CN115049170B - Method for debugging threading work of threading machine controller - Google Patents
Method for debugging threading work of threading machine controller Download PDFInfo
- Publication number
- CN115049170B CN115049170B CN202210978355.2A CN202210978355A CN115049170B CN 115049170 B CN115049170 B CN 115049170B CN 202210978355 A CN202210978355 A CN 202210978355A CN 115049170 B CN115049170 B CN 115049170B
- Authority
- CN
- China
- Prior art keywords
- sequence
- index
- threading
- difference
- threading machine
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 54
- 229910000831 Steel Inorganic materials 0.000 claims description 33
- 239000010959 steel Substances 0.000 claims description 33
- 229910001294 Reinforcing steel Inorganic materials 0.000 claims description 30
- 230000002159 abnormal effect Effects 0.000 claims description 25
- 230000032683 aging Effects 0.000 claims description 25
- VIEYMVWPECAOCY-UHFFFAOYSA-N 7-amino-4-(chloromethyl)chromen-2-one Chemical compound ClCC1=CC(=O)OC2=CC(N)=CC=C21 VIEYMVWPECAOCY-UHFFFAOYSA-N 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 19
- 230000006978 adaptation Effects 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 11
- 239000000126 substance Substances 0.000 claims description 9
- 230000008859 change Effects 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 2
- 230000002950 deficient Effects 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 description 25
- 230000003014 reinforcing effect Effects 0.000 description 8
- 230000003044 adaptive effect Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 239000000835 fiber Substances 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009435 building construction Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- Manufacturing & Machinery (AREA)
- Feedback Control In General (AREA)
- Debugging And Monitoring (AREA)
Abstract
The invention relates to the technical field of equipment parameter regulation and control, in particular to a method for debugging threading work of a threading machine controller.
Description
Technical Field
The invention relates to the technical field of equipment parameter regulation and control, in particular to a method for debugging the threading work of a threading machine controller.
Background
With the rapid development of the modern building industry, the quantity of processed reinforcing steel bar screws used by the modern building industry is larger and larger, and particularly, more threading screws are used on the building construction site. The threading machine is the equipment of processing reinforcing bar screw rod, the threading machine during operation, put into the pipe chuck earlier the pipe that will process the screw thread, strike the chucking, press the start switch, the pipe just rotates along with the chuck, adjust the die opening size on the die head, set for the screw mouth length, then pull the hand wheel of feed clockwise, make the die sword on the die head paste the tip of pivoted pipe with the constant force, the die sword is with the automatic cutting mantle fiber, but at the threading in-process, reinforcing bar fixed problem and cutter march problem all can influence reinforcing bar threading result.
At present, for reinforcing steel bars with different specifications, the requirement of reinforcing steel bar threading is realized through manual adjustment, but because the working environment and the use time of the threading machine are long, the performance of threading opportunity is gradually reduced, an adjustment error and a hysteresis phenomenon can occur by utilizing manual operation under the condition, the equipment parameters of the threading machine cannot be accurately adjusted in time, and then the reinforcing steel bar threading result is inconsistent.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a method for debugging the threading work of a threading machine controller, which adopts the following technical scheme:
acquiring the current of a motor and the advancing speed of a threading die cutter after the current steel bar is clamped based on sampling frequency to obtain a current sequence and an advancing speed sequence of the motor in the whole threading time period of the current steel bar; calculating a motor current stability index according to the difference between adjacent motor currents in the motor current sequence, and calculating a thread uniformity index of the current reinforcing steel bar surface by combining the motor current stability index and the travelling speed change corresponding to the travelling speed sequence;
obtaining the thread uniformity index of each reinforcing steel bar in the current batch under the current threading machine to obtain a thread uniformity index sequence; calculating a local abnormal factor of each thread uniformity index according to the difference of any two thread uniformity indexes in the thread uniformity index sequence, and acquiring the yield of the reinforcing steel bars in the current batch based on the local abnormal factor; obtaining historical steel bar yields of multiple batches to form a historical yield sequence, and performing continuous iterative training on a yield prediction network by using the historical yield sequence to confirm a target yield prediction network corresponding to the current threading machine according to the difference between input data and output data of the yield prediction network;
acquiring the target good product rate prediction network of each threading machine, and enabling the real-time good product rate sequence of each threading machine to pass through the corresponding target good product rate prediction network to obtain a good product rate prediction sequence of each threading machine; obtaining the total quantity of good products of each threading machine according to the prediction sequence of the good product rate and the total quantity of the reinforcing steel bars corresponding to the batch, calculating a good product adaptation index of the current threading machine based on the total quantity of the good products, and calculating a machine aging index of the current threading machine according to a difference value of two adjacent elements in the prediction sequence of the good product rate; and determining a target threading machine according to the good product adaptation index and the machine aging index, and adjusting the equipment parameters of each threading machine to the equipment parameters of the target threading machine.
