PCFC algorithm-based rotating machine blade dust deposition trend prediction method

Abstract

The invention discloses a rotating machine blade dust deposition trend prediction method based on a PCFC algorithm, relates to the field of mechanical equipment, solves the problem of poor prediction effect of the rotating machine blade dust deposition trend prediction method, and comprises the following steps: S1, carrying out vibration monitoring on a rotating machine blade in each historical monitoring time period, the method comprises the following steps: S1, performing blade dust deposition monitoring on a rotating machine blade in each historical monitoring time period, and obtaining a blade vibration monitoring value corresponding to each historical monitoring time period to obtain blade vibration monitoring data, S2, performing blade dust deposition monitoring on the rotating machine blade in each historical monitoring time period, and obtaining a blade dust deposition index value corresponding to each historical monitoring time period according to a monitoring result, s3, according to the blade dust deposition monitoring data and the blade vibration monitoring data, real-time dust deposition prediction is conducted on the rotating machine blade, and dust deposition early warning is issued according to the prediction result, and the accuracy of the dust deposition trend prediction result of the rotating machine blade is improved.

Description

Technical Field

[0001] The present invention belongs to the field of machinery and relates to image recognition technology, in particular to a method for predicting dust accumulation trends of rotating machinery blades based on a PCFC algorithm. Background Art

[0002] The existing early warning method for dust accumulation on rotating machinery blades has the following specific defects when monitoring dust accumulation on rotating machinery blades: 1. Existing early warning methods for dust accumulation on rotating machinery blades require the installation of high-precision probes and the use of spectrum analysis to determine vibration types, which increases equipment operation and maintenance costs and aggravates the operating burden of enterprises. 2. The existing dust accumulation warning method for rotating machinery blades cannot make real-time dust accumulation predictions on rotating machinery blades based on the blade dust accumulation monitoring data and blade vibration monitoring data of the historical period, and cannot issue dust accumulation warnings based on the prediction results, resulting in a lag in the warning process.

[0003] To this end, we propose a method for predicting the dust accumulation trend of rotating machinery blades based on the PCFC algorithm. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a method for predicting the dust accumulation trend of rotating machinery blades based on the PCFC algorithm. The present invention aims to improve the timeliness of the dust accumulation warning of rotating machinery blades and reduce the equipment operation and maintenance costs.

[0005] In order to achieve the above object, the present invention adopts the following technical solution: a method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm, comprising the following specific steps: Step S1: acquiring historical vibration data of a rotating machinery blade, selecting a plurality of historical monitoring periods from the acquired historical vibration data of the rotating machinery blade, performing vibration monitoring on the rotating machinery blade in each historical monitoring period, and acquiring a blade vibration monitoring value corresponding to each historical monitoring period to obtain blade vibration monitoring data; Step S2: performing blade dust accumulation monitoring on the rotating machinery blades in each historical monitoring period based on the blade vibration monitoring data, obtaining a blade dust accumulation index value corresponding to each historical monitoring period based on the monitoring results, and obtaining blade dust accumulation monitoring data; Step S3: Perform real-time dust accumulation prediction on the blades of the rotating machinery based on the blade dust accumulation monitoring data and the blade vibration monitoring data, and issue a dust accumulation warning based on the prediction results.

[0006] Furthermore, the step S1 further includes the following specific steps: Step S11: acquiring blade vibration monitoring data corresponding to the rotating machinery blade to obtain historical vibration data of the rotating machinery blade, and marking the historical vibration data of the rotating machinery blade as a blade vibration history cycle; Step S12: Divide the blade vibration history period into a plurality of history monitoring periods, and mark the divided history monitoring periods as L1 history monitoring period to La history monitoring period in chronological order; Step S13: monitoring the blade historical vibration index of the rotating machinery blade in the L1 historical monitoring period, obtaining the blade vibration monitoring value corresponding to the L1 historical monitoring period according to the monitoring result, and naming it as the L1 blade vibration monitoring value; Step S14: acquiring the L2 blade vibration monitoring value to the La blade vibration monitoring value respectively; Step S15: defining the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 historical monitoring period to the La historical monitoring period as blade vibration monitoring data.

