file. First, zero-mean normalization is performed on the key data segments. The equipment automatically calculates the mean of all digital flow values ​​within the key data segment, and then subtracts this mean from each digital flow value to obtain the zero-mean normalized data stream, thereby eliminating the influence of the DC component on the Fourier transform result. After processing, select the Fourier transform function on the equipment's operation interface, set the number of transform points to an integer multiple of the number of data points in the key data segment, and start the transform operation. The equipment converts the time-domain signal of the key data segment into a frequency-domain signal and automatically generates a spectrum graph. The horizontal axis of the spectrum graph represents frequency, and the vertical axis represents signal amplitude. The fourth step is the extraction of signal energy distribution characteristics. First, determine the effective frequency range of the spectrum. According to the Nyquist criterion for signal acquisition, the effective frequency range is from 0 to half of the uniform sampling frequency set in step 222. For example, if the uniform sampling frequency is 1000Hz, then the effective frequency range is 0-500Hz. Frequency components outside this range are invalid interference signals and can be ignored. Divide the effective frequency range into multiple consecutive frequency intervals at 10Hz intervals. Read the signal amplitude corresponding to all frequency points in each frequency interval on the spectrum. Square each amplitude, and then add all the squared amplitudes in the same interval to obtain the energy value of each frequency interval. Add the energy values ​​of all frequency intervals to obtain the total energy value. Divide the energy value of each interval by the total energy value to obtain the energy percentage of each interval. For example, the energy value of a certain interval is 50 (L / s). 2 Total energy value = 1000 (L / s) 2 Then, the energy percentage of this interval = 50 ÷ 1000 = 5%. At the same time, find the frequency point with the largest amplitude on the spectrum graph. This frequency point is the main frequency. Record the frequency value corresponding to the main frequency and the energy percentage of the interval in which the main frequency is located. Organize the energy percentage of each frequency interval, the main frequency value, and the main frequency energy percentage of each key data segment into a table, which is the signal energy distribution characteristic of the key data segment. Fifth step, construct the second process parameter set. In the parameter set record document, create a new second process parameter set table. The table columns include the corresponding key data segment, the energy percentage of each frequency interval, the main frequency value, the main frequency energy percentage, and remarks. Fill in the signal energy distribution characteristics of all key data segments into the table in sequence. Mark the corresponding peak number in the remarks column to ensure that each feature can be clearly associated with the peak in the first process parameter set. After all the data is organized, the table is the second process parameter set.

[0048] Step 332 involves associating and combining the first process parameter set with the second process parameter set to form multiple process parameters with inherent correlations. Specifically, this includes: First, parameter correlation analysis. Simultaneously, open the first and second process parameter set tables in the parameter set record document, and systematically analyze the correlations of the parameters in both sets. For example, analyzing the peak value and the proportion of dominant frequency energy: the peak value reflects the instantaneous maximum value of the airflow; a larger peak value indicates a greater airflow impact intensity, and the corresponding dominant frequency energy proportion is usually higher, thus showing a positive correlation. Analyzing the rise time and the proportion of high-frequency range energy: the rise time reflects the flow rate rising from 10% peak value to 90% peak value. The rate of increase and the shorter the rise time indicate more drastic flow changes. Dramatic changes generate more high-frequency components, thus the energy proportion of the high-frequency range will be higher, and the two are negatively correlated. Analyzing the variance and the energy proportion distribution across the entire frequency range, the variance reflects the overall degree of flow fluctuation. The larger the variance, the more chaotic the flow fluctuation, and the more uniform the energy proportion distribution across each frequency range, with no obvious concentrated dominant frequency component. The two are correlated. Through this comparative analysis, the specific correlation logic between the parameters in the two sets is clarified. The second step is parameter correlation combination. Based on the correlation logic obtained from the analysis, the parameters with inherent correlation in the two sets are paired and combined to form new process parameters. The specific combination... The methods include: First, pairing time-domain parameters with frequency-domain parameters. For example, pairing the peak value of the first process parameter with the dominant frequency energy ratio of the second process parameter to form a combined parameter of peak value and dominant frequency energy ratio, which reflects the correspondence between the instantaneous peak value of the flow and the degree of energy concentration at the dominant frequency; pairing the rise time with the energy ratio of the high-frequency interval to form a combined parameter of rise time and high-frequency interval energy ratio, reflecting the correspondence between the flow rise rate and the distribution of high-frequency components; pairing the variance with the energy ratio distribution of the entire frequency interval to form a combined parameter of variance and the energy distribution of the entire frequency, reflecting the correspondence between the overall fluctuation of the flow and the energy distribution of each frequency component; Second, retaining independent core parameters for two sets. Parameters that have a clear physical meaning and directly reflect the core characteristics of flow changes are retained as independent process parameters without needing to be combined. The third step involves organizing combined parameters. A new combined process parameter list table is created in the parameter set record document. The table includes columns for parameter name, parameter composition, physical meaning, and correlation basis. All newly combined parameters and retained independent parameters are sequentially filled into the table. For example, for the parameter composition of peak value and dominant frequency energy ratio, the peak value of the first process parameter and the dominant frequency energy ratio of the second process parameter are entered. The physical meaning reflects the correspondence between the instantaneous peak flow and the degree of concentration of dominant frequency energy. The correlation basis states that the larger the peak value, the higher the dominant frequency energy ratio. After filling in the list, each parameter is reviewed, and duplicates or highly correlated parameters are deleted, ultimately forming multiple process parameters with inherent correlation.

[0049] Step 333: Select process parameters reflecting torque changes from engine operating data and extract the original data sequence of each parameter's continuous change over time. Specifically, this includes: First, determining the selection criteria for parameters reflecting torque changes. Based on the working principle of a single-cylinder rotary engine, the selection criteria are clarified. The magnitude of the engine output torque is directly related to the changes in gas pressure inside the compression chamber, the rate of change of airflow, and the degree of fluctuation. The larger the peak pressure in the compression chamber and the faster the pressure change rate, the greater the engine output torque. The larger the instantaneous peak airflow and the more drastic the flow rate change, the more obvious the corresponding torque fluctuation. Simultaneously, referring to historical data from engine bench tests, torque change curves are extracted from the historical data. By comparing the change curves of different process parameters, it is clear that the changing trends of parameters such as peak value and main frequency energy ratio, rise time and high frequency range energy ratio, mean, and variance are completely consistent with the torque change trend. Therefore, the changing characteristics of these parameters are used as the core basis for screening. The second step is parameter screening. Based on the above screening basis, each line in the combined process parameter list formed in step 332 is selected. The parameters that combine peak value and main frequency energy ratio, rise time and high frequency range energy ratio, mean independent parameter, and variance independent parameter are screened. For each selected parameter, it is marked in the list as selected and reflects torque change. Unselected parameters are marked as unselected and have a weak correlation with torque to ensure that the screening process is traceable.

[0050] The third step is the extraction of the raw data sequence. For each selected process parameter reflecting torque changes, the value of the parameter changing continuously over time is extracted from the raw flow time series data table and signal analysis results. For the parameter combining peak value and main frequency energy ratio, the timestamp of the peak value in the first process parameter is used as the timestamp of the combined parameter, and the corresponding value is the data corresponding to the peak value and the main frequency energy ratio. For the parameter combining rise time and high frequency range energy ratio, the timestamp of the peak value corresponding to the rise time is used as the timestamp, and the value is the data corresponding to the rise time and the high frequency range energy ratio. For the mean independent parameter, if it is a global mean, the time range is marked as the entire collection period, and the value is the global mean. If it is a moving mean, the timestamp of the middle sampling point of each moving window is used as the timestamp, and the value is the moving mean of the corresponding window. For the variance independent parameter, the timestamp and value of the global variance or moving variance are extracted in the same way. The timestamp and value data of each parameter are organized into a separate Excel table in chronological order of the timestamps to form the raw data sequence of each parameter changing continuously over time.

