Intelligent pressure regulating control system for pumping high-concentration ultra-fine tailings filling slurry

Abstract

The application provides a high-concentration ultra-fine tailings filling slurry pumping intelligent pressure regulating control system and relates to the technical field of filling slurry control. The system comprises a sensor module, a data acquisition console, an industrial computer and a digital-to-analog converter. The data acquisition console controls and receives the pressure, temperature and flow rate collected by the sensor module in the filling pipeline and the pressure difference at the inflection point and the middle section of the pipeline. The data acquisition console collects the data collected by the sensor module into the industrial computer. The industrial computer learns the relationship between the internal pressure, temperature, flow rate and filling pump rotating speed of the pipeline by using a deep learning algorithm, automatically selects a suitable filling pump output power, determines a control signal and sends the control signal to the digital-to-analog converter, which controls the filling pump. The system realizes the pressure regulation of the whole pipeline by controlling the filling pump, greatly reduces the filling cost and can control the possible pipe explosion and blockage in the filling pipeline transportation process in advance.

Description

Technical Field

[0001] This invention relates to the field of filling slurry control technology, and in particular to an intelligent pressure regulation control system for pumping high-concentration ultrafine tailings filling slurry. Background Technology

[0002] In the mining of metal mines, the high volume of underground goafs after mining makes them highly susceptible to safety problems such as roof falls, water inrushes, collapses, surface collapses, and underground tunnel collapses. To maximize the recovery of mineral resources, protect the underground and surface environment, and fundamentally solve the safety problems of underground mining operations, especially in recent years, with the continuous advancement of backfilling materials, backfilling processes, pipeline transportation equipment, and technology, backfilling mining methods have been widely used in non-ferrous metal mines and precious metal mines. Due to its irreplaceable advantages, the backfilling method is also increasingly being used in mines such as coal mines and iron mines, which are traditionally unsuitable for backfilling.

[0003] With the development of materials science, the diverse types and varieties of cementitious materials, coupled with the diversification of admixtures, have led to increasingly complex and diverse compositions of colloids, slurries, or pastes. Furthermore, the development of slurry pumping technology has made cemented backfilling the mainstream backfilling technology in mines. Paste backfilling technology involves deeply thickening tailings slurry into a non-stratified, non-segregated, and non-dehydrated paste, which is then filled into the underground goaf. This reduces the ash-sand ratio in backfilling, lowers material consumption, maximizes the utilization of tailings resources, and achieves "one waste, two hazards" treatment, reducing environmental pollution. It boasts significant advantages in safety, environmental protection, economy, and efficiency. Although cemented backfilling offers significant advantages in terms of materials, some problems still exist during slurry transportation. During backfilling mining, the negative pressure generated when the backfill slurry enters the downward pumping section causes accelerated segregation, leading to sedimentation and blockage of the pipeline at the bottom. Excessive acceleration and high concentration flow rates during slurry transportation can also easily cause blockages within the pipeline, especially at inflection points. Even in areas with severe blockages or excessive pressure, pipe bursts can occur, damaging the entire filling system and causing safety hazards and economic losses to mining operations.

[0004] Most mine backfilling control systems control the strength and flowability of the backfill material by regulating the material ratio and concentration of the slurry, thereby controlling the entire backfilling system. However, this method requires a large amount of additional equipment, such as deep cone thickeners and flocculant addition control devices. Furthermore, insufficient control over pipeline pressure during transport can lead to pipeline damage and even pipeline bursts. Currently, the main backfilling methods on the market include gravity backfilling, hydraulic backfilling, and pumped backfilling. Due to the long backfilling pipelines and large backfilling ratios, most mines use pumped backfilling. However, when using high-concentration ultrafine tailings slurry for long-distance transport, without real-time monitoring and pump pressure adjustment, pipe blockage and bursts are highly likely. Pumped transport is difficult, and downward pumping has always been a technical challenge for backfill slurry transport, with few scholars researching this topic. There is an urgent need to research and develop an automatic control system to control and monitor the internal pressure of pipelines to prevent pipe bursts and achieve safe and efficient filling. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an intelligent pressure regulation control system for pumping high-concentration ultrafine tailings backfill slurry, which addresses the shortcomings of the prior art and realizes intelligent pressure regulation control for pumping high-concentration ultrafine tailings backfill slurry.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: an intelligent pressure regulation control system for high-concentration ultrafine tailings backfill slurry pumping, including a sensor module, a data acquisition console, an industrial computer, and a digital-to-analog converter (DAC). The data acquisition console controls the sensor module to collect pressure, temperature, flow rate, and pressure difference at pipe bends and the middle section of the pipe. The data acquisition console then collects the data collected by the sensor module into the industrial computer. The industrial computer uses a deep learning algorithm to learn the relationship between the internal pressure, temperature, flow rate, and backfill pump speed, and automatically selects an appropriate backfill pump output power to determine the control signal. The control signal is then sent to the DAC, which controls the backfill pump. Preferably, the sensor module includes a pressure sensor, a temperature sensor, and a flow sensor installed at the backfill pipe inlet, as well as pressure difference sensors installed at each bend and the middle section of the backfill pipe.

