An automated unmanned coal mine pump station fluid supply system

Abstract

The present invention discloses an automated unmanned coal mine pump station liquid supply system, comprising: a multi-stage purified water treatment controller for controlling and estimating pure water production, RO membrane life, pure water production volume, and the optimal water level in the pure water tank; an automatic liquid distribution station controller for emulsion configuration, optimal liquid distribution concentration estimation, oil consumption estimation, liquid consumption estimation, and optimal liquid level estimation in the emulsion tank; a pump station controller for controlling the opening or closing of each pump body in the pump station and estimating the real-time liquid output of the pump station; an automatic return liquid filtration control for detecting return liquid flow rate and return liquid concentration and estimating the real-time liquid consumption difference; a system controller for data collection, data storage, data upload, and data estimation of each controller; and a centralized controller for receiving relevant data from each sub-controller sent by the system controller. The present invention improves the real-time data efficiency of the overall pump station liquid supply, instantly regulates the liquid supply pressure and liquid supply volume, and improves the utilization efficiency of the liquid supply center.

Description

Technical Field

[0001] The invention belongs to the technical field of automatic control of mining equipment, and in particular relates to an automated unmanned coal mine pump station fluid supply system. Background Art

[0002] In the field of mining, fully mechanized mining faces require the use of large quantities of high-pressure emulsions as the working medium for equipment such as hydraulic supports. During coal mining, inaccurate emulsion concentrations used in hydraulic supports can affect the support's lifespan and reduce its service life. Therefore, it is necessary to regularly and automatically adjust the emulsion ratio based on water inflow and return flow to ensure the emulsion concentration is within the hydraulic support's application range. Intelligent emulsion distribution and automatic output control are performed based on the support's liquid consumption, pump station output, high-pressure centralized liquid supply station output pressure, pipeline return liquid filtration station return flow, pure water quality, and distribution oil flow to ensure the emulsification pump produces sufficient high-pressure emulsion. Intelligent big data comprehensive calculations ensure maximum efficiency in distribution and supply, improve the utilization rate of distribution emulsion oil, reduce emulsion oil consumption, and extend the service life of RO membranes in water purification stations.

[0003] In the existing emulsion pump station liquid supply method, the steps are as follows: 1) a multi-stage purification water supply station produces pure water, 2) a distribution liquid concentration detector is installed at the outlet of the distribution station, and the distributed emulsion is stored in a liquid tank and supplied to the emulsification pump station in real time, 3) a pressure sensor is installed at the outlet pipe of the emulsification pump station, and the outlet pressure is obtained through the monitoring value of the sensor, and the emulsion pump station is controlled according to the monitored outlet pressure. When the outlet pressure is less than a specific threshold (such as 31.5MPa), the emulsification pump is continuously pressurized, that is, the frequency of the inverter in the pump station is increased. If the pressure is not enough within the set time (no external pipeline problem), the auxiliary pump is started to pressurize and increase the liquid flow rate. 4) After the emulsification pump discharges the liquid, the emulsion is filtered through a high-pressure filter station and then supplied to the bracket; 5) The return liquid from the bracket pipeline is returned to the liquid tank through the return liquid filter station.

[0004] The above solutions in the prior art have the following problems:

[0005] 1) During the pure water production process, there is no big data analysis of the RO membrane maintenance process, and the service life of the RO membrane cannot be determined during use, which easily affects the utilization rate of the support equipment;

[0006] 2) The emulsion dispensing station does not control the amount of liquid discharged during dispensing. The existing method easily causes excessive waste of emulsified oil, and the dispensed emulsion easily forms sediment, which shortens the service life of the bracket.

[0007] 3) The high-pressure emulsion produced by the emulsification pump changes through the outlet pressure value, and the auxiliary pump and standby pump are adjusted to start and stop. Frequent start and stop of the power frequency consumes a lot of power, has high power harmonics, and the supply pressure and flow rate are unstable, which will inevitably affect the production efficiency of the working face equipment and the service life of the bracket;

[0008] 4) The pipeline returns the liquid to the liquid tank. The emulsion concentration is not detected during the return liquid, and is not compared with the dispensed emulsion, which can easily affect the concentration of the emulsion in the liquid tank. The return liquid flow is not detected, the emulsion utilization rate cannot be known, and the dispensed liquid volume cannot be accurately controlled, resulting in excessive use of emulsified oil and excessive waste. Summary of the Invention

[0009] In view of this, the purpose of the implementation of the present invention is to provide an intelligent liquid supply and liquid preparation method, through TCP / IP, PROFINET field bus, electronic equipment and control system, edge computing, etc., to solve various problems in the existing technology in the processes of pure water production, emulsion preparation, emulsification pump pressurization and pressure replenishment, return liquid filtration, etc., so as to make the entire liquid supply process more digital, more accurate, more energy-saving and efficient.

