Elevator performance analysis system and method based on mixed signals

Abstract

The invention relates to an elevator performance analysis system and method based on mixed signals, and the system comprises a sensing subsystem which is used for collecting original mixed signals in the elevator operation process; the main control subsystem is used for inputting the original mixed signal into an elevator performance analysis model to obtain an elevator analysis and evaluation result; the elevator performance analysis model is obtained by training a neural network model through standard performance indexes and a characteristic signal set; the characteristic signal set is obtained by processing an original mixed signal set. The method can be suitable for all elevators, and the problem that old elevators are not matched with current standards is effectively solved; in addition, the system is easy to operate, has no professional threshold limitation on users, has a relatively high intelligent level, remarkably reduces waste of time and labor cost, and has a wide application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field, and in particular to an elevator performance analysis system and method based on mixed signals. Background Art

[0002] With the advancement of science and technology, my country's construction industry has developed rapidly. Today, the use of elevators in medium and high-rise buildings is becoming more and more common. At present, my country has a large number of elevators. As special equipment, elevators have high maintenance costs. At the same time, old elevators do not meet current standards and have low intelligence levels. There are no clear regulations for elevator use. The actual use of elevators seriously exceeds the design requirements. It is impossible to predict elevator failures and troubleshoot them in advance. Usually, maintenance is carried out after the failure occurs, which greatly increases the probability of serious accidents caused by elevator failures. Elevator operating performance includes basic function realization, operating quality, etc. Elevator operating performance evaluation combined with the elevator's operating status can solve the above problems to a certain extent. Summary of the Invention

[0003] In order to solve the problems existing in the prior art, the purpose of the present invention is to provide an elevator performance analysis system and method based on mixed signals to overcome the problem that old elevators do not match the current standards. The system is simple to operate and has a high level of intelligence, which greatly reduces the loss of time and labor costs.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] An elevator performance analysis system based on mixed signals, comprising:

[0006] The sensor subsystem is used to collect the original mixed signals during the operation of the elevator;

[0007] The main control subsystem is used to input the original mixed signal into the elevator performance analysis model to obtain elevator analysis and evaluation results; the elevator performance analysis model is obtained by training a neural network model using standard performance indicators and a characteristic signal set; the characteristic signal set is obtained by processing the original mixed signal set.

[0008] Optionally, the sensing subsystem includes:

[0009] Capacitive acceleration sensor module, used to collect the original car acceleration signal;

[0010] Piezoelectric acceleration sensor module, used to collect original car vibration signals;

[0011] Acoustic signal sensing module, used to collect original car operation sound signals;

[0012] Current sensing module, used to collect the original door machine current signal;

[0013] The encoding module is used to collect the original traction sheave and wire rope operation data.

[0014] Optionally, the main control subsystem includes:

[0015] A program running module, configured to process the original mixed signal set to obtain the characteristic signal set;

[0016] a neural network processing module, configured to input the original mixed signal into an elevator performance analysis model to obtain elevator analysis and evaluation results; the elevator performance analysis model is obtained by training a neural network model using standard performance indicators and a characteristic signal set;

[0017] The information storage module is used to store the original mixed signal set, the characteristic signal set and other redundant information.

[0018] Optionally, the characteristic signal set includes: speed variation mean, vertical acceleration peak, running vertical vibration frequency peak, running vertical vibration kurtosis peak, vertical running cepstral coefficient, door opening and closing cepstral coefficient, door machine current peak, no-load downward unit displacement, no-load upward relative braking distance, door lock engagement depth, door clearance, and elevator standard indicators.

[0019] Optionally, the program execution module includes:

[0020] The mean value acquisition unit is used to set the elevator rated speed and the rated acceleration time to reach the rated speed of the elevator, and to determine the speed variation mean value in combination with the original car acceleration signal:

[0021]

[0022]

[0023] Among them, Q i is the mean value of speed variation, j is the state variable, 0 / 1 represents the rated state and actual state respectively, ΔT i 1 is the acceleration time from 5% speed to i characteristic speed in actual state, ΔT i 0 ΔT is the acceleration time from 5% speed to i characteristic speed under rated state. i j T is the acceleration time from 5% speed to characteristic speed in actual state, i j is the acceleration time from rest to characteristic speed i in state j, is the acceleration time from standstill to 5% speed characteristic in state j;

[0024] The first peak value acquisition unit is used to determine the vertical acceleration according to the original car acceleration signal. and when When , the vertical jerk is taken as the vertical jerk peak value;

[0025] Where a is the vertical running acceleration, t is the selected sensor sampling time, and T is the elevator car running time.