[0007] Furthermore, the step S13 further includes the following specific steps: Step S131: Divide the L1 historical monitoring period into a number of consecutive blade vibration time points, and mark the marked blade vibration time points as Z1 blade vibration time point to Zb blade vibration time point in chronological order; Step S132: acquiring blade vibration values ​​corresponding to the blade vibration time point Z1 to the blade vibration time point Zb of the rotating machinery through a vibration sensor, thereby obtaining blade vibration values ​​Z1 to Zb; Step S133: creating a blade vibration period line graph using the Z1 blade vibration value to the Zb blade vibration value and the Z1 blade vibration time point to the Zb blade vibration time point.

[0008] Furthermore, the step S13 further includes the following specific steps: Step S134: In the blade vibration period line graph, the maximum ordinate value is marked as the first blade vibration peak value, the ordinate value whose ordinate value is just smaller than the first blade vibration peak value is marked as the second blade vibration peak value, the minimum ordinate value is marked as the first blade vibration valley value, and the ordinate value whose ordinate value is just larger than the first blade vibration valley value is marked as the second blade vibration valley value; Step S135: Calculate the average value of the blade vibration values ​​from Z1 to Zb to obtain the average vibration value of the time period; Step S136: Calculating the first blade vibration peak value, the second blade vibration peak value, the first blade vibration valley value, the second blade vibration valley value, and the time period average vibration value to obtain the L1 blade vibration monitoring value; Calculate the L1 blade vibration monitoring value.

[0009] Furthermore, the step S133 further includes the following specific steps: In the existing plane rectangular coordinate system, the blade vibration time point is used as the horizontal coordinate and the blade vibration value is used as the vertical coordinate to create a blade vibration plane rectangular coordinate system; In the rectangular coordinate system of the blade vibration plane, the Z1 blade vibration time point is used as the horizontal coordinate and the Z1 blade vibration value is used as the vertical coordinate, and the coordinate point is marked as the Z1 vibration coordinate point. The Z2 blade vibration time point is used as the horizontal coordinate and the Z2 blade vibration value is used as the vertical coordinate, and so on. The Zb blade vibration time point is used as the horizontal coordinate and the Zb blade vibration value is used as the vertical coordinate, and the coordinate point is marked as the Zb vibration coordinate point. The Z1 vibration coordinate point to the Zb vibration coordinate point are connected in sequence to obtain a blade vibration period line graph.

[0010] Furthermore, the step S2 further includes the following specific steps: Step S21: Obtain blade dust accumulation monitoring data, and obtain historical monitoring periods L1 to La based on the blade dust accumulation monitoring data; Step S22: collecting front images of the rotating machinery blades in the historical monitoring period L1 to the historical monitoring period La, and obtaining front images of the blades in the L1 period to the La period; Step S23: collecting side images of the rotating machinery blades in the L1 historical monitoring period to the La historical monitoring period, and obtaining the side images of the blades in the L1 period to the La period; Step S24: performing dust accumulation analysis on the front image of the blade during the L1 period, and obtaining the L1 blade front dust area ratio according to the analysis result; Step S25: performing dust accumulation area analysis on the front image of the blade from the L2 period to the La period, respectively, to obtain the dust accumulation ratio of the front of the blade from L2 to the front of the blade from La; Step S26: Analyze the dust accumulation thickness of the blade side image during the L1 period, and obtain the dust accumulation thickness of the blade side during the L1 period according to the analysis result; Step S27: performing dust accumulation thickness analysis on the blade side images from the L2 period to the La period, respectively, to obtain the dust accumulation thickness on the blade side from the L2 period to the La period; Step S28: Calculate the product of the dust thickness on the side of the L1 blade and the dust ratio on the front of the L1 blade to obtain the dust index value of the L1 blade; calculate the product of the dust thickness on the side of the L2 blade and the dust ratio on the front of the L2 blade to obtain the dust index value of the L2 blade; and so on, calculate the product of the dust thickness on the side of the La blade and the dust ratio on the front of the La blade to obtain the dust index value of the La blade; Step S29: defining the blade dust accumulation index value L1 to the blade dust accumulation index value La as blade dust accumulation monitoring data.