[0010] The present invention discloses an automated unmanned coal mine pump station fluid supply system, comprising:

[0011] Multi-stage purified water treatment controller, used for controlling and estimating pure water production, RO membrane life, pure water production volume, and optimal water level in the pure water tank;

[0012] Automatic liquid dispensing station controller, used for emulsion configuration, optimal liquid concentration estimation, oil usage estimation, liquid usage estimation, and optimal liquid level estimation of the emulsion tank;

[0013] Pump station controller, which controls the start or stop of each pump in the pump station and estimates the real-time liquid output of the pump station;

[0014] Automatic return liquid filtration control, used for return liquid flow rate and return liquid concentration detection, and estimation of real-time liquid usage difference;

[0015] System controller, used for data collection, data storage, data upload, and data estimation of each controller;

[0016] The centralized controller is set on the ground; it is used to receive relevant data of each sub-controller sent by the system controller, and is used for real-time monitoring on the ground, making reports, and troubleshooting.

[0017] Furthermore, the system controller obtains various data values of pure water production, liquid preparation, emulsification pump discharge, liquid return, and liquid consumption equipment collected during a historical period, and associates the parameters detected per unit time with pure water production volume, liquid preparation oil volume, liquid preparation concentration, liquid preparation volume, and emulsification pump discharge volume as data base samples;

[0018] The data are used as variables to calculate the water quality before water production, the pure water quality after water production, the emulsion output, the emulsion return, the pure water tank level, the emulsion tank level, the hydraulic equipment control parameters, the optimal liquid level H 液佳 Input into the comprehensive calculation model of liquid level and pure water level to calculate and obtain the real-time pure water production volume, thereby obtaining the optimal water level value of the pure water tank and reducing the real-time working time of RO membrane reverse osmosis;

[0019] The collected data are used as variables, and the emulsified oil consumption, emulsified pump output, emulsified liquid return volume, emulsified liquid concentration, emulsified liquid return concentration, liquid tank level, hydraulic equipment control parameters, and optimal liquid concentration H are used as variables. 配佳 The data is input into the comprehensive calculation model of liquid distribution concentration to obtain the real-time liquid distribution concentration, automatically adjust the concentration value to the optimal usage value, obtain the real-time liquid output, and thus obtain the optimal liquid level value of the emulsion tank to prevent the emulsion in the liquid tank from deteriorating.

[0020] Furthermore, the comprehensive calculation model of liquid level and pure water level specifically includes:

[0021] 1) Assume that the real-time liquid outflow rate is E1, the real-time liquid return rate is E2, the real-time liquid distribution rate is E3, the bottom area of the liquid tank is S, h n is the sampling time, H 基 is the basic liquid level of the liquid tank, and n is the number of sampling times;

[0022] The instantaneous liquid level △H1 is calculated as follows:

[0023]

[0024] 2) Calculate the instantaneous average liquid level When there is no liquid backflow

[0025] Calculate the instantaneous average liquid level When there is liquid backflow,

[0026] Calculate the instantaneous average liquid level △H n :

[0027] When there is liquid backflow;

[0028] E 1i , E 2i , E 3i They are respectively the real-time liquid outflow E1, real-time liquid return flow E2, and real-time liquid distribution flow E3 collected for the i-th time;

[0029] 3.) After a certain period of accumulation, the optimal liquid level H for production liquid is obtained 液佳 ;

[0030] H液佳 =H 液基 +△H n +ε, where ε is the liquid correction factor;

[0031] 4) Calculate the pure water level value H 水佳

[0032] K is the mixing ratio.

[0033] Furthermore, the liquid correction factor ε is determined by the following method:

[0034] The collected emulsification pump outlet pressure value, high pressure filter station system pressure value, liquid tank level, liquid outflow, liquid return flow, hydraulic equipment control parameters, and the optimal liquid level H 液佳 Composed of feature vectors [X1,X2,…X n ] as input and the liquid correction factor ε as output to form the test set of monitoring data of the pump station liquid supply system;

[0035] The test set of monitoring data of the pumping station liquid supply system is input into the MLP-based convolutional memory residual neural network model for calculation. The learning rate is set to 0.002, and the network is iteratively trained using the gradient descent method until the loss function converges. The real-time maximum liquid supply flow rate under the set pressure is obtained to control the liquid discharge adjustment value of the emulsification pump.

[0036] Furthermore, the MLP-based convolutional memory residual neural network includes two MLPs and a convolutional memory residual neural network. Each multi-layer perceptron MLP includes an input layer, a hidden layer, and an output layer. One of the MLPs performs self-fusion on each dimension of data, and the other MLP communicates between different channels to fuse each dimension of data with each other.

[0037] The input eigenvalues are formed into vectors, and the eigenvectors are cross-correlated with the convolution kernels before being input into the MLP. After the MLP extracts the underlying features of the data, it passes through the nonlinear activation layer of the convolutional memory residual neural network, the global average pooling layer, the convolution layer, and finally the fully connected layer to predict the adjustment value of the emulsification pump.

[0038] Furthermore, the comprehensive calculation model for the solution concentration specifically includes:

[0039] The concentration of the collected solution is N 配 , H n is the liquid level height of the liquid tank, and the transient concentration value of the liquid outflow is N n , the transient concentration value of the backflow liquid is M n , real-time liquid outflow E1, real-time liquid return flow E2, real-time liquid distribution flow E3;

[0040] Calculate the instantaneous concentration of the solution:

[0041] N 配瞬时 =(N n *(S*H n )+E 3n *h n +E 2n *h n -E 1n *h n -M n (E 2n *h n )) / (S*H n +E 3n *h n -E 1n *h n )

[0042] Calculate the average optimal solution concentration

[0043] δ is the correction coefficient.

