A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump

Abstract

A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump belongs to the technical field of centrifugal pump status monitoring. It includes the following steps: 1. Embedding a fully solid-state laser radar array on the inner wall of the pump casing; 2. Embedding a thermal infrared camera array on the inner wall of the pump casing; 3. Calculating and generating a position information map of each blade; 4. Calculating and generating a vibration information map of each blade; 5. Calculating and generating a temperature information map of each blade; 6. Adaptively diagnosing the vibration health status of the blade; 7. Adaptively diagnosing the temperature health status of the blade. A non-contact centrifugal pump impeller health status monitoring method based on vision is realized, which overcomes the shortcomings of the contact measurement method in this technical field; a non-contact centrifugal pump impeller health status monitoring method without constraints is realized, which overcomes the shortcomings of the existing non-contact centrifugal pump impeller health status monitoring method based on vision that is limited by constraints.

Description

Technical Field

[0001] The invention belongs to the technical field of centrifugal pump status monitoring, and in particular relates to a method for monitoring the health status of a fully open impeller in an aviation centrifugal pump. Background Art

[0002] The aviation centrifugal pump is one of the key equipment in the aerospace, and the operating status of the aviation centrifugal pump will directly determine the reliability and stability of the aerospace operation. The impeller is the core key component of the aviation centrifugal pump, and the operating status of the impeller directly determines the health status of the aviation centrifugal pump. Therefore, online real-time monitoring of the operating status (such as vibration, temperature, etc.) of the impeller in the aviation centrifugal pump is a technology with great value. On the other hand, the monitoring big data of the operating status of the impeller in the aviation centrifugal pump can also be used to optimize the impeller structure of the aviation centrifugal pump.

[0003] However, the impeller of the aviation centrifugal pump is completely wrapped inside the metal pump casing of the aviation centrifugal pump, and it is difficult to obtain the operating status data of the impeller of the aviation centrifugal pump. One of the more commonly used methods at present is to use a transparent material to make the pump casing, and then use optical photography to take pictures of the operating status data of the impeller from the outside of the pump casing. The advantage of this method is that it uses a non-contact measurement method based on vision, which will not affect the operating status of the pump itself. However, the biggest disadvantage of this method is that it is only applicable to laboratory scenarios and cannot be used for aviation centrifugal pumps that run in real time on aerospace vehicles. Another more common method is to install some sensors inside the pump casing that can measure the operating status of the impeller. The sensors installed by this method usually have simple performance, so the impeller operating status parameters that can be obtained are also relatively rough, and cannot be accurate to specific blades, let alone accurately detect the status of specific parts on the blades.

[0004] In fact, with the rapid development of modern sensor technology, more and more high-precision sensors have been miniaturized. For example, solid-state Lidar technology has long been used in iPhones. Although its module size is only a few millimeters, it can quickly laser scan physical scenes within 5 meters and complete digital 3D modeling. For another example, industrial-grade miniature thermal infrared camera technology has also been widely used. Some thermal infrared camera lenses directly integrated on handheld devices such as mobile phones are only a few millimeters in diameter. Both of the above non-contact sensors use light as the sensing parameter, so they have extremely high detection response frequencies, which are particularly suitable for real-time status monitoring of the impeller of an aviation centrifugal pump with ultra-high speed. Summary of the invention

[0005] In view of the above-mentioned problems existing in the prior art, the purpose of the present invention is to provide a method for monitoring the health status of a fully-open impeller in an aviation centrifugal pump. By continuously and adaptively monitoring two parameters, namely, the vibration and temperature of each blade on the centrifugal pump impeller, the real-time operating status of the impeller can be grasped in detail, so as to facilitate early warning of abnormal conditions of the impeller. At the same time, big data analysis of the impeller operating status can also provide a reliable basis for improving the impeller design parameters.

[0006] The present invention provides the following technical solutions:

[0007] A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump, which realizes early warning of abnormal conditions of the impeller by continuously and adaptively monitoring the vibration and temperature parameters of each blade on the impeller of the centrifugal pump; the method comprises the following steps:

[0008] Step 1: embed a full solid-state laser radar array on the inner wall of the pump casing to scan the spatial structure of the front of the impeller and measure the distance of each point on the front of the impeller relative to the laser radar array;

[0009] Step 2: embed a thermal infrared camera array on the inner wall of the pump casing to measure the temperature of various locations on the front face of the impeller. The measurement range of the thermal infrared camera array covers the entire front face of the impeller.

