CN114894289A - Large-mass comparator based on data fusion algorithm - Google Patents
Large-mass comparator based on data fusion algorithm Download PDFInfo
- Publication number
- CN114894289A CN114894289A CN202210694262.7A CN202210694262A CN114894289A CN 114894289 A CN114894289 A CN 114894289A CN 202210694262 A CN202210694262 A CN 202210694262A CN 114894289 A CN114894289 A CN 114894289A
- Authority
- CN
- China
- Prior art keywords
- data
- sensor
- sensors
- weighing
- mass comparator
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000004927 fusion Effects 0.000 title claims abstract description 39
- 238000005303 weighing Methods 0.000 claims abstract description 49
- 238000000034 method Methods 0.000 claims abstract description 30
- 238000013528 artificial neural network Methods 0.000 claims abstract description 24
- 238000001914 filtration Methods 0.000 claims abstract description 23
- 230000003044 adaptive effect Effects 0.000 claims abstract description 7
- 230000006870 function Effects 0.000 claims description 35
- 238000006243 chemical reaction Methods 0.000 claims description 20
- 238000004891 communication Methods 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000003491 array Methods 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 claims 1
- 210000002569 neuron Anatomy 0.000 claims 1
- 238000005070 sampling Methods 0.000 description 5
- 230000002596 correlated effect Effects 0.000 description 3
- 230000000875 corresponding effect Effects 0.000 description 3
- 230000035945 sensitivity Effects 0.000 description 3
- 229910000831 Steel Inorganic materials 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000010959 steel Substances 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01G—WEIGHING
- G01G23/00—Auxiliary devices for weighing apparatus
- G01G23/01—Testing or calibrating of weighing apparatus
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01G—WEIGHING
- G01G3/00—Weighing apparatus characterised by the use of elastically-deformable members, e.g. spring balances
- G01G3/12—Weighing apparatus characterised by the use of elastically-deformable members, e.g. spring balances wherein the weighing element is in the form of a solid body stressed by pressure or tension during weighing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H21/00—Adaptive networks
- H03H21/0012—Digital adaptive filters
- H03H21/0025—Particular filtering methods
- H03H21/0029—Particular filtering methods based on statistics
- H03H21/003—KALMAN filters
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention discloses a data fusion algorithm-based large-mass comparator, a mass comparator scale body and an instrument; the instrument is provided with an initialization module for initializing and comparing the data of the weighing sensor obtained in real time with the pre-stored horizontal state data to determine whether the instrument is in a horizontal state; the fault identification module is used for fusing data of the symmetrical retransmission sensors and identifying the weighing sensors with faults according to the correlation among the sensors; adaptive filtering module for adaptively determining Kalman filtering by standard deviation of data from multiple sensorsRValue sumQA value; number ofAccording to the fusion module: the method comprises the steps of taking multiple paths of weighing sensor signals as input, establishing a mass comparator error compensation model by using an error compensation method of radial basis function neural network multi-sensor information fusion, and obtaining a weighing result by taking a compensated value as output.
Description
Technical Field
The invention belongs to the field of quality measurement, and particularly relates to a large-mass comparator based on a data fusion algorithm.
Background
The mass comparator is widely applied to mass measurement, is a special electronic balance, is mainly used for mass comparison of a standard weight and a detected weight, and is a very important quantity value transmission tool. Its main principle of tracing to the source does, will be surveyed the weight and carry out a lot of comparisons with standard weight through the quality comparator, reachs the difference of being surveyed weight and standard weight, owing to the quality of known standard weight, alright reachs the value of being surveyed the weight. At present, the mass comparator is mainly divided into an electromagnetic force sensor and a strain type sensor. The electromagnetic force sensor has high precision and is mainly used for transferring the quantity value of small-mass and high-precision (E and F grades) weights, has high requirements on the use environment and cannot move once installation and debugging are completed. The mass comparator of the strain sensor is mainly used for magnitude transmission of large-mass (more than 1 t) and low-precision (M grade) working weights, has relatively low requirements on the use environment, can be used by adjusting the level after being moved, is called as the large-mass comparator for short, and plays a very important role in the aspects of weighing machine manufacturing, traffic safety and the like.
