CN113448250B - Dyeing machine auxiliary agent intelligent distribution system and control method thereof - Google Patents
Dyeing machine auxiliary agent intelligent distribution system and control method thereof Download PDFInfo
- Publication number
- CN113448250B CN113448250B CN202111001290.8A CN202111001290A CN113448250B CN 113448250 B CN113448250 B CN 113448250B CN 202111001290 A CN202111001290 A CN 202111001290A CN 113448250 B CN113448250 B CN 113448250B
- Authority
- CN
- China
- Prior art keywords
- distribution
- auxiliary agent
- stop value
- value
- valve
- 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
- 239000012752 auxiliary agent Substances 0.000 title claims abstract description 206
- 238000009826 distribution Methods 0.000 title claims abstract description 202
- 238000004043 dyeing Methods 0.000 title claims abstract description 70
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000003860 storage Methods 0.000 claims abstract description 35
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 29
- 239000000463 material Substances 0.000 claims abstract description 23
- 238000011010 flushing procedure Methods 0.000 claims abstract description 17
- 238000003062 neural network model Methods 0.000 claims description 61
- 238000002156 mixing Methods 0.000 claims description 41
- 239000007788 liquid Substances 0.000 claims description 31
- 238000012549 training Methods 0.000 claims description 29
- 239000000243 solution Substances 0.000 claims description 23
- 238000011217 control strategy Methods 0.000 claims description 16
- 239000000654 additive Substances 0.000 claims description 12
- 230000000996 additive effect Effects 0.000 claims description 9
- 210000002569 neuron Anatomy 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 239000011259 mixed solution Substances 0.000 claims description 5
- 239000000126 substance Substances 0.000 claims description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 4
- 210000005036 nerve Anatomy 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 2
- 239000013589 supplement Substances 0.000 claims description 2
- 230000008569 process Effects 0.000 description 22
- 238000013528 artificial neural network Methods 0.000 description 14
- 230000006870 function Effects 0.000 description 10
- 230000000694 effects Effects 0.000 description 9
- 230000009471 action Effects 0.000 description 8
- 238000013135 deep learning Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 238000004519 manufacturing process Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 239000012530 fluid Substances 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000007704 transition Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 239000002699 waste material Substances 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 238000009792 diffusion process Methods 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 239000002253 acid Substances 0.000 description 1
- 239000003513 alkali Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000003153 chemical reaction reagent Substances 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 210000001503 joint Anatomy 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 239000008237 rinsing water Substances 0.000 description 1
- 239000010865 sewage Substances 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000005406 washing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
-
- D—TEXTILES; PAPER
- D06—TREATMENT OF TEXTILES OR THE LIKE; LAUNDERING; FLEXIBLE MATERIALS NOT OTHERWISE PROVIDED FOR
- D06B—TREATING TEXTILE MATERIALS USING LIQUIDS, GASES OR VAPOURS
- D06B23/00—Component parts, details, or accessories of apparatus or machines, specially adapted for the treating of textile materials, not restricted to a particular kind of apparatus, provided for in groups D06B1/00 - D06B21/00
- D06B23/20—Arrangements of apparatus for treating processing-liquids, -gases or -vapours, e.g. purification, filtration or distillation
- D06B23/205—Arrangements of apparatus for treating processing-liquids, -gases or -vapours, e.g. purification, filtration or distillation for adding or mixing constituents of the treating material
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Textile Engineering (AREA)
- Treatment Of Fiber Materials (AREA)
Abstract
The invention discloses an intelligent distribution system of dyeing machine auxiliaries and a control method thereof, wherein the system comprises an auxiliary storage tank, an auxiliary distribution valve, a flushing water valve, a distribution rotor pump, a reverse forced pipeline mixer, a flowmeter, a blanking valve, a dyeing machine material cylinder and a control system, the control system comprises an upper computer intelligent PC system, a PLC control system, a compressed air barometer, a frequency converter and a field signal collector, and the upper computer intelligent PC system acquires the names, the dosages and the numbers of the blanking valves to form an auxiliary distribution order; and acquiring a recommended pre-stop value according to the current auxiliary agent distribution data, superposing the recommended pre-stop value with the empirical pre-stop value according to a variable proportion to acquire a final pre-stop value, acquiring distribution parameters by the PLC control system, judging the residual distribution quantity according to the final pre-stop value, controlling the opening and closing of an auxiliary agent distribution valve, a blanking valve, a distribution rotor pump and a flushing water valve, and adjusting the output frequency of the frequency converter. The invention automatically adjusts the pre-stop value according to the historical distribution data, thereby improving the distribution precision.
Description
Technical Field
The invention relates to the technical field of material reagent distribution, in particular to an intelligent distribution system for an auxiliary agent of a dyeing machine and a control method thereof.
Background
In the process of distributing the auxiliary agent of the dyeing machine, the viscosity and the flowability of different auxiliary agents, the lengths of distribution pipelines of different blanking ports, the operation frequency of a distribution pump, the pressure intensity of compressed air, the distribution amount and other factors can cause different volumes of the auxiliary agent flowing through a distribution valve after the distribution valve is closed, and the factors comprehensively influence the distribution precision of the auxiliary agent. At present, the precision control of the distribution of the auxiliary agent needs technicians to carry out distribution tests for multiple times, record the errors of an order distribution value and an actual distribution value and then adjust the distribution precision of the auxiliary agent. Since an auxiliary agent distribution system of a dyeing machine usually has tens of auxiliary agents and tens of dye vats, if these conditions are combined one-to-one, there will be thousands of combinations, and the work for adjusting the distribution pre-stop value according to these combinations is very large.
Some auxiliary agents distributed by the system are strong acid, strong alkali or high-corrosivity solution, and if the auxiliary agents are directly discharged during debugging, the auxiliary agents cause great harm to the environment; if the container is taken up and the container is put back to the storage tank, the skin of an operator is easily stained in the transferring process, and the human body is injured. At present, when an auxiliary agent distribution system of a dyeing machine is debugged, a common method is to take water for distribution precision test, set different pre-stop values according to different blanking ports, and directly discharge the water during the test. Although the simplified setting method has better performability, the influence of factors such as the variety of the auxiliary agent, the running frequency of the distribution pump, the pressure of compressed air and the like is not considered, so that the distribution precision is poorer in actual running. In order to ensure that the actual delivery volume reaches the delivery volume required by an order, the pre-stop value is usually adjusted to be smaller, so that the actual delivery volume is larger during actual delivery, waste is caused, redundant auxiliaries cannot be fully utilized during dyeing, and the waste is discharged to a sewage pipe after dyeing is finished, so that the environment is polluted.
When a plurality of auxiliary agents are conveyed by sharing one flowmeter in a time-sharing way, due to the characteristics of some auxiliary agents, such as the auxiliary agents with larger specific gravity and poor diffusion capacity, when the pipeline is flushed by water after the auxiliary agents are conveyed, the mixed liquid flowing through the flowmeter is very uneven. Because the measurement of the flowmeter needs the plasma of the solution to be relatively uniform so as to accurately measure, the mixture of the non-uniform auxiliary agent and the water flows through the flowmeter timing, so that the metering error is large, and the metering of the rinsing water quantity is influenced. On the other hand, when the same dyeing machine needs a plurality of auxiliary agents, in the conversion process of distributing one auxiliary agent and then distributing the second auxiliary agent, the mixed auxiliary agents flowing through the flow meter are very uneven due to the fact that specific gravity, diffusion capacity and conductivity of different auxiliary agents are different greatly, the metering value of the flow meter is also fluctuated severely, and the metering error is very large. Based on the problems existing at present, an intelligent high-precision distribution scheme of the auxiliary agent of the dyeing machine is urgently needed to be researched.