[0051] Step 334: Based on the original data sequence, construct the continuous dynamic change characteristics of each process parameter over time, and identify and mark extreme points and inflection points as feature points on the dynamic change characteristics of each parameter. Specifically, this includes: First, constructing the dynamic change characteristics: Import the original data sequence Excel spreadsheet for each parameter into the data processing software. Create a new plotting project in the software, using the timestamp as the x-axis and the parameter value as the y-axis. Select the line graph type and plot a continuous curve of each parameter changing over time. After plotting, smooth the curve by averaging the values ​​of five adjacent data points. Replace the value of each data point on the curve with the average of the two data points before and after it. This eliminates curve fluctuations caused by minor noise interference while preserving the overall trend of the curve. The smoothed continuous curve represents the continuous dynamic change characteristics of each process parameter over time. The first step is to identify and label extreme points. For each smoothed dynamic change feature curve, the data point viewing function is enabled in the data processing software. Click on each data point on the curve to view its corresponding timestamp and parameter value. Starting from the beginning of the curve, traverse point by point, comparing the value of the current data point with the values ​​of the five adjacent data points before and after it. If the value of the current data point is greater than the values ​​of the five adjacent data points before and after it, the point is determined to be a maximum point. If the value of the current data point is less than the values ​​of the five adjacent data points before and after it, the point is determined to be a minimum point. The identified maximum and minimum points are uniformly labeled as extreme points. In the software, maximum points are marked with red triangles and minimum points are marked with blue circles. At the same time, the timestamp and parameter value of the point are marked next to each mark. After the marking is completed, the information of all extreme points is organized into a table to form an extreme point list.

[0052] The third step is inflection point identification and calibration. An inflection point is a point where the trend of a curve changes. Inflection points are identified by analyzing changes in the curve's slope. First, the slope of each data point on the curve is calculated in the data processing software. The current data point and its next value are selected, and the value of the current data point is subtracted from the value of the next data point. Then, this subtracted value is divided by the time difference between the two data points (i.e., slope = (next point value - current point value) ÷ (next point timestamp - current point timestamp)). All the slope values ​​are then organized into a slope sequence. Next, the slope sequence is iterated through, checking the sign of each slope value relative to its adjacent slope values. If the current slope value is positive and the adjacent slope values ​​are negative, or vice versa, then the slope has a positive sign. When the sign changes, the curve data point corresponding to the current slope value is the inflection point. For example, if the slope of the current data point is 5 (positive), the slope of the previous data point is -3 (negative), and the slope of the next data point is -4 (negative), then the current data point is the inflection point. Mark the inflection point with a green square, and label its timestamp and parameter value. Organize all inflection point information into a table to form an inflection point list. The fourth step is to summarize the feature points by merging the extreme point list and the inflection point list to form a total feature point list. The table items include parameter name, feature point type, timestamp, parameter value, and marking symbol. Review the list one by one to ensure that there are no duplicate markings or omissions of feature points, and that the timestamp and value of each feature point accurately correspond to the curve data. Finally, the identification and calibration of feature points are completed.

[0053] Step 335: For each calibrated feature point, calculate the tangent direction angle representing the instantaneous change trend, and pair the tangent direction angles of the dynamic change characteristics of different process parameters at the feature points corresponding to the time, calculate the angle between each pair of direction angles, and quantify the cooperative relationship between the change directions of each parameter; specifically, this includes: First, tangent direction angle calculation: For each feature point in the feature point list, select one adjacent data point before and after the feature point on its corresponding dynamic change characteristic curve. The time interval between the adjacent data point and the feature point should be the same to ensure that the two selected points accurately reflect the tangent direction of the curve at the feature point. For example, if the feature point timestamp = 0.1s, select a data point 0.099s forward (point A) and a data point 0.101s backward (point B). In the data processing software, connect point A and point B with a straight line; this straight line is the approximate tangent at the feature point, with the horizontal direction to the right as the reference direction, i.e., the direction of time increase. The first step involves using the angle measurement tool in the software to measure the angle between the tangent and the reference direction, which is the tangent direction angle. If the tangent slopes upward, the angle is positive; if the tangent slopes downward, the angle is negative. For example, if the tangent slopes upward by 30°, the direction angle is +30°; if it slopes downward by 20°, the direction angle is -20°. The tangent direction angle of each feature point is then added to the new column "Tangent Direction Angle" in the feature point master list. The second step is to pair the direction angles. The feature point master list is sorted by timestamp from smallest to largest. Each feature point's timestamp is examined row by row to find feature points with identical timestamps or a time difference of less than 0.001s. These feature points represent characteristic changes in different process parameters occurring at or near the same time, corresponding to consistent operating conditions. For example, the peak value and the main frequency energy ratio parameter have a maximum point at 0.1s, and the rise time and the high-frequency range energy ratio parameter have an inflection point at 0.1005s, with a time difference of 0.0005s, which is less than 0.In 001s, the tangent direction angles of the two feature points can be paired. The paired feature point information is organized into a direction angle pairing list. The table includes the pairing number, parameter 1 name, parameter 1 feature point information, parameter 1 direction angle, parameter 2 name, parameter 2 feature point information, parameter 2 direction angle, and time difference, ensuring that the correspondence of each pair of direction angles is clear. The third step is to calculate the included angle and quantify the cooperative relationship. For each pair of tangent direction angles in the direction angle pairing list, the included angle is calculated in two cases. First, if both direction angles are positive or both are negative, the included angle is obtained by subtracting the smaller direction angle value from the larger one. For example, if direction angle 1 = +30° and direction angle 2 = +15°, the included angle = 30° - 15° = 15°; if direction angle 1 = -25° and direction angle 2 = -10°, the included angle = (-10° - 15°) / ... (°) - (-25°) = 15°; The second method involves two direction angles, one positive and one negative. The absolute values ​​of the two direction angles are added together to obtain the included angle. For example, direction angle 1 = +20°, direction angle 2 = -10°, included angle = 20° + 10° = 30°. The included angle directly reflects the synergistic relationship between the parameter changes. The smaller the included angle, the closer the changes in the two parameters are, and the stronger the synergy, meaning the two parameters change in the same direction with the operating conditions. The larger the included angle, the greater the difference in the changes in the two parameters, and the weaker the synergy, meaning the two parameters change in different directions with the operating conditions. Add columns for included angle size and synergy level to the direction angle pairing list. The synergy level is divided according to the included angle size, such as included angle < 30° for strong synergy, 30° ≤ included angle < 60° for medium synergy, and included angle ≥ 60° for weak synergy, thus quantifying the synergistic relationship.

[0054] Step 336: Based on the cooperative relationship, distinguish between the engine's steady-state and transient operating conditions, and establish parameterized correction rules for the correlation angle and torque detection error based on the differences in the cooperative relationship angle under different operating conditions. Specifically, this includes: First, determining the basis for distinguishing operating conditions. Combining the operating characteristics of a single-cylinder rotary engine, clarifying the core differences between steady-state and transient operating conditions. Under steady-state conditions, the engine's speed, load, and other operating parameters remain stable, and the change trends of each process parameter are smooth. Therefore, the cooperative relationship between the different parameter change directions is strong, and the corresponding angles are generally small and change smoothly. Under transient operating conditions, the operating parameters fluctuate rapidly, and the change trends of each process parameter are drastic and unstable. The cooperative relationship between the parameter change directions is weak, and the corresponding angles will suddenly increase and change significantly. Therefore, the size of the cooperative relationship angle and the amplitude of the angle change are used as the core basis for distinguishing operating conditions. Second, steady-state and transient operating conditions... Transient operating conditions are distinguished by sorting the azimuth angle pairing list by timestamp. An analysis window is defined as 10 rotor rotation cycles (rotation cycle = 60 ÷ engine speed; for example, if engine speed = 10000 r / min, rotation cycle = 0.006 s, 10 cycles = 0.06 s). The angle size and variation range are analyzed window by window. If all paired angles within a window are less than 30°, and the variation range between two adjacent paired angles is less than 5°, the engine is considered to be in steady-state operating condition within the corresponding time period. If, at a certain moment, multiple paired angles suddenly increase to over 60°, and the variation range between two adjacent paired angles is greater than 20°, the engine is considered to have transitioned from steady-state to transient operating condition. A new operating condition type column is added to the azimuth angle pairing list, marking the operating condition corresponding to each pairing. The start and end timestamps of each operating condition are also recorded to form an operating condition classification list.