The currently used large-mass comparator mostly adopts a type of 3 or 4 sensors, and signals of all the sensors are connected in parallel and then converted into a mass value after being processed by signal amplification, filtering and the like. The problems that exist are mainly:
1. the balance body level is difficult to adjust. Big quality comparator must adjust the level after the removal, makes its sensor be in same horizontal plane just can normal use, and the screw on the mode adjustment every angle that adopts the horizontal bubble of visual observation at present mostly makes it be in horizontal position, and this kind of mode visual observation error is great, is difficult to adjust horizontal position.
2. The unbalance loading error and the nonlinear error of the large-mass comparator are difficult to solve. All sensors of the large-mass comparator are connected in parallel through a junction box to form a signal and then subjected to AD conversion, and because the sensitivity and other parameters of each sensor are different, a potentiometer on the junction box needs to be adjusted before each sensor signal is connected in parallel, so that the sensitivity of the potentiometer is consistent as much as possible, and the quality values displayed by weights placed at different positions are consistent, so that the problem of unbalance loading errors is solved, but the actual adjustment operation is very difficult, and the manual adjustment of the potentiometer wastes time and labor. Meanwhile, the nonlinear error caused by the problems of scale body deformation, stress and the like cannot be solved at present, and the precision of the large-mass comparator is influenced.
Due to the defects, the use effect of a plurality of large-mass comparators in the practical application process is not ideal, and the mass comparators are difficult to adjust manually and cannot meet the requirements of clients for field use.
Disclosure of Invention
The invention aims to provide a large-mass comparator based on a data fusion algorithm, so as to solve the problems of scale body horizontal adjustment, unbalance loading errors and nonlinear errors of the large-mass comparator in actual use.
The technical solution for realizing the purpose of the invention is as follows:
a big mass comparator based on a data fusion algorithm comprises a mass comparator scale body provided with a plurality of paths of weighing sensors; the scale body of the mass comparator is connected with the instrument through a junction box; the instrument is provided with:
the initialization module is used for completing initialization of data parameters and comparing the data of the weighing sensor obtained in real time with the pre-stored horizontal state data so as to determine whether the scale body of the mass comparator is in a horizontal state;
the fault identification module is used for fusing data of the weighing sensors based on the Pearson correlation coefficient and identifying the weighing sensors with faults according to the correlation among the sensors;
an adaptive filtering module for adaptively determining Kalman filtering based on standard deviation of multiple sensor dataRValue sumQA value;
the data fusion module based on the radial basis function neural network comprises: and taking a plurality of paths of weighing sensor signals as input, establishing a mass comparator error compensation model by using an error compensation method of radial basis function neural network multi-sensor information fusion, and obtaining a weighing result by taking a compensated value as output.
Compared with the prior art, the invention has the following remarkable advantages:
(1) according to the invention, the AD conversion is carried out on each strain weighing sensor by adopting a digital junction box mode, the coefficient of each sensor is automatically calculated by adopting an artificial intelligent mode, the sensitivity of the sensor is prevented from being adjusted by adopting a potentiometer, and the method is more accurate and convenient.
(2) The invention designs a horizontal adjustment function, which judges whether the weighing sensor is in a horizontal state or not by comparing the numerical value of each weighing sensor in the idle state with the numerical value of the weighing sensor in the horizontal state, and can adjust the height of each angle according to the numerical value of each sensor so as to enable the angle to reach the horizontal state.
(3) The invention designs a data fusion algorithm, which is used for identifying faults of sensors through a data fusion algorithm based on Pearson correlation coefficients, judging whether the sensors are abnormal or not, and fusing data of four sensors through a neural network based on a radial basis function after self-adaption of TaKallman filtering so as to effectively solve unbalanced load errors and nonlinear errors.
Drawings
FIG. 1 is a block diagram of a data fusion algorithm based mass comparator.
FIG. 2 is a flow chart of an initialization module.
FIG. 3 is a flow chart of a data fusion algorithm.
FIG. 4 is a graph of a radial basis function neural network.
Detailed Description
For the purpose of illustrating the technical solutions and technical objects of the present invention, the present invention will be further described with reference to the accompanying drawings and specific embodiments.
With reference to fig. 1 to 4, the large mass comparator based on data fusion of the present invention is mainly divided into three parts: the device comprises a mass comparator scale body, an embedded digital junction box and an instrument. The scale body of the mass comparator consists of four sensors and a steel structure table board. The embedded digital junction box mainly comprises an AD conversion module and a communication module. The instrument mainly comprises an initialization module, a data fusion module and a display module.
