Method for monitoring and controlling equipment operation of aluminum oxide water taking station

Abstract

The invention discloses an aluminum oxide water taking station equipment operation monitoring control method which comprises the following steps: acquiring operation parameters of key positions of water taking station equipment in real time through a sensor network, performing primary processing, uploading processed data to a data analysis control module, performing deep analysis on the received data by the data analysis control module, and transmitting the analyzed data to a server; when the running state of the equipment is abnormal, alarm information is generated immediately, a user is notified through a user interaction interface, meanwhile, the data analysis control module generates a control instruction according to an analysis result and sends the control instruction to the control execution module, and the control execution module intelligently controls the water taking station equipment according to the received control instruction. And stable operation of equipment is ensured. According to the invention, comprehensive real-time monitoring, intelligent fault early warning and diagnosis and automatic optimization adjustment can be realized, the operation efficiency of equipment is improved, the service life of the equipment is prolonged, and the maintenance cost is reduced.

Description

Technical Field

[0001] The invention relates to an alumina water intake station equipment operation monitoring and control method, belonging to the technical field of water intake station monitoring. Background Art

[0002] In the alumina production process, the water intake station is a key water supply and treatment link, and the stable operation of its equipment is crucial to ensuring production efficiency and product quality. With the continuous development of industrial automation and intelligent technology, more and more water intake stations have begun to adopt advanced monitoring and control technologies to improve the operating efficiency and stability of equipment. In the alumina production process, the water intake station is a key water supply and treatment link, and the stable operation of its equipment is directly related to production efficiency and product quality. However, the traditional water intake station equipment monitoring method mainly relies on manual inspections and simple instrument indications, and there are problems such as incomplete monitoring, inaccurate data, and slow response speed. These problems not only affect the operating efficiency of the water intake station, but may also cause production interruptions due to equipment failure, resulting in economic losses.

[0003] At present, in the field of industrial automation and intelligent monitoring and control, some monitoring and control technologies for water intake station equipment have been applied. These technologies are mainly used in water intake stations of large-scale alumina production enterprises, and various sensors and instruments are installed to monitor the operating status and key parameters of the equipment in real time. At the same time, these technologies also combine data processing and intelligent control algorithms to achieve automatic adjustment and optimization of equipment. However, these technologies still have some problems and limitations in practical applications, such as incomplete monitoring, inaccurate data, and slow response speed.

[0004] Incomplete monitoring: Existing technologies usually only monitor some key parameters of water intake station equipment, while ignoring some other important status information, resulting in an inability to fully reflect the operating status of the equipment. Inaccurate data: Due to the accuracy limitations of sensors and instruments and errors in data collection and processing, the data collected by existing technologies often have certain inaccuracies, which affects the accurate judgment of the equipment status. Slow response speed: Existing technologies usually use fixed logical rules for control and cannot dynamically adjust and optimize according to the actual operating status of the equipment, resulting in slow response speed and inability to respond to equipment failures or abnormal situations in a timely manner. Poor user interactivity: The user interaction interface of existing technologies is usually complex and not friendly enough, which causes certain difficulties and inconveniences for users during use. Summary of the invention

[0005] In view of this, an object of the present invention is to provide an alumina water intake station equipment operation monitoring and control method to at least solve the problems of poor timeliness and difficulty in management and control caused by manual inspections mentioned in the background technology.

[0006] The objective of the present invention is achieved through the following technical solutions:

[0007] A method for monitoring and controlling the operation of equipment in an alumina water intake station comprises the following steps:

[0008] Step s1, the sensor network collects the operating parameters of the key positions of the water intake station equipment in real time;

[0009] Step s2: The industrial data acquisition gateway is responsible for receiving the data collected by the sensor network and performing preliminary processing;

[0010] Step s3: upload the processed data to the data analysis control module, which performs in-depth analysis on the received data to determine whether the operating status of the equipment is normal.

[0011] Yes, to maintain stable operation of the equipment;

[0012] No, the data analysis and control module will immediately generate an alarm message and notify the user through the user interaction interface. At the same time, the data analysis and control module will also generate control instructions based on the analysis results and send them to the control execution module. The control execution module will perform intelligent control of the water intake station equipment based on the received control instructions to achieve automatic optimization and adjustment of the equipment.

[0013] Furthermore, in step s1, the key positions of the alumina water intake station equipment include but are not limited to a water tank, a water pump, and an electric valve. The water tank is connected to various alumina production equipment through a water supply system. The water supply system includes a water pump outlet pipe and a water pump and an electric valve arranged on the water pump outlet pipe; a liquid level sensor is installed in the water tank, and the water pump is connected to a water pump current sensor, a water pump motor temperature sensor, a water pump motor vibration sensor, a water pump motor oil level monitoring sensor and a water intake controller. The water intake controller is used to realize water pump operation control, water pump stop control, and water pump remote control; a water pump outlet pipe pressure gauge and a water pump outlet pipe flow meter are installed on the water pump outlet pipe; the electric valve is connected to the water valve controller, and the water valve controller is used for electric valve operation control, electric valve stop control, electric valve flow regulation control, and electric valve remote control input.

[0014] Further, in step s2, preliminary data processing includes but is not limited to data filtering, amplification, conversion, and image processing;

[0015] In step s3, the data is transmitted to the server by accessing the industrial switch. The server stores the data in its own database and pushes the data to the intelligent analysis model and the equipment operation monitoring and control system respectively. The intelligent analysis model uses the intelligent analysis model to analyze and judge the real-time operation status, operation trend, and operation history comparison of each device according to the data source, and identifies the operation status and potential faults of each device through the equipment operation monitoring and control system.

[0016] The intelligent analysis model includes but is not limited to a liquid level analysis model, a current analysis model, a temperature analysis model, a vibration analysis model, an oil level video analysis model, a pipeline flow analysis model, and a pipeline pressure analysis model.

[0017] Furthermore, the specific steps of constructing the liquid level analysis model include:

[0018] Step a: Collect historical liquid level data, water inlet flow data, water outlet flow data, and weather data.

[0019] Step b: data preprocessing, including data cleaning, time synchronization, and unit unification;

[0020] Step c, algorithm steps,

[0021] 1) Initial condition setting: define the initial liquid level H0 and the time step Δt;

[0022] 2) Liquid level calculation: For each time step Δt, calculate the new liquid level:

[0023] Calculate the net flow: Net flow t = Q int -Q out , the net flow represents the net amount of water entering the reservoir at the tth time point;

[0024] Consider weather factors:

[0025] Rainfall: If P t > 0, the net flow increases P t ;

[0026] Evaporation: If E t > 0, the net flow decreases by E t ;

[0027] The updated net flow is: Net flow totalt =Net flow t+P t -E t ;

[0028] Calculate the level change: ΔH t represents the change in liquid level in the tth time step, and A is the bottom area of ​​the reservoir;

[0029] Update level: H t+1 =H t +ΔH t , H t+1 represents the liquid level at the t+1th time point;

[0030] By repeating the above steps, the liquid level change in the future (such as 24 hours, 48 ​​hours) can be predicted to obtain the model liquid level H model value;

[0031] Step d, early warning,

[0032] Set the safe liquid level range [H min , H max ], when the model liquid level H model When the value exceeds this range, the alarm system is triggered.