[0026] Optionally, the program execution module further includes:

[0027] a second peak value acquisition unit, configured to perform noise reduction processing on the original car vibration signal, perform Fourier transform on the vibration signal after noise reduction processing, and extract the frequency peak value of the high frequency band as the running vertical vibration frequency peak value;

[0028] The third peak value acquisition unit is used to use the vibration signal after the noise reduction processing to obtain the kurtosis of the vibration signal and the peak value corresponding to the kurtosis, and use the peak value corresponding to the kurtosis as the running vertical vibration kurtosis peak value.

[0029] Optionally, the program execution module further includes:

[0030] a fourth peak value acquisition unit, configured to perform noise reduction processing on the original door machine current signal, and obtain the current peak value during the door opening and closing phase according to the current signal after noise reduction processing;

[0031] The coefficient acquisition unit is used to perform wavelet packet decomposition and noise reduction on the original car operation sound signals in the elevator vertical operation stage and the elevator door opening and closing operation stage, perform Mel filtering on the highest frequency in the sound signal after wavelet packet decomposition and noise reduction, and perform logarithmic and discrete cosine transforms at the same time to obtain the vertical operation cepstral coefficients and the door opening and closing cepstral coefficients respectively.

[0032] Optionally, the program execution module further includes:

[0033] The displacement acquisition unit is used to obtain the displacement of the elevator's upward and downward wire ropes respectively based on the original traction sheave and wire rope operation data, and determine the acquisition time between two adjacent pulses of the traction sheave, divide the displacement of the elevator's upward and downward wire ropes by the acquisition time respectively to obtain the traction sheave speed, and further determine the displacement of the wire rope at the target moment and when it finally stops moving in combination with the time when the traction sheave speed starts to decrease, subtract the displacement at the target moment from the displacement when it finally stops moving, and respectively obtain the no-load downward unit displacement and no-load upward relative braking distance.

[0034] To achieve the above object, the present invention further provides an elevator performance analysis method based on mixed signals, comprising:

[0035] Original mixed signals are collected during elevator operation, and the original mixed signals are input into an elevator performance analysis model to obtain elevator analysis and evaluation results. The elevator performance analysis model is obtained by training a neural network model using standard performance indicators and a feature signal set. The feature signal set is obtained by processing the original mixed signal set.

[0036] The beneficial effects of the present invention are:

[0037] The present invention can comprehensively analyze and evaluate the basic functions, operating quality and other operating performance of an elevator. Relying on the mixed signal set generated during the operation of the elevator, the present invention extracts the characteristic signal set from the mixed signal, and uses a neural network to solve the optimal indicator weights for performance analysis based on standard rating indicators.

[0038] The present invention can be applied to all elevators, overcoming the problem that old elevators do not match current standards. The method of use is relatively simple, there is no threshold restriction for users, and it has a high level of intelligence, which greatly reduces the loss of time and labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 Schematic diagram of an elevator performance analysis system based on mixed signals according to an embodiment of the present invention;

[0041] Figure 2 Schematic diagram of the original mixed signal extraction process according to an embodiment of the present invention;

[0042] Figure 3 Schematic diagram of a feature signal set extraction process according to an embodiment of the present invention;

[0043] Figure 4 Schematic diagram of an elevator performance analysis method based on mixed signals according to an embodiment of the present invention;

[0044] Among them, 1. Main control system; 2. Power supply system; 3. Sensing system; 4. Information storage module; 5. Neural network processor; 6. Program running module; 7. Capacitive accelerometer; 8. Piezoelectric accelerometer; 9. Acoustic signal sensor; 10. Current sensor; 11. Pressure sensor; 12. Original mixed signal set; 13. Original car acceleration signal; 14. Original car vibration signal; 15. Original car operation sound signal; 16. Original door machine current signal; 17. Original traction sheave and wire rope operation data; 18. Door lock engagement depth; 19. Door clearance; 20. Elevator standard indicators; 21. Speed variation mean; 22. Vertical jerk peak; 23. Running vertical vibration frequency peak; 24. Running vertical vibration kurtosis peak; 25. Vertical operation cepstral coefficient; 26. Door opening and closing cepstral coefficient; 27. Door machine current peak; 28. No-load downward unit displacement; 29. No-load upward relative braking distance. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the 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.