[0007] Preferably, the built-in software program of the industrial control computer is used to determine the control signal, display pressure and flow rate in real time, draw curves, animate the filling process, provide early warning of filling pipeline blockage, control the filling pump power, and start the filling pump with one key.

[0008] Preferably, the industrial control computer's built-in software program is built using the Python standard graphical user interface library.

[0009] Preferably, the main interface of the industrial control computer's built-in software program includes a logo display area, a filling process animation display area, real-time text, a function button bar, and a gauge control. The industrial control computer models the filling system using the modeling software Cinema 4D, and then uses keyframes and cloning methods to realize the running animation of the mixing tank and pipe arrows, adjusts the lighting and rendering materials, and obtains the filling process animation in a specified format. The animation is then converted into a GIF format image and played in a loop using Tkinter.

[0010] Preferably, the industrial control computer decomposes the GIF image, traverses each frame of the image and outputs it on the canvas, and updates it using the update method; at the same time, it segments the GIF image so that the text can be displayed normally.

[0011] To enable GIF playback in a loop, a GIF playback task is submitted to a thread pool, and the task uses the `after` method to call itself to achieve loop playback.

[0012] Preferably, the specific method by which the industrial control computer uses deep learning algorithms to learn the relationship between the internal pressure, temperature, flow rate, and filling pump speed of the pipeline, and automatically selects the appropriate output power of the filling pump, is as follows:

[0013] (1) The pressure, temperature and flow signals inside the filling pipe are acquired in real time through the sensor module. These signals represent the state (degree of blockage) inside the pipe to a certain extent.

[0014] (2) Define the degree of blockage and establish a mapping function between pump speed and degree of blockage;

[0015] The degree of congestion is defined as follows: the entropy values ​​of pressure, temperature, and flow signals within a time interval t, and the total flow within that time interval are used as the measure of the degree of congestion. That is, S(pressure, temperature, flow) = (alpha*H(pressure) + beta*H(temperature) + gamma*H(flow)) / sum(flow) = degree of congestion, where H is the information entropy. The larger the total flow, the lower the degree of congestion. alpha, beta, and gamma are the weights of each loss.

[0016] The mapping function between pump speed and degree of blockage is established as follows: V(pump speed) = [pressure, temperature, flow rate], S(pressure, temperature, flow rate) = degree of blockage, that is: S(V(pump speed)) = degree of blockage;

[0017] (3) Design a system for converting pump speed into pipeline pressure. The input of the system is pressure and the output is pump speed. The input state of the system changes with time and is a function of time, so the system is a dynamic system. The current output of the dynamic system depends not only on the current input, but also on the input and output of the system in the past.

[0018] (4) The pump speed-blocking degree mapping function is fitted based on the dynamic neural network NARX model to obtain the pump speed-blocking degree mapping model; the NARX neural network structure includes an input layer, a hidden layer and an output layer; the number of nodes in the input layer is set according to the number of input values, and the number of nodes in the output layer is set according to the number of predicted values;

[0019] (5) Construct a pumping simulation system and collect pump speed simulation data to conduct neural network training experiments to verify the feasibility of the method; the pump speed simulation data is a single-input multi-output data pair, including pump speed, pressure, temperature and flow rate, where pump speed is the independent variable and pressure, temperature and flow rate are the dependent variables.