The scale body of the mass comparator is connected with the AD conversion module; the AD conversion module is connected with the communication module; the communication module is connected with an initialization module of the instrument, the initialization module is executed when the instrument is just started for use, and once the operation is finished, the module is not required to be executed in the use process. The communication module is directly connected with the data fusion module in the using process, and the data fusion module is connected with the display module.
The weight comparator scale body is used for bearing the weight and converting the mass of the weight into a voltage signal.
The working process of the scale body of the mass comparator is as follows:
the weighing platform is a steel platform with the platform surface of 1.2m multiplied by 1.2m, and four weighing sensors are respectively placed on four feet. When the weight calibrating device is used, weights to be calibrated are placed on the table board, the four weighing sensors are stressed, and the mass of the weights is converted into four paths of voltage signals which are output to a digital-to-analog conversion module (AD conversion module for short) of the embedded junction box in parallel.
The AD conversion module is used for converting the voltage signal of the weighing sensor into a digital signal so as to analyze and process the later-period signal.
The working process of the AD conversion module is as follows:
the AD conversion module consists of four AD7176 data conversion plates, and each AD data conversion plate is connected with a weighing sensor. And each AD data conversion board sequentially collects signals of each weighing sensor in a circulating manner according to preset settings, and transmits the data of the four sensors to the communication module after each cycle of collection is finished so as to send the data to the upper computer. And the STM32F103 single chip microcomputer in the embedded junction box is responsible for setting the sampling mode, the sampling rate and the like of the AD7176 conversion plate. In order to improve the precision of AD conversion, the sampling rate of four AD7176 data conversion boards is set to be 16 times/second, the sampling mode is set to be cyclic sampling, when the quality comparator is electrified and starts to be used each time, the STM32F103 single chip microcomputer in the embedded junction box firstly carries out parameter setting on the AD7176 data conversion module, and parameters can not be changed in the whole using process.
The communication module is used for transmitting the data converted by the AD7176 conversion plate to an upper computer instrument in real time.
The working process of the communication module is as follows:
the communication module adopts an RS422 communication mode, firstly judges whether the data of the four sensors are complete, respectively sends the data of the four sensors in sequence if the data of the four sensors are complete, and adds a data head and a data tail so as to facilitate the upper computer instrument to analyze the data. The specific data format is as follows:
AA ×× ×× ×× ×× ×× ×× ×× ×× ×× ×× ×× ×× FF
AA is the header, a set of data start identifiers. FF is data tail and a group of data end marks. The data of the weighing sensors are arranged in the middle, and each sensor occupies 24 bits bit.
The communication module sends the sensor data to the instrument, if the instrument is just started, an initialization process is needed, the communication data are transmitted to the initialization module, and the initialization module does not need to be executed after the initialization is finished.
The working process of the initialization module is as follows:
as shown in fig. 2, the initialization module mainly initializes the data parameters of the meter and adjusts the level of the quality comparator.
The initialization parameters mainly comprise parameters for radial basis function neural network training and data of a level state sensor prestored in a scale body of the mass comparator. The parameters of the radial basis function neural network training are mainly used for data fusion of four sensors. The scale body horizontal state sensor data of the mass comparator is the numerical values of four pre-stored sensors in the scale body horizontal state (sensor1, sensor2, sensor3, sensor4) This data is obtained by pre-calibration.
sensor1 refers to the value of the No. 1 sensor in the horizontal state;
sensor2 refers to the value of the No. 2 sensor in the horizontal state;
sensorno. 3 sensor in horizontal stateThe value of (d);
sensor4 refers to the value of sensor number 4 in the horizontal state;
during initialization, the scale body is kept in an idle state, and after the initialization of the initialization module is finished, the data of the four weighing sensors are read through the serial port (real1, real2, real3, real4)。
real1 refers to data read by a No. 1 sensor in real time;
real2 refers to data read by the No. 2 sensor in real time;
realthe 3-finger sensor reads data in real time;
real4 refers to data read by a No. 4 sensor in real time;
and then obtaining load cell data in real time (real1, real2, real3, real4) With prestored horizontal state data (sensor1, sensor2, sensor3, sensor4) Comparing, if the data of the sensor read in real time is smaller, the angular position of the sensor is lower; if the sensor data read in real time is larger, the position of the angle at which the sensor is positioned is higher. Simultaneously, the following requirements are met:
then the scale body of the mass comparator is in a horizontal state to meet the normal use requirement. If the condition that the scale body of the mass comparator is not in the horizontal state cannot be satisfied, the instrument displays alarm information. A is a horizontal state determination threshold, 1000 in this embodiment, which can be adjusted according to different sensors.