[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] like Figure 1 As shown, this embodiment discloses an elevator performance analysis system based on mixed signals, including: a sensing subsystem for collecting original mixed signals during elevator operation; a main control subsystem for inputting the original mixed signals into an elevator performance analysis model to obtain elevator analysis and evaluation results; the elevator performance analysis model is obtained by training a neural network model using standard performance indicators and a feature signal set; and the feature signal set is obtained by processing the original mixed signal set 12.

[0048] Specifically, the system includes a main control system 1 , a power supply system 2 and a sensor system 3 .

[0049] The main control system 1 includes an information storage module 4, a neural network processor 5, and a program execution module 6. The information storage module 4 stores system data, including the original mixed signal set 12, the characteristic signal set, and system operation information. The neural network processor 5 is used to deploy a neural network model. By importing standard elevator performance indicators 20 and training the characteristic signal set, the model obtains the optimal indicator weights for performance analysis. The program execution module 6 processes the original mixed signal set 12 and extracts the characteristic signal set.

[0050] The power supply system 2 is responsible for the power supply of the entire system.

[0051] The sensing system 3 is composed of five sensors: a capacitive acceleration sensor 7, a piezoelectric acceleration sensor 8, an acoustic signal sensor 9, a current sensor 10 and an encoder 11. It is used to collect the original mixed signals during the operation of the elevator and transmit the parameters to the information storage module.

[0052] Furthermore, the sensing subsystem includes: a capacitive acceleration sensing module for collecting the original car acceleration signal; a piezoelectric acceleration sensing module for collecting the original car vibration signal; an acoustic signal sensing module for collecting the original car operation sound signal; a current sensing module for collecting the original door machine current signal; and an encoding module for collecting the original traction wheel and wire rope operation data.

[0053] Specifically, the aforementioned sensing system 3 is installed at a designated position in the car to measure the original car acceleration signal 13, the original car vibration signal 14, the original car operation sound signal 15, the original door machine current signal 16, and the original traction sheave and wire rope operation data 17. The time domain T of the signals involved in the present invention is 0 to 30 seconds. Figure 2 shown.

[0054] Furthermore, the main control subsystem includes: a program running module, which is used to process the original mixed signal set 12 and obtain a characteristic signal set; a neural network processing module, which is used to input the original mixed signal into the elevator performance analysis model to obtain elevator analysis and evaluation results; the elevator performance analysis model is obtained by training the neural network model using standard performance indicators and characteristic signal sets; and an information storage module, which is used to store the original mixed signal set 12, the characteristic signal set and other redundant information.

[0055] Furthermore, the characteristic signal set includes: speed variation mean, vertical acceleration peak, running vertical vibration frequency peak, running vertical vibration kurtosis peak, vertical running cepstral coefficient, door opening and closing cepstral coefficient, door machine current peak, no-load downward unit displacement, no-load upward relative braking distance, door lock engagement depth, door clearance, and elevator standard indicators.

[0056] Specifically, the original mixed signal is obtained by extracting features through the program running module 6, including the speed variation mean 21, the vertical acceleration peak 22, the running vertical vibration frequency peak 23, the running vertical vibration kurtosis peak 24, the vertical running cepstral coefficient 25, the door opening and closing cepstral coefficient 26, the door machine current peak 27, the no-load downward unit displacement 28, the no-load upward relative braking distance 29, the door lock engagement depth 18, the door gap 19, and the elevator standard index 20, such as Figure 3 shown.

[0057] Furthermore, the program running module includes: a mean value acquisition unit for setting the elevator rated speed and the rated acceleration time to reach the elevator rated speed, and determining the speed variation mean value in combination with the original car acceleration signal. A first peak value acquisition unit for determining the vertical acceleration according to the original car acceleration signal. and when When , the vertical jerk is taken as the vertical jerk peak value.