[0055] The third step involves correlation analysis between the angle and torque detection error. The engine is mounted on a test bench, started, and runs according to a preset program. Simultaneously, the torque detection process of this invention and the bench-specific high-precision torque sensor are activated. During the test, the following data are recorded synchronously: the angle of coordination between parameters of each process under different operating conditions, the torque detection value calculated based on the original data of this invention, and the actual torque value measured by the bench torque sensor. After the test, the torque detection error for each operating condition is calculated as: Torque detection error = Torque detection value - Actual torque value. The torque detection error is then correlated with the corresponding angle of coordination. A comparative analysis of angles was conducted to statistically analyze the error patterns under different angle ranges and operating conditions. For example, under steady-state conditions, the torque detection error was small when the angle was <15°; the error increased slightly when the angle was 15°-30°; under transient conditions, the torque detection error corresponding to the same angle range was larger than that under steady-state conditions, and the greater the change in angle, the more significant the increase in error. The analysis results were compiled into an angle error correlation table to clarify the error variation patterns under different conditions. The fourth step was to establish parameterized correction rules. Based on the patterns in the angle error correlation table, parameterized correction rules were established. One rule was the steady-state correction rule, which was based on the angle... The first type of correction rule uses a fixed correction coefficient. The smaller the angle, the smaller the correction coefficient; the larger the angle, the larger the correction coefficient. For example, when the angle is <15°, the correction coefficient is 0.98, meaning the corrected torque value equals the original measured value × 0.98. When the angle is 15°-30°, the correction coefficient is 0.95. The second type is a transient condition correction rule, which uses a dynamic correction coefficient. In addition to considering the angle range, it also needs to add the correction increment corresponding to the angle change amplitude. The larger the angle change amplitude, the larger the correction increment. For example, when the angle is 30°-60° and the change amplitude is 20°-30°, the base correction coefficient is 0.92, and the superimposed correction... With an increment of 0.03, the final correction coefficient is 0.89. When the included angle is greater than 60° and the change range is greater than 30°, the basic correction coefficient is 0.85, and with the addition of a correction increment of 0.05, the final correction coefficient is 0.80. At the same time, the applicable conditions are clearly defined in the correction rules, such as the corresponding engine speed range (8000-15000r / min) and load range (20%-100%), to ensure the relevance of the rules. All correction rules are compiled into a parameterized correction rule manual, which includes the rule number, applicable operating conditions, included angle range, included angle change range, correction coefficient, correction formula, remarks, etc.

[0056] Step 337: Based on the parameterized correction rules, generate a set of correction parameters including filter coefficients for signal processing and phase compensation parameters for timing alignment. Specifically, this includes: First, determining the filter coefficients. According to the correction requirements for different operating conditions and angle ranges in the parameterized correction rule manual, determine the filter coefficients used for signal processing. A larger filter coefficient results in stronger filtering and better noise filtering, but may lose some signal details. Under steady-state conditions, parameter changes are stable, and noise interference is relatively small. Therefore, set a smaller filter coefficient according to the angle range. When the angle is <15°, the filter coefficient = 0.1 (weak filtering, retaining more original signal details); when the angle is 15°-30°, the filter coefficient... =0.2 (medium filtering, balancing noise filtering and signal preservation); Under transient conditions, parameter changes are drastic, noise interference is strong, and the larger the angle and the greater the change amplitude, the more noise there is. Therefore, a larger filter coefficient is set and an increment is added. When the angle is 30°-60° and the change amplitude is 20°-30°, the basic filter coefficient = 0.3, the increment is 0.05, and the final filter coefficient = 0.35; when the angle is >60° and the change amplitude is >30°, the basic filter coefficient = 0.4, the increment is 0.1, and the final filter coefficient = 0.5 (strong filtering, effectively filtering high-frequency noise). All filter coefficients are compiled into a filter coefficient table, and the corresponding applicable conditions, angle range, angle change amplitude and coefficient value are marked.

[0057] The second step is to determine the phase compensation parameters. Because of differences in sensor installation position, signal transmission path length, and sensor response speed for different process parameters, phase differences exist between signals of different parameters, affecting torque detection accuracy. Through bench test data, the phase difference between each parameter signal and the standard signal under different operating conditions is measured. Under steady-state conditions, the phase difference is relatively stable. A fixed phase compensation amount is set according to the corresponding angle range, i.e., the translation time of the signal on the time axis. For example, when the steady-state angle is <15°, the phase difference = 0.0002s, and the phase compensation amount is set to 0.0002s; when the steady-state angle is 15°-30°, the phase difference = 0.0003s, and the compensation amount = 0.0003s. Under transient conditions, the phase difference increases with the increase of the angle change amplitude. Therefore, the phase difference is dynamically adjusted according to the angle change amplitude. The compensation amount is set as follows: for a change range of 20°-30°, the compensation amount is 0.0005s; for a change range > 30°, the compensation amount is 0.0008s. The adjustment frequency of the compensation amount is set to be consistent with the frequency of the angle change. All phase compensation parameters are compiled into a phase compensation parameter table, indicating the applicable operating conditions, angle range, change range, compensation amount, and adjustment rules. The third step involves generating and organizing the correction parameter set. The filter coefficient table and phase compensation parameter table are integrated to generate the correction parameter set. The parameter set is organized in folders, containing two subfolders: steady-state operating condition parameters and transient operating condition parameters. These subfolders store the filter coefficients and phase compensation parameter tables for the corresponding operating conditions. Simultaneously, a user manual for the correction parameter set is compiled, clearly defining the meaning, calling conditions, usage methods, and precautions for each parameter. The parameter set and user manual are then packaged and stored.

[0058] By integrating time-domain and frequency-domain parameters to form correlated parameters, the intrinsic relationship between flow data and torque changes can be fully explored. Compared with single-dimensional analysis, it can more accurately capture the core driving factors of torque fluctuations. Based on the quantitative results of parameter synergy, steady-state and transient conditions can be distinguished, thus improving the adaptability of detection.