[0044] Furthermore, the correction coefficient δ is obtained by the following method:

[0045] The amount of emulsified oil used, the amount of emulsified pump output, the amount of emulsion returned, the concentration of emulsion output, the concentration of emulsion returned, the level of the liquid tank, the control parameters of the hydraulic equipment and the calculated optimal liquid concentration H are calculated. 配佳 The data is input into the MLP-based convolutional memory residual neural network model to calculate and obtain the correction value of the liquid concentration.

[0046] Furthermore, the loss function of the MLP-based convolutional memory residual neural network is a mean square error loss function.

[0047] Compared with the prior art, the above technical solution provided by the embodiment of the present invention has at least the following beneficial effects:

[0048] The embodiments of the present invention provide an intelligent liquid supply method, controller, system control, centralized controller and control system for a liquid supply center, and provide an intelligent digital control system for a coal mine pump station liquid supply center. The working status can be understood in real time through the sub-control, and the master control that obtains data uploaded from each distributed sub-control can achieve high-speed connection with the remote workbench via Ethernet TCP / IP, thereby realizing remote control, data collection, data monitoring, and data aggregation, realizing the overall connection of the intelligent Internet of Things, improving the real-time data efficiency of the overall pump station liquid supply, facilitating timely understanding of the situation at the liquid supply site, and timely adjusting the operating status to improve maximum efficiency. High-end intelligent modules are used to realize centralized data collection, centralized control, remote monitoring, unmanned operation, reduce labor costs, improve operational efficiency, and statistical big data, so that the liquid supply pressure and liquid supply volume can be adjusted in real time, greatly improving the utilization efficiency of the liquid supply center. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a schematic diagram of the working process of the liquid supply center according to one embodiment of the present invention;

[0050] Figure 2 This is a principle block diagram of a liquid supply center system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The present invention will be further described below with reference to the accompanying drawings, but the present invention is not limited in any way. Any changes or substitutions made based on the teachings of the present invention fall within the scope of protection of the present invention.

[0052] To solve the above technical problems, the present invention provides a high-intelligence, high-digital, and high-coordination method for a liquid supply center, comprising the following steps:

[0053] Test the water quality before and after water production, record the total water production time and the output of pure water. Test the output concentration, output volume, and emulsified oil consumption of the liquid distribution station.

[0054] Detect the pump station outlet pressure value, liquid flow rate, high-pressure filter station system pressure value, and liquid tank level.

[0055] Detect the concentration and flow rate of return liquid emulsion.

[0056] This data is collected at high speed, edge-calculated, and converted into Ethernet packets by distributed control boxes at the equipment end. This data is then transmitted to the central control and liquid supply center via the PROFINET industrial fieldbus at high speed and in real time as sampled data. This data is then recorded and accumulated in real time. Using an established comprehensive calculation model, the optimal outlet concentration and supply volume are calculated in real time. This effectively extends the life of the RO membrane, efficiently utilizes emulsified oil, and precisely controls concentration to improve the lubrication of the bracket, extending its service life.

[0057] The collected data are used as variables to calculate the water quality before water production, the pure water quality after water production, the emulsion output, the emulsion return, the pure water tank level, the emulsion tank level, the hydraulic equipment control parameters, the optimal liquid level H 液佳 This input is fed into a comprehensive calculation model for the liquid level and pure water level to determine the real-time pure water production volume, thereby determining the optimal water level in the pure water tank. If pure water demand is low, water production can be stopped immediately, thus reducing the real-time operating time of the RO membrane reverse osmosis system.

[0058] The collected data are used as variables, and the emulsified oil consumption, emulsified pump output, emulsified liquid return volume, emulsified liquid concentration, emulsified liquid return concentration, liquid tank level, hydraulic equipment control parameters, and optimal liquid concentration H are used as variables. 配佳Input into the comprehensive calculation model for liquid dosing concentration, the real-time concentration is calculated. The concentration is automatically adjusted to the optimal usage value, and the real-time liquid output is obtained, thereby determining the optimal liquid level in the emulsion tank. If the liquid usage is low, dosing can be stopped immediately, saving emulsified oil and effectively preventing deterioration of the emulsion in the tank.

[0059] Using these collected data as variables, the emulsification pump outlet pressure, high-pressure filter station system pressure, liquid tank level, liquid discharge flow, liquid return flow, and hydraulic equipment control parameters are input into a pre-established mathematical model to calculate the maximum liquid supply flow required in real time under the set pressure. This is used to control the liquid output of the emulsification pump and ensure the liquid supply requirements of equipment such as the bracket. At the same time, when the liquid supply is low, the liquid output of the emulsification pump can be controlled, the auxiliary pump can be stopped, and the frequency of the main pump can be reduced to achieve low-frequency constant-pressure liquid supply. This reduces the utilization rate of the pump station, increases the mechanical life of the pump station, and reduces equipment power consumption, achieving effective energy conservation and emission reduction.