Further, the method for obtaining the yield of the current batch of steel bars based on the local abnormal factor includes:
setting a local abnormal factor threshold, and when the local abnormal factor is smaller than the local abnormal factor threshold, determining that the steel bar after corresponding threading is good; and counting the first number of good products, calculating the ratio of the total number of the reinforcing steel bars in the current batch to the first number, and taking the ratio as the yield of the reinforcing steel bars.
Further, the method for obtaining the target yield prediction network includes:
based on an improved yield prediction network algorithm, performing one-time iterative training on a yield prediction network by using the historical yield sequence to obtain a yield prediction value corresponding to each historical yield in the historical yield sequence, and forming a historical yield prediction sequence;
respectively calculating first difference values between two elements at corresponding positions in the historical yield prediction sequence and the historical yield sequence, and performing control adjustment on network parameters of a yield prediction network according to the first difference values to obtain a new yield prediction network; the network parameters include: the input space of the yield prediction network and each parameter in the improved yield prediction network algorithm;
performing one-time iterative training on the new yield prediction network by using the historical yield sequence to obtain a new historical yield prediction sequence; calculating a second difference value between two elements at corresponding positions in the historical yield prediction sequence and the fresh historical yield prediction sequence, when the second difference value meets a difference value threshold value, calculating a third difference value between two elements at corresponding positions in the fresh historical yield prediction sequence and the historical yield sequence, accumulating the third difference value to obtain a third difference value accumulated value, and when the third difference value accumulated value is smaller than the accumulated value threshold value, determining that the new yield prediction network is the target yield prediction network.
Further, the improved yield prediction network algorithm is as follows:
wherein the content of the first and second substances,as a weight value, the weight value,andin order to count the number of times of learning,andfor a certain memory cell to be activated,in order to learn the rate of speed,in order to be the desired value,in order to output the value of the output,in order to generalize the parameters of the process,is as followsAn activated memory cellThe number of learned times at the time of the sub-learning,is as followsAn activated memory cellThe number of learned times at the time of the sub-learning,is as followsAn activated memory cellThe number of learned times at the time of the sub-learning,to balance the learning constants.
Further, the method for controlling and adjusting the network parameters of the yield prediction network according to the first difference value includes:
when the first difference is larger than or equal to the first difference threshold, based on an expert control system, introducing extremum control for quick correction; when the first difference is less than or equal to the negative of the first difference threshold, then control adjustments are made in conjunction with proportional control and CMAC control.
Further, the method for calculating the good product adaptation index of the current threading machine based on the total number of the good products includes:
and counting the number of good products matched with the standard nuts, and taking the ratio of the number of the good products to the total number of the good products as a good product matching index.
Further, the method for calculating the machine aging index of the current threading machine according to the difference value of two adjacent elements in the yield prediction sequence comprises the following steps:
respectively calculating the predicted value difference between two adjacent predicted values of the good product rate in the good product rate prediction sequence, and accumulating the predicted value difference to obtain a predicted value difference accumulated value; and calculating the ratio between the combined quantity of two adjacent elements in the yield prediction sequence and the difference value accumulated value of the prediction value, and taking the ratio as the machine aging index.
Further, the method for determining the target threading machine according to the good product adaptation index and the machine aging index comprises the following steps:
calculating the ratio between the good product adaptation index and the machine aging index, and taking the ratio as the working state index of the corresponding threading machine; the working state index and the machine aging index are in a negative correlation relationship, and the working state index and the good product adaptation index are in a positive correlation relationship;
and acquiring the working state index of each threading machine, and taking the threading machine corresponding to the largest working state index as the target threading machine.
The embodiment of the invention at least has the following beneficial effects: the method is characterized in that the proprietary yield prediction network of each threading machine is trained by using historical data of the threading machines, so that the proprietary yield prediction network can be used for accurately predicting the yield prediction value of each threading machine in a targeted manner, and then the optimal equipment parameters of the threading machines are confirmed according to the prediction result of each threading machine, so that the equipment parameters of the global threading machines are adjusted in time, each threading machine is in the optimal working state, the consistency of the steel bar threading results is improved, and the yield of the steel bar threading is also improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart illustrating steps of a method for debugging threading operation of a threading machine controller according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of the debugging method of the threading operation of the threading machine controller according to the present invention with reference to the accompanying drawings and preferred embodiments shows the following detailed descriptions of the specific implementation manner, structure, features and effects thereof. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of a method for debugging threading work of a threading machine controller provided by the invention in detail with reference to the accompanying drawings.
The embodiment of the invention aims at the following specific scenes: training a good product rate prediction network of a corresponding threading machine according to historical operation data of each threading machine, predicting the good product rate by using the good product rate prediction network of each threading machine, acquiring the threading machine with the optimal working state based on a prediction result, and adjusting equipment parameters of other threading machines based on the equipment parameters of the threading machine.