[0059] In a preferred embodiment of the present invention, step 4 above, using a set of correction parameters, synchronously filters and compensates the flow signal in the original flow time series data with the real-time speed and angular position signals of the single-cylinder rotor; by identifying and fitting the key extreme points of the flow signal waveform within a cycle, constructing a feature polygon and obtaining the geometric axis of symmetry, and determining the flow characteristic parameters corresponding to the rotor's periodic motion based on the position of the axis of symmetry, may include:

[0060] In this embodiment of the invention, step 440 involves constructing a corresponding digital low-pass filter based on the preset filter coefficients in the correction parameter set, and using the digital low-pass filter to filter the original flow time series data to obtain the filtered flow signal. Specifically, this includes: First, extracting the preset filter coefficients. Using a data processing device, the previously stored correction parameter set document is opened. The current engine operating condition is first determined as either steady-state or transient, as determined by the multi-dimensional feature analysis described earlier. Then, based on the operating condition type, the corresponding preset filter coefficients are found and extracted from the correction parameter set. During extraction, the coefficient identification information needs to be verified to ensure that the extracted coefficients completely match the current operating condition, avoiding confusion between steady-state and transient coefficients. Second, constructing the digital low-pass filter. Using the extracted preset filter coefficients as the core basis, a digital low-pass filter is constructed using the data processing device. During construction, the cutoff frequency of the filter must first be determined. Specifically, the current real-time rotational speed of the single-cylinder rotor is read from the engine ECU (Electronic Control Unit). This real-time rotational speed is divided by 60 to obtain the fundamental frequency of the pressure signal under the current operating conditions. Since the pressure signal generated by the compression chamber completes a full cycle change with each rotation of the rotor, the fundamental frequency of the pressure signal is consistent with the rotor rotation frequency. Then, the calculated fundamental frequency of the pressure signal is multiplied by 1.3 to obtain the cutoff frequency of the digital low-pass filter. The purpose of setting this cutoff frequency is to ensure that the filter can filter out high-frequency interference signals higher than the fundamental frequency of the pressure signal, while fully preserving the effective components of the flow signal corresponding to the pressure signal. In addition, the order of the filter also needs to be set. The order is set to 4th order according to the filtering accuracy requirements to balance the filtering effect and signal delay. Finally, the construction of the digital low-pass filter is completed.

[0061] The third step involves filtering. Using data processing equipment, the original flow time-series data is retrieved from the storage unit. Digital flow value sequences are filtered out, and synchronously acquired digital pressure value sequences are discarded, retaining only flow-related data. This digital flow value sequence is then used as input data and imported into the constructed digital low-pass filter. The filtering process is then initiated, allowing the filter to remove high-frequency interference signals from the digital flow value sequence. The fourth step verifies the filtering effect and determines the filtered signal. After filtering, the data processing equipment displays the digital flow value sequences before and after filtering as waveforms, providing a visual comparison. The comparison criteria for the periodic characteristics of the two waveforms are as follows: the filtered waveform should clearly retain the periodic trend of the original waveform, the attenuation of the peak and trough values ​​in each period relative to the original waveform should not exceed 5%, and the waveform should not have obvious distortion, discontinuities, or additional noise peaks. If this standard is not met, it indicates that the filter coefficient or cutoff frequency setting is unreasonable, and the operator needs to readjust the filter coefficient. If it is a steady-state condition, the filter coefficient can be appropriately increased; if it is a transient condition, the cutoff frequency can be finely adjusted, the filter can be reconstructed, and the filtering process can be repeated until a filtered flow signal that meets the judgment criteria is obtained.

[0062] Step 441: Calculate the time compensation amount of the filtered flow signal relative to the real-time speed and angular position signal based on the phase compensation parameters in the correction parameter set. Specifically, this includes: First, acquiring basic data. Using a data processing device, the current real-time speed of the single-cylinder rotor is read from the engine ECU. The timestamp of the data is recorded during reading to ensure that the speed data and the currently processed filtered flow signal belong to the same time interval. Second, the correction parameter set is reopened, and the corresponding phase compensation parameters are extracted according to the current operating condition. After extraction, these parameters are checked against the operating condition identifier of the filtering coefficient to ensure that the phase compensation parameters and filtering coefficients belong to the same operating condition, avoiding parameter mismatch. Second, calculating the time compensation amount. According to the preset calculation logic, the time compensation amount is calculated using the data processing device. Specifically, 60 is divided by the real-time speed of the single-cylinder rotor read from the engine ECU to obtain the time compensation amount for one complete rotation of the single-cylinder rotor. The time required for one revolution is calculated. Then, this time is multiplied by the phase compensation parameter extracted from the correction parameter set to obtain the time compensation amount. The physical meaning of this time compensation amount is the deviation of the filtered flow signal from the real-time speed and angular position signal on the time axis. This deviation can be eliminated through compensation. The third step is to verify the rationality of the time compensation amount. After the calculation is completed, the rationality of the time compensation amount needs to be verified. The verification standard is that the absolute value of the time compensation amount cannot exceed one-tenth of the time required for one revolution of a single cylinder rotor. If it exceeds this range, it means that either the phase compensation parameter is extracted incorrectly or the real-time speed is not read accurately. At this time, the phase compensation parameter in the correction parameter set needs to be rechecked to confirm whether the parameter corresponding to the working condition has been extracted. At the same time, the real-time speed is read again from the engine ECU to check whether there are abnormal fluctuations in the speed data. After correction based on the verification results, the calculation process is executed again until a time compensation amount that meets the rationality requirements is obtained.

[0063] Step 442: Based on the time compensation amount, perform time-axis translation and interpolation resampling operations on the filtered flow signal to generate a phase-compensated flow signal. Specifically, this includes: First, performing a time-axis translation operation. Using a data processing device, retrieve the filtered flow signal obtained in step 440, observe its time-axis distribution, and based on the time compensation amount calculated in step 441, perform an overall translation of the filtered flow signal. If the time compensation amount is positive, it indicates that the filtered flow signal lags behind the real-time rotational speed and angular position signals, requiring the entire filtered flow signal to be translated backward. The shift duration equals the time compensation amount. If the time compensation amount is negative, it indicates that the filtered flow signal leads the real-time speed and angle position signals. The filtered flow signal needs to be shifted forward as a whole, with the shift duration equal to the absolute value of the time compensation amount. During the shift, it is necessary to ensure that the waveform shape of the signal remains unchanged, only changing the signal's position on the time axis. The second step is to perform interpolation resampling. Because data at some time points will be missing after the signal shift, for example, when shifting forward, the data at the very beginning will exceed the original time range; when shifting backward, the data at the very end will exceed the original time range. To ensure signal integrity and time resolution consistency, the shifted flow signal needs to be interpolated and resampled using data processing equipment. The resampling frequency must match the sampling frequency of the original flow time-series data, which was set in the signal acquisition stage. Linear interpolation is used; specifically, for time points with missing data, the flow data for that time point is calculated using linear fitting based on the values ​​of two adjacent valid data points before and after that time point, filling in the missing data. The third step is to verify the phase synchronization effect. After resampling, the data processing equipment... The compensated flow signal is overlaid with the waveforms of the real-time speed and angle position signals. The characteristic points of the two signals, such as the peak point of the speed signal and the 0° moment point of the angle position signal, are compared with the timestamps of the characteristic points of the flow signal. The verification standard is that the timestamp deviation of the corresponding characteristic points does not exceed one-tenth of a single sampling period. If this standard is not met, it indicates that the translation amount is not set accurately. The translation direction or translation duration needs to be readjusted, and the translation and interpolation resampling operations are performed again until the compensated flow signal is synchronized with the real-time speed and angle position signals in time, generating a phase-compensated flow signal.

[0064] Step 443: Based on the real-time speed and angular position signals of the single-cylinder rotor, extract a periodic flow waveform segment corresponding to a complete rotor rotation cycle from the filtered flow signal. Specifically, this includes: First, determining the extraction time range. Using a data processing device, synchronously read the real-time speed and angular position signals of the single-cylinder rotor from the engine ECU, ensuring the timestamps of the two sets of signals are perfectly aligned. First, calculate the time required for the single-cylinder rotor to complete a complete rotation cycle based on the real-time speed signal, calculated by dividing 60 by the real-time speed. Then, find the 0° mark in the angular position signal, using this 0° mark as the starting time point for extraction, and add the time of a complete rotation cycle to the 0° mark as the ending time point for extraction, thus determining the time range corresponding to a complete rotor rotation cycle. Second, performing waveform segment extraction. Using a data processing device, retrieve the phase-compensated flow signal obtained in step 442, and extract the waveform segment according to the determined extraction time range. The process involves filtering all flow data within the corresponding time range from the flow signal to form an independent flow waveform segment. This segment represents a periodic flow waveform segment corresponding to a complete rotor rotation cycle. During the extraction process, it is crucial to ensure that the flow data at both the start and end time points are fully included, without omitting any sampling point. The third step is to verify the completeness of the extraction. After extraction, the time length of the extracted flow waveform segment is checked using data processing equipment and compared with the calculated time of a single rotor rotation cycle. The verification standard is that the deviation between the extracted waveform segment time length and the calculated rotation cycle time should not exceed 5%. If the deviation exceeds this range, it indicates either an error in the real-time rotation speed calculation or inaccurate positioning of the 0° moment point of the corner position signal. In this case, the operator needs to recalculate the rotation cycle time or recheck the 0° moment point marker of the corner position signal, correct the error, and then perform the extraction operation again until a periodic flow waveform segment with the required time length is obtained.