[0033] Furthermore, the construction of the current analysis model includes the following steps:

[0034] Step a: current collection,

[0035] The real-time current signal of the motor is measured by current sensors and current transformers to ensure accurate and stable data;

[0036] Step b: current signal processing, including denoising and filtering, the formula is:

[0037] i(t)=(x(t)*h(t));

[0038] Among them, x(t) is the original current signal, h(t) is the filter impulse response, and i(t) is the denoised signal;

[0039] Step c: feature value extraction,

[0040] Extract the effective value of the current signal I RMS , the formula is:

[0041]

[0042] Among them, i(t) is the instantaneous current, T is the period; the effective value reflects the average power of the motor current;

[0043] Peak: Formula I peak =max(i(t)) The peak value represents the maximum instantaneous value of the current and is used to detect overload conditions;

[0044] Mean: Formula The average value is used to evaluate the average load on the motor;

[0045] Harmonic content: The spectrum of the current signal is analyzed by Fourier transform to extract the harmonic components. The harmonic content reflects the degree of distortion of the current waveform.

[0046] Step d: Fault diagnosis,

[0047] Threshold comparison method: involves the comparison of the preset threshold value with the actual feature value;

[0048] Pattern recognition method: Use machine learning algorithms to establish a fault pattern library, match real-time feature parameters with the patterns in the library, and achieve fault diagnosis.

[0049] Furthermore, a temperature analysis model is constructed to predict the temperature distribution and change trend of the motor during operation through mathematical modeling and numerical calculation. The specific steps include:

[0050] Step a: Calculate the heat source loss, including but not limited to the copper loss and iron loss inside the motor. The calculation formula is:

[0051] Q=P cu +P Fe ,

[0052] Where P cu is the copper loss, P cu =I 2 R, where I is the current and R is the winding resistance

[0053] Q Fe is the iron loss, P Fe =k1f2B 2 V, where k1 is the empirical coefficient, f is the frequency, B is the magnetic flux density, and V is the core volume;

[0054] Step b: Take the heat source loss as the heat source for temperature field calculation and solve the temperature field. The formula is:

[0055]

[0056] Where k is the thermal conductivity, T is the temperature to be determined, and Q is the heat source loss intensity;

[0057] Step c, perform heat dissipation analysis and set boundary conditions;

[0058] Step d, using finite element method or finite volume method to solve the temperature field equation, and obtain the temperature distribution of different heat dissipation methods relative to the inside of the motor;

[0059] Step e: predict the temperature distribution and change trend of the water pump motor under different working conditions, monitor the real-time temperature status of the motor, and detect and alarm in time.

[0060] Furthermore, the specific steps of constructing the vibration analysis model include:

[0061] Step a: signal acquisition,

[0062] The vibration sensor is used to continuously collect the vibration signal of the water pump during operation and record it as a time series x(t).

[0063] Step b: Data cleaning, the formula is:

[0064] x(t)=(y*h)(t);

[0065] Where y(t) is the original signal, h(t) is the impulse response of the low-pass filter, and x(t) is the filtered signal;

[0066] Step c, feature extraction, performing time domain, frequency domain, and time-frequency domain analysis on the preprocessed vibration signal to extract time domain features, frequency domain features, and time-frequency features;

[0067] Step d, constructing a support vector machine model, a forest regression model, and a long short-term memory network model to perform fault diagnosis and classification during the operation of the water pump;

[0068] Step e, early warning, real-time analysis of the vibration signal of the water pump. When abnormal vibration is detected, the system sends out an early warning signal to help maintenance personnel discover and handle potential faults in a timely manner.

[0069] Furthermore, the specific steps of constructing the oil level video analysis model include:

[0070] Step a, collecting an oil level image, and preprocessing the image, wherein the preprocessing includes graying, denoising, and contrast enhancement;

[0071] Step b: feature extraction, using edge detection algorithm to extract oil level boundary features;

[0072] Step c: after extracting the oil level boundary features, construct a support vector machine model and a convolutional neural network model to perform fault diagnosis classification and determine the oil level height;

[0073] Step d: using threshold comparison and condition judgment methods to automatically trigger an early warning or alarm mechanism.

[0074] Furthermore, the specific steps of constructing the pipeline flow analysis model include:

[0075] Step a: Data collection, including historical flow data, meteorological data, and equipment operating parameters of the water intake station;

[0076] Step b: data preprocessing, including data cleaning, data standardization and data conversion;

[0077] Step c: feature extraction: extract features useful for prediction tasks from the original data, and use PCA and t-SNE methods to reduce the number of features and reduce the complexity of the model;

[0078] Step d: Based on the extracted features, a linear regression model, a convolutional neural network or a recurrent neural network model is constructed to obtain prediction results and performance indicators;

[0079] Step e: Using the threshold comparison method, an alarm is triggered when the model performance is lower than a specific threshold or an anomaly is detected.

[0080] Furthermore, the specific steps of constructing the pipeline pressure analysis model include:

[0081] Step a, data collection: basic information of water intake station, pressure detection data, pump station operation status, water flow, pipeline friction coefficient, connection method, and valve and pump station location information, etc.

[0082] Step b, data preprocessing: including data cleaning, feature selection and feature extraction;

[0083] Step c: Based on the extracted features, a multivariate linear regression model or a neural network model is used to monitor the pipeline pressure. The model construction formula is:

[0084] P=β0+β1Q+β2D+β3L+β4ρ+∈,

[0085] Where P is pressure, Q is flow rate, D is pipe diameter, L is pipe length, ρ is fluid density, β is regression coefficient, and ∈ is error term;

[0086] Step d: Using the threshold comparison method, when the monitoring result is lower than a specific threshold or an abnormality is detected, an alarm is triggered.

[0087] Compared with the prior art, the beneficial effects of the present invention are:

[0088] (1) Comprehensive real-time monitoring. The present invention realizes comprehensive real-time monitoring of key components such as water pumps, valves, and pipelines in the water intake station, ensuring comprehensive coverage of the equipment operation status. At the same time, the system can obtain and record key operating parameters such as water pressure, flow, and liquid level in real time, providing an accurate data basis for subsequent data analysis and control decisions;

[0089] (2) Intelligent fault warning and diagnosis. The system's built-in intelligent analysis model can conduct in-depth mining and intelligent analysis of real-time monitored data. By comparing historical data and identifying abnormal patterns, it can promptly discover potential equipment failures. Once an abnormality or fault warning signal is detected, the system can immediately send an alarm to the operation and maintenance personnel and provide the possible cause of the fault and preliminary diagnosis results, helping the operation and maintenance personnel to quickly locate the problem and reduce troubleshooting time.

[0090] (3) Automatic optimization and adjustment. The present invention can automatically fine-tune and optimize the water intake station equipment based on real-time monitoring data and intelligent analysis results, which not only improves the operating efficiency of the equipment, but also helps to extend the service life of the equipment and reduce maintenance costs;

[0091] (4) Improve production stability and efficiency. The present invention can ensure the stable operation of water intake station equipment through real-time monitoring and intelligent control, avoiding production interruption or efficiency reduction caused by equipment failure. At the same time, the system can also dynamically adjust the operating status of the equipment according to production needs to ensure the continuity and efficiency of the alumina production process;

[0092] (5) Reduce energy consumption and costs. The system can reduce the energy consumption of equipment as much as possible while meeting production needs by intelligently adjusting the operating status of the equipment. At the same time, by timely warning and diagnosing equipment failures, additional costs such as downtime and maintenance caused by failures are avoided, which helps to reduce overall operating costs;

[0093] (6) Friendly user interaction experience. The monitoring and control system involved in the present invention provides an intuitive and easy-to-use user interaction interface, which makes it convenient for operation and maintenance personnel to view the real-time operating status of the equipment, receive alarm information, set parameters and perform remote control operations. The interface design focuses on user experience, allowing operation and maintenance personnel to quickly get started and become proficient in operating the system, thereby improving work efficiency and satisfaction.