[0011] Furthermore, the step S24 further includes the following specific steps: Obtain an image of the front side of the leaf during the L1 period, use an image recognition algorithm to identify the dust accumulation area in the image of the front side of the leaf during the L1 period, obtain a first image feature area, and mark the non-dust accumulation area in the image of the front side of the leaf during the L1 period as a second image feature area; Acquiring area values ​​of the first image feature region and the second image feature region respectively to obtain an area value of the first image region and an area value of the second image region; The area value of the first image region and the area value of the second image region are calculated to obtain the gray ratio of the front surface of the L1 leaf; The dust coverage ratio of the L1 leaf was calculated.

[0012] Furthermore, the step S26 further includes the following specific steps: The image recognition algorithm is used to identify the leaf edge in the leaf side image during the L1 period, and a number of edge feature points are selected on the leaf edge, and a sample edge feature point is randomly selected from the multiple edge feature points obtained; Draw a straight line perpendicular to the blade edge through the sample edge feature point to obtain the blade side marking line, identify the dust accumulation area in the blade side marking line, obtain the marking line dust accumulation area, obtain the length of the line segment occupied by the marking line dust accumulation area in the blade side marking line, and obtain the dust accumulation thickness value of the area corresponding to the sample edge feature point; The dust thickness value of the region corresponding to each edge feature point is obtained respectively, and the average value of the obtained dust thickness values ​​of the multiple regions is calculated to obtain the dust thickness on the side of the L1 blade.

[0013] Furthermore, the step S3 further includes the following specific steps: Step S31: obtaining blade vibration monitoring data, and obtaining blade vibration monitoring values ​​L1 to La according to the blade vibration monitoring data; Step S32: obtaining blade dust accumulation monitoring data, and obtaining blade dust accumulation index values ​​L1 to La according to the blade dust accumulation monitoring data; Step S33: performing polynomial fitting based on the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust accumulation index value to the La blade dust accumulation index value, and creating a dust accumulation prediction model based on the fitting results; Step S34: Marking a blade dust accumulation real-time prediction period with the time point corresponding to the current moment as the end time point of the period, acquiring the blade vibration monitoring value corresponding to the blade dust accumulation real-time prediction period to obtain a real-time blade vibration monitoring value, substituting the real-time blade vibration monitoring value into the dust accumulation prediction model, acquiring the blade dust accumulation index value output by the dust accumulation prediction model to obtain a real-time blade dust accumulation index value; Step S35: Obtain the blade dust accumulation index reference interval. If the real-time blade dust accumulation index value is within the blade dust accumulation index reference interval, there is no need to issue a blade cleaning warning. If the real-time blade dust accumulation index value is not within the blade dust accumulation index reference interval, a blade cleaning warning is issued.

[0014] Furthermore, the step S33 further includes the following specific steps: A polynomial fitting function is established by converting the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust accumulation index value to the La blade dust accumulation index value; Substitute the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust index value to the La blade dust index value into the polynomial fitting function respectively, and calculate the residual sum of squares function RSS of the blade dust index value; The coefficients a0 to a1 of the polynomial fitting function in the residual sum of squares function RSS are respectively n Calculate partial derivatives and get n numbers containing unknown numbers a0 to a n Function expression, and convert n unknown numbers a0 to a n The function expressions of are combined to obtain n groups containing a0 to a n The equations of a0 to a are obtained by solving the equations. n Specific value of Change a0 to a n The specific value of is substituted back into the polynomial fitting function to obtain the dust accumulation prediction model.

[0015] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention creates a dust accumulation prediction model by analyzing historical data to provide early warning of dust accumulation in mechanical equipment. This eliminates the need to use multiple high-precision probes to analyze blade vibration spectrum data, effectively saving equipment operation and maintenance costs. 2. The present invention predicts dust accumulation on rotating machinery blades in real time based on blade dust accumulation monitoring data and blade vibration monitoring data in historical periods, and issues dust accumulation warnings based on the prediction results, thereby improving the timeliness of the warning process. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0017] Figure 1 It is a diagram of the implementation steps of the present invention; Figure 2 It is a line graph of the blade vibration period of the present invention. DETAILED DESCRIPTION