[0058] Specifically, the original car acceleration signal 13 is obtained after feature extraction by the program running module 6 .

[0059] The speed variation mean 21 and the vertical acceleration peak 22 reflect the traction performance of the elevator car. Assume that the rated speed of the elevator is V, 25% V, 50% V, and 75% V are the speed characteristic points, the starting speed is 5% V, and the rated acceleration time to reach the characteristic point is T i 0 , the actual acceleration time is T i 1 , the process of solving the velocity variation:

[0060] Take i=25%V, 50%V, 75%V, The variability

[0061] The process of calculating the vertical jerk peak value:

[0062] Vertical jerk 22 when When a 1 is the peak value.

[0063] Furthermore, the program running module also includes: a second peak acquisition unit, which is used to perform noise reduction processing on the original car vibration signal, perform Fourier transform on the vibration signal after noise reduction processing, and extract the frequency peak of the high-frequency band as the running vertical vibration frequency peak; a third peak acquisition unit, which is used to use the vibration signal after noise reduction processing to obtain the kurtosis of the vibration signal and the peak value corresponding to the kurtosis, and use the peak value corresponding to the kurtosis as the running vertical vibration kurtosis peak value.

[0064] Specifically, the original car vibration signal 14 is obtained after feature extraction by the program running module 6 .

[0065] The operating vertical vibration frequency peak 23 and the operating vertical vibration kurtosis peak 24 reflect the overall operating performance of the elevator. To calculate the operating vertical vibration frequency peak 23, the original elevator car vibration signal 14 during the vertical operation phase in the time domain T is taken, subjected to noise reduction processing, and Fourier transform to extract the frequency peak in the high-frequency band. To calculate the operating vertical vibration kurtosis peak 24, the original elevator car vibration signal 14 during the vertical operation phase in the time domain T is taken, subjected to noise reduction processing, and then its kurtosis is calculated and its peak value is obtained to obtain the operating vertical vibration kurtosis peak 24.

[0066] Furthermore, the program running module also includes: a fourth peak acquisition unit, which is used to perform noise reduction processing on the original door machine current signal, and obtain the current peak value of the door opening and closing stage based on the current signal after noise reduction processing; a coefficient acquisition unit, which is used to perform wavelet packet decomposition and noise reduction on the original car operation sound signal of the elevator vertical operation stage and the elevator door opening and closing operation stage respectively, and perform Mel filtering on the highest frequency in the sound signal after wavelet packet decomposition and noise reduction, and perform logarithmic and discrete cosine transform at the same time to obtain the vertical operation cepstral coefficient and the door opening and closing cepstral coefficient respectively.

[0067] Specifically, the original gate motor current signal 16 is obtained after feature extraction by the program execution module 6. The gate motor current peak value 27 reflects the operating performance of the gate system. The solution is as follows: the original gate motor current signal 16 is subjected to noise reduction processing and the current peak value during the gate opening and closing phases in the time domain T is obtained.

[0068] The original car operation sound signal 15 is obtained after feature extraction by the program operation module 6. The vertical operation cepstral coefficient 25 solution process: take the original car operation sound signal 15 of the elevator vertical operation phase in the time domain T, perform wavelet packet noise reduction, and then perform Mel filtering on the highest frequency while performing logarithmic and discrete cosine transform to obtain the cepstral coefficient. The door opening and closing cepstral coefficient 26 takes the original car operation sound signal 15 of the elevator door opening and closing phase in the time domain T, and the remaining solution process is the same as above. Among them, Mel filtering:

[0069]

[0070] Furthermore, the program running module also includes: a displacement acquisition unit, which is used to obtain the displacement of the elevator's upward and downward wire ropes respectively based on the original traction wheel and wire rope operation data, and determine the acquisition time between two adjacent pulses of the traction wheel, divide the displacement of the elevator's upward and downward wire ropes by the acquisition time respectively, to obtain the traction wheel speed, and further combine the traction wheel speed start to decrease time to determine the displacement of the wire rope at the target moment and when it finally stops moving, subtract the displacement at the target moment from the displacement when it finally stops moving, and obtain the no-load downward unit displacement and the no-load upward relative braking distance respectively.