[0065] Step 444: Perform extreme value analysis on the periodic flow waveform segment to identify all local maxima and minima in the waveform. Specifically, this includes: First, setting the sliding window parameters. Using the data processing device, open the extreme value analysis tool to analyze the captured periodic flow waveform segment. First, set the sliding window size. Based on the original sampling frequency and the fluctuation characteristics of the flow signal, set the sliding window size to 6 consecutive sampling points, and the window's sliding step size to 1 sampling point, meaning that after each slide, the window only moves backward by 1 sampling point. Second, sliding window traversal and extreme value judgment. Start the sliding window traversal program, allowing the window to slide sequentially from the first sampling point of the periodic flow waveform segment. For each of the 6 sampling points covered by the current window, focus on comparing the two middle sampling points with the points before and after the third and fourth sampling points within the window, using the position between the third and fourth sampling points as the center. The values ​​of adjacent sampling points are compared. If the value of the third sampling point in the window is greater than the values ​​of all the sampling points before and after it, then the third sampling point is determined to be a local maximum. If the value of the third sampling point in the window is less than the values ​​of all the sampling points before and after it, then the third sampling point is determined to be a local minimum. Similarly, the same judgment logic is applied to the fourth sampling point. The third step is to record the extreme point information. After completing the extreme value judgment for each window, if a local maximum or local minimum is identified, the specific information of the extreme point is recorded by the data processing device, including the timestamp corresponding to the extreme point, the corresponding flow value, and the position number of the extreme point in the periodic flow waveform segment. The process of sliding window traversal, extreme value judgment and information recording is repeated until the sliding window covers the entire periodic flow waveform segment to ensure that no local maximum or local minimum is missed.

[0066] Step 445: Merge all local maxima and minima into a group of key extrema, and sort them according to their chronological order of appearance in the waveform segment to generate an ordered sequence of extrema. Specifically, this includes: First, merging extrema: Using a data processing device, summarize the information of all local maxima and minima recorded in step 444 into a single data table. Remove any duplicate data from the table. Define all extrema without duplicates as a single group of key extrema. After definition, verify the total number of key extrema to ensure it matches the total number of extrema identified in step 444. To avoid data loss during the merging process, the second step is to sort and generate an ordered sequence. Using the timestamps of the key extreme points recorded in the data table as the sorting basis, all key extreme points are sorted in ascending order. Specifically, the key extreme point with the smallest timestamp is selected from the data table and used as the first point in the ordered extreme point sequence; then, the point with the smallest timestamp is selected from the remaining key extreme points and used as the second point in the sequence; and so on, until all key extreme points are sorted. After sorting, an ordered extreme point sequence is generated. The data processing device needs to record the arrangement order, timestamp, flow rate value, and corresponding corner position of each point in the sequence.

[0067] Step 446: Connect the points in the ordered extreme point sequence sequentially, and connect the first two points of the sequence to form a closed planar feature polygon. Specifically, this includes: First, establishing the connection relationships between points. Using a data processing device, retrieve the generated ordered extreme point sequence. Based on the sequence's order, establish the connection relationships between adjacent extreme points sequentially. Specifically, first select the first point in the sequence and connect it to the second point with a straight line; then select the second point and connect it to the third point with a straight line; following this logic, sequentially complete the straight line connections between the (n-1)th point and the nth point in the sequence, where n is the total number of points in the ordered extreme point sequence. During the connection process, ensure that each connection line is a straight line and only connects two adjacent points, without any cross-point connections. The second step involves closing the polygon to form a complete planar figure. After connecting adjacent points sequentially, the last point of the ordered extreme point sequence is selected and connected to the first point of the sequence with a straight line. This connection closes the previously formed polygonal figure, creating a complete planar figure. The third step verifies the validity of the planar feature polygon. The formed planar figure is viewed using data processing equipment to verify its validity. The verification criteria are: the figure has no breaks, no intersecting or overlapping lines, and the figure is a closed polygonal structure. If there are breaks, the connection relationship of adjacent points needs to be re-examined, and missing lines need to be added. If there are intersecting or overlapping lines, the sorting order of the ordered extreme point sequence needs to be rechecked, the sorting corrected, and the connection repeated until a closed planar feature polygon that meets the verification criteria is formed.

[0068] Step 447: Perform geometric analysis on the closed planar feature polygon, calculate and determine one geometric axis of symmetry of the feature polygon; specifically including: First, calculate the geometric center of the planar feature polygon. Using a data processing device, retrieve the coordinate data of all vertices of the planar feature polygon. First, summarize the X-axis coordinates of all vertices, divide the sum by the total number of vertices to obtain the X-axis coordinate of the geometric center; then, summarize the Y-axis coordinates of all vertices, divide the sum by the total number of vertices to obtain the Y-axis coordinate of the geometric center; combine the X-axis and Y-axis coordinates to obtain the geometric center of the closed planar feature polygon. After calculation, the results of coordinate summation and division operations need to be checked again to avoid calculation errors; Second, use the rotation method to find the geometric axis of symmetry. Using the calculated geometric center as the rotation center, perform a rotation operation on the planar feature polygon using the data processing device. The specific logic of the rotation operation is as follows: Each time, the polygon is rotated 1° clockwise. After rotation, the area of ​​the overlapping region between the rotated polygon and the original polygon is calculated. The symmetry is obtained by dividing the area of ​​the overlapping region by the total area of ​​the original polygon. The symmetry value corresponding to this rotation angle is recorded. Then, the operation of rotating, calculating symmetry, and recording data is repeated until the rotation angle covers all angles from 0° to 180°. The third step is to determine the geometric axis of symmetry. Using a data processing device, the set of data with the largest symmetry value is selected from the 180 sets of rotation angle symmetry data. The rotation angle corresponding to this set of data is the rotation angle of the geometric axis of symmetry. With the geometric center as the reference, a straight line is drawn according to this rotation angle. This straight line is the geometric axis of symmetry of the planar feature polygon. If there are multiple sets of data with the same symmetry value and all of them are the maximum value, the straight line that is consistent with the rotation direction of the rotor is selected as the final geometric axis of symmetry.