Disclosure of Invention
In order to overcome the defects and shortcomings in the prior art, the invention provides an intelligent distribution system of dyeing machine auxiliaries and a control method thereof, the system learns from distribution historical experience, collects a large amount of actual distribution data, trains a pre-stop value neural network model and a predicted time-use neural network model through a multilayer fully-connected neural network to obtain an optimized control strategy, inputs main influence factors into the pre-stop value neural network model during transmission to calculate a recommended pre-stop value, superposes the preset empirical pre-stop value and the recommended pre-stop value in a variable proportion to calculate a final pre-stop value, and controls the closing time of a distribution valve according to the final pre-stop value to achieve the effect of improving distribution precision; in addition, when in distribution, the main influence data is input into a distribution time prediction neural network model, and the distribution predicted time of the auxiliary agent is calculated, the actual distribution time is subtracted by the field PLC control system from the distribution predicted time, and if the distribution predicted time exceeds a set proportion, the PLC control system sends out distribution overtime alarm and information to remind personnel to check the reason of overtime; the problem of large metering error of a flowmeter during the process of auxiliary agent conversion during time-sharing conveying of various auxiliary agents is solved through a pipeline reverse forced mixer.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention provides an intelligent distribution system for an auxiliary agent of a dyeing machine, which comprises: the system comprises an auxiliary agent storage tank, an auxiliary agent distribution valve, a flushing water valve, a distribution rotor pump, a pipeline reverse forced mixer, a flow meter, a blanking valve, a dyeing machine material cylinder and a control system, wherein the pipeline reverse forced mixer is provided with a reverse mixing pump, and the reverse mixing pump is used for mixing the auxiliary agent and water in the pipeline uniformly in a circulating manner;
the auxiliary agent storage tank is connected with an auxiliary agent distribution valve, the auxiliary agent distribution valve is connected with a distribution rotor pump, the distribution rotor pump is connected with a flowmeter, a connecting pipeline between the distribution rotor pump and the flowmeter is provided with a reverse circulation branch, the reverse mixing pump is arranged in the reverse circulation branch, the input end of the reverse mixing pump is connected with the input end of the flowmeter, the output end of the reverse mixing pump is connected with the output end of the distribution rotor pump, the flowmeter is connected with a blanking valve, the blanking valve is arranged corresponding to a material cylinder of the dyeing machine, and the flushing water valve is connected with the auxiliary agent distribution valve;
the control system comprises an upper computer intelligent PC system, a PLC control system, a compressed air barometer, a frequency converter and a field signal collector;
the compressed air barometer is used for acquiring a valve core compressed air pressure value, the frequency converter is used for driving the distribution rotor pump, and the field signal collector is used for acquiring a valve on-off state signal and a liquid level state of the auxiliary agent storage tank;
the upper computer intelligent PC system is provided with an auxiliary agent distribution order acquisition module and a pre-stop value neural network model construction module;
the auxiliary agent distribution order acquisition module is used for acquiring the name and the dosage of the auxiliary agent and the cylinder number of the dyeing machine to form an auxiliary agent distribution order;
the pre-stop value neural network model building module is used for building a pre-stop value neural network model, the pre-stop value neural network model obtains a recommended pre-stop value according to current auxiliary agent distribution data, the recommended pre-stop value and an empirical pre-stop value are superposed according to a variable proportion to obtain a final pre-stop value, the variable proportion is increased progressively along with the increase of training iteration times, and the auxiliary agent distribution data comprises an auxiliary agent number, a blanking valve number, a compressed air pressure value, a frequency converter frequency and an ordered target distribution quantity;
and the final pre-stop value is obtained by superposing the empirical pre-stop value according to a variable proportion, and is specifically represented as follows:
wherein the content of the first and second substances,
a recommended pre-stop value is indicated,
expressing a preset experience pre-stop value, K expressing a variable proportion value, dividing the variable proportion value into equal parts according to the set training iteration times, increasing the equal parts after each training iteration, and enabling K to be more than or equal to 0 and less than or equal to 1;
the PLC control system is used for collecting the real-time flow of the flow meter, the running frequency of the frequency converter, the air pressure value of compressed air, the switching state of each valve and the liquid level state of the auxiliary agent storage tank, judging the residual delivery amount according to the final pre-stop value, controlling the opening and closing of the auxiliary agent delivery valve, the blanking valve, the delivery rotor pump and the flushing water valve according to a variable proportion control strategy, and adjusting the output frequency of the frequency converter.
According to a preferable technical scheme, an ultrasonic liquid level meter is arranged in the auxiliary agent storage tank, a high-level float switch is arranged at the top of the auxiliary agent storage tank, the ultrasonic liquid level meter is used for detecting the liquid level height value of the auxiliary agent storage tank, and the high-level float switch is used for outputting the switching value for stopping the auxiliary agent supplement when the liquid level of the auxiliary agent storage tank exceeds the upper limit value.
As a preferable technical scheme, a balance cover type liquid level meter is arranged in the material cylinder of the dyeing machine and used for detecting the liquid level and the upper limit of the liquid level in the material cylinder of the dyeing machine.
As a preferred technical scheme, the upper computer intelligent PC system is also provided with a time-for-prediction neural network model building module;
the prediction time-use neural network model building module is used for building a prediction time-use neural network model, the prediction time-use neural network model is used for calculating the prediction time-use, the PLC control system multiplies the prediction time-use by a set proportion to obtain the longest time limit allowed by the delivery of the auxiliary agent, and if the actual time-use of each scanning period is greater than the longest time limit, an alarm control signal is output.