After initialization is completed, the large-mass comparator enters a normal working state, and the data fusion module reads data of the four weighing sensors. The data fusion module workflow diagram is shown in fig. 3.
The data fusion module is mainly divided into three parts: the system comprises a fault identification module based on Pearson correlation coefficient data fusion, a Kalman self-adaptive filtering module and a data fusion module based on a radial basis function neural network.
Firstly, fault identification based on data fusion of Pearson correlation coefficients is carried out on data of four sensors read by a serial port, and the specific idea is as follows:
in the process of loading and unloading the weight, the change rules of the four sensors are the same, data should be increased or decreased simultaneously, the relationship between every two sensors is strongly correlated, and if the data of the weighing sensor is not strongly correlated with the data of other weighing sensors, the sensor is in a fault, is not in a normal working state and needs to be adjusted or maintained.
The method comprises the following specific steps
1) The four sensor data are stored in an array of 20 data, respectively
、
、
、
The standard deviation of each array was calculated separately using equation (1) and recorded as
：
Wherein
Is corresponding to the first in the sensor arraylThe number of the data is one,l=1,2，…20；
is as followsiStandard deviation of individual sensors when
Get
When the temperature of the water is higher than the set temperature,
standard deviation for the 1 st sensor; in the same way, when
Get
When the utility model is used, the water is discharged,
standard deviation for the 2 nd sensor;
nis corresponding to the number of data in the sensor array at this timen=20
2) Then, the Pearson correlation coefficient between every two sensors is calculated by using a formula (2)
In the formulaXAndYtwo sensor arrays are needed to calculate the pearson correlation coefficient.
wherein cov (X，Y) Is composed ofXAndYthe covariance of (a);
Eis a mathematical expectation.
3) If it is not
When the value is smaller than the first set threshold B1, in this embodiment B1 is 100, which indicates that the scale body is in an empty state or the weight has been placed, the weighing sensor should be in an irregular fluctuation state at this time, it is determined that the four sensors are irrelevant, and at this time, the pearson correlation coefficient is not calculated, and no fault identification is performed.
4) If it is not
Greater than or equal to the first set threshold B1, indicating that the sensor data is in a severe change, i.e., loaded and unloaded states, it is determined that the four sensors are strongly correlated at this time,
should be between 0.5 and 1. If the value is less than 0.5 or a negative value, the correlation of the sensor is not strong, and accordingly, the faulty load cell can be analyzed and identified.
5) Reading the next data and repeating the above process.
If no fault exists, the adaptive Kalman filtering based on the standard deviation is carried out, and the standard deviation of the data of the four sensors is passed
From adaptive Kalman filteringRValue sumQThe value is obtained.
Covariance parameters for Kalman filteringRValue sumQThe value affects the convergence speed of the filtering and the filtering model error. To not only ensure the collectionThe convergence speed can obtain a numerical value with a small model error, and the large-mass comparator adopts a Kalman adaptive filtering algorithm based on standard deviation. The main idea is as follows:
when the weight is loaded or unloaded, the data of the weighing sensor can be rapidly increased or reduced, the filtering following performance is required to be better,Rvalue sumQThe value may take a larger value in order to save convergence time; and after no-load or weight placement is finished, the data needing filtering can be kept stable, relevant interference factors are filtered out to obtain stable and accurate numerical values, and smaller numerical values need to be set at the momentRValue sumQThe value is obtained. The standard deviation of the data can just reflect the change rate of the data, and the standard deviation is very large when the data of the loading and unloading weights are increased or reduced rapidly; when the weight or no load is stable, the standard deviation of the data of the weighing sensor is smaller, and the weight is adjusted according to the standard deviationRValue sumQThe value is obtained.
The method comprises the following specific steps:
1) reading in the standard deviation of the four sensor data calculated above
；
2) If it is not
The values are all larger than or equal to a first set threshold value B1 (100), which indicates that the weights are in the loading or unloading process and do not carry out Kalman filtering.