[0069] Step 448: Map the geometric symmetry axis to a coordinate system based on the single-cylinder rotor angle, and determine the angular orientation value of the geometric symmetry axis in the angular coordinate system as the final flow characteristic parameter used to characterize the current rotor rotation cycle. Specifically, this includes: First, constructing an angular reference coordinate system. Using data processing equipment, construct a two-dimensional coordinate system based on the single-cylinder rotor angle. The origin of this coordinate system is set to the position corresponding to the 0° moment of the single-cylinder rotor angle. The X-axis of the coordinate system is the direction of rotor angle increase, i.e., extending from the origin to the 360° angle direction, and the scale on the X-axis indicates the angle. The Y-axis of the coordinate system is the amplitude direction of the flow signal, and the scale on the Y-axis indicates the flow value. After the coordinate system is constructed, the coordinate axis names and scales must be clearly labeled to ensure accurate mapping. Second, mapping the geometric symmetry axis to the angular reference coordinate system. Using data processing equipment, retrieve the coordinate data of the geometric symmetry axis determined in step 447, and the timestamp-rotation value recorded in step 443. The angular position correspondence is derived from the real-time rotation speed: angular position = (timestamp difference ÷ time for one rotation) × 360°. Based on this correspondence, the timestamp X-axis coordinate of each point on the geometric symmetry axis is converted into the corresponding angular angle X-axis coordinate, while the flow rate value coordinate on the Y-axis remains unchanged. Through this conversion, the geometric symmetry axis is mapped from the time-flow rate coordinate system to the angular-flow rate coordinate system, obtaining the specific position of the geometric symmetry axis in the angular reference coordinate system. The third step is to calculate the angular azimuth value and determine the flow rate characteristic parameter. Using data processing equipment, the angle between the mapped geometric symmetry axis and the X-axis of the angular reference coordinate system is measured. This angle is the angular azimuth value of the geometric symmetry axis in the angular coordinate system. After the measurement is completed, the accuracy of the measurement results is checked to ensure that the angle measurement deviation does not exceed 0.5°. Finally, this angular azimuth value is determined as the flow rate characteristic parameter used to characterize the current rotor rotation cycle, and the rotor rotation cycle number, timestamp, and other information corresponding to this parameter are recorded.

[0070] By performing layered and progressive signal optimization and feature extraction operations, the reliability of flow characteristic parameters is improved. By setting filtering parameters, high-frequency interference components in the original flow signal can be specifically removed, avoiding the influence of interference signals on feature extraction and improving signal purity.

[0071] In a preferred embodiment of the present invention, step 5 above, which calculates the real-time output torque value of a single-cylinder rotor based on flow characteristic parameters and combined with the intake manifold absolute pressure and intake temperature under the current engine operating conditions, using a torque mapping model to achieve online torque monitoring of the engine operating state, may include:

[0072] In this embodiment of the invention, step 550 involves acquiring the real-time intake manifold absolute pressure and real-time intake temperature values ​​of the engine during the current working cycle. Specifically, this includes: First, confirming the current working cycle time range. Using a data processing device, the rotor rotation cycle information corresponding to the flow characteristic parameters recorded in step 448 is retrieved to clarify the timestamp range of the current working cycle. This serves as the time matching benchmark for acquiring pressure and temperature data, ensuring that the acquired data strictly corresponds to the current working cycle. Second, acquiring the real-time intake manifold absolute pressure value. Using a data processing device, data communication is established with the intake manifold absolute pressure sensor installed in the engine's intake system. The real-time pressure data output by the sensor is read. During the reading process, the timestamp corresponding to each set of pressure data needs to be recorded. Simultaneously, the sensor's operating status is checked to confirm that the sensor has no fault alarms and that the power supply is stable. Subsequently, from the read pressure data, pressure values ​​whose timestamps fall within the current working cycle time range are selected. If multiple pressure data exist within this time range, the average of these data is taken as the real-time intake manifold absolute pressure value for the current working cycle. If no pressure data exists within the time range, the pressure data from adjacent working cycles is used. The first step involves linear interpolation to supplement the pressure data, ensuring its integrity. The second step involves acquiring real-time intake air temperature data. This is done by establishing data communication with the intake air temperature sensor installed at the engine intake manifold inlet using data processing equipment. The real-time temperature data output by this sensor is read, and the timestamp corresponding to each set of temperature data is recorded. The sensor's operating status is checked, and temperature values ​​whose timestamps fall within the current working cycle time range are selected from the read temperature data. If multiple temperature data exist within this time range, their average value is taken as the real-time intake air temperature value for the current working cycle. If no corresponding temperature data exists, the same linear interpolation method as used to supplement the pressure data is employed to obtain the supplementary temperature value. The third step involves verifying the validity of the acquired real-time intake manifold absolute pressure and real-time intake air temperature values ​​using data processing equipment. The verification standard is that the intake manifold absolute pressure must be within the normal operating pressure range of the engine, and the intake air temperature must be within the normal operating temperature range of the engine. If the data exceeds the normal range, it is considered abnormal data, and the sensor connection status and data communication link must be checked again, and data is acquired again. If the data is still abnormal after re-acquisition, the abnormal information is recorded and a preliminary alarm is triggered.

[0073] Step 551: Combine the flow characteristic parameters, real-time intake manifold absolute pressure value, and real-time intake temperature value to form a set of multi-dimensional real-time operating condition parameters. Specifically, this includes: First, summarizing core parameters: Using data processing equipment, retrieve the current working cycle flow characteristic parameters determined in step 448, and the valid real-time intake manifold absolute pressure value and real-time intake temperature value obtained and verified in step 550. Summarize the specific values ​​of these three parameters, along with their corresponding current working cycle number and timestamp information, into a single data table to ensure that each parameter is accurately associated with the current working cycle. Second, verifying parameter correlation: Verify the correlation of the three summarized parameters. The verification standard is that the timestamps of all three parameters fall within the time range of the current working cycle, and the working cycle numbers are consistent with no cross-correlation. If abnormal parameter correlation is found, the parameters need to be readjusted. The first step involves retrieving the parameter data from the corresponding steps, correcting it, and then summarizing and verifying it again until all three parameters are fully correlated with the current work cycle. The second step involves constructing a multi-dimensional real-time operating condition parameter group. This group is formed by combining the flow characteristic parameters, real-time intake manifold absolute pressure, and real-time intake temperature values ​​in a fixed order: flow characteristic parameters - intake manifold absolute pressure - intake temperature. This creates a real-time operating condition parameter group containing three dimensions of information. After combination, a unique identifier code is assigned to this parameter group. The third step involves processing abnormal parameter groups. If a parameter is found to be missing or invalid during the combination process, the multi-dimensional real-time operating condition parameter group is determined to be an abnormal parameter group. The abnormal identifier code and the type of missing / invalid parameter are recorded. This parameter group is discarded, not included in subsequent torque calculations, and an alarm is triggered to remind the user to check the corresponding sensor or data processing stage.

[0074] Step 552: Input the multi-dimensional real-time operating condition parameters into the pre-calibrated torque mapping model to calculate the real-time output torque value of the single-cylinder rotor corresponding to the current working cycle. Specifically, this includes: First, retrieving the pre-calibrated torque mapping model. Using data processing equipment, the pre-calibrated torque mapping model is retrieved from the storage unit. During retrieval, the version information and calibration operating condition range of the torque mapping model must be verified to ensure that the retrieved model matches the current engine model and operating conditions. If the torque mapping model does not match, the corresponding version of the torque mapping model must be retrieved again. If no matching torque mapping model is found, the calculation is stopped and fault information is recorded. Second, the pre-calibration construction process of the torque mapping model. First, the single-cylinder rotary aero engine to be tested is fixedly mounted on the test bench platform, ensuring the engine is securely installed and precisely connected to the dynamometer. A high-precision intake manifold absolute pressure sensor is installed near the throttle valve on the engine's intake manifold, and a high-precision intake air temperature sensor is installed at the intake manifold inlet. A dynamic pressure sensing hole is then drilled on the axial end face of the single-cylinder rotor as described above, and a one-way pressure valve and a miniature flow sensor are installed. Simultaneously, it is ensured that all sensors communicate normally with the data acquisition equipment and the engine ECU. Furthermore, a test condition scheme covering the entire operating range of the engine is prepared, including different speed levels and different load levels to ensure comprehensive calibration conditions. The test is then conducted according to the pre-set test condition scheme. The engine is started sequentially and entered various test conditions. Under each condition, the engine is allowed to run stably for a preset time. Under each stable condition, three sets of core data are simultaneously recorded using data acquisition equipment: first, the flow characteristic parameters extracted in steps 440-448 above; second, the average values ​​of the intake manifold absolute pressure and intake temperature under that condition, collected by the intake manifold absolute pressure sensor and intake temperature sensor; and third, the actual output torque value of the single-cylinder rotor under that condition, directly measured by a dynamometer. After each test condition is completed, the condition number, all collected parameter values, and the actual torque value are recorded, forming a complete set of calibration data. This process is repeated until all pre-test conditions are met. The test and data acquisition under the established operating conditions yielded a calibration dataset covering all operating conditions. The acquired calibration dataset underwent preprocessing. First, outlier data was removed; if a set of calibration data contained parameter values ​​exceeding the sensor's measurement range or excessive fluctuations in actual torque values, it was identified as outlier and removed. Second, missing data was supplemented; if some parameter data was missing under a certain operating condition, two adjacent normal operating condition data points were selected nearby, and the missing data was supplemented using linear interpolation. Finally, all valid calibration data were normalized, converting flow characteristic parameters, intake manifold absolute pressure values, intake temperature values, and actual output torque values ​​to fixed numerical ranges, resulting in the preprocessed calibration dataset.