The invention also provides a control method of the intelligent distribution system of the dyeing machine additive, which comprises the following steps:
the upper computer intelligent PC system obtains the name and the dosage of the auxiliary agent and the cylinder number of the dyeing machine, forms an auxiliary agent distribution order and sends the order to the PLC control system;
the PLC control system receives an auxiliary agent distribution order, controls a distribution rotor pump to rotate, controls the opening of a corresponding auxiliary agent distribution valve according to the auxiliary agent number, and controls the opening of a corresponding blanking valve according to the dyeing machine cylinder number;
the reverse forced mixer mixes the solution flowing into the reverse mixing pump with the solution flowing into the reverse mixing pump at the next moment and outputs the mixed solution to the flowmeter;
the intelligent PC system of the upper computer substitutes the serial number of the auxiliary agent, the serial number of the blanking valve, the air pressure value of compressed air, the frequency of the frequency converter and the target distribution quantity of the order into the trained neural network model of the pre-stop value, calculates to obtain a recommended pre-stop value and stores the recommended pre-stop value into a register of the PLC control system;
superposing a preset empirical pre-stop value and a preset recommended pre-stop value according to a variable proportion to calculate a final pre-stop value, subtracting a metering value of a flow meter from the delivery quantity of an order in each scanning period to obtain a residual delivery quantity, and closing an auxiliary agent delivery valve and simultaneously closing a blanking valve and a delivery rotor pump by a PLC (programmable logic controller) control system when the residual delivery quantity is less than or equal to the final pre-stop value;
the preset empirical pre-stop value and the preset recommended pre-stop value are superposed according to a variable proportion to calculate a final pre-stop value, which is specifically represented as:
wherein the content of the first and second substances,
a recommended pre-stop value is indicated,
expressing a preset experience pre-stop value, K expressing a variable proportion value, dividing the variable proportion value into equal parts according to the set training iteration times, increasing the equal parts after each training iteration, and enabling K to be more than or equal to 0 and less than or equal to 1;
the PLC control system starts a flushing water valve, a blanking valve and a distribution rotor pump, and conveys the auxiliary agent remained in the pipeline into a material cylinder of the dyeing machine;
the PLC control system stores the actual delivery amount of the auxiliary agent into a register and sends a delivery completion mark;
the intelligent PC system of the upper computer acquires actual distribution quantity, and the actual distribution quantity, the auxiliary agent number, the blanking valve number, the compressed air pressure value, the frequency of the frequency converter, the ordered target distribution quantity and the final pre-stop value are combined and stored in a database.
As a preferred technical solution, the training step of the pre-stop value neural network model includes:
taking the additive number, the blanking valve number, the compressed air pressure value, the frequency of a frequency converter and the ordered target distribution quantity in the database as the input of the pre-stop value neural network model;
the pre-stop value neural network model adopts a full-connection neural network model, and the structure of the full-connection neural network model sequentially comprises the following steps: the device comprises an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer;
the calculation formula of each layer of the fully-connected neural network model is as follows:
whereiniIs the first layer of the layeriThe number of the nerve cells is one,jis the first layer corresponding to the previous layer of the present layerjThe output of the first and second processors is,mthe number of neurons in the layer above the current layer,nthe second few nerve layers are represented,yin order to be output, the output is,xin order to be an input, the user can select,θis the weight value of the current layer,bis the offset value of the present layer,fis the activation function of the layer;
taking the deviation value of the distribution historical data used for training as the output of the pre-stop neural network model;
iteratively training weights of a first hidden layer of a prestack-valued neural network model using an Adam optimizer
And bias
The weight of the second hidden layer
And bias
The weight of the third hidden layer
And bias
And obtaining a pre-stop value neural network model.
As a preferred technical solution, the cost function formula of the fully-connected neural network model is as follows:
where C is the cost, L represents the total number of samples, S represents the number of samples,
a target delivery amount for the order is indicated,
the actual amount of delivery is indicated,zrepresenting the actual pre-stop value.
As a preferred technical scheme, the dyeing machine auxiliary agent intelligent distribution system is provided with a predicted time-of-use neural network model, and the cost function formula of the predicted time-of-use neural network model is as follows:
where C is the cost, L represents the total number of samples, S represents the number of samples,
in order to start the delivery of the auxiliary agent,
and the end time of the additive distribution.
Compared with the prior art, the invention has the following advantages and beneficial effects:
(1) the traditional dyeing machine auxiliary agent system is generally based on a preprogramming mode, distribution parameters are set according to distribution data during testing, the mode is basically fixed after debugging is finished, and a control strategy cannot be adjusted according to a large amount of actually operated data; and the closing time point of the auxiliary agent distribution valve is controlled after comprehensive calculation according to the recommended pre-stop value and the variable proportion control strategy, so that the introduction of a new auxiliary agent or a new dye vat can be completed under the condition of not influencing the overall control strategy while the intelligent distribution precision control is realized, and the requirements of factory production expansion and variety expansion can be well met.
(2) The invention adopts the predicted time-use neural network model to predict the time length required by the delivery of the auxiliary agent, dynamically monitors the delivery process, judges the actual time-out, finds out the process abnormality in time, reminds the personnel to process, reduces the waiting time, solves the technical problem that the time span of the auxiliary agent delivery process is too large and the monitoring is not good, and achieves the technical effect of carrying out the time-out monitoring and alarming on the auxiliary agent delivery process of each order.
(3) The traditional pipeline mixer is unpowered, the solution in the pipeline is axially and rotationally mixed by depending on the flow of the fluid, the front and back mixing effects are not obvious, the reverse forced mixer for the pipeline disclosed by the invention forces the solution to be circularly mixed by depending on the action of a pump, so that the front and back sections of the solution in the pipeline are mixed, the solution proportion transition during mixing can be smoother and more uniform, the mixed solution flowing out of the reverse forced mixer for the pipeline flows through a flowmeter, the additional metering fluctuation of the flowmeter cannot be caused, a more accurate metering value can be obtained, and the problem of large metering error of the flowmeter during the time-sharing conveying of various additives in the additive conversion process is solved.
Drawings
Fig. 1 is a schematic structural diagram of an intelligent auxiliary distribution system of a dyeing machine, which is disclosed by the invention;
FIG. 2 is a schematic diagram of the structure of the reverse forced mixer of the present invention;
FIG. 3 is a schematic diagram of the control system of the present invention;
FIG. 4 is a schematic diagram of a fully-connected neural network according to the present invention;
fig. 5 is a flow chart of the additive dispensing control of the present invention.
The system comprises an auxiliary agent storage tank 1, an auxiliary agent distribution valve 2, a flushing water valve 3, a distribution rotor pump 4, a reverse mixing pump 5, a flow meter 6, a blanking valve 7, a dyeing machine material cylinder 8, an inflow pipeline 9, a horizontal mixing pipeline 10, an outflow pipeline 11, a downward mixing pipeline 12, a reverse horizontal mixing pipeline 13 and an upward mixing pipeline 14.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
As shown in fig. 1, the present embodiment provides an intelligent distribution system for dyeing machine auxiliaries, including: the system comprises an auxiliary agent storage tank 1, an auxiliary agent distribution valve 2, a flushing water valve 3, a distribution rotor pump 4, a pipeline reverse forced mixer, a flowmeter 6, a blanking valve 7, a dyeing machine material cylinder 8 and a control system;
the auxiliary agent storage tank 1 is connected with an auxiliary agent distribution valve 2, the auxiliary agent distribution valve 2 is connected with a distribution rotor pump 4, the distribution rotor pump 4 is connected with a flow meter 6, the flow meter 6 is connected with a blanking valve 7, the blanking valve 7 is arranged corresponding to a material cylinder 8 of a dyeing machine, a flushing water valve 3 is connected with the auxiliary agent distribution valve 2, and a reverse blending pump 5 is arranged in the pipeline reverse forced blending device;
the system comprises an auxiliary agent storage tank, a PLC (programmable logic controller) and a liquid level height value, wherein the auxiliary agent storage tank is used for storing auxiliary agents to be dispensed to the dyeing machine, the serial numbers of the auxiliary agent storage tanks correspond to the serial numbers of the auxiliary agents of the dyeing machine one by one, an ultrasonic liquid level meter is adopted for liquid level detection of the auxiliary agent storage tank, and the PLC reads the value of the ultrasonic liquid level meter through a modulus input port and converts the value into the liquid level height value; and the system automatically sends out the prompt of supplementing the auxiliary agent according to the height of the residual liquid level. When the auxiliary agent is supplemented into the auxiliary agent storage tank, the detection value and the upper limit setting value of the ultrasonic liquid level meter are first-layer protection for preventing the auxiliary agent from overflowing; and a high-level float switch is arranged at the top of the auxiliary agent storage tank and used as a second layer of protection of the system, and when the liquid level reaches the high-level float switch, the system automatically stops the operation of the feeding pump, so that safety accidents are prevented.