[0094] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0096] Figure 1 A flowchart of the equipment operation monitoring and control of an alumina water intake station provided in one embodiment of the present invention.

[0097] In the figure: 1. Water tank; 2. Water pump; 3. Water pump outlet pipe; 4. Electric valve; 5. Liquid level height sensor; 6. Water pump current sensor; 7. Water intake controller; 8. Water pump motor temperature sensor; 9. Water pump motor vibration sensor; 10. Water pump motor oil level monitoring sensor; 11. Water pump outlet pipe pressure gauge; 12. Water pump outlet pipe flow meter; 13. Water valve controller; 14. Industrial data acquisition gateway; 15. Industrial switch; 16. Server 16; 17. Equipment operation monitoring and control system; 18. Liquid level analysis model; 19. Current analysis model; 20. Temperature analysis model; 21. Vibration analysis model; 22. Oil level video analysis model; 23. Pipeline flow analysis model; 24. Pipeline pressure analysis model. DETAILED DESCRIPTION

[0098] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0099] like Figure 1 As shown, the first embodiment of the present invention provides an alumina water intake station equipment operation monitoring and control method, which includes the following steps:

[0100] s1. The sensor network collects the operating parameters of key locations of water intake station equipment in real time;

[0101] s2. The industrial data acquisition gateway is responsible for receiving the data collected by the sensor network and performing preliminary processing;

[0102] s3. Upload the processed data to the data analysis control module, which conducts in-depth analysis on the received data to determine whether the operating status of the equipment is normal.

[0103] Yes, to maintain stable operation of the equipment;

[0104] If no, the data analysis control module will immediately generate an alarm message and notify the user through the user interface. At the same time, the data analysis control module will also generate control instructions based on the analysis results and send them to the control execution module. The control execution module will intelligently control the water station equipment according to the received control instructions to achieve automatic optimization and adjustment of the equipment. For example, the valve opening can be automatically adjusted according to the changes in water pressure and flow, and the working frequency of the water pump can be automatically adjusted according to the changes in its efficiency.

[0105] In step s1, the key positions of the alumina water intake station equipment include but are not limited to the water pool 1, water pump 2, and electric valve 4. The water pool 1 is connected to various alumina production equipment through a water supply system. The water supply system includes a water pump outlet pipe 3 and a water pump 2 and an electric valve 4 arranged on the water pump outlet pipe. A liquid level sensor 5 is installed in the water pool 1, and the water pump 2 is connected to a water pump current sensor 6, a water pump motor temperature sensor 8, a water pump motor vibration sensor 9, a water pump motor oil level monitoring sensor 10 and a water intake controller 7. The water intake controller 7 is used to realize water pump operation control, water pump stop control, and water pump remote control.

[0106] A water pump outlet pipe pressure gauge 11 and a water pump outlet pipe flow meter 12 are installed on the water pump outlet pipe 3. The electric valve 4 is connected to a water valve controller 13, which is used for electric valve operation control, electric valve stop control, electric valve flow adjustment control, and electric valve remote control input.

[0107] The liquid level sensor 5, the water pump current sensor 6, the water pump motor temperature sensor 8, the water pump motor vibration sensor 9, the water pump motor oil level monitoring sensor 10, the water intake controller 7, the water pump outlet pipe pressure gauge 11, the water pump outlet pipe flow meter 12, and the water valve controller 13 are all connected to the industrial data acquisition gateway 14. The data analysis control module includes an intelligent analysis model set on the server 16, and the industrial data acquisition gateway 14 is connected to the server 16 through the industrial switch 15.

[0108] In this embodiment, the relevant data of the water pool 1 includes but is not limited to historical liquid level data, water inlet flow data, water outlet flow data, and weather data; the relevant data of the water pump 2 includes but is not limited to the current data of the water pump motor, that is, the current sensor and the current transformer measure the real-time current of the motor; the temperature data of the water pump motor; the vibration signal of the water pump when it is running, recorded as the time series x(t), this step is the basis of the data, ensuring that we obtain the real vibration situation of the water pump when it is running; the oil level monitoring image data of the water pump motor;

[0109] Basic information of the water pump outlet pipe 3 and the electric valve 4 includes but is not limited to the length, diameter, material, pressure detection data of the pipe, operation status of the pump station, water flow, pipe friction coefficient, connection method, and valve and pump station location information; historical flow data, meteorological data, equipment operation parameters, etc. of the water intake station;

[0110] In step s2, the collected data is sent to the data acquisition module, and the data acquisition module performs preliminary processing on the received data, such as data filtering, amplification, conversion, etc.

[0111] In step s3, the data is transmitted to the server 16 by accessing the industrial switch 15. The server 16 stores the data in its own database and pushes the data to the intelligent analysis model and the equipment operation monitoring and control system 17. The intelligent analysis model uses the intelligent analysis model to analyze and judge the real-time operation status, operation trend, operation history comparison and other aspects of each device according to the data source, and identifies the operation status and potential faults of each device through the equipment operation monitoring and control system.

[0112] Specifically, the intelligent analysis model includes but is not limited to a liquid level analysis model 18, a current analysis model 19, a temperature analysis model 20, a vibration analysis model 21, an oil level video analysis model 22, a pipeline flow analysis model 23, and a pipeline pressure analysis model 24. Each model analyzes and judges the real-time operation status, operation trend, operation history comparison, etc. of each device, identifies the operation status and potential faults of the device, and once a problem is found, the module will analyze the abnormal liquid level data, water pump current, water pump motor temperature, water pump motor vibration, water pump oil level, water pump outlet pipe flow, and water pump outlet pipe pressure, and output them to the equipment operation monitoring and control system, and immediately generate control instructions, and adjust and optimize the equipment through water pump operation control, water pump stop control, water pump remote control, electric valve operation control, electric valve stop control, and electric valve flow regulation control. The equipment operation monitoring and control system 17 will display the abnormal alarm, liquid level height, water pump current, water pump motor temperature, water pump motor vibration, water pump oil level, water pump outlet pipe pressure, water pump outlet pipe flow, electric valve power input and other information pushed in real time. It can also remotely control the equipment through remote start and stop of the water pump, remote switch of the regulating valve, and remote adjustment of the regulating valve. It provides a friendly human-computer interaction interface, which is convenient for users to view the operating status of the equipment, receive alarm information, set parameters and perform remote control and other operations.

[0113] The liquid level analysis model 18 is constructed to monitor the liquid level of the pool in real time and plot the change trend of the liquid level in the past and future. When the liquid level exceeds the safe range, an alarm message is output. The specific construction steps of the liquid level analysis model 18 include:

[0114] (1) Model input

[0115] Input historical liquid level data, water inlet flow data, water outlet flow data, weather data, etc.

[0116] 1) Historical liquid level data is expressed as: H hist =[H1, H2, ..., H n ], where Hn represents the liquid level at the nth time point;

[0117] 2) The water inlet flow data is expressed as: Q in =[Q in1 , Q in2 , ..., Q in n ], where Qinn represents the water inflow at the nth time point;

[0118] 3) The water flow rate data is expressed as: Q out =[Q out1 , Q out2 , ..., Q out n], where Qoutn represents the outflow rate at the nth time point;

[0119] 4) Weather data (optional): including rainfall P = [P1, P2, ..., P n ] and evaporation E = [E1, E2, ..., E n ], where Pn represents the rainfall at the nth time point; En represents the evaporation at the nth time point.

[0120] (2) Data preprocessing

[0121] 1) Data cleaning: remove outliers and fill in missing values;

[0122] 2) Time synchronization: ensure that the timestamps of all input data are consistent;

[0123] 3) Unit consistency: Ensure that all data are in consistent units, such as liquid level in meters (m), flow in cubic meters per hour (m 3 / h) as the unit.

[0124] (3) Algorithm steps

[0125] 1) Initial condition setting: define the initial liquid level H0 and the time step Δt (for example, 1 hour)

[0126] 2) Liquid level calculation: For each time step Δt, calculate the new liquid level:

[0127] Calculate Net Flow: Net Flow t =Q in t -Q out t Net flow represents the net amount of water entering the reservoir at time point t

[0128] Consider weather factors:

[0129] Rainfall: If P t > 0, the net flow increases P t

[0130] Evaporation: If E t > 0, the net flow decreases by E t

[0131] The updated net flow is: Net flow total t =Net flow t+P t -E t

[0132] Calculate the level change: ΔH t represents the change in liquid level in the tth time step, and A is the bottom area of ​​the reservoir;

[0133] Update level: Ht+1 =H t +ΔH t , H t+1 represents the liquid level at the t+1th time point;

[0134] By repeating the above steps, the liquid level change in the future (such as 24 hours, 48 ​​hours) can be predicted to obtain the model liquid level H model value.

[0135] (4) Model verification

[0136] Use historical data to backtest and compare the liquid level H calculated by the model model Compared with the actual recorded liquid level H actual , evaluate the accuracy of the model.

[0137] (5) Alarm system

[0138] Set the safe liquid level range [H min , H max ], when the model liquid level H model When the value exceeds this range, the alarm system is triggered.

[0139] By collecting, processing and analyzing the current data of the water pump motor, a current analysis model 19 is constructed to evaluate its working status and performance. The construction of the current analysis model 19 includes the following steps:

[0140] (1) Current collection

[0141] The real-time current signal of the motor is measured by current sensors and current transformers to ensure accurate and stable data.

[0142] (2) Current signal processing

[0143] Denoising: Use filtering techniques, such as low-pass filters, to remove high-frequency noise;

[0144] i(t)=(x(t)*h(t));

[0145] Among them, x(t) is the original current signal, h(t) is the filter impulse response, and i(t) is the denoised signal;

[0146] Filtering: Extract the current signal in a specific frequency range by selecting a bandpass or lowpass filter, where the filter selection should be based on the motor operating frequency and noise characteristics.

[0147] (3) Feature value extraction

[0148] Extract the effective value of the current signal I RMS , the formula is:

[0149]

[0150] Where i(t) is the instantaneous current, T is the period; the effective value reflects the average power of the motor current;

[0151] Peak value: The formula is: I peak =max(i(t)) The peak value represents the maximum instantaneous value of the current and is used to detect overload conditions;

[0152] Average: The formula is: The average value is used to evaluate the average load on the motor;

[0153] Harmonic content: The spectrum of the current signal is analyzed through Fourier transform to extract the harmonic components. The harmonic content reflects the degree of distortion of the current waveform.

[0154] (4) Fault diagnosis

[0155] Threshold comparison method: involves the comparison of the preset threshold value with the actual feature value;

[0156] Pattern recognition method: Use machine learning algorithms (such as SVM and neural networks) to establish a fault pattern library, match real-time feature parameters with the patterns in the library, and realize fault diagnosis.

[0157] (5) Model application

[0158] Real-time monitoring: The model can be used to monitor the motor current in real time and understand the working status of the motor;

[0159] Early warning and alarm: When the current characteristic parameters are abnormal, the model issues an early warning or alarm to prevent potential failures;

[0160] Fault diagnosis: accurately determine the type and location of motor faults and guide maintenance work;

[0161] Efficiency evaluation: Evaluate motor operating efficiency through current analysis and make optimization suggestions.

[0162] A temperature analysis model 20 is constructed to predict the temperature distribution and change trend of the motor during operation through mathematical modeling and numerical calculation. The specific construction steps include:

[0163] (1) Heat source calculation

[0164] The heat sources mainly include copper loss, iron loss, mechanical loss, etc. inside the motor. These losses will be converted into heat energy during the operation of the motor, causing the motor temperature to rise. The heat source calculation usually uses the following formula:

[0165] Q=P cu +P Fe ,

[0166] Where P cu is the copper loss, Pcu =I 2 R, where I is the current and R is the winding resistance;

[0167] Q Fe Iron loss is a complex calculation method and is usually estimated using experimental data or empirical formulas: Fe =k1f 2 B 2 V, where k1 is the empirical coefficient, f is the frequency, B is the magnetic flux density, and V is the core volume.

[0168] (2) Temperature field solution

[0169] The temperature field is usually numerically calculated using the finite element method or the finite volume method, which uses the heat source loss as the heat source for temperature field calculation and solves the temperature field. The formula is:

[0170]

[0171] Among them, k is the thermal conductivity, T is the temperature to be determined, and Q is the intensity of the heat source. This formula describes the law of conservation of energy in the process of heat conduction.

[0172] (3) Perform heat dissipation analysis and set boundary conditions

[0173] Heat dissipation analysis includes but is not limited to convection heat dissipation and radiation heat dissipation on the motor surface. Convection heat dissipation is related to factors such as the flow rate, temperature, and heat dissipation coefficient of the cooling medium; radiation heat dissipation is related to factors such as the motor surface temperature, ambient temperature, and radiation coefficient. Heat dissipation analysis uses the following formula:

[0174] Convection cooling: Q conv =hA(TT amb ),

[0175] Where h is the convection heat dissipation coefficient, A is the heat dissipation area, T is the motor surface temperature, T amb is the ambient temperature;

[0176] Radiative cooling:

[0177] Where ∈ is the emissivity, σ is the Stefan-Boltzmann constant, A is the heat dissipation area, T is the motor surface temperature (Kelvin), T amb is the ambient temperature (Kelvin);

[0178] Set boundary conditions: Set boundary conditions according to the actual working conditions of the motor, such as inlet temperature, outlet temperature, heat dissipation coefficient, etc.

[0179] (4) Solving the temperature field: Use the finite element method or finite volume method to solve the temperature field equation and obtain the temperature distribution inside the motor.

[0180] (5) Model application

[0181] Predicting the temperature distribution and change trend of the water pump motor under different working conditions provides an important reference for the design, optimization and operation of the motor. At the same time, the model can also be used to monitor the real-time temperature status of the motor, timely detect and alarm potential overheating problems, and ensure the safe operation of the motor.

[0182] Construct a vibration analysis model 21 to analyze the vibration signal of the water pump in real time. The specific steps of constructing the vibration analysis model 21 include:

[0183] (1) Signal acquisition

[0184] The vibration sensor is used to continuously collect vibration signals during the operation of the water pump and record them as a time series x(t) to ensure that the actual vibration conditions of the water pump are obtained.

[0185] (2) Data cleaning

[0186] Filtering technology is used to remove noise and irrelevant components in the signal. The formula is:

[0187] x(t)=(y*h)(t),

[0188] Among them, y(t) is the original signal, h(t) is the impulse response of the low-pass filter (or called the kernel of the filter), and x(t) is the filtered signal.

[0189] (3) Feature extraction: perform time domain, frequency domain, and time-frequency domain analysis on the preprocessed vibration signal to extract time domain features, frequency domain features, and time-frequency features;

[0190] 1) Time domain characteristics:

[0191] Mean (\mu): reflects the average level of the signal, that is, the central trend of the signal. The formula is:

[0192]

[0193] In the formula, mu is the mean value of the signal. All values ​​of the signal in the time period T are integrated and divided by T to obtain the average value of the signal;

[0194] Variance ((\sigma^2)): reflects the discreteness of the signal, that is, the distribution of signal values ​​around the mean. The formula is:

[0195] Variance is the average of the squares of the differences between the signal value and the mean, which measures the degree of fluctuation of the signal value;

[0196] Peak value (P): The maximum absolute value of the signal, indicating the maximum amplitude of the signal during the observation period. The formula is:

[0197] The peak value directly reflects the maximum strength or amplitude of the signal.

[0198] 2) Frequency domain characteristics

[0199] Perform Fourier transform on the signal to convert it from the time domain to the frequency domain to analyze the frequency components of the signal. The formula is:

[0200]

[0201] Where X(ω) is the frequency domain representation of the signal, ω is the angular frequency, and x(t) is the original time domain signal. The Fourier transform decomposes the signal into a superposition of sine waves of different frequencies, thereby revealing the frequency characteristics of the signal.

[0202] 3) Time-frequency characteristics

[0203] Use wavelet transform to obtain the energy distribution of the signal at different times and frequencies. Wavelet transform is a time-frequency analysis method that can provide signal information in both the time domain and the frequency domain. The formula is:

[0204]

[0205] Among them, ψ(t) is the wavelet function, which determines the time-frequency resolution of the transformation; a is the scale parameter, which controls the expansion and contraction of the wavelet function; b is the translation parameter, which controls the position of the wavelet function in the time domain.

[0206] (4) Construct a support vector machine model, a random forest regression model, or a long short-term memory network model to perform fault diagnosis and classification during pump operation;

[0207] The extracted time domain features, frequency domain features and time-frequency features are marked as training data, and then the marked training data is input into a support vector machine model, a random forest regression model or a long short-term memory network model for training and optimization.

[0208] Support Vector Machine Model (SVM)

[0209] The decision function of the support vector machine model is:

[0210]

[0211] where α i is the Lagrange multiplier, which reflects the influence of each support vector on the decision function; K(x i , x) is the kernel function, which calculates the input sample x and the support vector x i The similarity between them; b is the bias term, which determines the intercept of the decision function.

[0212] Random Forest Regression Model (RandomForest)

[0213] The random forest regression model is used to build multiple decision trees to improve the stability and accuracy of the model. Each tree is built based on a different training subset and feature subset, and classified through a voting mechanism. The advantage of random forest is that it can handle high-dimensional data and is robust to noise and outliers in the training set.

[0214] Long Short-Term Memory Network Model (LSTM)

[0215] The long short-term memory network model solves the gradient vanishing and gradient exploding problems that are prone to occur in the traditional RNN training process by introducing mechanisms such as forget gate, input gate, and output gate. The update formula of LSTM includes:

[0216] Forget gate: f t =σ(W f ·[h t-1 , x t ]+b f ), which determines how much of the cell state at the previous moment should be retained;

[0217] Input gate: i t =σ(W i ·[h t-1 , x t ]+b i ), which determines how much of the current input information should be added to the cell state;

[0218] Output gate: o t =σ(W o ·h t-1 , x t ]+b o ), which determines how much the cell state at the current moment should output;

[0219] Cell state update: C t =f t ·C t-1 +i t tanh(W C ·[h t-1 , x t ]+b C ), which integrates the information of the forget gate and the input gate and updates the cell state;

[0220] Hide status update:h t =o t tanh(C t ), which updates the hidden state based on the information from the cell state and the output gate.

[0221] Model training and validation

[0222] Data partitioning: Divide the dataset into training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters and select the best model, and the test set is used to evaluate the final performance of the model.

[0223] Model training: Use the training set to train the model and minimize the loss function through optimization algorithms (such as gradient descent). The loss function measures the difference between the model's predicted value and the true value, and the optimization algorithm reduces this difference by continuously adjusting the model parameters.

[0224] Model validation and testing: Use validation sets and test sets to evaluate the performance of the model, such as accuracy, recall, F1 score, etc. These indicators reflect the performance of the model in the classification task and help us understand the advantages and disadvantages of the model.

[0225] Model optimization: Adjust the feature extraction method, algorithm parameters, or model structure based on the evaluation results to improve the performance and stability of the model.

[0226] (5) Model application

[0227] The new vibration signal is classified through the trained model, and the extracted feature data is input into the model to obtain the fault classification result. When abnormal vibration is detected, the system sends out a warning signal to help maintenance personnel find and deal with potential faults in time, ensuring the safe and stable operation of the water intake station.

[0228] Construct an oil level video analysis model 22 to monitor the oil level of the water pump in real time. The specific construction steps are as follows:

[0229] (1) Collect oil level images and pre-process the images captured by the camera to improve image quality and recognition accuracy.

[0230] Image preprocessing:

[0231] Grayscale: Convert a color image to a grayscale image to reduce the amount of data while retaining the image's contour and texture information. The formula is: textGray(i,j)=0.299×R(i,j)+0.587×G(i,j)+0.114×B(i,j); where R(i,j), G(i,j), and B(i,j) are the red, green, and blue component values ​​of the color image at position (i,j), respectively, and Gray(i,j) is the converted grayscale value;

[0232] Denoising: Remove noise points in the image through median filtering, Gaussian filtering and other methods;

[0233] Enhance contrast: Enhance the contrast of the image to make the oil level boundary clearer through methods such as histogram equalization. Histogram equalization is a nonlinear stretching technique that redistributes image pixel values ​​so that they are more evenly distributed across the grayscale.

[0234] (2) Feature extraction: edge detection algorithm is used to extract oil level boundary features.

[0235] Edge detection: Use edge detection algorithms such as the Sobel operator and the Canny operator to extract the oil level boundary. These algorithms detect edges by calculating the magnitude and direction of the image grayscale gradient. The gradient calculation formula of the Sobel operator in the x and y directions is:

[0236] Gx=[-1,0,1;-2,0,2;-1,0,1]*I

[0237] Gy=[-1,-2,-1;0,0,0;1,2,1]*I

[0238] Where I is the input image, Gx and Gy are the gradient images in the x and y directions respectively. The edge strength can be obtained by calculating the Euclidean norm of Gx and Gy:

[0239] Morphological transformation: The edge detection results are further processed using morphological transformation algorithms such as corrosion, dilation, opening, closing, etc. to remove noise points, fill holes, or separate connected objects. These algorithms usually involve set operations such as union, intersection, difference, etc.

[0240] Oil level calculation: After the oil level boundary is extracted, the oil level can be determined by calculating the distance between the boundary and the preset reference line. This usually involves geometric transformation and coordinate calculation.

[0241] (3) After extracting the oil level boundary features, a support vector machine model and a convolutional neural network model are constructed to perform fault diagnosis classification and determine the oil level height;

[0242] Support Vector Machine Model (Support Vector Machine, SVM)

[0243] SVM is a supervised learning algorithm used for classification and regression problems. In oil level identification, SVM can be used to distinguish between oil level areas and non-oil level areas. The core idea of ​​SVM is to find an optimal hyperplane that maximizes the sample intervals of different categories.

[0244] The decision function of SVM can be expressed as: f(x)=sign(w T x+b),

[0245] Among them, w is the weight vector, b is the bias term, and x is the input feature vector.

[0246] The training process of SVM involves solving the following optimization problem:

[0247]

[0248] texts.t.y i (w T x i +b)≥1-ξ i ,ξ i ≥0, i=1, 2, ..., N

[0249] Among them, C is the regularization parameter and ξi is the slack variable used to deal with the case of linear inseparability.

[0250] Convolutional Neural Network Model (CNN)

[0251] CNN is a deep learning model that is particularly suitable for image processing. In oil level recognition, CNN can be used to automatically extract oil level features in images and accurately identify the oil level height.

[0252] The basic structure of CNN includes convolutional layer, pooling layer, fully connected layer and activation function, etc. The convolutional layer extracts image features through convolution operation, the pooling layer reduces the amount of data and retains important features through downsampling operation, and the fully connected layer maps the extracted features to the output space.

[0253] The formula for the convolution operation can be expressed as:

[0254]

[0255] in, is the jth feature map of the lth layer, is the convolution kernel connecting the i-th feature map of the l-1th layer and the j-th feature map of the lth layer, is the bias term, f is the activation function (such as ReLU, sigmoid, etc.), and M is the set of input feature maps.

[0256] (4) Early warning and early warning mechanism

[0257] According to the extracted characteristic data, it is input into the model to determine the oil level height, and logical operations such as threshold comparison and condition judgment are used to realize early warning. For example, a minimum oil level threshold and a maximum oil level threshold can be preset. When the detected oil level is lower than the minimum threshold or higher than the maximum threshold, the system automatically triggers the early warning or alarm mechanism.

[0258] The specific construction steps of the pipeline flow analysis model 23 include:

[0259] (1) Data Collection

[0260] Collect historical flow data, meteorological data, equipment operating parameters, etc. of the water intake station.

[0261] (2) Data preprocessing

[0262] Data cleaning: remove invalid, erroneous or duplicate data;

[0263] Data standardization: Scale the data to the same range, such as using normalization or standardization (\(X_{\text{standardized}}=\frac{X-\mu}{\sigma}\), where \(X\) is the original data point, (\mu) is the mean of the data, and \(\sigma) is the standard deviation of the data);

[0264] Data conversion: convert non-numeric data into numerical data, such as converting weather conditions into corresponding numerical codes (One-Hot Encoding: For categorical features, One-Hot encoding can be used to convert categories into binary vectors; formula: \(x_i=[0,0,...,1,...,0]\), where the \(j\)th bit is 1, indicating that the \(i\)th instance belongs to the \(j\)th category);

[0265] Data binning: Binning continuous variables, such as dividing traffic data into three levels: high, medium, and low.

[0266] (3) Feature selection: extract features useful for prediction tasks from the original data, and use PCA and t-SNE methods to reduce the number of features and reduce the complexity of the model.

[0267] Extract features useful for prediction tasks from raw data; use correlation coefficients to evaluate the relationship between features and target variables, formula:

[0268] \(r_{X,Y}=\frac{\sum{(X-\mu_X)(Y-\mu_Y)}}{\sqrt{\sum{(X-\mu_X)^2}\sum{(Y-\mu_Y)^2}}}\)

[0269] -\(r_{X,Y}\) is the correlation coefficient between the feature \(X\) and the target variable \(Y\);

[0270] Feature construction: create new features, such as time series features (such as trend, seasonality, periodicity) (time series decomposition: for time series data, additive decomposition or multiplicative decomposition can be used to extract trends, seasonality, and residuals; formula: \(Y=T+S+R\) (additive decomposition), or \(Y=\frac{T\cdotS}{R}\) (multiplicative decomposition));

[0271] Feature dimensionality reduction: Use methods such as PCA and t-SNE to reduce the number of features and reduce the complexity of the model.

[0272] (4) Based on the extracted features, a linear regression model, a convolutional neural network or a recurrent neural network model is constructed to perform fault diagnosis classification.

[0273] Linear regression model: Use the least squares method to fit the linear model, formula:

[0274] \(\min_{\theta}\sum{(h_{\theta}(x)-y)^2}\)where \(h_{\theta}(x)\) is the predicted value and \(y\) is the actual traffic value.

[0275] Convolutional neural network (CNN) or recurrent neural network (RNN) models are used to capture complex patterns in time series data; back propagation algorithms are used to optimize model parameters.

[0276] formula:

[0277] \(\delta_{l}^{i}=\frac{\partial{C}}{\partial{z_{l}^{i}}}\cdot\frac{\partial{z_{l}^{i}}}}{\partial{w_{l-1}^{i}}}); This formula is part of the backpropagation algorithm and is used to calculate the gradient of the neural network to update the weights \(w_{l-1}^{i}\) and biases \(b_{l}^{i}\). \(C\) is the loss function and \(z_{l}^{i}\) is the activation value of the neuron.

[0278] Data segmentation: Divide the data set into training set, validation set and test set. Usually 70% of the data is used as the training set, 15% of the data is used as the validation set, and 15% of the data is used as the test set.

[0279] Cross-validation: Divide the dataset into K subsets and perform K training and validation cycles.

[0280] (5) Model evaluation

[0281] Performance indicators: Use mean square error (MSE) to evaluate the prediction accuracy of the model. Formula: \(MSE=\frac{1}{N}\sum{(y_i-\hat{y}_i)^2}\); MSE is used to measure the average square error between the model's predicted value \(\hat{y}_i\) and the actual value \(y_i\). \(N\) is the number of data points.

[0282] Coefficient of determination (R 2 ): used to evaluate the goodness of fit of the model; formula:

[0283] \(R^2=1-\frac{\sum{(y_i-\hat{y}_i)^2}}{\sum{(y_i-\bar{y})^2}}\); R 2 Indicates how well the model fits the data, and its value is between 0 and 1. \(\bar{y}\) is the mean of the actual values.

[0284] (6) Model application

[0285] Based on the extracted feature data, it is input into the model to obtain the prediction results and performance indicators.

[0286] Performance monitoring: Regularly monitor the model’s predictive performance to ensure that the model has not experienced performance degradation.

[0287] Anomaly detection: Monitor the model’s predicted output to detect any anomalies or unusual traffic patterns.

[0288] Data quality check: Ensure that there are no issues with the quality of the data used for model training.

[0289] Model stability: Check if the model is stable and shows no signs of overfitting or underfitting.

[0290] Logging: Record the model’s input data, prediction results, and performance metrics.

[0291] Alarm system: Alarms are triggered when model performance falls below a certain threshold or anomalies are detected.

[0292] (7) Model update

[0293] Collect new traffic data, including historical traffic data and recent traffic data.

[0294] Preprocess new data, including cleaning, standardization, and feature engineering.

[0295] Retrain the model using the new dataset.

[0296] Evaluate the performance of the model on the validation set and tune hyperparameters as needed.

[0297] Use the test set to evaluate the performance of the final model to ensure its accuracy.

[0298] Constructing the pipeline pressure analysis model 24, including the model establishment and training process of the operation formula, specifically includes the following steps:

[0299] (1) Data collection: basic information of water intake stations (pipeline length, diameter, material), pressure detection data, pump station operation status, water flow, pipeline friction coefficient, connection method, and valve and pump station location information.

[0300] (2) Data preprocessing, including data cleaning, feature selection and feature extraction,

[0301] Data cleaning: remove outliers (such as values ​​that are significantly deviated from the normal range) and fill in missing values ​​(you can use interpolation (value_column, method = 'linear'), median (value_column) or mean (value_column)); Feature selection: select features that are closely related to pipeline pressure, such as flow rate, pipeline diameter, length, etc.; Feature extraction: transform the original features to extract more useful information. For example, you can calculate the flow rate (flow rate / pipeline cross-sectional area) as a new feature; Algorithm formula:

[0302] Bernoulli equation:

[0303] This formula describes the relationship between pressure, velocity and height at any point when a fluid flows in a pipe. Among them, p is the pressure at a point in the fluid, ρ is the fluid density, v is the fluid velocity, g is the acceleration due to gravity, h is the height of the point, and C is a constant.

[0304] Darcy-Weisbach formula:

[0305] This formula is used to calculate the head loss along the way when the fluid flows in the pipe. Among them, ΔP is the head loss along the way, λ is the resistance coefficient along the way, L is the pipe length, D is the pipe diameter, ρ is the fluid density, and v is the average flow velocity of the fluid.

[0306] Flow rate and pressure relationship formula: Q = kP n

[0307] This formula is used to describe the relationship between flow and pressure. Among them, Q is the flow, P is the pressure, and k and n are constants obtained through experiments or simulations.

[0308] (3) Based on the extracted features, a multivariate linear regression model or a neural network model is used to monitor pipeline pressure.

[0309] For the multivariate linear regression model, an equation of the following form can be constructed: P = β0 + β1Q + β2D + β3L + β4ρ + ∈, which describes the relationship between pressure P and flow Q, pipe diameter D, pipe length L, fluid density ρ and other factors. β is the regression coefficient, and ∈ is the error term. For the neural network model, it is necessary to design the network structure, including the input layer, hidden layer and output layer, as well as the connection weights and biases between the layers.

[0310] Feature engineering: During the model building process, continue to perform work such as feature selection, feature extraction, and feature construction to improve the predictive performance of the model (for example, the flow rate (flow rate / pipe cross-sectional area) can be calculated as a new feature input into the model); create new features, such as lagged variables of time series, local features of pipelines, etc. (lagged variables: Feature_t = Value_{tn} (where n is the number of lag steps)).

[0311] Model training:

[0312] For the multivariate linear regression model, an optimization algorithm such as the least squares method can be used to solve the regression coefficient (y = Xw + b, where y is the dependent variable, X is the independent variable matrix, w is the weight vector, and b is the intercept.)

[0313] For neural network models, optimization algorithms such as backpropagation and gradient descent can be used to update network weights and biases. Where α is the learning rate and J(w) is the loss function. )

[0314] During the training process, it is necessary to set appropriate hyperparameters such as the number of iterations and learning rate to ensure that the model can converge to a better solution.

[0315] Model Evaluation:

[0316] The mean square error (MSE), root mean square error (RMSE), and coefficient of determination (R 2 ) and other indicators to evaluate the model performance. (Mean square error: MSE = \frac{1}{N}\sum_{i=1}^{N}(y_i-\hat{y}_i)^2; root mean square error: RMSE = \sqrt{MSE}; coefficient of determination: R2 = 1-\frac{SS_res}{SS_tot}, where SS_res is the residual sum of squares and SS_tot is the total sum of squares).

[0317] Model optimization: Use grid search, random search, or Bayesian optimization to tune model parameters (grid search: find the best parameter combination by traversing a predefined parameter grid; random search: randomly sample parameter combinations from the parameter space to find the best parameters).

[0318] Continuous monitoring and updates

[0319] Monitor the performance of the model to ensure its accuracy and stability; regularly retrain the model with new data (Model_Performance = \frac{Predicted_Value-Actual_Value}{Actual_Value}).

[0320] (4) Model application

[0321] Based on the extracted feature data, it is input into the model to obtain the prediction result. The threshold comparison method is adopted. When the model performance is lower than a specific threshold or an anomaly is detected, an alarm is triggered.

[0322] The above description is only a preferred embodiment of the present invention and does not constitute any form of confidentiality restriction on the present invention. Any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the scope of the technical solution of the present invention.

Claims

1. A method for monitoring and controlling the operation of equipment in an alumina water intake station, characterized in that: The following steps are involved: Step s1, the sensor network collects the operating parameters of the key positions of the water intake station equipment in real time; Step s2: The industrial data acquisition gateway is responsible for receiving the data collected by the sensor network and performing preliminary processing; Step s3: upload the processed data to the data analysis control module, which performs in-depth analysis on the received data to determine whether the operating status of the equipment is normal. Yes, to maintain stable operation of the equipment; No, the data analysis and control module will immediately generate an alarm message and notify the user through the user interaction interface. At the same time, the data analysis and control module will also generate control instructions based on the analysis results and send them to the control execution module. The control execution module will perform intelligent control of the water intake station equipment based on the received control instructions to achieve automatic optimization and adjustment of the equipment.

2. The alumina water intake station equipment operation monitoring and control method according to claim 1 is characterized in that: In step s1, the key positions of the alumina water intake station equipment include but are not limited to a water pool (1), a water pump (2), and an electric valve (4). The water pool (1) is connected to various alumina production equipment through a water supply system. The water supply system includes a water pump outlet pipe (3) and a water pump (2) and an electric valve (4) arranged on the water pump outlet pipe. A liquid level sensor (5) is installed in the water pool (1). The water pump (2) is connected to a water pump current sensor (6), a water pump motor temperature sensor (8), a water pump motor vibration sensor (9), a water pump motor oil level monitoring sensor (10) and a water intake controller (7). The water intake controller (7) is used to realize water pump operation control, water pump stop control, and water pump remote control. A water pump outlet pipe pressure gauge (11) and a water pump outlet pipe flow meter (12) are installed on the water pump outlet pipe (3). The electric valve (4) is connected to a water valve controller (13). The water valve controller (13) is used for electric valve operation control, electric valve stop control, electric valve flow regulation control, and electric valve remote control input.

3. The alumina water intake station equipment operation monitoring and control method according to claim 1 or 2, characterized in that: In step s2, preliminary data processing includes but is not limited to data filtering, amplification, conversion, and image processing; In step s3, the data is transmitted to the server (16) by accessing the industrial switch (15). The server (16) stores the data in its own database and pushes the data to the intelligent analysis model and the equipment operation monitoring and control system (17) respectively. The intelligent analysis model uses the intelligent analysis model to analyze and judge the real-time operation status, operation trend, and operation history comparison of each device according to the data source, and identifies the operation status and fault potential of each device through the equipment operation monitoring and control system (17); The intelligent analysis model includes but is not limited to a liquid level analysis model (18), a current analysis model (19), a temperature analysis model (20), a vibration analysis model (21), an oil level video analysis model (22), a pipeline flow analysis model (23), and a pipeline pressure analysis model (24).