[0075] Using the preprocessed calibration dataset as input variables (flow characteristic parameters, intake manifold absolute pressure, and intake temperature) and the actual output torque as the output variable, a mapping relationship is constructed through data fitting. During the fitting process, different fitting methods are tried one by one, and the deviation between the fitting results and the actual torque values ​​is compared. The fitting method with the smallest deviation is selected as the final torque mapping model. The core logic of the fitting operation is to establish a mathematical correspondence rule between the input and output variables, ensuring that a specific set of flow characteristic parameters, intake manifold absolute pressure, and intake temperature values ​​can be used to calculate the torque mapping model according to this correspondence rule. The process involves: 1) Calculating a unique corresponding torque value; 2) Randomly selecting a portion of data from the preprocessed calibration dataset as validation data; 3) Substituting the input variables from the validation data into the constructed initial torque mapping model to calculate the torque value output by the model; 4) Comparing the calculated value with the actual torque value from the validation data to determine the degree of deviation; 5) If the deviation is within the preset allowable range, the torque mapping model accuracy meets the requirements, and the model construction is complete; 6) If the deviation exceeds the allowable range, the fitting method needs to be readjusted, or more test data from calibration conditions needs to be added, and the fitting and validation process repeated until the torque value is calculated. If the deviation between the calculated value output by the torque mapping model and the actual torque value is within the allowable range, the calibrated torque mapping model is obtained and stored in the storage unit of the data processing device for later use. The third step involves inputting parameters and executing calculations. Using the data processing device, the multi-dimensional real-time operating condition parameter set constructed in step 551 is input into the torque mapping model one by one according to the input order required by the model. The torque mapping model processes the input flow characteristic parameters, real-time intake manifold absolute pressure value, and real-time intake temperature value according to the pre-set correspondence rules between input and output variables, automatically calculating a corresponding torque value. The torque value is the calculated real-time output torque of the single-cylinder rotor under the current working cycle. The fourth step is to verify the rationality of the calculation result. Through data processing equipment, the calculated real-time output torque of the current working cycle is compared with the theoretical torque range under this condition. If the calculated value falls within the theoretical torque range, it is determined to be a reasonable result, and the calculated value is determined to be the real-time output torque of the single-cylinder rotor under the current working cycle. If the calculated value exceeds the theoretical torque range, it is determined to be an abnormal result. The operator needs to re-check the validity of the multi-dimensional real-time working condition parameter group and the matching of the torque mapping model, correct it, and substitute it into the calculation again until a reasonable real-time output torque value is obtained.

[0076] Step 553: Record the real-time output torque value of the single-cylinder rotor and perform parameter combination and torque mapping model calculations in each working cycle of the engine to achieve online torque monitoring of the engine's working state. Specifically, this includes: First, recording the real-time output torque value: Using a data processing device, the real-time output torque value of the single-cylinder rotor in the current working cycle, determined in step 552, along with the corresponding working cycle number, multi-dimensional real-time operating condition parameter group, and timestamp information, is stored in a local database. An index is created in chronological order during storage to facilitate querying and tracing historical data. Simultaneously, the data is backed up to a remote server to prevent local data loss. Second, displaying torque data in real-time: Using a data processing device, the real-time output torque value of the current working cycle is transmitted to the engine's monitoring interface, and the display content is updated in real-time. The display interface must clearly indicate the current torque value, the corresponding working cycle number, and the real-time time. Simultaneously, a waveform graph is used to synchronously display the torque change trend of multiple recent working cycles for intuitive understanding of the dynamic changes in torque. Third, setting an abnormal alarm mechanism: Using a data processing device, a torque abnormality threshold range is preset in the monitoring system. The threshold range is based on the engine... The rated torque and safe operation requirements are determined by setting a maximum allowable torque value and a minimum allowable torque value. The system compares the real-time output torque value of the current working cycle with the preset threshold range in real time. If the torque value is greater than the maximum allowable torque value or less than the minimum allowable torque value, an alarm mechanism is immediately triggered. The monitoring interface displays a red alarm prompt and sounds an alarm. At the same time, the abnormal torque value, the corresponding working cycle information, and the time of the abnormality are recorded to provide data support for fault diagnosis. The fourth step is to achieve continuous online monitoring. Through data processing equipment, a cyclic execution command is set. In each working cycle of engine operation, the complete process of step 550 (obtaining the absolute pressure value of the intake manifold and the intake temperature value of the current working cycle), step 551 (combining multi-dimensional real-time operating parameters), step 552 (substituting into the torque mapping model to calculate the real-time output torque value), and step 553 (recording and displaying the torque value) is automatically repeated without manual intervention. This ensures that the system can continuously and in real time acquire and monitor the torque data of each working cycle of the engine, realizing online torque monitoring of the engine's working state. If the engine stops running, the monitoring process automatically terminates; if the engine restarts, the monitoring process automatically restarts.

[0077] Accurate acquisition and validity verification of intake manifold absolute pressure and temperature ensure the reliability of operating parameters. Verification of parameter combinations and correlations avoids parameter confusion or mismatch. Real-time display and abnormal alarm mechanisms enable full traceability of torque data, improving engine operation safety and maintenance efficiency.

[0078] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0079] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0080] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for detecting torque in a single-cylinder rotary aero engine, characterized in that, The method includes: Step 1: Open a dynamic pressure sensing hole on the axial end face of the single-cylinder rotor of the engine, which communicates with the internal compression chamber, and integrate and install a one-way pressure valve at the dynamic pressure sensing hole. Step 2: Based on the dynamic pressure sensing orifice and the one-way pressure valve, the airflow flow and pressure coupling signal flowing through the valve body are collected in real time during engine operation to obtain the raw flow time series data. Step 3: Perform multi-dimensional feature analysis on the original flow time series data to obtain multiple process parameters with intrinsic correlation; quantify the cooperative relationship of different parameter change directions by calculating the angle between the tangents of the process parameter change trajectory curves at feature points, and analyze the steady-state and transient response characteristics based on the cooperative relationship, and establish parameterized correction rules to generate the corresponding correction parameter set. Step 4: Using the correction parameter set, synchronous filtering and phase compensation are performed on the flow signal in the original flow time series data and the real-time speed and angular position signal of the single-cylinder rotor; by identifying and fitting the key extreme points of the flow signal waveform within a cycle, a feature polygon is constructed and the geometric symmetry axis is obtained, and the flow characteristic parameters corresponding to the rotor periodic motion are determined according to the position of the symmetry axis. Step 5: Based on the flow characteristic parameters, combined with the intake manifold absolute pressure and intake temperature under the current engine operating conditions, the real-time output torque value of a single-cylinder rotor is calculated using a torque mapping model to achieve online torque monitoring of the engine operating state.