[0018] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] Example 1 See also Figure 1 The blade dust accumulation monitoring device in the present invention belongs to the image recognition technology in computer vision. Now a technical solution is provided: a method for predicting the dust accumulation trend of rotating machinery blades based on the PCFC algorithm, including the following specific steps: Step S1: acquiring historical vibration data of a rotating machinery blade, selecting a plurality of historical monitoring periods from the acquired historical vibration data of the rotating machinery blade, performing vibration monitoring on the rotating machinery blade in each historical monitoring period, and acquiring a blade vibration monitoring value corresponding to each historical monitoring period to obtain blade vibration monitoring data; The step S1 further includes the following specific steps: Acquiring blade vibration monitoring data corresponding to the rotating machinery blade to obtain historical vibration data of the rotating machinery blade, and marking the historical vibration data of the rotating machinery blade as a blade vibration history cycle; It should be noted here that: In the present application, the duration corresponding to the blade vibration history cycle involved here is specifically three months.

[0020] The blade vibration history period is divided into a number of history monitoring periods, and the divided history monitoring periods are marked as L1 history monitoring period to La history monitoring period in chronological order; It should be noted here that: In this application, L mentioned here is specifically the sign symbol corresponding to the historical monitoring period, and a mentioned here is the quantity value corresponding to the historical monitoring period, and a is an integer greater than 0; In this application, the durations corresponding to the L1 historical monitoring period to the La historical monitoring period are equal, and the corresponding durations are all 24 hours.

[0021] Perform blade historical vibration index monitoring on the rotating machinery blade in the L1 historical monitoring period, obtain the blade vibration monitoring value corresponding to the L1 historical monitoring period according to the monitoring result, and name it as the L1 blade vibration monitoring value; The details are as follows: The L1 historical monitoring period is divided into several consecutive blade vibration time points, and the marked blade vibration time points are marked as Z1 blade vibration time point to Zb blade vibration time point in chronological order; It should be noted here that: In the present application, Z referred to here is the sign symbol corresponding to the blade vibration time point, b referred to here is the quantity value corresponding to the blade vibration time point, and b is an integer greater than 0.

[0022] The vibration sensor is used to obtain the blade vibration values ​​of the rotating machine blade corresponding to the blade vibration time point Z1 to the blade vibration time point Zb, and the blade vibration values ​​Z1 to Zb are obtained; In the existing plane rectangular coordinate system, the blade vibration time point is used as the horizontal coordinate and the blade vibration value is used as the vertical coordinate to create a blade vibration plane rectangular coordinate system; See also Figure 2 In the rectangular coordinate system of the blade vibration plane, the Z1 blade vibration time point is used as the abscissa and the Z1 blade vibration value is used as the ordinate, and the coordinate point is marked as the Z1 vibration coordinate point. The Z2 blade vibration time point is used as the abscissa and the Z2 blade vibration value is used as the ordinate, and the coordinate point is marked as the Z2 vibration coordinate point. Similarly, the Zb blade vibration time point is used as the abscissa and the Zb blade vibration value is used as the ordinate, and the Zb vibration coordinate point is connected from the Z1 vibration coordinate point to the Zb vibration coordinate point in sequence to obtain a blade vibration period line graph. In the blade vibration period line graph, the maximum ordinate value is marked as the first blade vibration peak value, the ordinate value whose ordinate value is only less than the first blade vibration peak value is marked as the second blade vibration peak value, the minimum ordinate value is marked as the first blade vibration valley value, and the ordinate value whose ordinate value is only greater than the first blade vibration valley value is marked as the second blade vibration valley value; It should be noted here that: In this application, the vibration value involved here is specifically the vibration frequency value.