In the embodiment, the auxiliary agent distribution valve uses a pneumatic actuator to control the action of the valve and is used for controlling the opening and closing of each auxiliary agent passage in the pipeline; the flushing water valve is used for opening water to flush the auxiliary agent remained in the pipeline into the material cylinder of the dyeing machine after the auxiliary agent distribution valve is closed; the distribution rotor pump is used for pushing the auxiliary agent or the water forwards in the pipeline;
in this embodiment, the reverse mixing pump is used for uniformly mixing the auxiliary agent and water in the pipeline; the flow meter is used for metering the amount of the dispensed assistant agent and the amount of the washing water, the output of the flow meter adopts a pulse sending mode, each pulse represents each unit volume of solution flowing through the flow meter, and the PLC reads the pulse sent by the flow meter through the high-speed pulse input port and converts the pulse into the actual dispensing amount; the blanking valve is used for controlling the auxiliary agent and water to fall into which dyeing machine material cylinder; the dyeing machine material jar is the target jar of auxiliary agent delivery, and the system needs to deliver the auxiliary agent ration to the jar in, installs balanced cover formula level gauge in the dyeing machine material jar for liquid level display and liquid level upper limit control when the auxiliary agent is delivered, in case surpass the upper limit value, the system can automatic pause the operation and send audible and visual alarm information, suggestion personnel handle the trouble. In this embodiment, the auxiliary dispensing valve and the flushing water valve are three-way valves.
Under the action of the rotating thrust of the rotor pump, the auxiliary agent in the auxiliary agent storage tank flows to a corresponding material cylinder of the dyeing machine through a pipeline, and the flowing process is as follows: auxiliary agent storage tank → pipeline → auxiliary agent dispensing valve → pipeline → rotor pump → pipeline → reverse mixing pump → pipeline → flowmeter → pipeline → blanking valve → pipeline → dyeing machine material jar;
as shown in fig. 2, the distributing rotor pump 4 feeds the inhomogeneous mixed solution from the inflow pipe 9, the back mixing pump 5 is turned on, and due to the effect of the back mixing pump, a part of the solution flowing through the horizontal mixing pipe 10 flows to the outflow pipe 11, and another part of the solution flows to the down mixing pipe 12, passes through the back horizontal mixing pipe 13 and the up mixing pipe 14, and then returns to the horizontal mixing pipe 10 to be mixed with the solution coming from the inflow pipe 9 at the next moment, so that the front and back mixing of the solution in the pipes is realized. In the process of switching the auxiliary agents, the frequency of the distribution rotor pump is reduced, the flow flowing out of the pipeline of the reverse forced mixer is far smaller than the flow circulating in the pipeline of the reverse forced mixer, the proportion transition of the two solutions is smoother when the distribution is switched, and a good proportion gradual-change mixing effect can be achieved. The mixed solution flowing out of the reverse forced mixer through the pipeline flows through the flowmeter, so that extra metering fluctuation of the flowmeter is avoided, a more accurate metering value can be obtained, and the problem of large metering error of the flowmeter in the process of converting the auxiliary agents during time-sharing conveying of various auxiliary agents is solved.
The difference between the reverse forced mixer for the pipeline and the traditional pipeline mixer is as follows: 1. the traditional pipeline mixer is unpowered, and achieves the mixing effect by the flow of fluid; the pipeline reverse forced mixer is powered and forces the solution to be circularly mixed under the action of a pump; 2. the traditional pipeline mixer axially and rotatably mixes the solution in the pipeline, and the front and back mixing effects are not obvious; the pipeline reverse forced mixer mixes the solution in the pipeline in the front section and the rear section, and the solution proportion transition during mixing can be smoother and more uniform.
As shown in fig. 3, the control system comprises an upper computer intelligent PC system, a PLC control system, a relay, an electromagnetic valve, a compressed air barometer, a frequency converter, and a field signal collector;
the PLC control system outputs DC24V to control a relay (for protecting a PLC output point), the relay outputs DC24V to control an electromagnetic valve, the electromagnetic valve converts an electric signal into compressed air, the compressed air controls the action of a pneumatic actuator of a distribution valve, and the pneumatic actuator of the valve controls the opening and closing of the valve in a mechanical connection mode.
The relay is connected with the electromagnetic valve and used for amplifying a signal of the PLC to control the action of the electromagnetic valve and protect an output point of the PLC, the PLC outputs a DC24V signal to control a coil of the relay, a DC24V of a power supply is electrically connected to a contact of the relay, and the power supply is led out from the other end of the contact to obtain an amplified control current signal. Each flushing water valve, each distribution valve and each blanking valve are connected through an electromagnetic valve, the electromagnetic valves convert electrical signals into compressed gas signals to control the action of the fluid valves, the PLC outputs 24V direct current to drive coils of the electromagnetic valves, magnetic fields generated by the coils drive valve cores to act, and the valve cores act to enable the output of the compressed gas to be reversed to drive the fluid valves (the auxiliary agent distribution valves, the blanking valves and the like) to be switched.
The frequency converter is used for changing the frequency of three-phase alternating current and driving the rotation of the motor of the distribution rotor pump, and the main purpose of using the frequency converter is to control the rotating speed of the motor of the rotor pump by controlling the output frequency of the frequency converter, so as to control the rotating speed of the rotor pump, finally achieve the purposes of controlling the distribution speed of the auxiliary agent and improving the distribution precision. During the dispensing of the auxiliary agent, the frequency converter 45HZ is operated, and when the actual dispensing amount is smaller than the set dispensing amount (for example 3000 ml), the frequency converter is switched to low frequency operation (for example, the low frequency set value is 10 HZ), and when the auxiliary agent dispensing valve is closed, a large amount does not flow so as to cause an excessive error. The frequency converter of the rotor pump motor adopts a Mitsubishi frequency converter, carries out data communication with a PLC through a Modbus RTU, reads a state signal of the frequency converter, and writes the operation frequency of the frequency converter; in order to ensure the safety of the system, key signals of starting, stopping, alarming and the like of the frequency converter are butted with the PLC in a hardware wiring mode.
The compressed air barometer is arranged on a total air inlet triplet, the pressure reducing device is also a part of the triplet, compressed air passes through the triplet and then flows to each electromagnetic valve, the compressed air barometer is an electric signal device for converting the pressure of the compressed air after pressure reducing into 4-20mA analog current for output, and the output is connected to an FX5U-4AD analog input module of the PLC, so that the system can acquire the real-time pressure value of the compressed air. The closing time of the valve is influenced by the pressure value of the compressed air, and the higher the pressure value is, the faster the valve is closed, and the smaller the amount of the auxiliary agent flowing through the valve is when the valve is closed. In this embodiment, a specific usage of SMC barometer with analog output is current type, and the output current is 4-20 ma.
The PLC control system collects state information of various sensors, such as real-time flow of a flowmeter, running frequency of a frequency converter, a compressed air pressure value, switching states of various valves, liquid level values of an auxiliary agent storage tank, high and low liquid level protection switching states of the auxiliary agent storage tank, manual/automatic switching states, switching states of an emergency stop device, alarm reset switching states and the like, and outputs control information of various actuating mechanisms, such as electrification of the flowmeter, starting and frequency adjustment of the frequency converter, switching of the valves, switching of alarm color lamps, switching of a buzzer and the like, through the state information and order information sent by an upper computer.
The on-site signal collector is responsible for collecting on-off state signals of a valve, liquid level values of the auxiliary agent storage tank, high-low liquid level protection on-off state signals of the auxiliary agent storage tank and the like.
The upper computer intelligent PC system is in butt joint with a factory ERP and a central control computer of a dyeing machine, an auxiliary agent distribution order is obtained from the ERP, when the upper computer intelligent PC system receives the auxiliary agent distribution order, information in the auxiliary agent distribution order comprises an auxiliary agent number, a dosage and a dyeing machine cylinder number, the upper computer intelligent PC system sends the data to the PLC control system in an OPC communication mode, after the PLC control system receives the data, the corresponding auxiliary agent distribution valve is opened according to the auxiliary agent number, the corresponding blanking valve is opened according to the dyeing machine cylinder number, and meanwhile, a frequency converter is started to control the rotation of a distribution rotor pump motor and drive the distribution rotor pump to operate;
in the auxiliary agent distribution process, the frequency of a frequency converter of a distribution rotor pump is gradually increased until a set value is reached, after stable feeding is carried out for a period of time, when the actual distribution quantity is close to the target distribution quantity (the difference value between the actual distribution quantity and the target distribution quantity can be set), the frequency of the frequency converter is reduced, and after the actual distribution quantity is stable for a short period of time, a PLC control system sends the collected frequency of the frequency converter and the collected air pressure value of compressed air to an intelligent PC system of an upper computer;
the intelligent PC system of the upper computer substitutes the serial number of the auxiliary agent, the serial number of the blanking valve, the air pressure value of the compressed air, the frequency of the frequency converter and the distribution amount of the order into the pre-stop value neural network model and the predicted time-use neural network model, and calculates to obtain the recommended pre-stop value
When the time is expected, the time is sent to a PLC control system;
the PLC control system pre-stops the experience value
And recommending a pre-stop value
And calculating a final pre-stop value according to variable proportion superposition, wherein the calculation formula of the variable proportion control strategy is as follows:
In the initial stage of project implementation, K takes a very small value, and the variable ratio control strategy mainly plays a role in a control mode based on empirical value setting, so that the system cannot cause large distribution errors due to poor network parameters, and catastrophic events are avoided; with the gradual increase of the data quantity collected by operation, the training times of the neural network are increased, the network model parameters are closer to the optimal state, the K value is gradually improved, the K value can be divided into equal parts according to the set training iteration times, then the adjustment quantity of the equal parts is increased once the training is finished, the influence of the upper computer intelligent system on the comprehensive decision is increased, when the network is trained to the optimal state, K is equal to 1, and the variable ratio control strategy is completely determined by the upper computer intelligent PC system.
And subtracting the metering value of the flow meter from the ordered delivery amount in each scanning period to obtain the residual delivery amount, and when the residual delivery amount is less than or equal to the final pre-stop value, immediately sending an auxiliary agent delivery valve closing signal by the PLC control system, closing the auxiliary agent delivery valve, and simultaneously closing the blanking valve and the rotor pump motor.
The greatest benefit of the variable ratio control of the final pre-stop value in this embodiment is that the fastest introduction of the newly added auxiliary agent types and the newly added dye vat can be realized. After a factory uses the existing auxiliary agents and dye vats for production for a period of time, sometimes the types of the auxiliary agents need to be increased because the requirements for dyeing cloth in a new order are different; sometimes, because the order quantity is increased, a dye vat needs to be added, an auxiliary blanking port needs to be added when the dye vat is added, if a variable proportion control strategy is not available, an original trained neural network model cannot be applied when the auxiliary is added or the dye vat is added every time, a large amount of distribution data needs to be collected again from the beginning to perform network model training, and a large amount of time is delayed. After the variable proportion control strategy is adopted, the whole network model does not need to be retrained, a small value can be taken for the K value of an order with new additives or new blanking ports in distribution, and the K value is correspondingly increased along with the gradual increase of the training times of the new additives and the new blanking ports. When it is observed that the recommended pre-stop value is continuously better than the pre-stop value, K is again taken to be 1. The variable proportion control strategy can complete the introduction of new auxiliary agents or new dye vats under the condition of not influencing the overall control strategy, and can well meet the requirements of factory production expansion and variety expansion.
In this embodiment, in order to ensure that the system does not generate a transition risk when the intelligent distribution mode is started, after the neural network model is trained for the first time, the upper computer intelligent PC system does not send the recommended pre-stop value to the PLC when the order is distributed, the upper computer intelligent PC system records data such as the pre-stop value, the recommended pre-stop value, the order distribution quantity, the actual distribution quantity, the auxiliary agent number, the blanking port number, and the like, and the personnel rechecks whether the recommended pre-stop value in the multiple distribution orders is more reasonable than the actual pre-stop value. If a certain number of delivery orders are observed, the recommended pre-stop value is more reasonable than the actual pre-stop value, and the intelligent system is formally used online.
In this embodiment, the PLC control system multiplies the predicted time by a set ratio, for example, by 120%, to obtain the maximum time limit allowed for the current auxiliary agent delivery, subtracts the maximum time limit from the actual time in each scanning period, and sends an audible and visual alarm and displays an alarm message to prompt a person to process if the obtained data is greater than or equal to 0.
The intelligent PC system of the upper computer trains the neural network model according to the collected historical distribution data set at regular intervals of acquisition times, and automatically corrects the network model.
After model training is completed, the upper computer intelligent PC system sends related distribution parameters into a trained neural network respectively during each distribution, calculates a recommended pre-stop value and a predicted use time, and sends the recommended pre-stop value and the predicted use time to the PLC control system, the pre-stop value is intelligently adjusted in a deep learning mode, the distribution precision of the auxiliary agent is improved, waste is reduced, the environment is better protected, and meanwhile, the dyeing quality stability is improved.
As shown in fig. 4, the pre-stop value neural network model and the predicted time-use neural network model both adopt a fully-connected neural network with a five-layer structure, the first layer is an input layer, the second layer is a first hidden layer, the third layer is a second hidden layer, the fourth layer is a third hidden layer, and the fifth layer is an output layer; the number of the neurons of the first hidden layer is 200, the number of the neurons of the second hidden layer is 200, and the number of the neurons of the third hidden layer is 100; the Relu function is used for the activation function of the 3 hidden layers. In training the network, the Adam algorithm is used as a gradient descent loss function optimizer.
The calculation formula of each layer of the fully-connected neural network model is designed as follows:
whereiniThe number of the neurons in this layer,jto the output of the first layer above the current layer,mthe number of neurons in the layer above the current layer,nthe second few nerve layers are represented,yin order to be output, the output is,xin order to be an input, the user can select,θis the weight value of the current layer,bis the offset value of the present layer,fis the activation function of this layer.
The output of each layer being the input of a lower layer, i.e.
。
After the model training is finished, when the model is used for predicting the recommended pre-stop value, the output of the last layer is the recommended pre-stop value,namely:
= y；
the cost function formula of the pre-stopping value full-connection neural network model is as follows:
where C is the cost, L represents the total number of samples, S represents the number of samples,
a target delivery amount for the order is indicated,
the actual amount of delivery is indicated,zrepresenting the actual pre-stop value.
The cost function formula of the full-connection neural network model in expected time is designed as follows:
wherein the content of the first and second substances,
in order to start the delivery of the auxiliary agent,
and the end time of the additive distribution.
Gradually solving partial derivative of each layer by a cost function in a gradient descent and back propagation mode to obtain each weight
Value and offset of
The fully-connected neural network model is updated.
Wherein θ is
X is
T denotes a transpose of the matrix.
Building the multilayer fully-connected neural network in a deep learning platform TensorFow, inputting 5000 recorded distribution data, training, and calculating the distribution data of each layer by the software by using a gradient reverse transfer rule
And
these are
、
And the network structure forms a trained network model for calculating the recommended pre-stop value.
The model structure and gradient back propagation of the neural network are similar to those of the recommended pre-stop value when in prediction, the main difference is that the output of the model is different and a variable proportion control strategy is not adopted, the structure and the principle are basically the same, and the description is not repeated here.
In the embodiment, the influence of factors such as the viscosity and the fluidity of different auxiliaries, the lengths of distribution pipelines of different blanking ports, the operation frequency of a distribution pump, the pressure intensity of compressed air, the distribution amount and the like on the amount of the auxiliaries flowing through at the closing moment of the valve is comprehensively reflected through a neural network model; by adopting a multilayer full-connection neural network, the distribution error characteristics of each auxiliary agent and each blanking port (one blanking port corresponds to one dye vat) and the distribution error characteristics of each auxiliary agent, the frequency of a frequency converter (the frequency converter of a distribution pump) and the air pressure value of compressed air are automatically learned through training;
the characteristics can be automatically contained in the weight of the neural network model, and the numerical values such as the additive number, the blanking port number, the frequency of the frequency converter, the air pressure value of compressed air and the like are substituted into the neural network model for calculation during distribution to obtain a recommended pre-stop value, so that debugging is not needed in each combination, and the PLC controls the closing time of the distribution valve according to the recommended pre-stop value, thereby achieving the effect of improving the distribution precision.
Example 2
As shown in fig. 5, the present embodiment provides a method for controlling an intelligent distribution system of an auxiliary agent for a dyeing machine, including the following steps:
step 1: the upper computer intelligent PC system reads the name and dosage of the auxiliary agent and the cylinder number of the dyeing machine from the ERP system, controls and reads the control instruction information of the auxiliary agent distribution order from the dyeing machine, and sends the auxiliary agent distribution order to the PLC control system after comprehensive processing;
step 2: the PLC control system receives the auxiliary agent distribution order, opens the corresponding auxiliary agent distribution valve according to the auxiliary agent number, opens the corresponding blanking valve according to the dyeing machine cylinder number, simultaneously starts the frequency converter, controls the rotation of the rotor pump motor, drives the distribution rotor pump to operate, and sets the state of the distribution completion register to be OFF;
step 3: under the action of rotation of the rotor pump, the auxiliary agent in the auxiliary agent storage tank flows to a corresponding material cylinder of the dyeing machine through a pipeline, and the flowing process is as follows: auxiliary agent storage tank → pipeline → auxiliary agent dispensing valve → pipeline → rotor pump → pipeline → reverse mixing pump → pipeline → flowmeter → pipeline → blanking valve → pipeline → dyeing machine material jar;
step 4: in the process of distributing the auxiliary agent, the frequency of a frequency converter of the rotor pump reaches a set value, and after the frequency of the frequency converter of the rotor pump is stabilized for a short period of time, the PLC control system sends the collected frequency of the frequency converter and the collected air pressure value of the compressed air to an intelligent PC system of an upper computer;
step 5: the intelligent PC system of the upper computer substitutes the serial number of the auxiliary agent, the serial number of a blanking port (blanking valve), the air pressure value of compressed air, the frequency of a frequency converter and the ordered target delivery quantity into the trained fully-connected neural network prediction model, calculates to obtain a recommended pre-stop value and writes the recommended pre-stop value into a register of the PLC control system;
step 6: the PLC control system calculates a final pre-stop value according to a certain proportion of a preset experience pre-stop value and a preset recommendation pre-stop value, subtracts a metering value of a flow meter from an ordered delivery quantity to obtain a residual delivery quantity in each scanning period, and immediately sends an auxiliary agent delivery valve closing signal to close an auxiliary agent delivery valve and close a blanking valve and a rotor pump motor when the residual delivery quantity is less than or equal to the final pre-stop value;
step 7: the PLC control system opens a flushing water valve, opens the blanking valve again, starts a frequency converter at the same time, controls the rotation of a rotor pump motor, drives a distribution rotor pump to operate, and pushes all auxiliary agents remained in a pipeline into a dye vat of the dyeing machine by water;
step 8: the PLC control system stores the final actual delivery amount of the auxiliary agent into a register and sends a delivery completion mark to the upper computer intelligent PC system;
step 9: and the intelligent PC system of the upper computer reads the actual distribution amount in the PLC register, and adds distribution data such as an auxiliary agent number, a blanking valve number, a compressed air pressure value, a frequency converter frequency, an ordered target distribution amount, a pre-stop value and the like to form an array to be stored in an MS-SQL database.
The training step and the using step of the neural network model in the expected use are similar to those of the pre-stop value, the main difference is that the output of the model is different and a variable proportion control strategy is not adopted, the PLC control system monitors whether the auxiliary agent distribution process is overtime after obtaining the expected use, most processes in the middle of the steps are the same, and the description is not repeated.