If it is not
Respectively less than a first set threshold B1 (100) and greater than or equal to a second set threshold B2 (80), which indicates that the weight loading or unloading is completed, but the scale body is in a violent vibration state, the following steps are set:
if it is not
Respectively less than the second set threshold B2 (80) and greater than or equal to the third set threshold B3 (50), which indicates that the weight loading or unloading is completed, but the scale body is in a large vibration state, the following steps are set:
if it is not
Respectively less than the third set threshold value B3 (50) and greater than or equal to the fourth set threshold value B4 (30), which indicates that the scale body is in a small vibration state after the weight loading or unloading is completed, the following steps are set:
if it is not
Respectively less than a fourth set threshold value B4 (30), which indicates that the scale body is basically stable after the weight loading or unloading is finished, and the following steps are set:
3) and (3) reading a new datum in each sensor array queue again to serve as a queue head, discarding the last datum at the tail of the queue, and repeating the step 1) for circular filtering.
The filtered data is based on a radial basis function neural network algorithm and is mainly used for solving the problem of unbalance loading errors of the quality comparator. The main idea is as follows: the output of the four sensors of the traditional quality comparator adopts a parallel circuit mode, a potentiometer in a junction box is repeatedly adjusted, and the gain of each channel is adjusted to reduce the offset load error. However, the unbalance loading error of the mass comparator is influenced by the rigidity and strength of the scale body, the internal stress in the processing and mounting process, the mechanical deformation of the scale body during bearing and other nonlinear factors, and the required precision can be achieved only by adjusting the potentiometer only, so that the mass comparator is an important factor influencing the precision of the mass comparator. The radial basis function neural network is widely applied to multi-sensor information fusion, and can well approach a nonlinear function. The algorithm is characterized in that four paths of weighing sensor signals are used as input, a mass comparator error compensation model is established by using an error compensation method of radial basis function neural network multi-sensor information fusion, a compensated value is used as output, and the compensation model is shown in figure 4, wherein:
in the formula:
is the compensated output value. Number of hidden layer neuronsmThe experiment confirms that after a plurality of experiments,m=20；Wis the weight vector of the radial basis function neural network,
wherein
；HIn the form of a vector of radial basis functions,
wherein
；dIs the input layer bias value.
The algorithm adopts a Gaussian function as a basic function of the network, and then:
in the formula:
is an input vector in which
，
Are respectively the firsti（iValues of =1,2,3, 4) sensors;
is a sensorjAn expansion constant of each node;
is a sensorjThe center vector of each of the nodes is,
，
are respectively the firsti（i=1,2,3, 4) number of sensorsjThe center vector of each node.
The method comprises the following specific steps:
firstly, training the error compensation basis function neural network of the quality comparator is carried out. By utilizing different masses, the intrinsic mass comparator respectively loads three values of 500kg, 1t and 2t to different positions of a scale body of the mass comparator, and acquires 60 groups of signals of 4 paths of weighing sensors to obtain data after normalization processing
And is used for training the basis function neural network. The training of the radial basis function is realized by adopting a gradient training method, and the learning objective function is as follows:
in the formula
In order to be a forgetting factor,
is an error signal.
Wherein
Is the mass of the weight to be placed, i.e. the true value.
To minimize using the objective function, the amount of correction of each parameter should be proportional to its negative gradient. The correction quantity of each parameter can be obtained:
wherein
Is the learning rate.
Objective functionEIs set to 1 × 10 -8 Learning rate
All are 0.015, forgetting factor
Are all 0.6. The parameters can be obtained by training 60 groups of data according to formula (5), formula (6), formula (7), formula (8), formula (9) and formula (10)
Central vector of
Weight matrixWAnd output layer biasing
The value of (c). This process only need carry out data acquisition and training when quality comparator uses for the first time, in case after the training of basis function neural network is good, just need not train again, and the direct use parameter of later stage can.
The display module is mainly used for displaying relevant fault information and weighing results in real time.
According to the invention, the data of the four sensors of the large-mass comparator are analyzed and processed through the initialization module, so that the horizontal adjustment of the large-mass comparator can be conveniently completed, the influence of unbalance loading and nonlinear errors can be reduced through a data fusion algorithm, and the precision of the large-mass comparator is improved.