Referring to fig. 1, a flowchart illustrating a method for debugging threading operation of a threading machine controller according to an embodiment of the present invention is shown, where the method includes the following steps:
step S001, acquiring the current of the motor after the current steel bar is clamped and the advancing speed of the die cutter based on the sampling frequency to obtain a current sequence and an advancing speed sequence of the motor in the whole threading period of the current steel bar; and calculating a motor current stability index according to the difference between adjacent motor currents in the motor current sequence, and calculating the thread uniformity index of the current reinforcing steel bar surface by combining the motor current stability index and the corresponding travelling speed change of the travelling speed sequence.
Specifically, when the mantle fiber machine controls the rotary table to clamp, the motor for clamping almost stops rotating, the current is increased greatly, and when the current is increased to a certain value, the system judges that clamping is finished and sends a clamping finishing instruction. Because the reinforcing bar is influenced by the die cutter when rotating, slight shaking can occur to cause that the threading result of the reinforcing bar is not perfect, so the motor electricity for monitoring the instantaneous clamping of the reinforcing bar during threading is neededThe industrial ammeter is used for measuring the current of the motor in real time during clamping through a circuit connected near the motor, the sampling frequency is set, a plurality of motor currents in the whole threading time period of a reinforcing steel bar are collected based on the sampling frequency, and a motor current sequence is obtainedWherein the content of the first and second substances,for the motor current sampled at the 1 st time,is as followsThe sub-sampled motor current is then used,is the number of samples.
Preferably, in the embodiment of the present invention, the sampling frequency is 5s.
After the reinforcing steel bar is fixed, the cutting feed by using the die cutter is started, the cutting feed speed of the die cutter directly influences the threading result of the reinforcing steel bar to meet the requirement, so that the travelling speed of the die cutter can be obtained in real time by using a speed measurement sensor, namely the travelling speed is the increment of displacement in unit time, and a plurality of travelling speeds in the whole threading time period of the reinforcing steel bar are obtained on the basis of the set sampling frequency to obtain a travelling speed sequenceWherein the content of the first and second substances,is the travel speed of the 1 st sample,is as followsThe speed of travel of the sub-sample,is the number of samples.
Further, in the whole threading process of the steel bar, instability occurs in steel bar tightening due to shaking, and further the motor current changes, so that the motor current stability index is calculated according to the difference between adjacent motor currents in the motor current sequence, and the calculation formula of the motor current stability index is as follows:
wherein the content of the first and second substances,the motor current stability index is obtained;for the second in the motor current sequenceA motor current;for the front of the motor current sequenceAverage motor current between individual motor currents;for the front in the motor current sequenceAverage motor current between individual motor currents;for the second in the motor current sequenceCurrent of motor andthe difference between the individual motor currents.
It should be noted that if no shaking occurs during threading of the steel bar, the magnitude of the motor current when the steel bar is clamped is almost kept unchanged, that is, the difference between the motor current at the current moment and the motor current at the previous moment is zero, and the difference between the motor current mean values is also zero, so that the motor current stability index is kept to be 1; if the motor current changes in the threading process of the reinforcing steel bar, the motor current stability index is smaller than 1, the larger the motor current change degree is, and the closer the motor current stability index is to 0.
Because at the mantle fiber in-process, thereby the cutter can receive the resistance of thread groove and lead to the travelling speed of die cutter to receive the influence, control the travelling speed between every moment and the previous moment and can let reinforcing bar screw thread interval more even and the screw depth also can keep unanimous, consequently combine the travelling speed change that motor current stability index and travelling speed sequence correspond to calculate the screw thread homogeneity index on present reinforcing bar surface, concrete method is: obtaining the maximum advancing speed and the minimum advancing speed in the advancing speed sequence and the average advancing speed, and similarly, obtaining the minimum motor current in the motor current sequence, and calculating the thread uniformity index of the surface of the steel bar according to the difference between the maximum advancing speed, the minimum advancing speed and the average advancing speed, the minimum motor current and the motor current stability index, wherein the calculation formula of the thread uniformity index is as follows:
wherein, the first and the second end of the pipe are connected with each other,is an index of thread uniformity;is the maximum travel speed;is the average travel speed;is the minimum travel speed;is the minimum motor current;is a hyperbolic tangent function and is a normalization means;is an arcsine function and is a normalization means.
It should be noted that the more unstable the motor current is, the less ideal the steel bar threading result is, the worse the thread uniformity index of the corresponding steel bar is, so that the motor current stability index and the thread uniformity index are in a positive correlation;the difference value between the maximum value of the advancing speed of the die cutter and the mean value can show whether the die cutter has sudden change behavior at a certain moment in the advancing process and the amplitude value of the maximum sudden change value compared with the mean value, the larger the amplitude value is, the larger the thread pitch is, the deviation occurs, the pitch is different in size, and further the thread uniformity on the surface of the steel bar is poor, otherwise, the smaller the amplitude value is, the uncontrollable deviation occurs in the threading work, the deviation is smaller, the deviation is closer to the mean value and can be ignored, and the thread uniformity on the surface of the steel bar is better at the moment;the numerical value can represent the similarity of each thread formed by threading and the previous thread in the threading process for the relative change condition of the advancing speed of the die cutter, and can represent whether the threading machine keeps the same working state to a certain extent, the larger the numerical value is, the smaller the thread uniformity index is, and the smaller the numerical value is, the larger the thread uniformity index is.