The training method of the fully-connected neural network prediction model for recommending the pre-stop value comprises the following steps:
step 1: the upper computer intelligent PC system reads the name and dosage of the auxiliary agent and the cylinder number of the dyeing machine from the ERP system, controls and reads the control instruction information of the auxiliary agent distribution order from the dyeing machine, and sends the auxiliary agent distribution order to the PLC control system after comprehensive processing;
step 2: the PLC control system starts a rotor pump and a corresponding valve according to the received auxiliary agent distribution order, starts the distribution of the auxiliary agent and sets the state of a distribution completion register to be OFF;
step 3: the auxiliary agent delivery volume reaches the value of stopping in advance, and PLC control system closes delivery valve, blanking valve and rotor pump motor to collect the actual delivery data of auxiliary agent delivery, include: the PLC control system stores the data into a register of the PLC and sets the state of a distribution completion register to be ON;
step 4: the intelligent PC system of the upper computer reads the value of a register in the PLC in an OPC communication mode, and when the condition of a distribution completion register is detected to be changed from OFF to ON, the serial number of an auxiliary agent, the serial number of a blanking port, the air pressure value of compressed air, the frequency of a frequency converter, the ordered target distribution quantity, the final pre-stop value and the actual distribution quantity in the auxiliary agent distribution register in the PLC control system are stored in an MS-SQL database of the intelligent PC system of the upper computer;
step 5: continuously distributing the auxiliary agents in the production of the system, and continuously repeating the processes from Step1 to Step4 until the distribution records of the MS-SQL database reach the specified quantity;
step 6: taking the assistant number, the blanking valve number, the compressed air pressure value, the frequency of the frequency converter and the ordered target distribution amount recorded in the database as the input of the multilayer fully-connected neural networkx；
Step 7: the actual delivery amount, the order target delivery amount and the actual pre-stop value are used as the output of the multi-layer fully-connected neural network, namely, the deviation value of the delivery history data used for trainingy；
Step 8: by inputtingxOutput ofyUsing Adam optimalIterating the weights of the first hidden layer of the multi-layer fully-connected neural network
And bias
The weight of the second hidden layer
And bias
The weight of the third hidden layer
And bias
Obtaining a fully-connected neural network prediction model of the recommended pre-stop value;
the method adopts an automatic control technology and a deep learning technology to control the strategy method according to the variable proportion, realizes the innovation of the auxiliary agent distribution technology of the dyeing machine, and ensures the basic operation of the auxiliary agent distribution system by using a control mode mainly based on the traditional automatic control technology in the early stage; the intelligent control and intelligent process monitoring of the distribution of the auxiliary agent are realized by using a control mode mainly based on deep learning in the later period; the introduction of new auxiliary agents or new dye vats is completed under the condition of not influencing the overall control strategy, and the requirements of factory production expansion and variety expansion can be well met; the auxiliary agent distribution process of each order is monitored overtime by adopting a deep learning technology, system faults are found in time, and waiting time when the faults occur is shortened; the most appropriate recommended pre-stop value is predicted by using a deep learning technology, the distribution precision of the auxiliary agent can be improved, the effects of saving the auxiliary agent, protecting the environment and improving the product quality are achieved, and the method has better social value and economic value.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.
Claims (8)
1. The utility model provides a dyeing machine auxiliary agent intelligence delivery system which characterized in that includes: the system comprises an auxiliary agent storage tank, an auxiliary agent distribution valve, a flushing water valve, a distribution rotor pump, a pipeline reverse forced mixer, a flow meter, a blanking valve, a dyeing machine material cylinder and a control system, wherein the pipeline reverse forced mixer is provided with a reverse mixing pump, and the reverse mixing pump is used for mixing the auxiliary agent and water in the pipeline uniformly in a circulating manner;
the auxiliary agent storage tank is connected with an auxiliary agent distribution valve, the auxiliary agent distribution valve is connected with a distribution rotor pump, the distribution rotor pump is connected with a flowmeter, a connecting pipeline between the distribution rotor pump and the flowmeter is provided with a reverse circulation branch, the reverse mixing pump is arranged in the reverse circulation branch, the input end of the reverse mixing pump is connected with the input end of the flowmeter, the output end of the reverse mixing pump is connected with the output end of the distribution rotor pump, the flowmeter is connected with a blanking valve, the blanking valve is arranged corresponding to a material cylinder of the dyeing machine, and the flushing water valve is connected with the auxiliary agent distribution valve;
the control system comprises an upper computer intelligent PC system, a PLC control system, a compressed air barometer, a frequency converter and a field signal collector;
the compressed air barometer is used for acquiring a valve core compressed air pressure value, the frequency converter is used for driving the distribution rotor pump, and the field signal collector is used for acquiring a valve on-off state signal and a liquid level state of the auxiliary agent storage tank;
the upper computer intelligent PC system is provided with an auxiliary agent distribution order acquisition module and a pre-stop value neural network model construction module;
the auxiliary agent distribution order acquisition module is used for acquiring the name and the dosage of the auxiliary agent and the cylinder number of the dyeing machine to form an auxiliary agent distribution order;
the pre-stop value neural network model building module is used for building a pre-stop value neural network model, the pre-stop value neural network model obtains a recommended pre-stop value according to current auxiliary agent distribution data, the recommended pre-stop value and an empirical pre-stop value are superposed according to a variable proportion to obtain a final pre-stop value, the variable proportion is increased progressively along with the increase of training iteration times, and the auxiliary agent distribution data comprises an auxiliary agent number, a blanking valve number, a compressed air pressure value, a frequency converter frequency and an ordered target distribution quantity;
and the final pre-stop value is obtained by superposing the empirical pre-stop value according to a variable proportion, and is specifically represented as follows:
wherein the content of the first and second substances,
a recommended pre-stop value is indicated,
expressing a preset experience pre-stop value, K expressing a variable proportion value, dividing the variable proportion value into equal parts according to the set training iteration times, increasing the equal parts after each training iteration, and enabling K to be more than or equal to 0 and less than or equal to 1;
the PLC control system is used for collecting the real-time flow of the flow meter, the running frequency of the frequency converter, the air pressure value of compressed air, the switching state of each valve and the liquid level state of the auxiliary agent storage tank, judging the residual delivery amount according to the final pre-stop value, controlling the opening and closing of the auxiliary agent delivery valve, the blanking valve, the delivery rotor pump and the flushing water valve according to a variable proportion control strategy, and adjusting the output frequency of the frequency converter.
2. The dyeing machine auxiliary intelligent distribution system according to claim 1, characterized in that an ultrasonic level meter is arranged in the auxiliary storage tank, a high float switch is arranged on the top of the auxiliary storage tank, the ultrasonic level meter is used for detecting the liquid level height value of the auxiliary storage tank, and the high float switch is used for outputting a switching value for stopping the auxiliary supplement when the liquid level of the auxiliary storage tank exceeds an upper limit value.
3. The dyeing machine auxiliary intelligent distribution system of claim 1, wherein a balance cover type liquid level meter is arranged in the dyeing machine material cylinder and used for detecting the liquid level and the upper liquid level limit in the dyeing machine material cylinder.
4. The dyeing machine auxiliary intelligent distribution system of claim 1, wherein the upper computer intelligent PC system is further provided with an estimated time neural network model building module;
the prediction time neural network model building module is used for building a prediction time neural network model, the prediction time neural network model is used for calculating the prediction time, the PLC control system multiplies the prediction time by a set proportion to obtain the longest time limit allowed by the delivery of the auxiliary agent, and if the actual time of each scanning period is greater than the longest time limit, an alarm control signal is output.