[0060] Based on the same inventive concept, the present invention also provides a fully intelligent and fully digital liquid supply system for a liquid supply center, comprising:

[0061] Multi-stage purified water treatment controller, used for controlling and estimating pure water production, RO membrane life, pure water production volume, and optimal water level in the pure water tank;

[0062] Automatic liquid dispensing station controller, used for emulsion configuration, optimal liquid concentration estimation, oil usage estimation, liquid usage estimation, and optimal liquid level estimation of the emulsion tank;

[0063] Pump station controller, which controls the start or stop of each pump in the pump station and estimates the real-time liquid output of the pump station;

[0064] Automatic return liquid filtration control, used for return liquid flow rate and return liquid concentration detection, and estimation of real-time liquid usage difference;

[0065] System controller, used for data collection, data storage, data upload, and data estimation of each controller, including estimation of liquid flow, liquid concentration, and pure water production volume;

[0066] The centralized controller is set on the ground; it is used to receive relevant data of each sub-controller sent by the system controller, and is used for real-time monitoring on the ground, making reports, troubleshooting, etc.

[0067] The system controller uses collected data as variables, inputting sample data such as pre-production water quality, post-production pure water quality, emulsion output, emulsion return, pure water tank level, emulsion tank level, and hydraulic equipment control parameters into a comprehensive liquid and pure water level calculation model to calculate the real-time pure water production volume and, consequently, the optimal pure water tank water level. If pure water demand is low, water production can be stopped promptly, thus reducing the real-time operating time of the RO membrane reverse osmosis system.

[0068] Using these collected data as variables, sample data such as emulsified oil usage, emulsified pump output, emulsified return volume, emulsified output concentration, emulsified return concentration, tank level, and hydraulic equipment control parameters are input into a pre-established mathematical model to calculate the real-time dispensing concentration. This automatically adjusts the concentration to the optimal usage value, obtains the real-time output volume, and thus determines the optimal emulsion tank level. If the liquid usage is low, dispensing can be stopped immediately, saving emulsified oil and effectively preventing the deterioration of the emulsion in the tank.

[0069] Using the collected data as variables, the outlet pressure of the emulsification pump, the pressure of the high-pressure filter station system, the liquid tank level, the liquid discharge flow rate, the liquid return flow rate, and the hydraulic equipment control parameters are input into the comprehensive calculation model of the liquid distribution concentration to obtain the maximum liquid supply flow rate required in real time under the set pressure. This is used to control the liquid output of the emulsification pump and ensure the liquid supply requirements of equipment such as the bracket. At the same time, when the liquid supply volume is small, the liquid output of the emulsification pump can be controlled, the auxiliary pump can be stopped, and the frequency of the main pump can be reduced to achieve low-frequency constant-pressure liquid supply. This reduces the utilization rate of the pump station, increases the mechanical life of the pump station, and reduces the power consumption of the equipment. This achieves effective energy conservation and emission reduction.

[0070] This embodiment provides an intelligent liquid supply method for a liquid supply center, which is applied to an intelligent liquid supply system. Figure 1 As shown, the following steps are included:

[0071] S01: Collect various data values of pure water production, liquid distribution, emulsification pump output, return liquid, and liquid use equipment (including water quality, concentration, liquid output flow, return liquid flow, pipeline length, outlet pressure value, system pressure value), detect the water quality of water before water production and the water quality of pure water after water production, and record the total water production time and pure water output. Detect the liquid output concentration, liquid output, and emulsified oil consumption of the liquid distribution station. Detect the pump station outlet pressure value, liquid output flow, high-pressure filter station system pressure value, liquid tank level, and detect the return liquid emulsion concentration and return liquid flow. The detection cycle of each parameter is real-time or at very short intervals. Each detection data is stored on time for calculation and use, as shown in Table 1, Table 2, and Table 3. Variable relationship correspondence table.

[0072] Table 1 Parameter detection and water level control model of multi-stage water purification supply station

[0073]

[0074]

[0075] The data in the above table is a group of 2S intervals, and this time interval can be adjusted according to actual needs. Table 2 Automatic liquid dispensing station parameter detection and liquid dispensing control liquid level value and emulsion concentration value

[0076]

[0077] Table 3 Emulsification pump station parameter detection and liquid output and liquid output pressure control

[0078]

[0079] Table 2 is a relationship table of liquid dosage variables. Sample data such as emulsified oil consumption, emulsification pump output, emulsion return volume, emulsion output concentration, emulsion return concentration, liquid tank level, and hydraulic equipment control parameters are input into a pre-established mathematical model for calculation to obtain real-time liquid dosage concentration and real-time liquid output.

[0080] Table 3 is a variable relationship table of the emulsification pump discharge volume. The outlet pressure value of the emulsification pump, the pressure value of the high-pressure filter station system, the liquid level of the liquid tank, the liquid discharge flow rate, the liquid return flow rate, and the control parameters of the hydraulic equipment are input into the pre-established mathematical model to calculate the maximum liquid supply flow rate required in real time under the set pressure to control the liquid discharge volume of the emulsification pump.

[0081] The data in the above table is a set of 2S intervals. This time interval can be adjusted according to actual needs. The estimated optimal value can be 1 day, 1 week, or 1 month of normal operation.

[0082] All collected data is time-stamped, and all data detected at the same moment corresponds to an actual value. By correlating a set of data at the same moment as sample data, a large amount of sample data can be determined over a historical period. In practical applications, a larger amount of sample data improves the accuracy of this intelligent liquid supply method. The use of a high-performance 32-bit processor provides powerful data processing capabilities. Distributed sampling, computing, and storage provide powerful comprehensive and scalable capabilities.