[0071] Specifically, the unloaded downward unit displacement 28 and the unloaded upward relative braking distance 29 are used to detect the braking performance of the elevator car. The original traction sheave and wire rope operating data 17 are obtained after feature extraction by the program operation module 6. The unloaded downward relative braking distance 28 is solved by taking the encoder pulse signal during the downward phase and calculating the wire rope displacement S. Then, the acquisition time interval between two adjacent traction sheave pulses is obtained and divided to obtain the traction sheave speed V. Assuming the time when the traction sheave speed begins to decrease is T, the wire rope displacements at time t and when it finally stops are S1 and S2, respectively, and the braking distance is S2-S1. The unloaded upward relative braking distance 29 is solved in the same way as above.

[0072] like Figure 4 As shown, this embodiment also provides an elevator performance analysis method based on mixed signals, including: collecting original mixed signals during elevator operation, inputting the original mixed signals into an elevator performance analysis model, and obtaining elevator analysis and evaluation results; the elevator performance analysis model is obtained by training a neural network model using standard performance indicators and a feature signal set; and the feature signal set is obtained by processing the original mixed signal set.

[0073] Specifically, this embodiment discloses an elevator operation performance analysis system based on mixed signals, comprising the following steps:

[0074] S1. Install the sensor system at the designated location of the elevator;

[0075] S2, start the power supply system, detect and store the original mixed signal set;

[0076] S3, start the program running module to extract the characteristic signal set from the original mixed signal set;

[0077] S4. Importing characteristic signal set parameters and elevator standard indicators into a neural network processor, and obtaining parameter distribution weights for performance analysis after training;

[0078] S5. Calculate the overall operating performance parameters of the elevator based on the parameter distribution weights;

[0079] S6. Analyze and evaluate the elevator's operating performance based on the calculated operating performance indicators. If the elevator does not meet the minimum standards, the substandard parts should be maintained and replaced.

[0080] S7. Performance analysis is completed.

[0081] This embodiment provides an elevator performance analysis system based on mixed signals, including a main control system, a power supply system, and a sensor system.

[0082] The main control system consists of an information storage module, a neural network processor, and a program execution module. The information storage module stores system data, including the original mixed signal set, feature signal set, and other redundant information. The neural network processor is used to deploy the neural network model, which is trained by importing standard elevator performance indicators and feature signal sets to obtain the optimal indicator weights for performance analysis. The program execution module processes the original mixed signal set and extracts the feature signal set. The power supply system is responsible for powering the entire system. The sensing system consists of five sensors: a capacitive accelerometer, a piezoelectric accelerometer, an acoustic signal sensor, a current sensor, and an encoder. These sensors are used to collect the original mixed signals during elevator operation and transmit the parameters to the information storage module.

[0083] Further optimization schemes were implemented to obtain the original mixed signal set: a capacitive accelerometer and a piezoelectric accelerometer were installed at the bottom center of the car to accurately collect the original car acceleration and vibration signals; an acoustic signal sensor was magnetically fixed vertically above the door of the running car to accurately collect the original car operation sound signal; a current sensor was installed at the door motor position at the top of the car door to accurately collect the original door machine current signal; an encoder was attached to the wire rope via a rubber U-groove pulley, and the entire encoder was fixed to the base using a magnetic base to accurately collect the original traction sheave and wire rope operation data;

[0084] Further optimization scheme, for the feature extraction of the original mixed signal set, signal preprocessing methods include removing zero drift, Kalman filtering, etc.

[0085] To further optimize the solution, the speed variation mean and vertical acceleration peak are obtained by extracting the original car acceleration signal through the program running module. The speed variation mean and vertical acceleration peak reflect the traction performance of the elevator car. Assume that the rated speed of the elevator is V, 25%V, 50%V, and 75%V are the speed characteristic points, the starting speed is 5%V, and the rated acceleration time to reach the characteristic point is T i 0 , the actual acceleration time is T i 1 , the process of solving the speed variation: take i = 25% V, 50% V, 75% V, The variability The process of calculating the vertical jerk peak value: vertical jerk when When a 1 is the peak value.