S002, obtaining a thread uniformity index of each reinforcing steel bar in the current batch under the current threading machine to obtain a thread uniformity index sequence; calculating a local abnormal factor of each thread uniformity index according to the difference of any two thread uniformity indexes in the thread uniformity index sequence, and acquiring the yield of the reinforcing steel bars in the current batch based on the local abnormal factor; obtaining the yield of a plurality of batches of historical reinforcing steel bars, forming a historical yield sequence, and performing continuous iterative training on the yield prediction network by using the historical yield sequence so as to confirm the target yield prediction network corresponding to the current threading machine according to the difference between the input data and the output data of the yield prediction network.
Specifically, the method in step S001 is used to obtain the thread uniformity index of each reinforcing steel bar in the next batch of the threading machine, and form a thread uniformity index sequence, that is, one batch corresponds to one thread uniformity index sequence.
Preferably, the embodiment of the invention uses 100 steel bars as a batch.
Calculating the yield of the corresponding reinforcing steel bars in one batch according to the difference of any two thread uniformity indexes in the thread uniformity index sequence, and the specific method comprises the following steps: taking any one thread uniformity index in the thread uniformity index sequence as a target thread uniformity index, respectively calculating thread uniformity index difference values between the target thread uniformity index and other thread uniformity indexes, accumulating the thread uniformity index difference values to obtain a difference accumulated value, calculating a ratio between the difference accumulated value and the number of elements in the thread uniformity index sequence, and taking the reciprocal of the ratio as the local reachable density of the target thread uniformity index; obtaining local reachable density of each thread uniformity index in a thread uniformity index sequence to form a local reachable density set, taking any one local reachable density in the local reachable density set as a target local reachable density, respectively calculating local reachable density difference values between the target local reachable density and other target local reachable densities, accumulating the local reachable density difference values to obtain a local reachable density difference value accumulated value, calculating a ratio between the local reachable density difference value accumulated value and the number of elements in the local reachable density set, and taking a first ratio between the ratio and the target local reachable density as a local abnormal factor corresponding to the target local reachable density, namely the local abnormal factor corresponding to the thread uniformity index, so as to obtain a local abnormal factor of each thread uniformity index; setting a local abnormal factor threshold, when the local abnormal factor is smaller than the local abnormal factor threshold, confirming that the steel bars after corresponding threading are good products, counting a first quantity of the good products, calculating a ratio between the total quantity and the first quantity of the steel bars in the current batch, and taking the ratio as the yield of the steel bars.
Preferably, in the embodiment of the present invention, the local abnormal factor threshold is an empirical value, and then the local abnormal factor threshold is 0.8.
By using the method for acquiring the good product rate of the steel bars, the good product rates of a plurality of batches of historical steel bars under one threading machine are acquired, and a historical good product rate sequence is formed. Continuously and iteratively training a good product rate prediction network of the threading machine by utilizing a historical good product rate sequence, and confirming a target good product rate prediction network corresponding to the threading machine according to the difference between input data and output data of the good product rate prediction network, preferably, the good product rate prediction network in the embodiment of the invention is a CMAC neural network, and the specific process is as follows:
(1) Based on the idea of reliability distribution, the training algorithm of the conventional CMAC neural network is improved, and the improved CMAC neural network algorithm is as follows:
wherein the content of the first and second substances,as a weight value, the weight value,andin order to do the number of learning times,andfor a certain memory cell to be activated,in order to learn the rate of speed,in order to be the desired value,in order to output the value of the output,in order to generalize the parameters of the process,is as followsAn activated memory cellThe number of learned times at the time of the sub-learning,is as followsIs activated to storeStorage unit 1The number of learned times at the time of the sub-learning,is as followsAn activated memory cellThe number of learned times at the time of the sub-learning,to balance the learning constants.
It should be noted that the improved training algorithm of the CNAC neural network performs statistics on the number of learning times of each activated memory cell in the training process, the statistics not only includes changes of subsequent learning samples to the same number of activation times of the memory cell, but also includes changes of subsequent training to the number of activation times of the memory cell, and then when the weight is updated, the error is distributed according to the percentage of the number of learning times of the activated memory cell to the sum of the number of learning times of all activated memory cells, and the larger the percentage is, the smaller the distribution error is.