Claims (8)
1. A big mass comparator based on a data fusion algorithm comprises a mass comparator scale body provided with a plurality of paths of weighing sensors; the mass comparator is characterized in that the scale body of the mass comparator is connected with the instrument through a junction box; the instrument is provided with:
the initialization module is used for completing initialization of data parameters and comparing the data of the weighing sensor obtained in real time with the pre-stored horizontal state data so as to determine whether the scale body of the mass comparator is in a horizontal state;
the fault identification module is used for fusing data of the weighing sensors based on the Pearson correlation coefficient and identifying the weighing sensors with faults according to the correlation among the sensors;
an adaptive filtering module for adaptively determining Kalman filtering based on standard deviation of multiple sensor dataRValue sumQA value;
the data fusion module based on the radial basis function neural network comprises: and taking a plurality of paths of weighing sensor signals as input, establishing a mass comparator error compensation model by using an error compensation method of radial basis function neural network multi-sensor information fusion, and obtaining a weighing result by taking a compensated value as output.
2. The data fusion algorithm-based big mass comparator according to claim 1, wherein the engineering process of the initialization module is as follows:
during initialization, the scale body is kept in an idle state, and after the initialization of data parameters is completed, the data of a plurality of weighing sensors are read;
then, the data of the weighing sensor is obtained in real time and compared with the pre-stored horizontal state data, and if the data of the sensor read in real time is small, the position of the sensor is low; if the data of the sensor read in real time is larger, the position of the angle at which the sensor is positioned is higher;
simultaneously, the following requirements are met:
the scale body of the mass comparator is in a horizontal state to meet the normal use requirement, otherwise, the scale body of the mass comparator does not meet the use requirement;
whereinrealiIs as followsiThe data read by each weighing sensor in real time,sensoriis in a horizontal stateiThe numerical values of the sensors are prestored by the weighing sensors; a is a horizontal state determination threshold.
3. The data fusion algorithm-based big mass comparator as claimed in claim 1, wherein the specific process of identifying the faulty weighing sensor by the fault identification module is as follows:
storing the sensor data into an array respectively, calculating the standard deviation of each array respectively, and calculating the standard deviation of each array respectively:
wherein
Is corresponding to the first in the sensor arraylA piece of data;
is as followsiStandard deviation of individual sensors;
the average value of the data of the corresponding sensor array is taken as the average value;nthe number of data in the corresponding sensor array is obtained;
the pearson correlation coefficient between two sensors is then calculated:
in the formulaXAndYtwo sensor arrays for which Pearson correlation coefficients need to be calculated;
is thatXThe standard deviation of (a) is determined,
is thatYStandard deviation of (d); cov (X，Y) Is composed ofXAndYthe covariance of (a); Eis a mathematical expectation;
if it is not
If the value is less than the set threshold value, the scale body is in a no-load state or the weight is placed completely, the four sensors are judged to be irrelevant, the Pearson correlation coefficient is not calculated, and fault identification is not carried out;
if it is not
If the Pearson correlation coefficient between every two sensors is smaller than a set value, the correlation of the sensors is not strong, and accordingly, the weighing sensor with a fault is identified.
4. Mass comparator based on data fusion algorithm according to claim 1 or 3, characterized by an adaptive filtering module determining Kalman filteringRValue sumQValues include the following conditions:
if the standard deviation of each sensor is greater than or equal to a first set threshold value, the weight is indicated to be in the loading or unloading process, and Kalman filtering is not performed; and is set differently according to the following different working conditionsRValue sumQThe value:
(1) if the standard deviation of each sensor is respectively less than the second set threshold and more than or equal to a third set threshold B2, the weight loading or unloading is finished, but the scale body is in a larger vibration state;
(2) if the standard deviation of each sensor is respectively smaller than the third set threshold and larger than or equal to the fourth set threshold B2, the scale body is in a smaller vibration state after the weight loading or unloading is finished;
(3) and if the standard deviation of each sensor is respectively smaller than the fourth set threshold value, indicating that the weight loading or unloading is finished.