[0086] To further optimize the solution, the collected vibration signals were divided into the elevator operation phase and the elevator door opening and closing phase, given that the main working components and vibration sources of the elevator are different during these two phases. The original elevator car vibration signal during the vertical operation phase, in the time domain T, was taken and subjected to noise reduction. After Fourier transform, the high-frequency peak and the peak kurtosis of the vertical vibration during operation were extracted. The noise reduction method for the vertical vibration signal during operation was variational mode decomposition.

[0087] Further optimization scheme, vertical running cepstral coefficients and switch gate cepstral coefficients: the original sound signal is decomposed and reconstructed by wavelet packet, and its square value is obtained after fast Fourier transform, and then it is passed through Mel filter, and its value is subjected to logarithmic discrete cosine transform to finally obtain its cepstral coefficients, where Mel filter:

[0088]

[0089] To further optimize the scheme, the peak current of the gate machine is obtained by performing noise reduction processing on the original gate machine current signal and solving the peak current of the gate opening and closing stage in the time domain T.

[0090] To further optimize the solution, the unloaded downward unit displacement and unloaded upward relative braking distance are derived from the original traction sheave and wire rope operating data after feature extraction by the program execution module. The wire rope displacement S is calculated by taking the encoder pulse signal during the downward phase. The acquisition time T1 between two adjacent traction sheave pulses is then divided to obtain the traction sheave velocity V. Assuming the time at which the traction sheave velocity begins to decrease is T, and the wire rope displacements S1 and S2 at time t and final stop are respectively, the braking distance is calculated as S2-S1. The unloaded upward relative braking distance is calculated using the same process as above.

[0091] To further optimize the solution, the door lock engagement depth and door clearance can be directly detected by the sensing system, reflecting the comprehensive performance of the car door opening and closing.

[0092] This embodiment also discloses a method for analyzing elevator performance based on mixed signals, comprising the following steps:

[0093] A method for analyzing elevator operation performance comprises the following steps:

[0094] S1. Install the sensor system at the designated location of the elevator;

[0095] S2, start the power supply system, detect and store the original mixed signal set;

[0096] S3, start the program running module to extract the characteristic signal set from the original mixed signal set;

[0097] S4. Importing the characteristic signal set parameters and the elevator standard index into the neural network processor, and iteratively training the characteristic signal set to obtain the distribution weight of the optimal index for performance analysis;

[0098] S5. Calculate the overall operating performance parameters of the elevator based on the parameter distribution weights;

[0099] S6. Analyze and evaluate the elevator's operating performance based on the calculated operating performance parameters. If the elevator does not meet the minimum standards, the substandard parts should be maintained or replaced.

[0100] S7. Performance analysis is completed.

[0101] The working process of the present invention is as follows:

[0102] First, the measurement personnel conduct a preliminary inspection of the elevator that needs performance analysis and install all sensors of the sensing system at the designated locations of the tested elevator.

[0103] Secondly, the measurement personnel start the power supply system, run the elevator multiple times, gradually collect the original mixed signal set, and store the measured parameters.

[0104] Then, the program is started to run the module to extract the key information of the original mixed signal set and obtain the characteristic signal set.

[0105] Finally, the characteristic signal set and standard elevator index parameters are imported into the neural network processor. After training, the optimal parameter distribution weights are obtained for performance analysis. The elevator's operating performance parameters are then calculated based on the parameter distribution weights and analyzed and evaluated. Based on the results, if the elevator does not meet the minimum standards, the substandard parts are maintained or replaced, completing the performance analysis.

[0106] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. An elevator performance analysis system based on mixed signals, characterized in that: include: The sensor subsystem is used to collect the original mixed signals during the operation of the elevator; The main control subsystem is used to input the original mixed signal into the elevator performance analysis model to obtain elevator analysis and evaluation results; the elevator performance analysis model is obtained by training a neural network model using standard performance indicators and a characteristic signal set; the characteristic signal set is obtained by processing the original mixed signal set.

2. The elevator performance analysis system based on mixed signals according to claim 1, characterized in that: The sensing subsystem includes: Capacitive acceleration sensor module, used to collect the original car acceleration signal; Piezoelectric acceleration sensor module, used to collect original car vibration signals; Acoustic signal sensing module, used to collect original car operation sound signals; Current sensing module, used to collect the original door machine current signal; The encoding module is used to collect the original traction sheave and wire rope operation data.