[0083] Since the various independent variable data in the above-mentioned sampling table are real-time data collected at adjustable fixed time intervals, the output dependent variables required by the system must also be output at fixed time intervals. In this way, for an operating control system such as a mining emulsion system that does not need to pursue an excessively high refresh rate, this set of mathematical models can well meet the control requirements and actual production applications of the invention.

[0084] In this embodiment, statistical analysis of actual production data for 8 consecutive days shows that the two pumps were running simultaneously for a total of 7869 minutes, and the running time of one pump was approximately half of that, totaling 4316 minutes.

[0085] In actual engineering applications, equipment and its sensors come from different manufacturers, and the data communication protocols and sampling frequencies are different. This has an impact on data cleaning and data deduplication, resampling, alignment, and merging.

[0086] S02: The parameters detected per unit time are associated with the pure water production volume (pure water level), the amount of oil used for liquid preparation, the concentration of liquid preparation, and the output of the emulsification pump as the data base samples. Sample data such as the water quality before water production, the pure water quality after water production, the emulsion output volume, the emulsion return volume, the pure water tank level, the emulsion tank level, and the hydraulic equipment control parameters are input into the comprehensive calculation model of the liquid level and pure water level, and iterative training is carried out to obtain the real-time water production volume and produce the required pure water.

[0087] The comprehensive calculation model of liquid level and pure water level specifically includes:

[0088] 1) Assume that the real-time liquid outflow rate is E1, the real-time liquid return rate is E2, the real-time liquid distribution rate is E3, the bottom area of the liquid tank is S, h n is the sampling time, H 基 is the basic liquid level of the liquid tank, and n is the number of sampling times.

[0089] The instantaneous liquid level △H1 is calculated as follows:

[0090]

[0091] 2)a) Calculate the instantaneous average liquid level When there is no liquid backflow

[0092] b) Calculate the instantaneous average liquid level When there is liquid backflow,

[0093] c) Calculate the instantaneous average liquid level △H n :

[0094] When there is liquid backflow

[0095] E 1i , E 2i , E 3i They are respectively the real-time liquid outflow E1, the real-time liquid return flow E2, and the real-time liquid distribution flow E3 collected for the i-th time.

[0096] 3.) After a certain period of accumulation, the optimal liquid level H for production liquid is obtained液佳 .

[0097] H 液佳 =H 液基 +△H n +ε, where ε is the liquid correction factor.

[0098] where ε is determined by the following method:

[0099] First, the collected emulsification pump outlet pressure value, high-pressure filter station system pressure value, liquid tank level, liquid flow rate, return flow rate, hydraulic equipment control parameters, and the optimal liquid level H are calculated. 液佳 Composed of feature vectors [X1,X2,…X n ] as input and the liquid correction factor ε as output to form the pump station liquid supply system monitoring data test set.

[0100] The test set of monitoring data of the pumping station liquid supply system is input into the MLP-based convolutional memory residual neural network model for calculation. The learning rate is set to 0.002, and the network is iteratively trained using the gradient descent method until the loss function converges. The maximum liquid supply flow rate required in real time under the set pressure is obtained to control the liquid discharge adjustment value of the emulsification pump.

[0101] The MLP-based convolutional memory residual neural network consists of two MLPs (multi-layer perceptrons) and a convolutional memory residual neural network. Each MLP includes an input layer, a hidden layer, and an output layer. One MLP performs self-fusion within each dimension of data, while the other MLP communicates between different channels and fuses the data in each dimension.

[0102] The input eigenvalues are formed into vectors, and the eigenvectors are cross-correlated with the convolution kernels before being input into the MLP. After the MLP extracts the underlying features of the data, it passes through the nonlinear activation layer of the convolutional memory residual neural network, the global average pooling layer, the convolution layer, and finally the fully connected layer to predict the liquid output adjustment value of the emulsification pump.

[0103] The loss function of the MLP-based convolutional memory residual neural network is the mean square error loss function.

[0104] 4) Calculate the pure water level value H 水佳

[0105] K is the mixing ratio.

[0106] S03: Input sample data such as emulsified oil consumption, emulsification pump output, emulsion return volume, emulsion output concentration, emulsion return concentration, liquid tank level, and hydraulic equipment control parameters into a pre-established mathematical model for calculation to obtain real-time liquid concentration and real-time output volume to control the output volume of the emulsification pump.

[0107] The comprehensive calculation model of liquid concentration specifically includes:

[0108] 1. Set the solution concentration value to N 配 , H n is the liquid level height of the liquid tank, and the transient concentration value of the liquid outflow is N n , the transient concentration value of the backflow liquid is M n , real-time liquid outlet flow E1, real-time liquid return flow E2, real-time liquid distribution flow E3.

[0109] 2. Estimation of instantaneous solution concentration

[0110] N 配瞬时 =(N n *(S*H n )+E 3n *h n +E 2n *h n -E 1n *h n -M n (E 2n *h n )) / (S*H n +E 3n *h n -E 1n *h n )

[0111] 4. Calculate the average optimal solution concentration

[0112] δ is the correction coefficient.