The weight adjustment rule is as follows: if the iterative learning times are more, the reliability of the included information is high, and the adjustment amount is less; if the iterative learning times are few, the reliability of the included information is low, and the adjustment amount is large. Therefore, the learning interference of the subsequent learning samples on the previous learning samples can be reduced, and the learning interference of the subsequent training on the previous training can also be reduced. The algorithm is based on the credibility distribution error, the error is less corrected for the storage unit with more learning times, and the error is more corrected for the storage unit with less learning times, so that the learning interference is reduced.
(2) Inputting the historical yield sequence into an improved CMAC neural network to perform one-time iterative training to obtain a yield predicted value corresponding to each historical yield in the historical yield sequence, and forming a historical yield prediction sequence; and respectively calculating first difference values between two elements at corresponding positions in the historical yield prediction sequence and the historical yield sequence, and performing control adjustment on network parameters of the improved CMAC neural network according to the first difference values to obtain a new CMAC neural network, wherein the network parameters comprise input space, generalization parameters, learning rate, learning times and the like of the network and other parameters of the improved CMAC neural network algorithm.
Specifically, an expert coordinator is introduced, and the control strategy is switched according to a first difference value:
wherein the content of the first and second substances,in order to be controlled by the expert,it is shown that the CMAC control is,indicating proportional control, e indicating a first difference,is a first difference threshold.
The expert control means that extreme value control is introduced to quickly correct the neural network based on an expert control system, so that errors can be reduced in the next iterative training; CMAC control refers to the regulation corresponding to the self-learning of a CMAC neural network; the proportional control is an auxiliary controller of the CMAC neural network, because the independent reference of the proportional control can reduce the relative stability of the system and even cause the instability of the closed-loop system.
(3) Performing one-time iterative training on the new CMAC neural network by using the historical yield sequence to obtain a new historical yield prediction sequence; and calculating a second difference value between two elements at corresponding positions in the historical yield prediction sequence and the fresh historical yield prediction sequence, when the second difference value meets a difference value threshold value, calculating a third difference value between two elements at corresponding positions in the fresh historical yield prediction sequence and the historical yield sequence, accumulating the third difference value to obtain a third difference value accumulated value, and when the third difference value accumulated value is smaller than the accumulated value threshold value, determining the new CMAC neural network as the target CMAC neural network.
Specifically, second difference values between two elements at corresponding positions in the historical yield prediction sequence and the new historical yield prediction sequence are respectively calculated, average second difference values between all the second difference values are calculated, a second difference value range is set, when the average second difference value is not within the second difference value range, control adjustment of network parameters is immediately carried out on the new CMAC neural network, otherwise, when the average second difference value is within the second difference value range, the new CMAC neural network is better than the prediction result of the improved CMAC neural network, then a third difference value between the two elements at corresponding positions in the new historical yield prediction sequence and the historical yield sequence is calculated, the third difference value is accumulated to obtain a third difference value accumulated value, when the third accumulated value is smaller than an accumulated value threshold value, the new CMAC neural network is confirmed to be the target CMAC neural network, when the third difference value accumulated value is larger than or equal to the accumulated value threshold value, control adjustment of the network parameters is carried out on the new CMAC neural network, and then iterative neural network training is carried out again until the target CMAC neural network is obtained.
The target CMAC neural network is a neural network corresponding to the best generalization ability, and the prediction effect of the neural network is the best.
S003, acquiring a target yield prediction network of each threading machine, and enabling the real-time yield sequence of each threading machine to pass through the corresponding target yield prediction network to obtain a yield prediction sequence of each threading machine; obtaining the total quantity of good products of each threading machine according to the yield prediction sequence and the total quantity of reinforcing steel bars corresponding to batches, calculating a good product adaptation index of the current threading machine based on the total quantity of the good products, and calculating a machine aging index of the current threading machine according to a difference value of two adjacent elements in the yield prediction sequence; and determining a target threading machine according to the good product adaptation index and the machine aging index, and adjusting the equipment parameters of each threading machine to the equipment parameters of the target threading machine.
Specifically, the method in step S002 is used to obtain a target yield prediction network for each threading machine, that is, one threading machine corresponds to one dedicated target yield prediction network.
And respectively acquiring a real-time yield sequence of each threading machine based on the time sequence, inputting each real-time yield sequence into a corresponding target yield prediction network to obtain a yield prediction sequence of the corresponding threading machine, and in the same way, one threading machine corresponds to one yield prediction sequence. Obtaining the total quantity of good products of the corresponding threading machine according to the predicted value of each good product rate and the total quantity of the reinforcing steel bars corresponding to the batch in the good product rate prediction sequence, counting the number of the good products matched with the standard nuts, and taking the ratio of the number of the good products to the total quantity of the good products as a good product matching index; calculating a machine aging index according to the difference value of two adjacent elements in the yield prediction sequence, wherein the method for acquiring the machine aging index comprises the following steps: and respectively calculating the predicted value difference value between two adjacent predicted values of the good product rate in the good product rate prediction sequence, accumulating the predicted value difference values to obtain a predicted value difference value accumulated value, calculating the ratio between the combined number of two adjacent elements in the good product rate prediction sequence and the predicted value difference value accumulated value, and taking the ratio as a machine aging index.