5. The data fusion algorithm-based big-mass comparator as claimed in claim 1, wherein the process of the data fusion module based on the radial basis function neural network is as follows:
firstly, training an error compensation basis function neural network of a mass comparator, respectively loading different positions of a scale body of the mass comparator by using different masses, acquiring multiple groups of multipath weighing sensor signals, and obtaining data after normalization processing
For training a basis function neural network;
the training of the radial basis function is realized by adopting a gradient training method, and the learning objective function is as follows:
in the formula
In order to be a forgetting factor,
is an error signal;
correction amount of each parameter:
setting an objective functionELearning rate
Forgetting factor
Training with multiple sets of data to derive the sensor numberjSpreading constant of each node
The sensor isjCenter vector of individual node
Weight matrixWAnd output layer biasing
A value of (d);
is a weight matrixWTo (1) ajA vector number;
after the acquisition and filtering processing of the signals of the multi-channel weighing sensor are finished once, the signal data and all parameters of the radial basis function neural network are processed
、
、W、
Respectively substituting the compensation model and the basis function of the neural network, and calculating to obtain the compensated output value
And the weight is output in real time as a weighing result of the mass comparator.
6. The data fusion algorithm-based big-mass comparator according to claim 1, wherein the basis functions of the gratuitous model and the neural network are respectively:
and (3) compensation model:
in the formula:mthe number of the hidden layer neurons is the number,Wis a weight vector of a radial basis function neural network, wherein
；HIs a radial basis function vector;
the basis functions are:
in the formula:
is an input vector containing the values of the sensors;
is a sensorjThe spreading constant of each node.
7. The data fusion algorithm-based high-quality comparator according to claim 1, wherein the junction box is composed of an AD conversion module and a communication module;
the AD conversion module is used for converting the voltage signal of the weighing sensor into a digital signal;
the communication module is used for sending the converted data to an upper computer instrument in real time.
8. The data fusion algorithm-based big mass comparator according to claim 7, wherein the working process of the communication module is as follows:
firstly, judging whether the data of the multi-channel sensors are complete, if so, respectively sending the data of the plurality of slave sensors in sequence, and adding a data head and a data tail, wherein the specific data format is as follows:
AA ×× ×× ×× ×× ×× ×× ×× ×× ×× ×× ×× ×× FF
AA is a data header, a group of data start marks; FF is data tail, and a group of data end marks; the middle is the data of the weighing sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210694262.7A CN114894289B (en) | 2022-06-20 | 2022-06-20 | Large-mass comparator based on data fusion algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210694262.7A CN114894289B (en) | 2022-06-20 | 2022-06-20 | Large-mass comparator based on data fusion algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114894289A true CN114894289A (en) | 2022-08-12 |
| CN114894289B CN114894289B (en) | 2024-02-02 |
Family
ID=82727607
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210694262.7A Active CN114894289B (en) | 2022-06-20 | 2022-06-20 | Large-mass comparator based on data fusion algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114894289B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117782289A (en) * | 2024-02-26 | 2024-03-29 | 深圳市杰曼科技股份有限公司 | Digital weighing system based on Internet of things and weighing method thereof |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104330084A (en) * | 2014-11-13 | 2015-02-04 | 东南大学 | Neural network assisted integrated navigation method for underwater vehicle |
| KR101907589B1 (en) * | 2018-01-22 | 2018-10-12 | 연세대학교 산학협력단 | Structural system identification using extended kalman filter and genetic algorithm |
| WO2018198576A1 (en) * | 2017-04-28 | 2018-11-01 | 株式会社Ｓｏｋｅｎ | Torque detection device and sensor module |
| CN110954132A (en) * | 2019-10-31 | 2020-04-03 | 太原理工大学 | Method for carrying out navigation fault identification through GRNN (generalized regression neural network) assisted adaptive Kalman filtering |
| CN111259927A (en) * | 2020-01-08 | 2020-06-09 | 西北工业大学 | Rocket engine fault diagnosis method based on neural network and evidence theory |
| WO2021018160A1 (en) * | 2019-07-31 | 2021-02-04 | Mettler-Toledo (Changzhou) Measurement Technology Ltd. | Hysteresis compensation method for a weighing device |