[0113] The correction coefficient δ is obtained by the following method:

[0114] The emulsified oil consumption, emulsified pump output, emulsion return volume, emulsion output concentration, emulsion return concentration, tank level, hydraulic equipment control parameters and the calculated optimal liquid concentration H are calculated. 配佳 The sample data is input into the MLP-based convolutional memory residual neural network model to calculate and obtain the real-time liquid concentration and real-time liquid output.

[0115] First, the collected emulsification pump outlet pressure value, high-pressure filter station system pressure value, liquid tank level, liquid flow rate, return flow rate, hydraulic equipment control parameters, and the optimal liquid level H are calculated. 液佳 Composed of feature vectors [X1,X2,…X n ] as input and the liquid correction factor ε as output to form the pump station liquid supply system monitoring data test set.

[0116] The test set of monitoring data of the pumping station liquid supply system is input into the MLP-based convolutional memory residual neural network model for calculation. The learning rate is set to 0.002, and the network is iteratively trained using the gradient descent method until the loss function converges. The maximum liquid supply flow rate required in real time under the set pressure is obtained to control the liquid discharge adjustment value of the emulsification pump.

[0117] The MLP-based convolutional memory residual neural network consists of two MLPs (multi-layer perceptrons) and a convolutional memory residual neural network. Each MLP includes an input layer, a hidden layer, and an output layer. One MLP performs self-fusion within each dimension of data, while the other MLP communicates between different channels and fuses the data in each dimension.

[0118] The input eigenvalues are formed into vectors, and the eigenvectors are cross-correlated with the convolution kernels before being input into the MLP. After the MLP extracts the underlying features of the data, it passes through the nonlinear activation layer of the convolutional memory residual neural network, the global average pooling layer, the convolution layer, and finally the fully connected layer to predict the liquid output adjustment value of the emulsification pump.

[0119] S04: At a preset time, the current liquid outflow and inflow are input into the linear regression model to obtain real-time water production, oil dosage, dosage concentration, dosage volume, and emulsion pump output as reference values for the most economical water production and dosage volumes. Data accumulation reveals a linear correlation in actual applications. Overall, the liquid volume, dosage volume, and water production are cyclical. When liquid usage is high, water production, dosage, and pumping stations operate according to the desired dependent variable values derived from the machine regression algorithm. When liquid usage is low, water production, dosage, and pumping station activations are reduced, with variable frequency drive maintaining low-speed, constant-pressure liquid supply. After obtaining the adjusted independent variable X and corresponding numerical coefficients, the various independent variable parameters collected by the system are then entered into the linear regression mathematical model. The system can use the model's output dependent variable values to determine the desired emulsion concentration, appropriate liquid output, and other important values. This serves as a key basis for the automatic control system to adjust the corresponding mechanical actions in real time.

[0120] In addition, in order to be able to apply different device learning algorithms, it is also necessary to perform corresponding sampling and changes on the data. Using the above-mentioned scheme in this embodiment, the sample data is trained using a calculation model. Once the test data is obtained, the water production volume, liquid distribution volume, and liquid usage volume can be estimated in advance, thereby performing predictive control. This simplifies the structure of the system and reduces costs. The above schemes are all implemented in an automated manner and controlled by artificial intelligence, eliminating the need for human intervention and freeing up manpower. Moreover, the above-mentioned control logic obtains patterns from historical data, breaking away from the reliance on human experience.

[0121] In this embodiment, a comprehensive data estimation algorithm is used as an example for illustration. When using the comprehensive data estimation algorithm, the detection parameters are input as independent variables into the comprehensive data estimation algorithm for iterative training. Based on actual conditions, the comprehensive ability of the data can be continuously optimized to achieve the optimal data output value.

[0122] By adopting the above scheme, after inputting the detection data into the comprehensive data estimation model, the water production volume, liquid distribution volume, and liquid output volume can be estimated in advance, so that the liquid supply pressure, the number of pumps started, and the actual liquid supply volume can be adjusted in real time without the defect of lag, which greatly improves the production efficiency of the working face.

[0123] Example

[0124] like Figure 2 The distributed liquid supply center intelligent digital control system shown in the figure shows three spray pump groups. 4, 5, and 6 are the sub-control boxes for the three spray pumps, used to monitor the operating status of the spray pumps in real time. 7 and 8 are the spray pump water tanks, which supply water to the spray pumps. 9 is the inlet filtration station. 11 is the sub-control box for the inlet filtration station, which controls the inlet filtration station to achieve timed pressure differential backwashing. 10 is the multi-stage purification water supply station, which provides pure water to the automated liquid dispensing station. 12 is the sub-control box for the multi-stage purification station, which is used for pure water production, cleaning, and water quality testing. 13 is the automated liquid dispensing station, used for emulsion preparation. 14 is the sub-controller for the automated liquid dispensing station, which controls and collects data for automated liquid dispensing. 15 is the automated return filtration station, which filters the emulsion returning from the support. 16 is the sub-controller for the automated return filtration, which controls filtration and backwashing time and pressure differential. 17 and 18 are emulsion storage tanks A and B, used to store emulsion dispensed from dispensing station 15. 19, 20, 21, and 22 are four emulsion pumps, and 23, 24, 25, and 26 are four emulsion pump sub-controllers, controlling the start and stop of each pump and collecting and uploading relevant data. 35 is the flameproof control and case box for the liquid supply center, responsible for data collection and uploading for all sub-controllers, and also for centrally controlling the start, stop, and lockout of each sub-controller. 34 is the flameproof control power supply box, used to convert 127V to 24V to power the sub-controllers. 27 is a multi-channel flameproof inverter, used to drive the emulsion pump motors for constant-pressure liquid supply. 28 is the flameproof remote computer monitoring system for the chute, used for centralized underground monitoring of the operating status and test data of each device in the liquid supply center, and for centrally controlling the start, stop, and lockout of each sub-controller. 29 is a gigabit mining switch, connected to the master control, chute monitoring, and underground ring network for real-time data exchange. 30 is the underground mine ring network. 31 is the surface control center, which uploads data through the underground fluid supply center master control at 35. The data is then uploaded through the gigabit switch and the underground ring network to the surface control center's fluid supply control interface, displaying the control status of each device in the underground fluid supply center.

[0125] Compared with the prior art, the above technical solution provided by the embodiment of the present invention has at least the following beneficial effects:

[0126] The embodiments of the present invention provide an intelligent liquid supply method, controller, system control, centralized controller and control system for a liquid supply center, and provide an intelligent digital control system for a coal mine pump station liquid supply center. The working status can be understood in real time through the sub-control, and the master control that obtains data uploaded from each distributed sub-control can achieve high-speed connection with the remote workbench via Ethernet TCP / IP, thereby realizing remote control, data collection, data monitoring, and data aggregation, realizing the overall connection of the intelligent Internet of Things, improving the real-time data efficiency of the overall pump station liquid supply, facilitating timely understanding of the situation at the liquid supply site, and timely adjusting the operating status to improve maximum efficiency. High-end intelligent modules are used to realize centralized data collection, centralized control, remote monitoring, unmanned operation, reduce labor costs, improve operational efficiency, and statistical big data, so that the liquid supply pressure and liquid supply volume can be adjusted in real time, greatly improving the utilization efficiency of the liquid supply center.

[0127] As used herein, the word "preferred" is intended to serve as an example, instance, or illustration. Any aspect or design described herein as "preferred" is not necessarily to be construed as advantageous over other aspects or designs. Rather, the use of the word "preferred" is intended to present concepts in a concrete manner. As used in this application, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or." That is, unless otherwise specified or clear from the context, "X employs A or B" is intended to mean any of the naturally inclusive permutations. That is, if X employs A; X employs B; or X employs both A and B, then "X employs A or B" is satisfied in any of the foregoing examples.

[0128] Moreover, although the present disclosure has been shown and described with respect to one or implementation, those skilled in the art will think of equivalent variations and modifications based on reading and understanding of this specification and the accompanying drawings. The present disclosure includes all such modifications and variations and is limited only by the scope of the appended claims. In particular, with respect to the various functions performed by the above-mentioned components (such as elements, etc.), the terms used to describe such components are intended to correspond to any component (unless otherwise indicated) that performs the specified function of the component (such as it is functionally equivalent), even if structurally different from the disclosed structure that performs the function in the exemplary implementation of the present disclosure shown herein. In addition, although the specific features of the present disclosure have been disclosed with respect to only one of several implementations, such features can be combined with one or other features of other implementations that can be desired and advantageous for a given or specific application. Moreover, insofar as the terms "including", "having", "containing" or their variations are used in specific embodiments or claims, such terms are intended to be included in a manner similar to the term "comprising".

[0129] The functional units in the embodiments of the present invention may be integrated into a single processing module, or each unit may exist physically separately, or multiple or more units may be integrated into a single module. The aforementioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium. The aforementioned storage medium may be a read-only memory, a magnetic disk, or an optical disk, etc. The aforementioned devices or systems may execute the storage method in the corresponding method embodiment.

[0130] In summary, the above embodiment is one implementation method of the present invention, but the implementation method of the present invention is not limited to the described embodiment. Any other changes, modifications, substitutions, combinations, and simplifications that deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the scope of protection of the present invention.

Claims

1. An automated unmanned coal mine pump station liquid supply system, characterized in that: include: Multi-stage purified water treatment controller, used for controlling and estimating pure water production, RO membrane life, pure water production volume, and optimal water level in the pure water tank; Automatic liquid dispensing station controller, used for emulsion configuration, optimal liquid concentration estimation, oil usage estimation, liquid usage estimation, and optimal liquid level estimation of the emulsion tank; Pump station controller, which controls the start or stop of each pump in the pump station and estimates the real-time liquid output of the pump station; Automatic return liquid filtration control, used for return liquid flow rate and return liquid concentration detection, and estimation of real-time liquid usage difference; System controller, used for data collection, data storage, data upload, and data estimation of each controller; The centralized controller is set on the ground; it is used to receive the relevant data of each sub-controller sent by the system controller for real-time monitoring on the ground, making reports, and troubleshooting; The system controller obtains various data values of pure water production, liquid preparation, emulsification pump discharge, liquid return, and liquid consumption equipment collected during a historical period, and associates the parameters detected per unit time with pure water production volume, liquid preparation oil volume, liquid preparation concentration, liquid preparation volume, and emulsification pump discharge volume as data base samples; The data are used as variables to calculate the water quality before water production, the pure water quality after water production, the emulsion output, the emulsion return, the pure water tank level, the emulsion tank level, the hydraulic equipment control parameters, the optimal liquid level H 液佳 Input into the comprehensive calculation model of liquid level and pure water level to calculate and obtain the real-time pure water production volume, thereby obtaining the optimal water level value of the pure water tank and reducing the real-time working time of RO membrane reverse osmosis; The collected data are used as variables, and the emulsified oil consumption, emulsified pump output, emulsified liquid return volume, emulsified liquid concentration, emulsified liquid return concentration, liquid tank level, hydraulic equipment control parameters, and optimal liquid concentration H are used as variables. 配佳 Input the data into the comprehensive calculation model of liquid concentration to obtain the real-time liquid concentration, automatically adjust the concentration value to the optimal usage value, obtain the real-time liquid output, and thus obtain the optimal liquid level value of the emulsion tank to prevent the emulsion in the tank from deteriorating; The comprehensive calculation model of liquid level and pure water level specifically includes: 1) Assume that the real-time liquid outflow rate is E1, the real-time liquid return rate is E2, the real-time liquid distribution rate is E3, the bottom area of the liquid tank is S, h n is the sampling time, H 基 is the basic liquid level of the liquid tank, and n is the number of sampling times; The instantaneous liquid level △H1 is calculated as follows: 2) Calculate the instantaneous average liquid level When there is no liquid backflow Calculate the instantaneous average liquid level When there is liquid backflow, Calculate the instantaneous average liquid level △H n : When there is liquid backflow; E 1i , E 2i , E 3i They are respectively the real-time liquid outflow E1, real-time liquid return flow E2, and real-time liquid distribution flow E3 collected for the i-th time; 3) After a certain period of accumulation, the optimal liquid level H for production liquid is obtained 液佳 ; H 液佳 =H 液基 +△H n +ε, where ε is the liquid correction factor; 4) Calculate the pure water level value H 水佳 K is the mixing ratio.

2. The automated unmanned coal mine pump station liquid supply system according to claim 1, characterized in that: The liquid correction factor ε is determined by the following method: The collected emulsification pump outlet pressure value, high pressure filter station system pressure value, liquid tank level, liquid outflow, liquid return flow, hydraulic equipment control parameters, and the optimal liquid level H 液佳 Composed of feature vectors [X1,X2,…X n ] as input and the liquid correction factor ε as output to form the test set of monitoring data of the pump station liquid supply system; The test set of monitoring data of the pumping station liquid supply system is input into the MLP-based convolutional memory residual neural network model for calculation. The learning rate is set to 0.002, and the network is iteratively trained using the gradient descent method until the loss function converges. The real-time maximum liquid supply flow rate under the set pressure is obtained to control the liquid discharge adjustment value of the emulsification pump.

3. The automated unmanned coal mine pump station liquid supply system according to claim 2, characterized in that: The MLP-based convolutional memory residual neural network includes two MLPs and a convolutional memory residual neural network. Each multi-layer perceptron MLP includes an input layer, a hidden layer, and an output layer. One of the MLPs performs self-fusion on each dimension of data, and the other MLP communicates between different channels to fuse each dimension of data with each other. The input eigenvalues are formed into vectors, and the eigenvectors are cross-correlated with the convolution kernels before being input into the MLP. After the MLP extracts the underlying features of the data, it passes through the nonlinear activation layer of the convolutional memory residual neural network, the global average pooling layer, the convolution layer, and finally the fully connected layer to predict the adjustment value of the emulsification pump.

4. The automated unmanned coal mine pump station liquid supply system according to claim 3, characterized in that: The comprehensive calculation model for the solution concentration specifically includes: The concentration of the collected solution is N 配 , H n is the liquid level height of the liquid tank, and the transient concentration value of the liquid outflow is N n , the transient concentration value of the backflow liquid is M n , real-time liquid outflow E1, real-time liquid return flow E2, real-time liquid distribution flow E3; Calculate the instantaneous concentration of the solution: N 配瞬时 =(N n *(S*H n )+E 3n *h n +E 2n *h n -E 1n *h n -M n (E 2n *h n )) / (S*H n +E 3n *h n -E 1n *h n ) Calculate the average optimal solution concentration Among them E 1n is the nth flow rate of the real-time liquid flow E1, E 2n is the nth flow rate of the real-time return flow E2, E 3n is the nth flow rate of the real-time liquid distribution flow E3, N 配瞬时i is the i-th instantaneous solution concentration.

5. The automated unmanned coal mine pump station liquid supply system according to claim 4, characterized in that: The correction coefficient δ is obtained by the following method: The amount of emulsified oil used, the amount of emulsified pump output, the amount of emulsion returned, the concentration of emulsion output, the concentration of emulsion returned, the level of the liquid tank, the control parameters of the hydraulic equipment and the calculated optimal liquid concentration H are calculated. 配佳 The data is input into the MLP-based convolutional memory residual neural network model to calculate and obtain the correction value of the liquid concentration.

6. The automated unmanned coal mine pump station liquid supply system according to claim 3, characterized in that: The loss function of the MLP-based convolutional memory residual neural network is a mean square error loss function.

Images