[0020] (6) Train the neural network to achieve a mapping from pump speed to blockage level;

[0021] (7) For the pump speed-blockage degree mapping model obtained by training, use a genetic algorithm to search for a pump speed within a certain pump speed range that will result in a lower degree of pipe blockage.

[0022] The beneficial effects of adopting the above technical solution are as follows: The intelligent pressure regulation control system for high-concentration ultrafine tailings backfill slurry pumping provided by the present invention can regulate the pressure of the entire pipeline by controlling the backfill pump, which greatly reduces backfilling costs, reduces the labor costs of monitoring and regulating pipeline pressure during the backfilling process, reduces the probability of problems in the backfilling pipeline, and can prevent possible pipe bursts and blockages during the transportation of backfilling pipelines, and can provide alarms and controls to reduce the failure rate of the system itself. Attached Figure Description

[0023] Figure 1 This is a structural block diagram of the intelligent pressure regulating control system for pumping high-concentration ultrafine tailings backfill slurry provided in an embodiment of the present invention.

[0024] Figure 2 A schematic diagram of a pumping control physical model built in a laboratory, provided for an embodiment of the present invention;

[0025] Figure 3 This is a functional block diagram of the built-in software program of the industrial control computer provided in an embodiment of the present invention;

[0026] Figure 4 This is a schematic diagram of the main interface of the built-in software program of the industrial control computer provided in an embodiment of the present invention. Detailed Implementation

[0027] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0028] In this embodiment, a high-concentration ultrafine tailings backing slurry pumping intelligent pressure regulating control system is used, such as... Figure 1 As shown, the system includes a sensor module, a data acquisition console, an industrial computer, and a digital-to-analog converter (DAC). The data acquisition console controls the sensor module to collect data on pressure, temperature, flow rate, and differential pressure at pipe bends and mid-sections. The data acquisition console then collects the data from the sensor module into the industrial computer. The industrial computer uses a deep learning algorithm to learn the relationship between the internal pressure, temperature, flow rate, and filling pump speed of the pipe, automatically selects the appropriate filling pump output power, determines the control signal, and sends the control signal to the DAC, which then controls the filling pump.

[0029] In this embodiment, the sensor module includes a pressure sensor, a temperature sensor, and a flow sensor installed at the filling pipe inlet, as well as differential pressure sensors installed at each bend in the filling pipe and in the middle section of the pipe. The industrial control computer's built-in software program is used to determine the control signals and to display pressure and flow in real time, draw curves, animate the filling process, provide early warning of filling pipe blockage, control the filling pump power, and enable one-button start of the filling pump.

[0030] In this embodiment, the industrial control computer's built-in software program is built using the Python standard graphical user interface library. The main interface of the industrial control computer's built-in software program includes a logo display area, a filling process animation display area, real-time text, a function button bar, and a gauge control. The industrial control computer models the filling system using the modeling software Cinema 4D, and then uses keyframes and cloning methods to realize the animation of the mixing tank and pipe arrows running. It adjusts the lighting and rendering materials to obtain the filling process animation in a specified format. After converting the animation into a GIF format image, it uses Tkinter for loop playback.

[0031] Since Tkinter cannot directly play animations, they need to be converted into GIF format images. Tkinter's canvas control supports GIF format images, but not GIF playback; it can only display static images. The industrial control computer decomposes the GIF image, iterates through each frame of the image and outputs it to the canvas, updating it using the `update` method. Simultaneously, the GIF image is segmented to ensure text display correctly. To enable looping playback of the GIF, a GIF playback task is submitted to a thread pool, and the task uses the `after` method to call itself to achieve looping.

[0032] In this embodiment, the industrial control computer uses deep learning algorithms to learn the relationship between the internal pressure, temperature, flow rate, and filling pump speed of the pipeline, and automatically selects the appropriate filling pump output power. The specific method is as follows:

[0033] (1) The pressure, temperature and flow signals inside the filling pipe are acquired in real time through the sensor module. These signals represent the state (degree of blockage) inside the pipe to a certain extent.

[0034] (2) Define the degree of blockage and establish a mapping function between pump speed and degree of blockage;

[0035] Pumps deliver slurry into pipelines, but excessive pumping pressure can cause blockages. Since pumping pressure is directly linked to pump speed, and pump speed can be adjusted, the degree of blockage can be defined as follows: the entropy values ​​of pressure, temperature, and flow signals within a time step t, along with the total flow rate during that time step, are used as the measure of the degree of blockage. That is, S(pressure, temperature, flow rate) = (alpha*H(pressure) + beta*H(temperature) + gamma*H(flow rate)) / sum(flow rate) = degree of blockage, where H is the information entropy (the smaller the entropy value, the more stable the signal), the larger the total flow rate, the lower the degree of blockage, and alpha, beta, and gamma are the weights of each loss.

[0036] Establish a mapping function between pump speed and degree of blockage: V(pump speed) = [pressure, temperature, flow rate], S(pressure, temperature, flow rate) = degree of blockage, that is: S(V(pump speed)) = degree of blockage;

[0037] (3) Design a system for converting pump speed into pipeline pressure. The input of the system is pressure and the output is pump speed. Since the input state of the system changes with time and is a function of time, the system is a dynamic system. However, the current output (speed) of the dynamic system depends not only on the current input (pressure) but also on the input (pressure) and output (pump speed) of the system in the past.

[0038] (4) The pump speed-blocking degree mapping function is fitted based on the dynamic neural network NARX (nonlinear autoregressive network based on external input) model to obtain the pump speed-blocking degree mapping model; the NARX neural network structure includes an input layer, a hidden layer and an output layer; the number of nodes in the input layer is set according to the number of input values, and the number of nodes in the output layer is set according to the number of predicted values; the NARX neural network incorporates a delay and feedback mechanism, thus enhancing its ability to remember historical data;

[0039] (5) Since training a neural network requires a certain amount of data, a pumping simulation system (including pump, pipe and fluid, with fixed parameters for pipe and fluid, and only the pump speed is adjustable) can be constructed and pump speed simulation data can be collected to conduct neural network training experiments to verify the feasibility of the method; the pump speed simulation data is a single-input multi-output data pair, including pump speed, pressure, temperature and flow rate, where pump speed is the independent variable, and pressure, temperature and flow rate are the dependent variables;

[0040] (6) Train the neural network to achieve the mapping from pump speed (pump speed at time t, pump speed from t-1 to tn, and blockage level from tn to t-1) to blockage level (time t);

[0041] (7) For the pump speed-blockage degree mapping model obtained by training, use a genetic algorithm to search for a pump speed within a certain pump speed range that will result in a lower degree of pipe blockage.

[0042] In this embodiment, a complete pumping control physical model is built in the laboratory. Based on the overall system design, the selection and installation of a small filling pump, pressure sensor, flow sensor, temperature sensor, analog-to-digital converter, and digital-to-analog converter are performed. The selection and installation of the operating system for a small integrated industrial control computer are also performed. In this embodiment, the pumping control physical model mainly consists of three parts: a filling pipeline system, a circuit system, and a sensor system. The filling pipeline system mainly includes PVC pipes, PVC elbows, PVC tees, PVC stop valves, an acrylic mixing tank, high-concentration ultrafine tailings, and a JT-600 small submersible pump. The circuit system mainly includes a PWM DC motor speed controller, a current-to-voltage module, a power supply, wiring, and an industrial control computer. The sensor system mainly includes a pressure sensor, a flow sensor, a temperature sensor, a differential pressure sensor, and a data acquisition control console (acquisition card).

[0043] Hardware installation such as Figure 2As shown: The slurry tank is connected to the filling pump by the filling pipeline, and the filling pump is connected to the lower filling pipeline. Pressure sensor, temperature sensor and flow sensor are set at the pipeline inlet, and differential pressure sensors 2, 4 and 6 are set at each bend of the pipeline respectively. Differential pressure sensors 1, 3 and 5 are set in the middle section of the conveying pipeline. The data collected by each sensor is collected into the industrial control computer through the data acquisition console, and then the industrial control computer sends the control signal to the digital-to-analog converter, which controls the filling pump. Among them, the mixing tank is used to prepare filling material with tailings as aggregate and cement and fly ash as gelling agent; the filling pump transports the prepared filling material to the filling pipeline; the data acquisition console classifies and summarizes the signals collected by the flow sensor, temperature sensor and differential pressure sensor, and transmits them to the industrial control computer; the industrial control computer is used to realize: (1) control and adjustment of filling process parameters; real-time monitoring and automatic adjustment of process parameters such as pipeline pressure, filling flow rate and filling temperature. (2) recording and querying of filling operation data; generation of production reports, recording of equipment operation, recording of process parameters and recording of monitoring instrument data. (3) Message prompts and alarms; provide message prompts for process faults or operating status of the filling system, and set alarm limits for important parameters such as filling concentration, pipeline pressure, and filling flow rate in the filling system. (4) Data processing and display; process, analyze, and visualize the data in the filling operation in real time. The digital-to-analog converter converts the discrete signal in binary digital form transmitted by the industrial control computer into an analog quantity based on a standard quantity, thereby controlling the filling pump; the pressure sensor measures the pressure in the filling pump discharge pipe; the temperature sensor measures the temperature in the filling pump discharge pipe; the flow sensor measures the flow rate in the filling pump discharge pipe; differential pressure sensor 1 measures the pressure difference between the two ends of the horizontal pipe; differential pressure sensor 2 measures the pressure difference between the two ends of the pipe at one inflection point; differential pressure sensor 3 measures the pressure difference between the two ends of the middle section of pipe 1; differential pressure sensor 4 measures the pressure difference between the two ends of the pipe at the second inflection point; differential pressure sensor 5 measures the pressure difference between the two ends of the middle section of pipe 2; differential pressure sensor 6 measures the pressure difference between the two ends of the pipe at the third inflection point.

[0044] In this embodiment, the pressure sensor selected is the YC-P300 Hirschmann standard pressure transmitter. This pressure transmitter forms a Wheatstone bridge using four high-precision resistors on a silicon chip. Because the resistivity of a semiconductor changes due to external forces, the resistance values ​​change, leading to bridge imbalance and ultimately outputting an electrical signal corresponding to the pressure change.

[0045] The flow sensor mainly consists of a plastic valve body, a water flow rotor assembly, and a Hall effect sensor. Its function is to detect the inlet water flow rate. When water passes through the water flow rotor assembly, the magnetic rotor rotates with the change in flow rate, and its rotational speed changes accordingly. The Hall effect sensor detects the movement of the magnetic rotor and outputs a corresponding pulse signal. The pulse signal is converted into a current signal by a 4-20mA circuit board, and the data acquisition card determines the water flow rate by outputting an analog current signal. Therefore, the flow sensor can monitor the inlet water flow rate in real time, providing accurate flow information.

[0046] The data acquisition control console requires a data acquisition card capable of receiving and outputting analog signals to receive electrical signals transmitted from sensors and send control signals to the filling pump control module. Since the pressure transmitter outputs a 4-20mA current signal, and the flow meter also outputs a 4-20mA current signal after passing through the flow velocity converter's 4-20mA module, a data acquisition card with multiple analog inputs, capable of receiving 4-20mA current signals and converting them into digital signals via an AD converter, should be selected. After screening, this embodiment selects the DAM-10AIAO-RS232+485 ten-channel analog input / output data acquisition card.

[0047] In this embodiment, the implementation of the built-in software program of the industrial control computer mainly involves several aspects, including main interface design, implementation of filling process animation, serial port read / write operation, real-time curve plotting, and software operation flow. Specific functions are as follows: Figure 3 As shown.

[0048] The main interface of the software is as follows: Figure 4 As shown, it mainly includes a logo display area, an animation display area, real-time text, a function button bar, and a scale control.

[0049] Cinema 4D, a 3D modeling software developed by Maxon Computer in Germany, is similar to Maya and 3ds Max, primarily used in film and television production, industrial design, and other fields. It boasts powerful animation and rendering capabilities. The filling system is modeled using Cinema 4D, and then keyframes and cloning methods are used to animate the mixing tank and pipe arrows. Lighting and rendering materials are adjusted to obtain a filling process animation in a specified format. Since Tkinter cannot directly play animations, they need to be converted to GIF format. Tkinter's canvas control supports GIF images but not GIF playback; it can only display static images. The GIF image can be broken down, each frame iterated through, and output to the canvas, then updated using the `update` method. Simultaneously, the GIF image is segmented to ensure text display correctly. To enable looping playback of the GIF, a GIF playback task is submitted to a thread pool, and the `after` method within the task calls itself to achieve looping.

[0050] The function button bar includes serial port debugging, power control, and monitoring / alarm functions. Clicking a button executes a pre-defined action. By setting the creation of new sub-windows as button actions and then adding corresponding function controls to these sub-windows, the main interface becomes more streamlined and the functional divisions are clearer.

[0051] To achieve real-time updates of text and scale numbers, data needs to be acquired in real time, converted into its format, and finally displayed on the main interface. After receiving a return message from the serial port and obtaining data from the data bits, the real-time acquisition and updating of text and scale numbers can be achieved through the loop execution of a sub-thread. Text updates require certain conditions; they should only begin when serial communication is enabled and text updates are received. When the serial port is closed, text updates must end, and the text should be set to its initial value. Otherwise, if the data is empty, the program will stop and report an error. The data acquisition and processing methods will be explained in detail below.

[0052] To obtain digital sensor data converted by the data acquisition card, the industrial control computer needs to use the corresponding communication protocol. Devices using the RS485 interface typically use the Modbus communication protocol, and the data acquisition card used supports the standard Modbus communication protocol.

[0053] Modbus RTU is a serial communication protocol commonly used in industrial automation, primarily for communication between field devices and controllers. As a variant of the Modbus communication protocol, Modbus RTU uses binary transmission instead of ASCII transmission. The Modbus RTU protocol is based on the RS-485 serial communication standard and employs a master-slave architecture. In this architecture, a master device (typically a computer or PLC) communicates with multiple slave devices (typically sensors, actuators, or other controllers). The Modbus RTU protocol defines a set of data packet formats and protocol rules for communication between slave and master devices. These data packets include information such as the slave device address, function code, data field, and checksum. In the Modbus RTU protocol, the master device sends a request message containing the function code and data to be executed. Upon receiving the request message, the slave device executes the requested operation and then sends a response message to the master device, containing the requested data or status information.

[0054] The key to acquiring data from the acquisition card lies in generating a request message conforming to the Modbus protocol. This request message typically consists of a device address, function code, data area, and checksum. The number of bytes occupied by each part of the message is shown in Table 1.

[0055] Table 1. Number of bytes in a Modbus message

[0056] Address code 1 function code 1 Data bits N Check bit 2

[0057] The address code is the address of the slave device. Commonly used function codes and their functions are shown in Table 2.

[0058] Table 2 Common Modbus Function Codes

[0059] 0x01 Read coil status 0x02 Read discrete input state 0x03 Read holding register 0x04 Read input register 0x05 Write a single coil 0x06 Write a single holding register 0x0F Write multiple coil states 0x10 Write the value of multiple holding registers

[0060] Function code 04 can be used to read analog signals, and function code 10 can be used to output multiple analog signals. A message using function code 04 consists of the device address, function code 04, register start address, number of registers, and checksum. A message using function code 10 consists of the device address, function code 10, first analog register address, number of analog outputs, number of output bytes set, analog output value, and checksum. Message parsing is shown in Table 3.

[0061] Table 3 Modbus RTU Message Analysis

[0062]

[0063] The checksum uses CRC16 Modbus, a method that uses the CRC (Cyclic Redundancy Check) algorithm for verification and calculation, primarily used for error detection in the Modbus communication protocol. It employs a 16-bit CRC polynomial, effectively detecting most single-bit, double-bit, and some multi-bit transmission errors. The basic idea of ​​the CRC checksum is to use linear coding theory. At the sending end, based on the k bits of information to be transmitted, a checksum (i.e., the CRC code) of r bits is generated according to certain rules and appended to the information, forming a new binary code sequence of (k+r) bits, which is then sent out.

[0064] There are several CRC checksum implementation algorithms, such as the calculation method and the table lookup method. The calculation method algorithm is as follows:

[0065] (1) Initialize the CRC register to 0xFFFF.

[0066] (2) For the first data byte, perform an XOR operation with 0xFFFF and store the result in the CRC register.

[0067] (3) Shift the contents of the CRC register one bit to the right, fill the highest bit with 0, and check the value of the bit removed.

[0068] (4) If the removed value is 0, repeat step 3. If it is 1, perform an XOR operation between the register value and 0xA001, and store the result in the CRC register.

[0069] (5) Repeat steps 3 and 4 to verify individual data bytes.

[0070] (6) Repeatedly check all bytes.

[0071] (7) Invert the check result, then swap the lower 8 bits and the higher 8 bits to obtain the final CRC checksum.

[0072] Design a Combobox and Entry control to modify the serial port number, baud rate, slave address, register start address, and number of registers. Implement parameter settings by binding an action to the serial port start button's properties.

[0073] After generating the correct serial port message, it needs to be sent to the serial port. Import the Pyserial library, pass in the baud rate and serial port number, and instantiate the serial port. Use the `write` method to send the message and the `read` method to receive the return value. The 04 function code message enables real-time data reading; it should be sent and its return value received in a loop when no other statements are calling the serial port. When other statements call the serial port, such as sending a 10 function code message, the system should immediately perform a serial port write operation and clear the return value.

[0074] To ensure orderly message transmission and avoid receiving erroneous return values, a queue is used to store messages to be sent. Messages are sent to the serial port in sequence using a first-in, first-out (FIFO) queue. After a message is written to the queue, the serial port write function is called immediately to achieve timely message transmission and reception. To control the queue length and prevent newly added data from being unretrieved during subsequent serial port write operations, a condition for adding 04 function code messages to the queue can be set. Messages are added to the queue only when it is not empty, and serial port read / write operations are performed immediately after addition to ensure the queue is empty.

[0075] This embodiment designs a filling pump control signal transmission function, unifying the values ​​of the Scale and Entry controls by sharing a single variable. The Scale control allows users to adjust the current output value by sliding a slider, while the Entry control enables more precise data transmission. The transmitted message is also displayed on the right side for easy debugging.

[0076] The returned message consists of address code, function code, number of bytes, query analog value, and check code. You only need to verify the CRC check code first. After verification, extract the fourth and fifth bytes of data, convert them from Bytes type to decimal number, and use them as the return value.

[0077] The returned data has a temporal order, so a first-in-first-out queue can be used to store it to prevent data from being replaced by the next received data before it is used.

[0078] Displaying real-time data by plotting curves is not only more intuitive but also plays a crucial role in adjusting PID parameters and monitoring data. The curve plotting primarily utilizes the Matplotlib library. Time is obtained via `datetime` as the X-axis data, data retrieved from a queue is used as the Y-axis data, and the data is added to the last position of the list. The previous frame's content is then cleared, thus achieving curve plotting. Finally, the `after` method calls itself to update the curve, incorporating a condition that plotting only occurs when the serial port and window are open.

[0079] The program defines three sub-threads by submitting thread tasks to the thread pool. The first task is serial port reading and writing and text updating, the second task is control, and the last task is GIF playback.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope defined by the claims of the present invention.

Claims

1. A smart pressure regulating control system for pumping high-concentration ultrafine tailings backfill slurry, characterized in that: It includes a sensor module, a data acquisition console, an industrial computer, and a digital-to-analog converter; the data acquisition console controls the sensor module to collect pressure, temperature, flow rate, and pressure difference at pipe bends and in the middle of the pipe. The data acquisition console then collects the data collected by the sensor module into the industrial control computer; the industrial control computer uses deep learning algorithms to learn the relationship between the internal pressure, temperature, flow rate and filling pump speed of the pipeline and automatically selects the appropriate filling pump output power, thereby determining the control signal and sending the control signal to the digital-to-analog converter, which controls the filling pump. The industrial control computer uses deep learning algorithms to learn the relationship between the internal pressure, temperature, flow rate, and filling pump speed of the pipeline, and automatically selects the appropriate filling pump output power. The specific method is as follows: (1) The pressure, temperature and flow signals inside the filling pipeline are acquired in real time through the sensor module. These signals represent the state inside the pipeline to a certain extent. (2) Define the degree of blockage and establish a mapping function between pump speed and degree of blockage; The degree of congestion is defined as follows: the entropy values ​​of pressure, temperature, and flow signals within a time interval t, and the total flow within that time interval are used as the measure of the degree of congestion. That is, S(pressure, temperature, flow) = (alpha*H(pressure) + beta*H(temperature) + gamma*H(flow)) / sum(flow) = degree of congestion, where H is the information entropy. The larger the total flow, the lower the degree of congestion. alpha, beta, and gamma are the weights of each loss. The mapping function between pump speed and blockage level is established as follows: V(pump speed) = [pressure, temperature, flow rate], S(pressure, temperature, flow rate) = blockage level, that is: S(V(pump speed)) = blockage level; (3) Design a system for converting pump speed into pipeline pressure. The input of the system is pressure and the output is pump speed. The input state of the system changes with time and is a function of time. Therefore, the system is a dynamic system. The current output of the dynamic system depends not only on the current input, but also on the input and output of the system in the past. (4) The pump speed-blocking degree mapping function is fitted based on the dynamic neural network NARX model to obtain the pump speed-blocking degree mapping model; the NARX neural network structure includes an input layer, a hidden layer and an output layer; the number of nodes in the input layer is set according to the number of input values, and the number of nodes in the output layer is set according to the number of predicted values; (5) Construct a pumping simulation system and collect pump speed simulation data to conduct neural network training experiments to verify the feasibility of the method; the pump speed simulation data is a single-input multi-output data pair, including pump speed, pressure, temperature and flow rate, where pump speed is the independent variable and pressure, temperature and flow rate are the dependent variables. (6) Train the neural network to achieve a mapping from pump speed to blockage level; (7) For the pump speed-blockage degree mapping model obtained by training, use a genetic algorithm to search for a pump speed within a certain pump speed range that will result in a lower degree of blockage in the pipeline.

2. The intelligent pressure regulating control system for pumping high-concentration ultrafine tailings backfill slurry according to claim 1, characterized in that: The sensor module includes a pressure sensor, a temperature sensor, and a flow sensor installed at the filling pipe inlet, as well as differential pressure sensors installed at each bend of the filling pipe and in the middle section of the pipe.

3. The intelligent pressure regulating control system for pumping high-concentration ultrafine tailings backfill slurry according to claim 1, characterized in that: The industrial control computer has built-in software programs for determining control signals, real-time display of pressure and flow, plotting curves, animation of the filling process, early warning of filling pipeline blockage, power control of the filling pump, and one-button start of the filling pump.

4. The intelligent pressure regulating control system for pumping high-concentration ultrafine tailings backfill slurry according to claim 1, characterized in that: The industrial control computer's built-in software program is built using the Python standard graphical user interface library.

5. The intelligent pressure regulating control system for pumping high-concentration ultrafine tailings backfill slurry according to claim 1, characterized in that: The main interface of the industrial control computer's built-in software program includes a logo display area, a filling process animation display area, real-time text, a function button bar, and a gauge control.

6. The intelligent pressure regulating control system for pumping high-concentration ultrafine tailings backfill slurry according to claim 1, characterized in that: The industrial control computer models the filling system using the modeling software Cinema 4D. Then, it uses keyframes and cloning methods to create animations of the mixing tank and pipe arrows, adjusts lighting and rendering materials, and obtains a filling process animation in a specified format. The animation is then converted into a GIF image and played in a loop using Tkinter.

7. The intelligent pressure regulating control system for pumping high-concentration ultrafine tailings backfill slurry according to claim 6, characterized in that: The industrial control computer decomposes the GIF image, traverses each frame of the image and outputs it on the canvas, and updates it using the update method; at the same time, it segments the GIF image so that the text can be displayed correctly. To enable GIF playback in a loop, a GIF playback task is submitted to a thread pool, and the task uses the `after` method to call itself to achieve loop playback.

Images