The non-defective product adaptation index is whether the steel bar after threading is matched with the standard nut, namely whether the steel bar can be screwed with the standard nut; the machine aging index is the aging degree of the threading machine reflected according to the difference between the yield of each batch, namely the larger the difference is, the more the machine aging is aggravated by the overload work of the threading machine.
Calculating the ratio between the good product adaptive index and the machine aging index, and taking the ratio as the working state index of the corresponding threading machine, wherein the working state index and the machine aging index are in a negative correlation relationship, and the working state index and the good product adaptive index are in a positive correlation relationship; the method comprises the steps that a threading machine corresponds to a good product adaptive index and a machine aging index, so that the working state index of each threading machine can be obtained, the threading machine corresponding to the largest working state index serves as a target threading machine, the target threading machine is the threading machine with the best equipment parameters during threading, and therefore in order to ensure that the overall threading machine can achieve the optimal condition, the equipment parameters of each threading machine are adjusted to the equipment parameters of the target threading machine.
It should be noted that, when there are a plurality of target threading machines, the average device parameter of the target threading machine is obtained, and the device parameter of each threading machine is adjusted to the average device parameter.
In summary, the embodiments of the present invention provide a method for debugging threading work of a threading machine controller, where the method trains a proprietary yield prediction network of each threading machine by using historical data of the threading machine, so that the proprietary yield prediction network can predict a yield prediction value of each threading machine accurately, and further confirm an optimal equipment parameter of the threading machine according to a prediction result of each threading machine, so as to adjust the equipment parameter of the global threading machine in time, so that each threading machine is in an optimal working state, improve consistency of steel bar threading results, and improve yield of steel bar threading.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that are within the spirit of the present invention are intended to be included therein.
Claims (7)
1. A method for debugging the threading work of a threading machine controller is characterized by comprising the following steps:
acquiring the current of a motor and the advancing speed of a threading die cutter after the current steel bar is clamped based on sampling frequency to obtain a current sequence and an advancing speed sequence of the motor in the whole threading time period of the current steel bar; calculating a motor current stability index according to the difference between adjacent motor currents in the motor current sequence, and calculating a thread uniformity index of the current reinforcing steel bar surface by combining the motor current stability index and the travelling speed change corresponding to the travelling speed sequence;
obtaining the thread uniformity index of each reinforcing steel bar in the current batch under the current threading machine to obtain a thread uniformity index sequence; calculating a local abnormal factor of each thread uniformity index according to the difference of any two thread uniformity indexes in the thread uniformity index sequence, and acquiring the yield of the reinforcing steel bars in the current batch based on the local abnormal factor; obtaining the yield of a plurality of batches of historical reinforcing steel bars to form a historical yield sequence, and performing continuous iterative training on a yield prediction network by using the historical yield sequence to confirm a target yield prediction network corresponding to the current threading machine according to the difference between input data and output data of the yield prediction network;
acquiring the target good product rate prediction network of each threading machine, and enabling the real-time good product rate sequence of each threading machine to pass through the corresponding target good product rate prediction network to obtain a good product rate prediction sequence of each threading machine; obtaining the total quantity of good products of each threading machine according to the prediction sequence of the good product rate and the total quantity of the reinforcing steel bars corresponding to the batch, calculating a good product adaptation index of the current threading machine based on the total quantity of the good products, and calculating a machine aging index of the current threading machine according to a difference value of two adjacent elements in the prediction sequence of the good product rate; determining a target threading machine according to the good product adaptation index and the machine aging index, and adjusting the equipment parameters of each threading machine to the equipment parameters of the target threading machine;
the calculation formula of the motor current stability index is as follows:
wherein the content of the first and second substances,the motor current stability index is obtained;for the second in the motor current sequenceA motor current;for the front of the motor current sequenceAverage motor current between individual motor currents;for the front of the motor current sequenceAverage motor current between individual motor currents;for the second in the motor current sequenceCurrent of motor andthe difference between the individual motor currents;is a motor current sequence;is the length of the motor current sequence;
the method for acquiring the thread uniformity index comprises the following steps:
obtaining the maximum advancing speed and the minimum advancing speed in the advancing speed sequence and the average advancing speed, and similarly, obtaining the minimum motor current in the motor current sequence, and calculating the thread uniformity index of the surface of the steel bar according to the difference between the maximum advancing speed, the minimum advancing speed and the average advancing speed, the minimum motor current and the motor current stability index, wherein the calculation formula of the thread uniformity index is as follows:
wherein the content of the first and second substances,is an index of thread uniformity;is the maximum travel speed;is the average travel speed;is the minimum travel speed;is the minimum motor current;is a hyperbolic tangent function;is an arcsine function;is a sequence of travel speeds;