2. The torque detection method for a single-cylinder rotor aero-engine according to claim 1, characterized in that, A dynamic pressure sensing hole communicating with the internal compression chamber is opened on the axial end face of the single-cylinder rotor of the engine, and a one-way pressure valve is integrated and installed at the dynamic pressure sensing hole, including: The internal structure of the single-cylinder rotor of the engine is analyzed to determine the dynamic pressure region in which the compression chamber of the single-cylinder rotor generates regular high pressure during the rotation cycle, and the three-dimensional spatial coordinates of the dynamic pressure region are obtained. The three-dimensional spatial coordinates are projected onto the axial end face of the single-cylinder rotor. Based on the mapped position, a dynamic pressure sensing hole is formed on the axial end face of the single-cylinder rotor by machining. The dynamic pressure sensing hole is connected to the inside of the compression chamber, and the periodic pressure changes in the compression chamber are converted into periodic gas pressure signals exported through the dynamic pressure sensing hole. Extract the frequency and pressure amplitude characteristics of the periodic gas pressure signal, calculate the minimum gas flow capacity to transmit the pressure signal based on the frequency and pressure amplitude characteristics, and design and determine the final aperture size of the dynamic pressure sensing orifice based on the minimum gas flow capacity. Match the appropriate one-way pressure valve to the final orifice size, and integrate the one-way pressure valve into the external port of the dynamic pressure sensing orifice.

3. The torque detection method for a single-cylinder rotor aero-engine according to claim 2, characterized in that, Based on a dynamic pressure sensing orifice and a one-way pressure valve, the airflow and pressure coupling signals flowing through the valve body are collected in real time during engine operation to obtain raw flow time-series data, including: Based on the outlet specifications of the one-way pressure valve, a miniature flow sensor and a miniature dynamic pressure sensor are adapted and fixedly installed on the outlet end face. When the engine is started, the periodic pressure changes generated in the compression chamber drive the airflow through the installed sensor measurement end, simultaneously activating the miniature flow sensor and the miniature dynamic pressure sensor to sense the airflow in real time and generate two continuous analog electrical signals, namely the analog flow signal and the analog pressure signal. It receives two analog electrical signals, calculates the minimum sampling frequency that satisfies the signal based on the rated speed range of the engine rotor and the expected change period of the pressure signal, and sets a fixed value higher than the minimum frequency as the unified sampling frequency of the signal acquisition unit. Using a unified sampling frequency, the two analog electrical signals are synchronously and periodically sampled and converted from analog to digital by the signal acquisition unit to generate digital flow rate value sequences and digital pressure value sequences with strictly aligned timestamps, thus forming the original flow time series data.

4. The torque detection method for a single-cylinder rotary aero-engine according to claim 3, characterized in that, Multi-dimensional feature analysis of the raw traffic time-series data yielded several intrinsically related process parameters, including: Calculate the mean and variance of the raw flow time series data, and identify the peak value of the time series data as well as the corresponding rise time and fall time. Combine the mean, variance, peak value, rise time and fall time to form the first process parameter set. Based on the peak position in the first process parameter set and the rise time, the corresponding key data segments are extracted from the original flow time series data; Fourier transform is performed on the key data segments, and the signal energy distribution characteristics are extracted from the transformed spectrum to form the second process parameter set; The first set of process parameters is associated and combined with the second set of process parameters to form multiple process parameters with inherent relationships.

5. The torque detection method for a single-cylinder rotor aero-engine according to claim 4, characterized in that, By calculating the angle between the tangents to the trajectory curves of process parameter changes at characteristic points, the cooperative relationship of different parameter change directions is quantified. Based on this cooperative relationship, steady-state and transient response characteristics are analyzed, and parameterized correction rules are established to generate corresponding correction parameter sets, including: Select process parameters that reflect torque changes from engine operating data, and extract the raw data sequence of each parameter changing continuously over time; Based on the original data sequence, the continuous dynamic change characteristics of each process parameter over time are constructed, and extreme points and inflection points are identified and marked as feature points on the dynamic change characteristics of each parameter. For each calibrated feature point, the tangent direction angle representing the instantaneous change trend is calculated, and the dynamic change characteristics of different process parameters are paired at the feature points corresponding to the time. The angle between each pair of direction angles is calculated to quantify the cooperative relationship between the change directions of each parameter. Based on the cooperative relationship, the steady-state and transient operating conditions of the engine are distinguished, and parameterized correction rules for the correlation angle and torque detection error are established for the differences in the cooperative relationship angle under different operating conditions. Based on the parameterized correction rules, a set of correction parameters is generated, including filter coefficients for signal processing and phase compensation parameters for timing alignment.

6. The torque detection method for a single-cylinder rotor aero-engine according to claim 5, characterized in that, Using a set of correction parameters, synchronous filtering and phase compensation are performed on the flow signal in the original flow time series data and the real-time speed and angular position signals of the single-cylinder rotor, including: Based on the preset filtering coefficients in the correction parameter set, a corresponding digital low-pass filter is constructed, and the original flow time series data is filtered using the digital low-pass filter to obtain the filtered flow signal. Based on the phase compensation parameters in the correction parameter set, the time compensation amount of the filtered flow signal relative to the real-time speed and angle position signals is calculated. Based on the time compensation amount, the filtered flow signal is shifted and interpolated on the time axis and resampled to generate a phase-compensated flow signal.

7. The torque detection method for a single-cylinder rotor aero-engine according to claim 6, characterized in that, By identifying and fitting the key extreme points of the flow signal waveform within a cycle, a feature polygon is constructed and the geometric axis of symmetry is determined. Based on the position of the axis of symmetry, the flow characteristic parameters corresponding to the rotor's periodic motion are determined, including: Based on the real-time speed and angular position signals of a single-cylinder rotor, a periodic flow waveform segment corresponding to a complete rotor rotation cycle is extracted from the filtered flow signal. Extremum analysis is performed on periodic flow waveform segments to identify all local maxima and local minima in the waveform; All local maxima and local minima are merged and defined as a set of key extrema. The key extrema are then sorted according to the order in which they appear in the waveform segment to generate an ordered sequence of extrema. Connect the points in the ordered extreme point sequence in sequence, and connect the first two points of the sequence to form a closed planar feature polygon; Perform geometric analysis on a closed planar feature polygon, calculate and determine one of the geometric axes of symmetry of the feature polygon; The geometric symmetry axis is mapped to a coordinate system based on the rotation angle of the single-cylinder rotor, and the angular orientation value of the geometric symmetry axis in the rotation angle coordinate system is determined as the final flow characteristic parameter used to characterize the current rotor rotation cycle.

8. The torque detection method for a single-cylinder rotor aero-engine according to claim 7, characterized in that, Based on flow characteristic parameters, combined with the intake manifold absolute pressure and intake air temperature under the current engine operating conditions, a torque mapping model is used to calculate the real-time output torque value of a single-cylinder rotor, realizing online torque monitoring of the engine's operating state, including: Obtain the real-time intake manifold absolute pressure and real-time intake air temperature of the engine under the current working cycle; The flow characteristic parameters, real-time intake manifold absolute pressure value and real-time intake temperature value are combined to form a set of multi-dimensional real-time operating condition parameters. The multidimensional real-time operating parameters are input into the pre-calibrated torque mapping model to calculate the real-time output torque value of the single-cylinder rotor corresponding to the current working cycle. The system records the real-time output torque value of a single-cylinder rotor and performs parameter combination and torque mapping model calculations in each working cycle of the engine to achieve online torque monitoring of the engine's working state.

9. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in any one of claims 1 to 8.

Images