[0023] Calculate the average value of the Z1 blade vibration value to the Zb blade vibration value to obtain the average vibration value of the time period; The L1 blade vibration monitoring value is obtained by calculating the first blade vibration peak value, the second blade vibration peak value, the first blade vibration valley value, the second blade vibration valley value and the average vibration value of the time period; The L1 blade vibration monitoring value is calculated using the following formula: ; Among them, Zjl1 is the vibration monitoring value of the L1 blade, Zf1 is the vibration peak value of the first blade, Zf2 is the vibration peak value of the second blade, Zg1 is the vibration valley value of the first blade, Zg2 is the vibration valley value of the second blade, and Zzp is the average vibration value of the time period; Repeat the process of obtaining the blade vibration monitoring value corresponding to the L1 historical monitoring period to obtain the L2 blade vibration monitoring value to the La blade vibration monitoring value respectively; The L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 historical monitoring period to the La historical monitoring period are defined as blade vibration monitoring data; Step S2: performing blade dust accumulation monitoring on the rotating machinery blades in each historical monitoring period based on the blade vibration monitoring data, obtaining a blade dust accumulation index value corresponding to each historical monitoring period based on the monitoring results, and obtaining blade dust accumulation monitoring data; The step S2 further includes the following specific steps: Obtain blade dust accumulation monitoring data, and obtain historical monitoring periods L1 to La based on the blade dust accumulation monitoring data; Capturing frontal images of the rotating machinery blades during the L1 historical monitoring period to the La historical monitoring period, and obtaining frontal images of the blades during the L1 period to the La period; Acquire side images of the rotating machinery blades during the L1 historical monitoring period to the La historical monitoring period, and obtain side images of the blades during the L1 period to the La period; It should be noted here that: In this application, the front image of the blade in the L1 period to the front image of the blade in the La period involved here are all the front images of the blade collected from the L1 historical monitoring period to the end time point of the La historical monitoring period, and the front image of the blade in the L1 period to the front image of the blade in the La period involved here are all the side images of the blade collected from the L1 historical monitoring period to the end time point of the La historical monitoring period; In this application, the image shooting parameters corresponding to the front image of the blade in the L1 period to the front image of the blade in the La period are the same, and the image shooting parameters corresponding to the side image of the blade in the L1 period to the side image of the blade in the La period are the same. The image shooting parameters involved here include but are not limited to shooting angle, shooting focal length and image resolution; The dust accumulation area of ​​the leaf front image during the L1 period is analyzed, and the L1 leaf front dust area ratio is obtained based on the analysis results; The details are as follows: Obtain an image of the front side of the leaf during the L1 period, use an image recognition algorithm to identify the dust accumulation area in the image of the front side of the leaf during the L1 period, obtain a first image feature area, and mark the non-dust accumulation area in the image of the front side of the leaf during the L1 period as a second image feature area; Acquiring area values ​​of the first image feature region and the second image feature region respectively to obtain an area value of the first image region and an area value of the second image region; The area value of the first image region and the area value of the second image region are calculated to obtain the gray ratio of the front surface of the L1 leaf; The L1 leaf positive dust ratio is calculated using the following formula: ; Among them, Jmb1 is the gray area ratio of the leaf L1, Mj1 is the area value of the first image area, and Mj2 is the area value of the second image area; Repeat the process of obtaining the dust accumulation ratio of the leaf L1, and analyze the dust accumulation area of ​​the leaf front images from the L2 period to the La period, and obtain the dust accumulation ratio of the leaf L2 to the La leaf front. Analyze the dust accumulation thickness of the blade side image during the L1 period, and obtain the dust accumulation thickness of the blade side during the L1 period based on the analysis results; The details are as follows: The image recognition algorithm is used to identify the leaf edge in the leaf side image during the L1 period, and a number of edge feature points are selected on the leaf edge, and a sample edge feature point is randomly selected from the multiple edge feature points obtained; Draw a straight line perpendicular to the blade edge through the sample edge feature point to obtain the blade side marking line, identify the dust accumulation area in the blade side marking line, obtain the marking line dust accumulation area, obtain the length of the line segment occupied by the marking line dust accumulation area in the blade side marking line, and obtain the dust accumulation thickness value of the area corresponding to the sample edge feature point; Repeat the process of obtaining the dust thickness value of the area corresponding to the sample edge feature point, obtain the dust thickness value of the area corresponding to each edge feature point respectively, and calculate the average value of the obtained dust thickness values ​​of multiple areas to obtain the dust thickness on the side of the L1 blade; Repeat the process of obtaining the dust accumulation thickness on the L1 blade side, and analyze the dust accumulation thickness of the blade side images from the L2 period to the La period, and obtain the dust accumulation thickness on the L2 blade side to the La blade side; Calculate the product of the dust thickness on the side of the L1 blade and the dust ratio of the front of the L1 blade to obtain the dust index value of the L1 blade. Calculate the product of the dust thickness on the side of the L2 blade and the dust ratio of the front of the L2 blade to obtain the dust index value of the L2 blade. Similarly, calculate the product of the dust thickness on the side of the La blade and the dust ratio of the front of the La blade to obtain the dust index value of the La blade. The L1 blade dust accumulation index value to the La blade dust accumulation index value are defined as the blade dust accumulation monitoring data; Step S3: performing real-time dust accumulation prediction on the rotating machinery blades based on the blade dust accumulation monitoring data and the blade vibration monitoring data, and issuing a dust accumulation warning based on the prediction results; The step S3 further includes the following specific steps: Acquire blade vibration monitoring data, and acquire blade vibration monitoring values ​​L1 to La according to the blade vibration monitoring data; Obtain blade dust accumulation monitoring data, and obtain blade dust accumulation index values ​​L1 to La according to the blade dust accumulation monitoring data; Perform polynomial fitting based on the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust accumulation index value to the La blade dust accumulation index value, and create a dust accumulation prediction model based on the fitting results; The details are as follows: A polynomial fitting function is established by converting the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust accumulation index value to the La blade dust accumulation index value; The polynomial fitting function is as follows: ; Among them, y is the blade dust accumulation index value, x is the blade vibration monitoring value, a0 to a n The coefficients of the polynomial fitting function, n is the order of the polynomial fitting function; Substitute the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust index value to the La blade dust index value into the polynomial fitting function respectively, and calculate the residual sum of squares function RSS of the blade dust index value; The coefficients a0 to a of the polynomial fitting function in the residual sum of squares function RSS are respectively n Calculate partial derivatives and get n numbers containing unknown numbers a0 to a n Function expression, and convert n unknown numbers a0 to a n The function expressions of are combined to obtain n groups containing a0 to a n The equations of a0 to a are obtained by solving the equations. n Specific value of Change a0 to a n Substitute the specific value of into the polynomial fitting function to obtain the dust accumulation prediction model; The time point corresponding to the current moment is used as the end time point of the period to mark a blade dust accumulation real-time prediction period, and the blade vibration monitoring value corresponding to the blade dust accumulation real-time prediction period is obtained to obtain the real-time blade vibration monitoring value, and the real-time blade vibration monitoring value is substituted into the dust accumulation prediction model, and the blade dust accumulation index value output by the dust accumulation prediction model is obtained to obtain the real-time blade dust accumulation index value; Obtain the blade dust accumulation index reference interval. If the real-time blade dust accumulation index value is within the blade dust accumulation index reference interval, there is no need to issue a blade dust cleaning warning. If the real-time blade dust accumulation index value is not within the blade dust accumulation index reference interval, a blade dust cleaning warning needs to be issued. It should be noted here that: In this application, the situation where the blade dust cleaning warning is issued here includes the situation where the real-time blade dust accumulation index value is at the boundary of the blade dust accumulation index reference range; The blade dust accumulation index benchmark interval is obtained as follows: The lower limit of the blade dust accumulation index benchmark interval involved here is 0, that is, there is no dust accumulation on the blades of the rotating machinery. Several historical monitoring periods in which blade dust cleaning warnings have been carried out are obtained, and the blade dust accumulation index value corresponding to each historical monitoring period is obtained. The blade dust accumulation index value with the smallest value is marked as the upper limit of the blade dust accumulation index benchmark interval to obtain the blade dust accumulation index benchmark interval.