the method for acquiring the local abnormal factor comprises the following steps: taking any one thread uniformity index in the thread uniformity index sequence as a target thread uniformity index, respectively calculating thread uniformity index difference values between the target thread uniformity index and other thread uniformity indexes, accumulating the thread uniformity index difference values to obtain a difference accumulated value, calculating a ratio between the difference accumulated value and the number of elements in the thread uniformity index sequence, and taking the reciprocal of the ratio as the local reachable density of the target thread uniformity index; obtaining local reachable density of each thread uniformity index in a thread uniformity index sequence to form a local reachable density set, taking any one local reachable density in the local reachable density set as a target local reachable density, respectively calculating local reachable density difference values between the target local reachable density and other target local reachable densities, accumulating the local reachable density difference values to obtain a local reachable density difference value accumulated value, calculating a ratio between the local reachable density difference value accumulated value and the number of elements in the local reachable density set, and taking a first ratio between the ratio and the target local reachable density as a local abnormal factor corresponding to the target local reachable density, namely a local abnormal factor corresponding to the thread uniformity index;
the method for acquiring the target yield prediction network comprises the following steps:
based on the improved yield prediction network algorithm, performing one-time iterative training on a yield prediction network by using the historical yield sequence to obtain a yield prediction value corresponding to each historical yield in the historical yield sequence, and forming a historical yield prediction sequence;
respectively calculating first difference values between two elements at corresponding positions in the historical yield prediction sequence and the historical yield sequence, and performing control adjustment on network parameters of a yield prediction network according to the first difference values to obtain a new yield prediction network; the network parameters include: the input space of the yield prediction network and each parameter in the improved yield prediction network algorithm;
performing one-time iterative training on the new yield prediction network by using the historical yield sequence to obtain a new historical yield prediction sequence; calculating a second difference value between two elements at corresponding positions in the historical yield prediction sequence and the fresh historical yield prediction sequence, when the second difference value meets a difference value threshold value, calculating a third difference value between two elements at corresponding positions in the fresh historical yield prediction sequence and the historical yield sequence, accumulating the third difference value to obtain a third difference value accumulated value, and when the third difference value accumulated value is smaller than the accumulated value threshold value, determining that the new yield prediction network is the target yield prediction network.
2. The method for debugging threading work of the threading machine controller according to claim 1, wherein the method for obtaining the yield of the current batch of the steel bars based on the local abnormal factor comprises:
setting a local abnormal factor threshold, and when the local abnormal factor is smaller than the local abnormal factor threshold, determining that the steel bar after corresponding threading is good; and counting the first number of good products, calculating the ratio of the total number of the reinforcing steel bars in the current batch to the first number, and taking the ratio as the yield of the reinforcing steel bars.
3. The method for debugging the threading operation of the threading machine controller according to claim 1, wherein the improved yield prediction network algorithm comprises:
wherein the content of the first and second substances,as a weight value, the weight value,andin order to do the number of learning times,andfor a certain memory cell to be activated,in order to learn the rate of speed,in order to be the desired value,in order to output the value of the output,in order to generalize the parameters of the process,is as followsThe activated memory cellThe number of learned times at the time of the sub-learning,is as followsAn activated memory cellThe number of learned times at the time of the sub-learning,is as followsAn activated memory cellThe number of learned times at the time of the sub-learning,to balance the learning constants.
4. The method for debugging threading work of the threading machine controller according to claim 1, wherein the method for controlling and adjusting the network parameters of the yield prediction network according to the first difference comprises the following steps:
when the first difference is larger than or equal to the first difference threshold, based on an expert control system, introducing extremum control for quick correction; when the first difference is less than or equal to the negative of the first difference threshold, a control adjustment is made in conjunction with the proportional control and the CMAC control.
5. The method for debugging threading work of the threading machine controller according to claim 1, wherein the method for calculating the good product adaptation index of the current threading machine based on the total number of the good products comprises the following steps:
and counting the number of good products matched with the standard nuts, and taking the ratio of the number of the good products to the total number of the good products as a good product matching index.
6. The method for debugging threading work of the threading machine controller according to claim 1, wherein the method for calculating the machine aging index of the current threading machine from the difference value of two adjacent elements in the yield prediction sequence comprises the following steps:
respectively calculating predicted value difference values between two adjacent predicted values of the good product rate in the good product rate prediction sequence, and accumulating the predicted value difference values to obtain predicted value difference value accumulated values; and calculating the ratio between the combined quantity of two adjacent elements in the yield prediction sequence and the difference value accumulated value of the prediction value, and taking the ratio as the machine aging index.