4. The method for monitoring and controlling the operation of the alumina water intake station equipment according to claim 3, characterized in that: The specific steps of constructing the liquid level analysis model (18) include: Step a: Collect historical liquid level data, water inlet flow data, water outlet flow data, and weather data. Step b: data preprocessing, including data cleaning, time synchronization, and unit unification; Step c, algorithm steps, 1) Initial condition setting: define the initial liquid level H0 and the time step Δt; 2) Liquid level calculation: For each time step Δt, calculate the new liquid level: Calculate the net flow: Net flow t = Q int -Q outt , the net flow represents the net amount of water entering the reservoir at the tth time point; Consider weather factors: Rainfall: If P t > 0, the net flow increases P t ; Evaporation: If E t > 0, the net flow decreases by E t ; The updated net flow is: Net flow totalt = Net flow t + P t -E t ; Calculate the level change: ΔH t represents the change in liquid level in the tth time step, and A is the bottom area of ​​the reservoir; Update level: H t+1 =H t +ΔH t , H t+1 represents the liquid level at the t+1th time point; By repeating the above steps, the liquid level change in the future (such as 24 hours, 48 ​​hours) can be predicted to obtain the model liquid level H model value; Step d, early warning, Set the safe liquid level range [H min , H max ], when the model liquid level H model When the value exceeds this range, the alarm system is triggered.

5. The method for monitoring and controlling the operation of the alumina water intake station equipment according to claim 3, characterized in that: The construction of the current analysis model (19) comprises the following steps: Step a: current collection, The real-time current signal of the motor is measured by current sensors and current transformers to ensure accurate and stable data; Step b: current signal processing, including denoising and filtering, the formula is: i(t)=(x(t)*h(t)). Among them, x(t) is the original current signal, h(t) is the filter impulse response, and i(t) is the denoised signal; Step c: feature value extraction, Extract the effective value of the current signal I RMs , the formula is: Among them, i(t) is the instantaneous current, T is the period; the effective value reflects the average power of the motor current; Peak: Formula I peak =max(i(t)) The peak value represents the maximum instantaneous value of the current and is used to detect overload conditions; Mean: Formula The average value is used to evaluate the average load on the motor; Harmonic content: The spectrum of the current signal is analyzed by Fourier transform to extract the harmonic components. The harmonic content reflects the degree of distortion of the current waveform. Step d: Fault diagnosis, Threshold comparison method: involves the comparison of the preset threshold value with the actual feature value; Pattern recognition method: Use machine learning algorithms to establish a fault pattern library, match real-time feature parameters with the patterns in the library, and achieve fault diagnosis.

6. The method for monitoring and controlling the operation of the alumina water intake station equipment according to claim 3, characterized in that: A temperature analysis model (20) is constructed to predict the temperature distribution and change trend of the motor during operation through mathematical modeling and numerical calculation. The specific steps include: Step a: Calculate the heat source loss, including but not limited to the copper loss and iron loss inside the motor. The calculation formula is: Q=P cu +P Fe , Where P cu is the copper loss, P cu =I 2 R, where I is the current and R is the winding resistance Q Fe is the iron loss, P Fe =k1f 2 B 2 V, where k1 is the empirical coefficient, f is the frequency, B is the magnetic flux density, and V is the core volume; Step b: Take the heat source loss as the heat source for temperature field calculation and solve the temperature field. The formula is: Where k is the thermal conductivity, T is the temperature to be determined, and Q is the heat source loss intensity; Step c, perform heat dissipation analysis and set boundary conditions; Step d, using finite element method or finite volume method to solve the temperature field equation, and obtain the temperature distribution of different heat dissipation methods relative to the inside of the motor; Step e: predict the temperature distribution and change trend of the water pump motor under different working conditions, monitor the real-time temperature status of the motor, and detect and alarm in time.

7. The method for monitoring and controlling the operation of the alumina water intake station equipment according to claim 3, characterized in that: The specific construction steps of the vibration analysis model (21) include: Step a: signal acquisition, The vibration sensor is used to continuously collect the vibration signal of the water pump during operation and record it as a time series x(t). Step b: Data cleaning, the formula is: x(t)=(y*h)(t) Where y(t) is the original signal, h(t) is the impulse response of the low-pass filter, and x(t) is the filtered signal; Step c, feature extraction, performing time domain, frequency domain, and time-frequency domain analysis on the preprocessed vibration signal to extract time domain features, frequency domain features, and time-frequency features; Step d, constructing a support vector machine model, a forest regression model, and a long short-term memory network model to perform fault diagnosis and classification during the operation of the water pump; Step e, early warning, real-time analysis of the vibration signal of the water pump. When abnormal vibration is detected, the system sends out an early warning signal to help maintenance personnel discover and handle potential faults in a timely manner.

8. The method for monitoring and controlling the operation of the alumina water intake station equipment according to claim 3, characterized in that: The specific steps of constructing the oil level video analysis model (22) include: Step a, collecting an oil level image, and preprocessing the image, wherein the preprocessing includes graying, denoising, and contrast enhancement; Step b: feature extraction, using edge detection algorithm to extract oil level boundary features; Step c: after extracting the oil level boundary features, construct a support vector machine model and a convolutional neural network model to perform fault diagnosis classification and determine the oil level height; Step d: using threshold comparison and condition judgment methods to automatically trigger an early warning or alarm mechanism.

9. The method for monitoring and controlling the operation of the alumina water intake station equipment according to claim 3, characterized in that: The specific steps of constructing the pipeline flow analysis model (23) include: Step a: Data collection, including historical flow data, meteorological data, and equipment operating parameters of the water intake station; Step b: data preprocessing, including data cleaning, data standardization and data conversion; Step c: feature extraction: extract features useful for prediction tasks from the original data, and use PCA and t-SNE methods to reduce the number of features and reduce the complexity of the model; Step d: Based on the extracted features, a linear regression model, a convolutional neural network or a recurrent neural network model is constructed to obtain prediction results and performance indicators; Step e: Using the threshold comparison method, an alarm is triggered when the model performance is lower than a specific threshold or an anomaly is detected.

10. The method for monitoring and controlling the operation of the alumina water intake station equipment according to claim 3, characterized in that: The specific steps of constructing the pipeline pressure analysis model (24) include: Step a, data collection: basic information of water intake station, pressure detection data, pump station operation status, water flow, pipeline friction coefficient, connection method, and valve and pump station location information, etc. Step b, data preprocessing: including data cleaning, feature selection and feature extraction; Step c: Based on the extracted features, a multivariate linear regression model or a neural network model is used to monitor the pipeline pressure. The model construction formula is: P=β0+β1Q+β2D+β3L+β4ρ+∈, Where P is pressure, Q is flow rate, D is pipe diameter, L is pipe length, ρ is fluid density, β is regression coefficient, and ∈ is error term; Step d: Using the threshold comparison method, when the monitoring result is lower than a specific threshold or an abnormality is detected, an alarm is triggered.

Images