[0010] Step 3: Calculate and generate a position information map of each blade;

[0011] Step 4: Calculate and generate a vibration information map of each blade according to the position information map generated in step 3;

[0012] Step 5: Calculate and generate a temperature information map of each blade according to the position information map generated in step 3;

[0013] Step 6: Adaptively diagnose the vibration health status of the blade according to the vibration information graph of each blade generated in step 4;

[0014] Step 7: Adaptively diagnose the temperature health status of the blades based on the temperature information graph of each blade generated in step 5.

[0015] Furthermore, the specific process of step 1 is as follows:

[0016] Step 1.1: On the inner wall of the pump casing facing the front of the impeller, select an annular area around the inlet with the center of the inlet as the center of the circle, and the annular area faces the front of the impeller;

[0017] Step 1.2: Divide the selected annular area equally into a group of sector-shaped sub-areas, and define a group of sectors in sequence according to their spatial positions as sector 1...sector n;

[0018] Step 1.3: A miniature all-solid-state laser radar module is embedded at the midpoint of each sector angle bisector to form a laser radar array with a radiation surface of 360 degrees, and its measurement range covers the entire front face of the impeller; the laser radar module includes a laser transmitter and a laser receiver;

[0019] Step 1.4: Connect all power supply lines and signal lines of the laser radar array to the supporting power supply and information processing device located outside the pump housing.

[0020] Furthermore, the specific process of step 2 is as follows:

[0021] Step 2.1: On the angular bisector of each sector, a miniature thermal infrared camera is embedded in a position close to the lidar module and near the pump casing inlet, forming a thermal infrared camera array with a radiation surface of 360 degrees. Its measurement range covers the entire front face of the impeller and is used to measure the temperature of various locations on the front face of the impeller;

[0022] Step 2.2: Connect all power supply lines and signal lines of the infrared camera array to the supporting power supply and information processing device located outside the pump housing.

[0023] Furthermore, the specific process of step 3 is as follows:

[0024] Step 3.1: From the two blade tip angles at the end of each blade on the impeller, select the blade tip angle that will first enter sector 1 when the blade rotates as the blade positioning mark point;

[0025] Step 3.2: Use the laser radar in sector 1 to continuously detect whether there is a blade positioning mark passing through the angle bisector of sector 1. When a blade positioning mark passes through, a locking signal L will be triggered;

[0026] Step 3.3: When the impeller starts to rotate, the blade where the first detected blade positioning mark is located will be marked as blade No. 1, and the marking will be done in sequence until the Nth blade; according to the order of the blade numbers and the order of the sector positions, the sectors where the remaining blades are located can be accurately inferred, so as to record the position of each blade when the locking signal L is generated each time;

[0027] Step 3.4: The position information of N blades at different times is fused in turn, and the data layers are spliced ​​and integrated to generate N sets of position information maps containing only the corresponding blades. The position information map contains the spatiotemporal joint information of the corresponding blades; through the position information map, the sector position of the corresponding blade at a specific time can be known at any time.

[0028] Furthermore, the specific process of step 4 is as follows:

[0029] Step 4.1: Whenever the locking signal L is generated, the radars in the lidar array will synchronously measure the current distance information at all locations on the front of all blades in the corresponding sector;

[0030] Step 4.2: Based on step 3, at each time when the locking signal L is generated, the position and number of each blade are known, and the blade distance information data measured in step 4.1 is matched to the corresponding blade, so as to record the distance information of each blade at different times;

[0031] Step 4.3: Combined with the blade position information map generated in step 3, the blade distance information obtained in step 4.2 is superimposed on the position information map to generate a blade vibration information map.

[0032] Furthermore, the specific process of step 5 is as follows:

[0033] Step 5.1: Whenever the locking signal L is generated, the thermal infrared cameras in the thermal infrared camera array will synchronously measure the current temperature information at all locations on the front side of all blades in the corresponding sector;

[0034] Step 5.2: Based on step 3, at each time when the locking signal L is generated, the position and number of each blade are known, and the blade temperature information data measured in step 5.1 is matched to the corresponding blade, so as to record the temperature information of each blade at different times;

[0035] Step 5.3: Combined with the position information map generated in step 3, the blade temperature information obtained in step 5.2 is superimposed on the position information map to generate a blade temperature information map.

[0036] Furthermore, the specific process of step 6 is as follows:

[0037] Step 6.1: Use the blade vibration information graph to construct a blade vibration noise probability distribution model;

[0038] Step 6.2: Calculate the frequency of abnormal vibration on each blade;

[0039] Step 6.3: Combined with the blade position information map generated in step 3, the information of the blade abnormal vibration frequency calculated in step 6.2 is superimposed on the position information map, thereby generating a blade abnormal vibration information map.