5. A control method of an intelligent dosing system of an auxiliary agent for a dyeing machine according to any one of claims 1 to 4, characterized by comprising the following steps:
the upper computer intelligent PC system obtains the name and the dosage of the auxiliary agent and the cylinder number of the dyeing machine, forms an auxiliary agent distribution order and sends the order to the PLC control system;
the PLC control system receives an auxiliary agent distribution order, controls a distribution rotor pump to rotate, controls the opening of a corresponding auxiliary agent distribution valve according to the auxiliary agent number, and controls the opening of a corresponding blanking valve according to the dyeing machine cylinder number;
the reverse forced mixer mixes the solution flowing into the reverse mixing pump with the solution flowing into the reverse mixing pump at the next moment and outputs the mixed solution to the flowmeter;
the intelligent PC system of the upper computer substitutes the serial number of the auxiliary agent, the serial number of the blanking valve, the air pressure value of compressed air, the frequency of the frequency converter and the target distribution quantity of the order into the trained neural network model of the pre-stop value, calculates to obtain a recommended pre-stop value and stores the recommended pre-stop value into a register of the PLC control system;
superposing a preset empirical pre-stop value and a preset recommended pre-stop value according to a variable proportion to calculate a final pre-stop value, subtracting a metering value of a flow meter from the delivery quantity of an order in each scanning period to obtain a residual delivery quantity, and closing an auxiliary agent delivery valve and simultaneously closing a blanking valve and a delivery rotor pump by a PLC (programmable logic controller) control system when the residual delivery quantity is less than or equal to the final pre-stop value;
the preset empirical pre-stop value and the preset recommended pre-stop value are superposed according to a variable proportion to calculate a final pre-stop value, which is specifically represented as:
wherein the content of the first and second substances,
a recommended pre-stop value is indicated,
expressing a preset experience pre-stop value, K expressing a variable proportion value, dividing the variable proportion value into equal parts according to the set training iteration times, increasing the equal parts after each training iteration, and enabling K to be more than or equal to 0 and less than or equal to 1;
the PLC control system starts a flushing water valve, a blanking valve and a distribution rotor pump, and conveys the auxiliary agent remained in the pipeline into a material cylinder of the dyeing machine;
the PLC control system stores the actual delivery amount of the auxiliary agent into a register and sends a delivery completion mark;
the intelligent PC system of the upper computer acquires actual distribution quantity, and the actual distribution quantity, the auxiliary agent number, the blanking valve number, the compressed air pressure value, the frequency of the frequency converter, the ordered target distribution quantity and the final pre-stop value are combined and stored in a database.
6. The control method of an intelligent dosing system of auxiliaries for dyeing machines according to claim 5, characterized in that said training step of said pre-stop value neural network model comprises:
taking the additive number, the blanking valve number, the compressed air pressure value, the frequency of a frequency converter and the ordered target distribution quantity in the database as the input of the pre-stop value neural network model;
the pre-stop value neural network model adopts a full-connection neural network model, and the structure of the full-connection neural network model sequentially comprises the following steps: the device comprises an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer;
the calculation formula of each layer of the fully-connected neural network model is as follows:
whereiniIs the first layer of the layeriThe number of the nerve cells is one,jis the first layer corresponding to the previous layer of the present layerjThe output of the first and second processors is,mthe number of neurons in the layer above the current layer,nthe second few nerve layers are represented,yin order to be output, the output is,xin order to be an input, the user can select,θis the weight value of the current layer,bis the offset value of the present layer,fis the activation function of the layer;
taking the deviation value of the distribution historical data used for training as the output of the pre-stop neural network model;
iteratively training weights of a first hidden layer of a prestack-valued neural network model using an Adam optimizer
And bias
The weight of the second hidden layer
And bias
The weight of the third hidden layer
And bias
And obtaining a pre-stop value neural network model.
7. The control method of an intelligent distribution system of auxiliaries for dyeing machines according to claim 6, characterized in that the cost function formula of said fully connected neural network model is:
where C is the cost, L represents the total number of samples, S represents the number of samples,
a target delivery amount for the order is indicated,
the actual amount of delivery is indicated,zrepresenting the actual pre-stop value.
8. A control method for an intelligent distribution system of dyeing machine auxiliaries according to claim 5, characterized in that the intelligent distribution system of dyeing machine auxiliaries is provided with a time-of-use prediction neural network model, the cost function formula of which is:
where C is the cost, L represents the total number of samples, S represents the number of samples,
in order to start the delivery of the auxiliary agent,
and the end time of the additive distribution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111001290.8A CN113448250B (en) | 2021-08-30 | 2021-08-30 | Dyeing machine auxiliary agent intelligent distribution system and control method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111001290.8A CN113448250B (en) | 2021-08-30 | 2021-08-30 | Dyeing machine auxiliary agent intelligent distribution system and control method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113448250A CN113448250A (en) | 2021-09-28 |
| CN113448250B true CN113448250B (en) | 2021-12-21 |
Family
ID=77818979
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111001290.8A Active CN113448250B (en) | 2021-08-30 | 2021-08-30 | Dyeing machine auxiliary agent intelligent distribution system and control method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113448250B (en) |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102409497B (en) * | 2011-09-01 | 2013-10-16 | 浙江理工大学 | Dyeing assistant automatic weighting delivery system and method |
| CN110592855A (en) * | 2019-08-29 | 2019-12-20 | 颜英说 | Cloth dyeing device and cloth dyeing control method thereof |
| US11579575B2 (en) * | 2019-12-03 | 2023-02-14 | Baidu Usa Llc | Inverse reinforcement learning with model predictive control |
| CN112131785B (en) * | 2020-09-09 | 2022-04-12 | 太原理工大学 | Screening and optimizing method of methanation nickel-based catalyst formula based on ANN-NSGA-II |
-
2021
- 2021-08-30 CN CN202111001290.8A patent/CN113448250B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113448250A (en) | 2021-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6531140B2 (en) | Water supply control device for backwashing plated products intermittently and method of using the same | |
| CN102409497B (en) | Dyeing assistant automatic weighting delivery system and method | |
| CN107541463B (en) | A kind of full-automatic on-line early warning diagnosis and treatment apparatus of anaerobic fermentation | |
| WO2021212585A1 (en) | Sludge settling velocity automatic measurement system | |
| CN202151561U (en) | Proportioning detecting and controlling device for double-component glue | |
| US8402998B2 (en) | System and method for delivering chemicals | |
| CN106255570A (en) | Water-soluble metalworking liquid automatically feeds control system | |
| CN109647812B (en) | Equipment for cleaning pipeline in aerospace field | |
| CN102498326A (en) | Method for automatically controlling a tube network | |
| CN100546706C (en) | Automatic control method for hydrochloric acid solution dilution technique | |
| CN113448250B (en) | Dyeing machine auxiliary agent intelligent distribution system and control method thereof | |
| CN100429598C (en) | Method for inspecting and controlling dyeing liquid component of dyeing machine on-line | |
| CN102392339B (en) | System and method for automatically delivering assistants for dye vat by gravimetric method | |
| CN202220275U (en) | Automatic weighing and delivering system for dyeing auxiliaries | |
| CN100429597C (en) | System for inspecting and controlling dyeing liquid component of dyeing machine on-line | |
| CN202220274U (en) | Automatic delivery system for weighting-method dye vat auxiliaries | |
| CN111980026B (en) | Automatic slurry feeding and mixing system and process based on PLC | |
| Sener et al. | PIc-Based Automation System Integration in Textile Dye Dosing Machine and Environmental System Design | |
| CN209471114U (en) | A kind of water pollution index's detection device being able to achieve auto-manual measuring and calculating | |
| CN112811542A (en) | Automatic control device for chemical adding in water plant | |
| CN107158998B (en) | A kind of mixed glue special plane of intelligence | |
| CN220968985U (en) | Intelligent salt solution preparation and replenishment system | |
| CN206383292U (en) | A kind of cement counterweight materials filling measurement device | |
| CN112439736A (en) | Flushing system and flushing method | |
| CN213388921U (en) | Sodium hypochlorite generator and automatic chlorination 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 |