7. The method for debugging threading work of the threading machine controller according to claim 1, wherein the method for determining the target threading machine according to the good product adaptation index and the machine aging index comprises the following steps:
calculating the ratio between the non-defective product adaptation index and the machine aging index, and taking the ratio as the working state index of the corresponding threading machine; the working state index and the machine aging index are in a negative correlation relationship, and the working state index and the good product adaptation index are in a positive correlation relationship;
and acquiring the working state index of each threading machine, and taking the threading machine corresponding to the largest working state index as the target threading machine.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210978355.2A CN115049170B (en) | 2022-08-16 | 2022-08-16 | Method for debugging threading work of threading machine controller |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210978355.2A CN115049170B (en) | 2022-08-16 | 2022-08-16 | Method for debugging threading work of threading machine controller |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115049170A CN115049170A (en) | 2022-09-13 |
CN115049170B true CN115049170B (en) | 2022-10-25 |
Family
ID=83167689
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210978355.2A Active CN115049170B (en) | 2022-08-16 | 2022-08-16 | Method for debugging threading work of threading machine controller |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115049170B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106114111A (en) * | 2016-06-24 | 2016-11-16 | 江苏大学 | A kind of adjustable damping energy regenerative suspension system containing driving behavior forecast model and control method thereof |
CN110658774A (en) * | 2019-11-05 | 2020-01-07 | 百世康电气有限公司 | Controller for threading machine |
CN110794680A (en) * | 2019-11-14 | 2020-02-14 | 浙江工业大学 | Magnetic levitation ball system prediction tracking control method based on extended state observer |
CN112308298A (en) * | 2020-10-16 | 2021-02-02 | 同济大学 | Multi-scenario performance index prediction method and system for semiconductor production line |
CN113487140A (en) * | 2021-06-11 | 2021-10-08 | 田新忠 | Operation management platform and control method for intelligent construction equipment on building construction site |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113759868B (en) * | 2021-08-16 | 2023-05-02 | 浙江运达风电股份有限公司 | Industrial Ethernet fault prediction method based on neural network |
-
2022
- 2022-08-16 CN CN202210978355.2A patent/CN115049170B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106114111A (en) * | 2016-06-24 | 2016-11-16 | 江苏大学 | A kind of adjustable damping energy regenerative suspension system containing driving behavior forecast model and control method thereof |
CN110658774A (en) * | 2019-11-05 | 2020-01-07 | 百世康电气有限公司 | Controller for threading machine |
CN110794680A (en) * | 2019-11-14 | 2020-02-14 | 浙江工业大学 | Magnetic levitation ball system prediction tracking control method based on extended state observer |
CN112308298A (en) * | 2020-10-16 | 2021-02-02 | 同济大学 | Multi-scenario performance index prediction method and system for semiconductor production line |
CN113487140A (en) * | 2021-06-11 | 2021-10-08 | 田新忠 | Operation management platform and control method for intelligent construction equipment on building construction site |
Also Published As
Publication number | Publication date |
---|---|
CN115049170A (en) | 2022-09-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
AU665048B2 (en) | Apparatus and method for feedback-adjusting working condition for improving dimensional accuracy of processed workpieces | |
CN104709748B (en) | A kind of film take-up tension control system based on FPGA and method | |
CN110488754B (en) | Machine tool self-adaptive control method based on GA-BP neural network algorithm | |
CN115336780B (en) | Loose conditioning water-adding control system based on neural network model and double parameter correction | |
CN110893001A (en) | Method and system for controlling water content of outlet of loosening and dampening process | |
CN115049170B (en) | Method for debugging threading work of threading machine controller | |
CN115305526A (en) | Self-adaptive control method for consistency of copper foil thickness and surface density based on X-ray measurement | |
DE102010002174A1 (en) | Method for controlling an injection molding process | |
CN114253219A (en) | Grinding force self-adaptive control method and system based on end face grinding | |
CN108303952A (en) | A kind of rate curve planing method of small line segment suitable for numerical control device cutter | |
CN116352607A (en) | Centerless grinder compensation control system and method based on numerical control analysis | |
CN110193427B (en) | Automatic control method for lime addition amount in copper flotation process | |
CN109240203B (en) | Multi-model-based continuous casting billet constant weight control method | |
CN102999008B (en) | Method for optimizing parameters of overlap controller of edge trimming circle shear | |
CN117131314A (en) | Mixed material temperature monitoring and regulating system for asphalt pavement construction | |
CN116463691A (en) | Electrolytic copper foil accurate production system and electrolytic copper foil accurate production method | |
CN110764404B (en) | PID parameter self-adaptive wire drawing machine frequency converter debugging method | |
Miodragović et al. | The application of metaheuristic algorithms in multi-objective optimization of engineering problems | |
CN112434868A (en) | Sheet drying process accurate control intelligent prediction model and application | |
CN111482466A (en) | Method for setting acceleration of rolling mill | |
US5770830A (en) | Method and apparatus for wire-cutting curved workpieces by compensating machining parameters on wire-cut electric discharge machine | |
CN118060149B (en) | Global control method, device, equipment and medium for electrode coating surface density | |
CN112589354A (en) | System and method for automatically controlling welding deformation | |
CN113909318B (en) | Bus line compensation process method | |
CN111086268B (en) | Integral reinforcement learning control method of calcium carbide furnace lime powder ball pressing control 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 |