[0040] Step 6.4: Automatic alarm or automatic shutdown protection: When an abnormal vibration state occurs on a blade at a high frequency for a long time, it is automatically determined that an alarm or shutdown protection is required; that is, adaptive monitoring of the vibration health status of the impeller blades is achieved.

[0041] Furthermore, the specific process of step 7 is as follows:

[0042] Step 7.1: Using the blade temperature information graph, construct a blade temperature noise probability distribution model;

[0043] Step 7.2: Calculate the frequency of abnormal temperature on each blade;

[0044] Step 7.3: Combined with the blade position information map generated in step 3, the information of the blade abnormal temperature fluctuation frequency calculated in step 7.2 is superimposed on the position information map, thereby generating a blade abnormal temperature information map;

[0045] Step 7.4: Automatic alarm or automatic shutdown protection: When an abnormal temperature state appears on a blade at a high frequency for a long time, it is automatically determined that an alarm or shutdown protection is required; that is, adaptive monitoring of the temperature health status of the impeller blades is achieved.

[0046] By adopting the above technology, compared with the prior art, the beneficial effects of the present invention are as follows:

[0047] 1) The present invention realizes a non-contact centrifugal pump impeller health status monitoring method based on vision, which overcomes the shortcomings of the contact measurement method in the technical field;

[0048] 2) The present invention realizes a non-contact centrifugal pump impeller health status monitoring method without constraints, which overcomes the shortcomings of the existing vision-based non-contact centrifugal pump impeller health status monitoring method that is limited by constraints;

[0049] 3) The present invention realizes a centrifugal pump impeller health status monitoring method with synchronous measurement of multiple parameters (blade position, vibration, temperature), which overcomes the shortcomings of the technical field that a single measurement parameter or multiple measurement parameters cannot be synchronously associated;

[0050] 4) The present invention realizes an adaptive centrifugal pump impeller health status monitoring method, which overcomes the lack of adaptability of traditional monitoring methods in this technical field;

[0051] 5) The present invention can be used for online monitoring of the real-time health status of centrifugal pumps on in-service aircraft, and can also be used for offline analysis of whether there is room for optimization in the design parameters of aviation centrifugal pumps;

[0052] 6) The device used in the present invention is small in size, simple to install and does not affect the original flow diversion performance of the pump chamber;

[0053] 7) The method adopted by the present invention has the advantages of good universality and a wide range of applicable aviation centrifugal pump types. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flow chart of a method for monitoring the health status of a fully open impeller in an aviation centrifugal pump;

[0055] Figure 2 is an example of an aviation centrifugal pump using a fully open impeller design; Figure 2 a is the pump casing, Figure 2 b is the cross-sectional view of the impeller, Figure 2 c is the front view of the impeller, Figure 2 d is a front view of the inner wall of the pump casing facing the front of the impeller;

[0056] Figure 3 This is a schematic diagram of the installation of the laser radar array and the thermal infrared camera array; in the figure, numbers 1 to 8 represent sectors 1 to 8 respectively, number 9 represents the laser radar module, and number 10 represents the thermal infrared camera;

[0057] Figure 4 This is a schematic diagram for selecting blade positioning mark points. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments of the specification. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0059] On the contrary, the present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention as defined by the claims. Further, in order to make the public have a better understanding of the present invention, some specific details are described in detail in the detailed description of the present invention below. Those skilled in the art can fully understand the present invention without the description of these details.

[0060] Example:

[0061] See also Figure 1-4 A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump. By continuously and adaptively monitoring the vibration and temperature of each blade on the centrifugal pump impeller, the real-time operating status of the impeller can be grasped in detail, which is convenient for early warning of abnormal conditions of the impeller. At the same time, the big data analysis of the impeller operating status can also provide a reliable basis for improving the impeller design parameters. It includes the following steps:

[0062] Step 1: Embed a fully solid-state lidar array on the inner wall of the pump casing.

[0063] Step 1.1: Figure 3 As shown, on the inner wall of the pump casing facing the front of the impeller, an annular area is selected around the inlet. The center of the annular area is the center of the inlet, and the outer radius of the annular area is slightly larger than the impeller radius. The annular area faces the front of the impeller (i.e., the side with the blade reinforcement ribs).

[0064] Step 1.2: Divide the annular area described in step 1.1 into 8 sector-shaped sub-areas equally, that is, the angle of each sector is 45 degrees. Figure 2 The sectors shown are defined as sectors 1 to 8 in sequence.

[0065] Step 1.3: At the midpoint of the angular bisector of each sector described in step 1.2, a micro all-solid-state Lidar module is embedded, that is, a total of 8 Lidar modules are embedded, thereby forming a Lidar array with a radiation surface of 360 degrees, and its measurement range covers the entire front of the impeller. The function of the above-mentioned Lidar array is to scan the spatial structure of the front of the impeller and measure the distance of each point on the front of the impeller relative to the Lidar array. Each Lidar module contains a laser transmitter and a laser receiver. The laser transmitter uses an infrared laser with a wavelength of 850 nanometers as a light source.

[0066] Step 1.4: Connect all power supply lines and signal lines of the laser radar array described in step 1.3 to the matching power supply and information processing device located outside the pump housing.

[0067] Step 2: Embed a thermal infrared camera array on the inner wall of the pump housing.

[0068] Step 2.1: Figure 3 As shown, on the angular bisector of each sector described in step 1, a miniature thermal infrared camera is embedded in each position close to the laser radar module and closer to the pump casing inlet, that is, a total of 8 thermal infrared cameras are embedded, thereby forming a thermal infrared camera array with a radiation surface of 360 degrees, and its measurement range covers the entire impeller front. The function of the above thermal infrared camera array is to measure the temperature of various locations on the impeller front.

[0069] Step 2.2: Connect all power supply lines and signal lines of the infrared camera array described in step 2.1 to the matching power supply and information processing device located outside the pump housing.

[0070] Step 3: Calculate and generate a position information map of each blade.

[0071] Step 3.1: From the two blade tip angles at the end of each blade on the impeller, select the blade tip angle that will first enter sector 1 when the blade rotates as the blade positioning mark point. Its position is shown in Figure 4 The circled position. Figure 2 As shown, the impeller of the centrifugal pump has 8 blades, and there are correspondingly 8 blade positioning marking points.

[0072] Step 3.2: Since the laser radar can scan the outline of the blade, it is easy to identify the blade positioning mark points defined in step 3.1 based on prior knowledge. Use the laser radar in sector 1 to continuously detect whether there are blade positioning mark points passing through the angle bisector of sector 1. When the above event occurs, a lock signal L (Lock) will be triggered.

[0073] Step 3.3: When the impeller starts to rotate, the blade where the first detected blade positioning mark point is located will be marked as blade No. 1. The blade where the second detected blade positioning mark point is located will be marked as blade No. 2. And so on, until the eighth blade is marked in sequence. In this way, all blades are numbered sequentially. According to the sequence of blade numbers and the sequence of sector positions, once the blade number entering sector No. 1 at a certain moment is known, the sector where the remaining 7 blades are located can be accurately inferred. Through the above positioning method, the position of each blade can be recorded every time the locking signal L is generated.

[0074] Step 3.4: The position information of blade No. 1 at different times is separately fused at the data layer, and a set of position information graphs containing only blade No. 1 can be spliced ​​and integrated to generate a set of position information graphs containing the spatiotemporal joint information of blade No. 1. From the position information graph, it can be observed that blade No. 1 moves periodically in 8 sectors according to the rotation direction of the impeller, and the sector position of blade No. 1 at a specific time can be known at any time. Similarly, the corresponding position information graph can also be generated for the second blade. By analogy, 8 sets of position information graphs will be generated for 8 blades.

[0075] Step 4: Calculate and generate the vibration information map of each blade.

[0076] Step 4.1: Whenever the locking signal L is generated, the eight laser radars in the laser radar array will synchronously measure the current distance information (the distance from the laser radar) at each location on the front side of all blades in the eight sectors.

[0077] Step 4.2: Based on step 3, at each moment when the locking signal L is generated, the position and number of each blade are known, so the blade distance data measured in step 4.1 can be matched to the corresponding blade, thereby recording the distance information of each blade at different moments (referring to the distance from the laser radar).

[0078] Step 4.3: Combined with the blade position information map generated in step 3, the blade distance information obtained in step 4.2 is superimposed on the position information map to generate a blade vibration information map. The vibration here refers to the axial vibration of the impeller. If the distance of a blade relative to the laser radar changes significantly, it means that the blade has undergone axial vibration. The vibration information map of blade No. 1 contains the combined time, space and vibration information of blade No. 1, that is, the real-time vibration changes of blade No. 1 when it moves in different time and different spatial regions can be observed. Similarly, each blade can generate a corresponding vibration information map, from which the axial vibration state of any blade at different times and different pump chamber locations can be monitored in real time.

[0079] Step 5: Calculate and generate the temperature information map of each blade.

[0080] Step 5.1: Whenever the locking signal L is generated, the eight thermal infrared cameras in the thermal infrared camera array will synchronously measure the current temperature information at all locations on the front side of all blades in the eight sectors.

[0081] Step 5.2: Based on step 3, at each moment when a locking signal L is generated, the position and number of each blade are known, so the blade temperature data measured in step 5.1 can be matched to the corresponding blade, thereby recording the temperature information of each blade at different moments.

[0082] Step 5.3: Combined with the position information map generated in step 3, the blade temperature information obtained in step 5.2 is superimposed on the position information map to generate a blade temperature information map. The temperature information map of blade No. 1 contains the combined time, space and temperature information of blade No. 1, that is, the real-time temperature changes of blade No. 1 when it moves in different time and different spatial regions can be observed. Similarly, each blade can generate a corresponding temperature information map, from which the temperature state of any blade at different times and different pump cavity locations can be monitored in real time.

[0083] Step 6: Adaptively diagnose the vibration health status of the blade.

[0084] Step 6.1: Using the blade vibration information graph, construct a blade vibration noise probability distribution model, including the following steps:

[0085] 6.1.1. From the blade vibration information diagram, we can know that the distance value D measured by the laser radar at the vibration sampling point with coordinate position (x, y, z) on the blade at time t is t (x, y, z). Since there is inevitably inherent noise Z (most typically Gaussian noise) in any measurement process, D t(x, y, z) will inevitably experience amplitude fluctuations caused by noise at the next moment, that is, the distance value measured by the laser radar at the vibration sampling point with coordinate position (x, y, z) on the blade at time t+1 will become D t+1 (x,y,z).D t+1 (x,y,z) and D t The difference between (x, y, z) will obey the probability distribution of the inherent noise Z.

[0086] This embodiment adopts a zero-mean Gaussian white noise model as the basic model of the blade vibration noise probability distribution model, that is, it is assumed that the fluctuation noise of the distance value at each vibration sampling point (x, y, z) in the blade vibration information graph obeys a Gaussian distribution with a mean of zero. Since the Gaussian distribution has only two parameters, the mean and the variance, and the mean of the blade vibration noise probability distribution model is known to be zero, there is only one unknown parameter in the blade vibration noise probability distribution model, namely, the standard deviation of the distribution.

[0087] 6.1.2. Calculate the standard deviation of the blade vibration noise probability distribution model:

[0088] 6.1.2.1. Using the vibration information diagrams of the eight blades, calculate the standard deviation of the vibration noise probability distribution model of each blade using the following formula:

[0089]

[0090] D t (x,y,z) and D t+1 (x, y, z) represents the distance value of the vibration sampling point at the blade coordinate position (x, y, z) relative to the laser radar at time t and time t+1. T is the end time of vibration sampling. Θ i,t | vib represents the set of all vibration sampling points at the tth time in the i-th (i=1,...,8) vibration information graph; Represents Θ i,t | vib A subset of The distance value of all vibration sampling points in is d and d=D t (x,y,z); #() represents taking the cardinality of the set (that is, the total number of elements in the set). It represents the standard deviation of the blade vibration noise probability distribution model corresponding to all vibration sampling points on the i-th blade when the distance value is d.

[0091] 6.1.2.2、Get The maximum value in is taken as the standard deviation σ of the blade vibration noise probability distribution model corresponding to all vibration sampling points on all 8 blades when the distance value is d d |vib ,Right now:

[0092]

[0093] Thus, a probability distribution model of blade vibration noise for the entire impeller is constructed. For any vibration sampling point on any blade, its vibration noise has a mean of zero and a variance of σ d | vib Gaussian distribution.

[0094] Step 6.2: Calculate the frequency of abnormal vibration on each blade, including the following steps:

[0095] 6.2.1. At a certain observation time s+1, if D s+1 (x,y,z) and D s If the difference between (x, y, z) and (x, y, z) obeys the blade vibration noise probability distribution model calculated in step 6.1, it means that the vibration sampling point with coordinate position (x, y, z) on the blade is a normal vibration, which is mainly caused by the inherent noise of the system. On the contrary, if D s+1 (x,y,z) and D s If the difference of (x, y, z) does not obey the blade vibration noise probability distribution model calculated in step 6.1, it means that an abnormal vibration occurs at the vibration sampling point with coordinate position (x, y, z) on the blade. This vibration is not caused by the inherent noise of the system, but by an abnormal state in the centrifugal pump. This type of abnormal vibration may pose a potential hazard to the impeller or affect the working efficiency of the impeller, so it is necessary to quantitatively count this type of abnormal vibration. Specifically, the total number of abnormal vibrations occurring at each vibration sampling point on each blade is counted, and the formula is:

[0096]

[0097] in, Represents the total number of abnormal vibrations that have occurred at the vibration sampling point with coordinates (x, y, z) on the i-th blade during the period from the impeller startup time t = 0 to the specified observation time s + 1. λ is a variable whose value is as follows:

[0098] When |D t+1 (x,y,z)-D t (x,y,z)>β·σ d | vib , then λ=1;

[0099] When |D t+1 (x,y,z)-D t (x,y,z)≤β·σ d | vib, then λ=0.

[0100] Among them, β is a natural number, representing σ d | vib multiples of.

[0101] 6.2.2. Based on the results of step 6.2.1 above, calculate the frequency of abnormal vibration at each vibration sampling point on each blade. The formula is:

[0102]

[0103] Where φ represents the sampling frequency of the lidar, and its unit is Hertz. Represents the frequency of abnormal vibration that has occurred at the vibration sampling point with coordinate position (x, y, z) on the i-th blade during the period from the impeller startup time t=0 to the specified observation time s+1.

[0104] Step 6.3: Combined with the blade position information map generated in step 3, the information of the blade abnormal vibration frequency calculated in step 6.2 is superimposed on the position information map to generate the blade abnormal vibration information map. The abnormal vibration information map of blade No. 1 contains the joint information of time, space and abnormal vibration frequency of blade No. 1, that is, it can be observed from it that the frequency of abnormal vibration of blade No. 1 when it moves in different time and different spatial regions. By analogy, each blade has a corresponding blade abnormal vibration information map, from which the abnormal vibration state of any blade at different time and different pump cavity locations can be monitored in real time.

[0105] Step 6.4: Automatic alarm or automatic shutdown protection. When an abnormal vibration state occurs on a blade at a high frequency for a long time, it can be automatically determined that an alarm or shutdown protection is required. This can achieve adaptive monitoring of the vibration health status of the impeller blades.

[0106] Step 7: Adaptively diagnose the temperature health status of the blade.

[0107] Step 7.1: Using the blade temperature information graph, construct a blade temperature noise probability distribution model, including the following steps:

[0108] 7.1.1. From the blade temperature information diagram, we can know that the temperature value measured by the thermal infrared camera at the temperature sampling point with coordinate position (x, y, z) on the blade at time t is E t (x, y, z). Since there is inevitably inherent noise Y (most typically Gaussian noise) in any measurement process, E t(x, y, z) will inevitably experience amplitude fluctuations caused by noise at the next moment, that is, the temperature value measured by the thermal infrared camera at the temperature sampling point with coordinate position (x, y, z) on the blade at time t+1 will become E t+1 (x,y,z). t+1 (x,y,z) and E t The difference between (x, y, z) will obey the probability distribution of the inherent noise Y.

[0109] This embodiment uses a zero-mean Gaussian white noise model as the basic model of the blade temperature noise probability distribution model, that is, it is assumed that the fluctuation noise of the temperature value at each temperature sampling point (x, y, z) in the blade temperature information graph obeys a Gaussian distribution with a mean of zero. Since the Gaussian distribution has only two parameters, the mean and the variance, and the mean of the blade temperature noise probability distribution model is known to be zero, there is only one unknown parameter in the blade temperature noise probability distribution model, namely, the standard deviation of the distribution.

[0110] 7.1.2. Calculate the standard deviation of the blade temperature noise probability distribution model:

[0111] 7.1.2.1. Using the temperature information graphs of the eight leaves, calculate the standard deviation of the temperature noise probability distribution model of each leaf using the following formula:

[0112]

[0113] Among them, E t (x,y,z) and E t+1 (x, y, z) represents the temperature value of the temperature sampling point at the blade coordinate position (x, y, z) at time t and time t+1. T is the end time of temperature sampling. Θ i,t | temp Represents the set of all temperature sampling points at time t in the i-th (i=1,...,8) leaf temperature information graph; Represents Θ i,t | temp A subset of The temperature values ​​of all temperature sampling points in are e and e=D t (x,y,z); #() represents taking the cardinality of the set (that is, the total number of elements in the set). Represents the standard deviation of the blade temperature noise probability distribution model corresponding to all temperature sampling points on the i-th blade when the temperature value is e.

[0114] 7.1.2.2、Get The maximum value of is taken as the standard deviation σ of the blade temperature noise probability distribution model corresponding to all temperature sampling points on all 8 blades when the temperature value is e e |temp ,Right now:

[0115]

[0116] Thus, a blade temperature noise probability distribution model for the entire impeller is constructed. For any temperature sampling point on any blade, the temperature noise has a mean of zero and a variance of σ e | temp Gaussian distribution.

[0117] Step 7.2: Calculate the frequency of abnormal temperature on each blade, including the following steps:

[0118] 7.2.1. At a certain observation time s+1, if E s+1 (x,y,z) and E s If the difference of (x, y, z) obeys the blade temperature noise probability distribution model calculated in step 7.1, it means that the temperature sampling point with coordinate position (x, y, z) on the blade is a normal temperature fluctuation, which is mainly caused by the inherent noise of the system. On the contrary, if E s+1 (x,y,z) and E s If the difference of (x, y, z) does not obey the blade temperature noise probability distribution model calculated in step 7.1, it means that an abnormal temperature fluctuation occurs at the temperature sampling point with coordinate position (x, y, z) on the blade. This temperature fluctuation is not caused by the inherent noise of the system, but by an abnormal state in the centrifugal pump. This type of abnormal temperature fluctuation may pose a potential hazard to the impeller or affect the working efficiency of the impeller, so it is necessary to quantitatively count this type of abnormal temperature fluctuation. Specifically, the total number of abnormal temperature fluctuations occurring at each temperature sampling point on each blade is counted, and the formula is:

[0119]

[0120] in, Represents the total number of abnormal temperature fluctuations that have occurred at the temperature sampling point with coordinates (x, y, z) on the i-th blade during the period from the impeller startup time t = 0 to the specified observation time s + 1. γ is a variable whose value is as follows:

[0121] When|E t+1 (x,y,z)-E t (x,y,z)|>θ·σ e | temp , then γ=1;

[0122] When|E t+1 (x,y,z)-E t (x,y,z)|≤θ·σ e| temp , then γ=0.

[0123] Among them, θ is a natural number, representing σ e | temp multiples of.

[0124] 7.2.2. Based on the results of step 7.2.1 above, calculate the frequency of abnormal temperature fluctuations at each temperature sampling point on each blade. The formula is:

[0125]

[0126] in, Represents the sampling frequency of the thermal infrared camera, and its unit is Hz. It represents the frequency of abnormal temperature fluctuations occurring at the temperature sampling point with coordinates (x, y, z) on the i-th blade during the period from the impeller startup time t=0 to the specified observation time s+1.

[0127] Step 7.3: Combined with the blade position information map generated in step 3, the information of the blade abnormal temperature fluctuation frequency calculated in step 7.2 is superimposed on the position information map to generate the blade abnormal temperature information map. The abnormal temperature information map of blade No. 1 contains the joint information of time, space and abnormal temperature fluctuation frequency of blade No. 1, that is, it can be observed from it the frequency of abnormal temperature fluctuation when blade No. 1 moves in different time and different spatial regions. By analogy, each blade has a corresponding blade abnormal temperature information map, from which the abnormal temperature state of any blade at different time and different pump cavity locations can be monitored in real time.

[0128] Step 7.4: Automatic alarm or automatic shutdown protection. When an abnormal temperature state appears on a blade for a long time and at a high frequency, it can be automatically determined that an alarm or shutdown protection is required. This can realize adaptive monitoring of the temperature health status of the impeller blades.

[0129] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump, which realizes early warning of abnormal conditions of the impeller by continuously and adaptively monitoring the vibration and temperature parameters of each blade on the centrifugal pump impeller; the method is characterized by: The following steps are involved: Step 1: embed a full-solid-state laser radar array on the inner wall of the pump casing to scan the spatial structure of the front of the impeller and measure the distance of each point on the front of the impeller relative to the full-solid-state laser radar array; Step 2: embed a thermal infrared camera array on the inner wall of the pump casing to measure the temperature of various locations on the front face of the impeller. The measurement range of the thermal infrared camera array covers the entire front face of the impeller. Step 3: Calculate and generate a position information map of each blade; Step 4: Calculate and generate a vibration information map of each blade according to the position information map generated in step 3; Step 5: Calculate and generate a temperature information map of each blade according to the position information map generated in step 3; Step 6: Adaptively diagnose the vibration health status of the blade according to the vibration information graph of each blade generated in step 4; the specific process is as follows: Step 6.1: Use the blade vibration information graph to construct a blade vibration noise probability distribution model; Step 6.2: Calculate the frequency of abnormal vibration on each blade; Step 6.3: Combined with the blade position information map generated in step 3, the information of the blade abnormal vibration frequency calculated in step 6.2 is superimposed on the position information map, thereby generating a blade abnormal vibration information map; Step 6.4: Automatic alarm or automatic shutdown protection: When an abnormal vibration state occurs on a blade at a high frequency for a long time, it is automatically determined that an alarm or shutdown protection is required; that is, adaptive monitoring of the vibration health status of the impeller blades is achieved; Step 7: Adaptively diagnose the temperature health status of the blades according to the temperature information graph of each blade generated in step 5; the specific process is as follows: Step 7.1: Using the blade temperature information graph, construct a blade temperature noise probability distribution model; Step 7.2: Calculate the frequency of abnormal temperature on each blade; Step 7.3: Combined with the blade position information map generated in step 3, the information of the blade abnormal temperature fluctuation frequency calculated in step 7.2 is superimposed on the position information map, thereby generating a blade abnormal temperature information map; Step 7.4: Automatic alarm or automatic shutdown protection: When an abnormal temperature state appears on a blade at a high frequency for a long time, it is automatically determined that an alarm or shutdown protection is required; that is, adaptive monitoring of the temperature health status of the impeller blades is achieved.

2. A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump according to claim 1, characterized in that The specific process of step 1 is as follows: Step 1.1: On the inner wall of the pump casing facing the front of the impeller, select an annular area around the inlet with the center of the inlet as the center of the circle, and the annular area faces the front of the impeller; Step 1.2: Divide the selected annular area equally into a group of sector-shaped sub-areas, and define a group of sectors in sequence according to their spatial positions as sector 1...sector n; Step 1.3: A miniature all-solid-state laser radar module is embedded at the midpoint of each sector angle bisector to form an all-solid-state laser radar array with a radiation surface of 360 degrees, and its measurement range covers the entire front face of the impeller; the all-solid-state laser radar module includes a laser transmitter and a laser receiver; Step 1.4: Connect all power supply lines and signal lines of the all-solid-state laser radar array to the supporting power supply and information processing device located outside the pump housing.

3. A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump according to claim 2, characterized in that The specific process of step 2 is as follows: Step 2.1: On the angular bisector of each sector, a miniature thermal infrared camera is embedded in a position close to the all-solid-state laser radar module and close to the pump casing inlet, forming a thermal infrared camera array with a radiation surface of 360 degrees. Its measurement range covers the entire front face of the impeller and is used to measure the temperature of various locations on the front face of the impeller; Step 2.2: Connect all power supply lines and signal lines of the infrared camera array to the supporting power supply and information processing device located outside the pump housing.

4. A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump according to claim 3, characterized in that The specific process of step 3 is as follows: Step 3.1: From the two blade tip angles at the end of each blade on the impeller, select the blade tip angle that will first enter sector 1 when the blade rotates as the blade positioning mark point; Step 3.2: Use the all-solid-state laser radar in sector 1 to continuously detect whether there is a blade positioning mark passing through the angular bisector of sector 1. When a blade positioning mark passes through, a locking signal L will be triggered; Step 3.3: When the impeller starts to rotate, the blade where the first detected blade positioning mark is located will be marked as blade No. 1, and the marking will be done in sequence until the Nth blade; according to the order of the blade numbers and the order of the sector positions, the sectors where the remaining blades are located can be accurately inferred, so as to record the position of each blade when the locking signal L is generated each time; Step 3.4: The position information of N blades at different times is fused in turn, and the data layers are spliced ​​and integrated to generate N sets of position information maps containing only the corresponding blades. The position information map contains the spatiotemporal joint information of the corresponding blades; through the position information map, the sector position of the corresponding blade at a specific time can be known at any time.

5. A method for monitoring the health status of a fully open impeller in an aviation centrifugal pump according to claim 4, characterized in that The specific process of step 4 is as follows: Step 4.1: Whenever the locking signal L is generated, the radars in the all-solid-state laser radar array will synchronously measure the current distance information at all locations on the front side of all blades in the corresponding sector; Step 4.2: Based on step 3, at each time when the locking signal L is generated, the position and number of each blade are known, and the blade distance information data measured in step 4.1 is matched to the corresponding blade, so as to record the distance information of each blade at different times; Step 4.3: Combined with the blade position information map generated in step 3, the blade distance information obtained in step 4.2 is superimposed on the position information map to generate a blade vibration information map.

6. A method for monitoring the health status of a fully open impeller in an aerospace centrifugal pump according to claim 5, characterized in that The specific process of step 5 is as follows: Step 5.1: Whenever the locking signal L is generated, the thermal infrared cameras in the thermal infrared camera array will synchronously measure the current temperature information at all locations on the front side of all blades in the corresponding sector; Step 5.2: Based on step 3, at each time when the locking signal L is generated, the position and number of each blade are known, and the blade temperature information data measured in step 5.1 is matched to the corresponding blade, so as to record the temperature information of each blade at different times; Step 5.3: Combined with the position information map generated in step 3, the blade temperature information obtained in step 5.2 is superimposed on the position information map to generate a blade temperature information map.

Images