3. The elevator performance analysis system based on mixed signals according to claim 1, characterized in that: The main control subsystem includes: A program running module, configured to process the original mixed signal set to obtain the characteristic signal set; a neural network processing module, configured to input the original mixed signal into an elevator performance analysis model to obtain elevator analysis and evaluation results; the elevator performance analysis model is obtained by training a neural network model using standard performance indicators and a characteristic signal set; The information storage module is used to store the original mixed signal set, the characteristic signal set and other redundant information.

4. The elevator performance analysis system based on mixed signals according to claim 1, characterized in that: The characteristic signal set includes: speed variation mean, vertical acceleration peak, running vertical vibration frequency peak, running vertical vibration kurtosis peak, vertical running cepstral coefficient, door opening and closing cepstral coefficient, door machine current peak, no-load downward unit displacement, no-load upward relative braking distance, door lock engagement depth, door clearance, and elevator standard indicators.

5. The elevator performance analysis system based on mixed signals according to claim 3, characterized in that: The program running module includes: The mean value acquisition unit is used to set the elevator rated speed and the rated acceleration time to reach the rated speed of the elevator, and to determine the speed variation mean value in combination with the original car acceleration signal: Among them, Q i is the mean value of speed variation, j is the state variable and takes 0 / 1 to represent the rated state and actual state respectively, is the acceleration time from 5% speed to i characteristic speed under actual conditions, It is the acceleration time from 5% speed to i characteristic speed under rated state. is the acceleration time from 5% speed to i characteristic speed under actual conditions, is the acceleration time from rest to characteristic speed i in state j, is the acceleration time from standstill to 5% speed characteristic in state j; The first peak value acquisition unit is used to determine the vertical acceleration according to the original car acceleration signal. and when When , the vertical jerk is taken as the vertical jerk peak value; Where a is the vertical running acceleration, t is the selected sensor sampling time, and T is the elevator car running time.

6. The elevator performance analysis system based on mixed signals according to claim 3, characterized in that: The program running module also includes: a second peak value acquisition unit, configured to perform noise reduction processing on the original car vibration signal, perform Fourier transform on the vibration signal after noise reduction processing, and extract the frequency peak value of the high frequency band as the running vertical vibration frequency peak value; The third peak value acquisition unit is used to use the vibration signal after the noise reduction processing to obtain the kurtosis of the vibration signal and the peak value corresponding to the kurtosis, and use the peak value corresponding to the kurtosis as the running vertical vibration kurtosis peak value.

7. The elevator performance analysis system based on mixed signals according to claim 3, characterized in that: The program running module also includes: a fourth peak value acquisition unit, configured to perform noise reduction processing on the original door machine current signal, and obtain the current peak value during the door opening and closing phase according to the current signal after noise reduction processing; The coefficient acquisition unit is used to perform wavelet packet decomposition and noise reduction on the original car operation sound signals in the elevator vertical operation stage and the elevator door opening and closing operation stage, perform Mel filtering on the highest frequency in the sound signal after wavelet packet decomposition and noise reduction, and perform logarithmic and discrete cosine transforms at the same time to obtain the vertical operation cepstral coefficients and the door opening and closing cepstral coefficients respectively.

8. The elevator performance analysis system based on mixed signals according to claim 3, characterized in that: The program running module also includes: The displacement acquisition unit is used to obtain the displacement of the elevator's upward and downward wire ropes respectively based on the original traction sheave and wire rope operation data, and determine the acquisition time between two adjacent pulses of the traction sheave, divide the displacement of the elevator's upward and downward wire ropes by the acquisition time respectively to obtain the traction sheave speed, and further determine the displacement of the wire rope at the target moment and when it finally stops moving in combination with the time when the traction sheave speed starts to decrease, subtract the displacement at the target moment from the displacement when it finally stops moving, and respectively obtain the no-load downward unit displacement and no-load upward relative braking distance.

9. A method for analyzing elevator performance based on mixed signals, characterized in that: include: Collecting original mixed signals during elevator operation, inputting the original mixed signals into an elevator performance analysis model, and obtaining elevator analysis and evaluation results; The elevator performance analysis model is obtained by training a neural network model using standard performance indicators and a characteristic signal set; the characteristic signal set is obtained by processing an original mixed signal set.

Images