[0024] In this application, if a corresponding calculation formula appears, the above calculation formula is dimensionless and its numerical calculation is performed. The weight coefficient, proportional coefficient and other coefficients in the formula are set to a result value obtained by quantifying each parameter. Regarding the size of the weight coefficient and the proportional coefficient, as long as it does not affect the proportional relationship between the parameter and the result value, it is acceptable.

[0025] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm, characterized in that: include: Step S1: acquiring historical vibration data of a rotating machinery blade, selecting a plurality of historical monitoring periods from the acquired historical vibration data of the rotating machinery blade, performing vibration monitoring on the rotating machinery blade in each historical monitoring period, acquiring a blade vibration monitoring value corresponding to each historical monitoring period, and obtaining blade vibration monitoring data; Step S2: performing blade dust accumulation monitoring on the rotating machinery blades in each historical monitoring period based on the blade vibration monitoring data, obtaining a blade dust accumulation index value corresponding to each historical monitoring period based on the monitoring results, and obtaining blade dust accumulation monitoring data; Step S3: Perform real-time dust accumulation prediction on the blades of the rotating machinery based on the blade dust accumulation monitoring data and the blade vibration monitoring data, and issue a dust accumulation warning based on the prediction results.

2. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 1, characterized in that: The step S1 further includes the following specific steps: Step S11: acquiring blade vibration monitoring data corresponding to the rotating machinery blade to obtain historical vibration data of the rotating machinery blade, and marking the historical vibration data of the rotating machinery blade as a blade vibration history cycle; Step S12: Divide the blade vibration history period into L1 history monitoring period to La history monitoring period; Step S13: monitoring the blade historical vibration index of the rotating machinery blade in the L1 historical monitoring period, and obtaining the L1 blade vibration monitoring value according to the monitoring result; Step S14: acquiring the L2 blade vibration monitoring value to the La blade vibration monitoring value respectively; Step S15: defining the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 historical monitoring period to the La historical monitoring period as blade vibration monitoring data.

3. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 2, characterized in that: The step S13 further includes the following specific steps: Step S131: Divide the L1 historical monitoring period into the blade vibration time point Z1 to the blade vibration time point Zb; Step S132: obtaining blade vibration values ​​corresponding to the blade vibration time point Z1 to the blade vibration time point Zb of the rotating machinery, and obtaining blade vibration values ​​Z1 to Zb; Step S133: creating a blade vibration period line graph using the Z1 blade vibration value to the Zb blade vibration value and the Z1 blade vibration time point to the Zb blade vibration time point.

4. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 2, characterized in that: The step S13 further includes the following specific steps: Step S134: In the blade vibration period line graph, the maximum ordinate value is marked as the first blade vibration peak value, the ordinate value whose ordinate value is just smaller than the first blade vibration peak value is marked as the second blade vibration peak value, the minimum ordinate value is marked as the first blade vibration valley value, and the ordinate value whose ordinate value is just larger than the first blade vibration valley value is marked as the second blade vibration valley value; Step S135: Calculate the average value of the blade vibration values ​​from Z1 to Zb to obtain the average vibration value of the time period; Step S136: Calculate the first blade vibration peak value Zf1, the second blade vibration peak value Zf2, the first blade vibration valley value Zg1, the second blade vibration valley value Zg2 and the time period average vibration value Zzp to obtain the L1 blade vibration monitoring value Zjl1.

5. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 3, characterized in that: The step S133 further includes the following specific steps: The blade vibration time point is used as the horizontal coordinate and the blade vibration value is used as the vertical coordinate to create a blade vibration plane rectangular coordinate system; In the rectangular coordinate system of the blade vibration plane, the Z1 blade vibration time point is used as the horizontal coordinate, and the Z1 blade vibration value is used as the vertical coordinate, and the coordinate point is marked as the Z1 vibration coordinate point. Similarly, the Zb blade vibration time point is used as the horizontal coordinate, and the Zb blade vibration value is used as the vertical coordinate, and the coordinate point is marked as the Zb vibration coordinate point. The Z1 vibration coordinate point to the Zb vibration coordinate point are connected in sequence to obtain a blade vibration period line graph.

6. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 1, characterized in that: The step S2 further includes the following specific steps: Step S21: Obtain blade dust accumulation monitoring data, and obtain historical monitoring periods L1 to La based on the blade dust accumulation monitoring data; Step S22: collecting front images of the rotating machinery blades in the historical monitoring period L1 to the historical monitoring period La, and obtaining front images of the blades in the L1 period to the La period; Step S23: collecting side images of the rotating machinery blades in the L1 historical monitoring period to the La historical monitoring period, and obtaining the side images of the blades in the L1 period to the La period; Step S24: performing dust accumulation analysis on the front image of the blade during the L1 period, and obtaining the L1 blade front dust area ratio according to the analysis result; Step S25: performing dust accumulation area analysis on the front image of the blade from the L2 period to the La period, respectively, to obtain the dust accumulation ratio of the front of the blade from L2 to the front of the blade from La; Step S26: Analyze the dust accumulation thickness of the blade side image during the L1 period, and obtain the dust accumulation thickness of the blade side during the L1 period according to the analysis result; Step S27: performing dust accumulation thickness analysis on the blade side images from the L2 period to the La period, respectively, to obtain the dust accumulation thickness on the blade side from the L2 period to the La period; Step S28: Calculate the product of the dust thickness on the side of the L1 blade and the dust ratio on the front of the L1 blade to obtain the dust index value of the L1 blade. Similarly, calculate the product of the dust thickness on the side of the La blade and the dust ratio on the front of the La blade to obtain the dust index value of the La blade. Step S29: defining the blade dust accumulation index value L1 to the blade dust accumulation index value La as blade dust accumulation monitoring data.

7. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 6, characterized in that: The step S24 further includes the following specific steps: Obtain an image of the front side of the leaf during the L1 period, use an image recognition algorithm to identify the dust accumulation area in the image of the front side of the leaf during the L1 period, obtain a first image feature area, and mark the non-dust accumulation area in the image of the front side of the leaf during the L1 period as a second image feature area; Acquiring area values ​​of the first image feature region and the second image feature region respectively to obtain an area value of the first image region and an area value of the second image region; The first image region area value Mj1 and the second image region area value Mj2 are calculated to obtain the L1 leaf front gray area ratio Jmb1.

8. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 6, characterized in that: The step S26 further includes the following specific steps: The image recognition algorithm is used to identify the leaf edge in the leaf side image during the L1 period, and a number of edge feature points are selected on the leaf edge, and a sample edge feature point is randomly selected from the multiple edge feature points obtained; Draw a straight line perpendicular to the blade edge through the sample edge feature point to obtain the blade side marking line, identify the dust accumulation area in the blade side marking line, obtain the marking line dust accumulation area, obtain the length of the line segment occupied by the marking line dust accumulation area in the blade side marking line, and obtain the dust accumulation thickness value of the area corresponding to the sample edge feature point; The dust thickness value of the region corresponding to each edge feature point is obtained respectively, and the average value of the obtained dust thickness values ​​of the multiple regions is calculated to obtain the dust thickness on the side of the L1 blade.

9. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 1, characterized in that: The step S3 further includes the following specific steps: Step S31: obtaining blade vibration monitoring data, and obtaining blade vibration monitoring values ​​L1 to La according to the blade vibration monitoring data; Step S32: obtaining blade dust accumulation monitoring data, and obtaining blade dust accumulation index values ​​L1 to La according to the blade dust accumulation monitoring data; Step S33: performing polynomial fitting based on the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust accumulation index value to the La blade dust accumulation index value, and creating a dust accumulation prediction model based on the fitting results; Step S34: Marking a blade dust accumulation real-time prediction period with the time point corresponding to the current moment as the end time point of the period, acquiring the blade vibration monitoring value corresponding to the blade dust accumulation real-time prediction period to obtain a real-time blade vibration monitoring value, substituting the real-time blade vibration monitoring value into the dust accumulation prediction model, acquiring the blade dust accumulation index value output by the dust accumulation prediction model to obtain a real-time blade dust accumulation index value; Step S35: Obtain the blade dust accumulation index reference interval. If the real-time blade dust accumulation index value is within the blade dust accumulation index reference interval, there is no need to issue a blade cleaning warning. If the real-time blade dust accumulation index value is not within the blade dust accumulation index reference interval, a blade cleaning warning is issued.

10. The method for predicting dust accumulation trend of rotating machinery blades based on PCFC algorithm according to claim 9, characterized in that: The step S33 further includes the following specific steps: A polynomial fitting function is established by converting the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust accumulation index value to the La blade dust accumulation index value; Substitute the L1 blade vibration monitoring value to the La blade vibration monitoring value and the L1 blade dust index value to the La blade dust index value into the polynomial fitting function respectively, and calculate the residual sum of squares function RSS of the blade dust index value; The coefficients a0 to a1 of the polynomial fitting function in the residual sum of squares function RSS are respectively n Calculate partial derivatives and get n numbers containing unknown numbers a0 to a n Function expression, and convert n unknown numbers a0 to a n The function expressions of are combined to obtain n groups containing a0 to a n The equations of a0 to a are obtained by solving the equations. n Specific value of Change a0 to a n The specific value of is substituted back into the polynomial fitting function to obtain the dust accumulation prediction model.

Images