| CN112651456A (en) * | 2020-12-31 | 2021-04-13 | 遵义师范学院 | Unmanned vehicle control method based on RBF neural network |
-
2022
- 2022-06-20 CN CN202210694262.7A patent/CN114894289B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104330084A (en) * | 2014-11-13 | 2015-02-04 | 东南大学 | Neural network assisted integrated navigation method for underwater vehicle |
| WO2018198576A1 (en) * | 2017-04-28 | 2018-11-01 | 株式会社Ｓｏｋｅｎ | Torque detection device and sensor module |
| KR101907589B1 (en) * | 2018-01-22 | 2018-10-12 | 연세대학교 산학협력단 | Structural system identification using extended kalman filter and genetic algorithm |
| WO2021018160A1 (en) * | 2019-07-31 | 2021-02-04 | Mettler-Toledo (Changzhou) Measurement Technology Ltd. | Hysteresis compensation method for a weighing device |
| CN110954132A (en) * | 2019-10-31 | 2020-04-03 | 太原理工大学 | Method for carrying out navigation fault identification through GRNN (generalized regression neural network) assisted adaptive Kalman filtering |
| CN111259927A (en) * | 2020-01-08 | 2020-06-09 | 西北工业大学 | Rocket engine fault diagnosis method based on neural network and evidence theory |
| CN112651456A (en) * | 2020-12-31 | 2021-04-13 | 遵义师范学院 | Unmanned vehicle control method based on RBF neural network |
Non-Patent Citations (4)
| Title |
|---|
| SHARAF ETC.: "Sensor Integration for Satellite-Based Vehicular Navigation Using Neural Networks", vol. 18, no. 2, XP011172346, DOI: 10.1109/TNN.2006.890811 * |
| 孙勇;李宝聚;孙志博;李振元;张罗宾;: "融合RBF神经网络和集对分析的风电功率超短期预测", 昆明理工大学学报(自然科学版), no. 05 * |
| 林海军;滕召胜;迟海;吴阳平;易钊;: "基于信息融合的汽车衡称重传感器故障诊断", 控制理论与应用, no. 01 * |
| 林海军;滕召胜;迟海;易钊;邬蓉蓉;: "基于多传感器信息融合的汽车衡误差补偿", 仪器仪表学报, no. 06 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117782289A (en) * | 2024-02-26 | 2024-03-29 | 深圳市杰曼科技股份有限公司 | Digital weighing system based on Internet of things and weighing method thereof |
| CN117782289B (en) * | 2024-02-26 | 2024-05-14 | 深圳市杰曼科技股份有限公司 | Digital weighing system based on Internet of things and weighing method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114894289B (en) | 2024-02-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112903235B (en) | Multi-element calibration method for thrust balance capable of completely simulating test state | |
| CN112800633B (en) | Processing method for multivariate calibration wind tunnel balance data | |
| CN114894289A (en) | Large-mass comparator based on data fusion algorithm | |
| CN107741269B (en) | Weighing sensor test compensation method based on fuzzy recognition | |
| CN108614231B (en) | Rapid calibration method based on electronic calibration piece | |
| CN109471061B (en) | Received signal strength difference positioning method for robust processing model parameter error | |
| CN116678489B (en) | Quality control method, system, equipment and medium for force sensor | |
| CN112816001B (en) | Jacking process-based synchronous testing method for load displacement of bearings of multiple support shaft systems | |
| CN114722332B (en) | Wind tunnel balance calibration data processing method | |
| CN109115408B (en) | Dynamic balance test method of large hydroelectric generating set based on centrifugal force equation | |
| CN112816216A (en) | Rolling bearing performance test bench and identification and correction method of abnormal test sample | |
| CN117465135B (en) | Spray nozzle voltage output self-correction method, device and storage medium | |
| CN117091801B (en) | Balance calibration method based on two-degree-of-freedom calibration equipment | |
| CN112871759B (en) | Lithium battery voltage drop evaluation method and system | |
| CN106709175A (en) | Method for obtaining material mechanical property parameters of printed circuit board | |
| CN118050538A (en) | Three-dimensional wind direction detection method and system | |
| CN107621607A (en) | Battery performance test method, test terminal, readable storage medium storing program for executing and test system | |
| CN116561646A (en) | Rolling bearing multidimensional parameter automatic detection line and rolling bearing grade evaluation method | |
| CN111859679A (en) | Wind turbine generator test load obtaining method, simulation load comparing method and simulation load comparing device | |
| CN116643183A (en) | Lithium ion battery self-discharge test voltage compensation method | |
| CN114910146B (en) | Method for automatically measuring and calculating weight of pig farm material tower after weighing analog quantity sensor fails | |
| CN114577318B (en) | Vehicle-mounted weighing module and sensing method thereof | |
| CN104101418A (en) | Electronic analytical balance trace loading and drift discrimination method | |
| CN115911617A (en) | Passive equalization method of battery pack | |
| CN113945262B (en) | Offset load error correction method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |