Intelligent pressure maintaining system for concrete pouring of vault of storage tank

Abstract

The invention relates to the technical field of low-temperature storage tank dome concrete construction management, and particularly discloses a storage tank dome concrete pouring intelligent pressure maintaining system which comprises a sensor fusion acquisition module, a data preprocessing and feature extraction module and a trend prediction module. An adaptive control decision module, an execution and feedback closed loop module and a mobile terminal application module; according to the invention, an end-to-end intelligent closed loop is formed based on trend prediction of LSTM + Transformer and a decision algorithm of MPC + terminal stability punishment. The system not only can accurately pre-judge the pressure dynamic state, but also can adaptively balance the pressure maintaining precision, the energy consumption and the equipment wear, so that the whole-process online self-correction is realized, and the reliability and the construction efficiency of the system are improved; by monitoring current, vibration, temperature and other multi-dimensional states of key equipment such as a fan, an electric control cabinet and a valve, the system can be automatically switched to a standby scheme or a safe shutdown mode under extreme working conditions, it is ensured that basic safety can still be maintained under various emergencies, and construction continuity is guaranteed.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of low-temperature storage tank dome concrete construction management, and particularly relates to an intelligent pressure maintaining system for concrete pouring of a storage tank dome. BACKGROUND

[0002] The tooling and equipment used in the traditional pressure maintaining process mainly include two centrifugal ventilators (one standby and one in use), a connecting piece between the high-pressure vortex fan and the sealing plate of the gate hole, a venting and pressure relief device, and a U-shaped differential pressure gauge. The tooling is arranged on the temporary ramp at the gate hole and serves as the power output for pressure maintaining. Therefore, 4-6 skilled workers need to be arranged for 24-hour shift work to ensure the safe and continuous operation of pressure maintaining. The labor cost is high, and the workers have a certain safety risk when working near the continuously running fan and control box.

[0003] Since the pressure in the tank needs to be continuously monitored for a long time during the dome pouring process, and the power distribution cabinet is prone to overheating when powered for a long time, the fan is prone to fatigue failure when continuously running for a long time. Therefore, 4-6 skilled workers such as electricians and riveters need to be arranged for 24-hour shift work to ensure the safe and continuous operation of pressure maintaining in the pouring process of previous projects. The labor cost is high, and the workers have a certain safety risk when working near the continuously running fan and control box. Therefore, the purpose of this research is to develop a safe and efficient remote control system to reduce labor costs and eliminate safety risks caused by human factors such as manual supervision.

[0004] To solve the above problems, the application provides an intelligent pressure maintaining system for concrete pouring of a storage tank dome. SUMMARY

[0005] The application aims to provide an intelligent pressure maintaining system for concrete pouring of a storage tank dome to solve the problems in the background art.

[0006] To achieve the above-mentioned purpose, the application provides the following technical solutions:

[0007] An intelligent pressure maintaining system for concrete pouring of a storage tank dome comprises:

[0008] A sensor fusion acquisition module is configured to synchronously acquire pressure, temperature and vibration signals inside the dome of the storage tank to obtain original multi-dimensional time series data.

[0009] A data preprocessing and feature extraction module is configured to filter, denoise and perform principal component analysis on the original multi-dimensional time series data to extract purified key features.

[0010] A trend prediction module is configured to input the key features into a prediction model based on a long short-term memory network (LSTM) to predict future pressure-time variation trends.

[0011] An adaptive control decision module for matching the future pressure-time trend with a preset theoretical pressure maintaining curve, and generating an execution strategy based on a model predictive control (MPC) algorithm;

[0012] An execution and feedback closed loop module for driving the fan and valve actuator according to the execution strategy, and collecting execution feedback signals in real time to form feedback data, which is fed back to the data preprocessing and feature extraction module in a closed loop.

[0013] Preferably, in the adaptive control decision module, the model predictive control algorithm includes a set cost function, which optimizes the execution strategy between maintaining pressure maintaining quality and minimizing energy consumption.

[0014] Preferably, the formula of the set cost function is:

[0015]

[0016] Wherein, J is the cost function value;

[0017] N: prediction time domain length;

[0018] Scenario: the number of future time periods planned for the entire pressure maintaining stage of the vault pouring (e.g. 1 min per step, a total of N=30 periods). Used for the depth of future state prediction.

[0019] y t+i|t : the pressure prediction value of the future t+i moment under the reference at the t moment

[0020] Scenario: the value collected by the tank pressure sensor and calculated by the trend prediction module (LSTM+Transformer). Reflects the future tank pressure maintaining condition under the current control.

[0021] y ref (t+i): the theoretical pressure maintaining reference value at the t+i moment

[0022] Scenario: the optimal pressure maintaining curve preset according to the concrete solidification curve and structural safety standard.

[0023] Error weighted two norm

[0024] Wherein, e=y t+i|t -y ref (t+i)

[0025] Q∈R m×m : the importance weighting matrix (diagonal matrix) of the pressure deviation at different positions or different time periods.

[0026] Significance: ensure higher penalty for pressure deviation at critical locations.

[0027] u t+i|t : control amount prediction value at t+i moment;

[0028] Scenario: execute valve opening degree, fan rotating speed according to the instruction outputted by MPC.

[0029] Control increment weighted two norm;

[0030] Wherein, Δu = u t+i|t - u t+i-1|t : control increment weighted two norm

[0031] R∈R p×p : penalty matrix (diagonal matrix) for different actuator action rate changes.

[0032] For smoothing control action, avoiding frequent and large switching of valve or fan.

[0033] γ>0: control smoothing weighting coefficient

[0034] Scenario: increase γ appropriately to reduce action frequency at night or during sensitive equipment wear period. Used for balancing pressure maintenance accuracy and equipment life.

[0035] V(x t+N|t ): terminal stability penalty function

[0036] Form: V(x) = x T Px, wherein x contains terminal pressure error and critical state variables (such as structural strain).

[0037] P∈R n×n : stability weight matrix of terminal state (positive definite matrix).

[0038] Significance: ensure that the control sequence can still maintain system stability after the prediction terminal, so as to avoid "letting go and collapsing" phenomenon.

[0039] β>0: terminal stability penalty coefficient.

[0040] Preferably, the obtained execution strategy adopts a model predictive control method to match the future pressure-time change trend with a preset theoretical pressure maintenance curve;

[0041] The formula of the model predictive control method is:

[0042] x k+1 = Ax k + Bu k , y k = Cx k

[0043] where xk∈Rn: the system state vector at the kth moment (such as pressure deviation, temperature gradient, etc.);

[0044] uk∈Rm: the control vector at the kth moment (such as valve opening, fan speed);

[0045] yk∈Rp: the measurable output at the kth moment (corresponding to the pressure reading);

[0046] Matrices A, B, and C are derived from system identification or physical modeling.

[0047] Preferably, the sensor fusion acquisition module includes an acoustic vibration sensor, a strain gauge sensor, and an infrared temperature sensor, and the synchronous sampling frequency of the output raw multi-dimensional time series data is not less than 1 Hz.

[0048] Preferably, it further includes a mobile terminal application module, which includes:

[0049] Real-time acquisition and display of the pressure signal output by the sensor fusion acquisition module;

[0050] Automatic generation of fan start / stop instructions based on the pressure signal and issuance to the execution and feedback closed loop module through the network to trigger automatic start / stop of the fan;

[0051] For manual input of on / off fan instructions by a remote operator, issuance and execution through the network within no more than 3 seconds;

[0052] For real-time viewing and playback of historical pressure fluctuation curves.

[0053] Preferably, it further includes a local operation screen module, which is used for local operation screen and has the same function as the mobile terminal application module, and the two are configured through interlocking logic to ensure that only one control method is allowed to take effect at the same time;

[0054] For automatically shielding fan and valve control instructions from the mobile terminal when the local operation screen is in a control state;

[0055] For active switching of the automatic / manual control mode of the pressure relief valve by both the local operation screen and the mobile terminal.

[0056] Preferably, it further includes a three-level pressure alarm module, which includes:

[0057] Class A alarm (L): when the detected pressure is lower than 95% of the standard value, the mobile terminal pushes an alarm message to the responsible person, the on-site buzzer sounds, and the pressure relief valve is fully closed;

[0058] Class B alarm (H): when the detected pressure exceeds the standard value within 5%, the system only records the alarm log, without buzzer and valve action;

[0059] Class C alarm (HH): when the detected pressure exceeds the standard value by more than 5%, the mobile terminal pushes the alarm information, the on-site buzzer sounds, and the opening degree of the pressure relief valve is automatically adjusted according to the PID operation result to maintain the safe pressure;

[0060] After any alarm action is triggered, the person in charge needs to manually reset on the mobile terminal or on-site operation screen to restore normal control;

[0061] When the system is in a normal network and cold / hot / complete start mode, the total start-up time is not more than 10s, the control command response and execution time is not more than 3s, and the corresponding device state is fed back to the mobile terminal and on-site operation screen within 3s.

[0062] Preferably, it also includes a device and pressure monitoring and protection module, comprising:

[0063] Real-time monitoring of the running state of the fan, electric control cabinet and key actuator, including but not limited to current, voltage, vibration and temperature;

[0064] When extreme working conditions (such as power failure, communication interruption, abnormal vibration of the device) are detected, the system automatically switches to a backup control scheme and places the system in a "manual-on-site" or "safety shutdown" mode;

[0065] Both monitoring data and control data are stored on the chain for evidence, ensuring that any abnormality or switching operation can be traced back.

[0066] Compared with the prior art, the present application has the following advantages:

[0067] (1) The present application realizes a flexible control mode of "remote monitoring + on-site priority" through the double-channel interlocking design of the mobile terminal APP and the on-site operation screen. On the one hand, the manager can always master the pressure condition in the tank through the mobile phone and remotely issue start-stop instructions; on the other hand, the on-site operator only needs to switch to the on-site screen to take over the control right, ensuring that the construction demand can still be responded quickly in the case of network fluctuation or emergency.

[0068] (2) The system of the present application has a built-in three-level pressure alarm logic, which can automatically trigger multiple linkage measures such as buzzer, valve action and remote push at different overrun boundary points. The hierarchical alarm not only makes the early warning more accurate, but also ensures that there will be no over-intervention when there is a slight deviation, and when there is a major risk, the most stringent safety protection action will be automatically taken, and manual reset is required to avoid repeated false actions.

[0069] (3) The trend prediction based on LSTM+Transformer and the decision algorithm of MPC+terminal stability punishment constitute an end-to-end intelligent closed loop. The system can not only accurately predict the pressure dynamics, but also adaptively balance the pressure maintaining accuracy, energy consumption and equipment wear and tear, realize online self-correction in the whole process, and improve the system reliability and construction efficiency; through multi-dimensional state monitoring of the current, vibration and temperature of key equipment such as fan, electric control cabinet and valve, the system can automatically switch to standby scheme or safety shutdown mode under extreme working conditions, so as to ensure the basic safety under various emergencies and ensure the continuity of construction. BRIEF DESCRIPTION OF DRAWINGS

[0070] Fig. 1 A storage tank dome concrete pouring intelligent pressure maintaining system composition block diagram is provided for the application;

[0071] Fig. 2 A monitoring system operation composition block diagram is provided for the application;

[0072] Fig. 3 A storage tank dome concrete pouring intelligent pressure maintaining system topology diagram is provided for the application. DETAILED DESCRIPTION

[0073] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0074] Embodiment one:

[0075] Please refer to Figs. 1 to 3 A storage tank dome concrete pouring intelligent pressure maintaining system is shown in the figure, which comprises:

[0076] A sensor fusion acquisition module is used for multi-source synchronous acquisition of pressure, temperature and vibration signals inside the storage tank dome, to obtain original multi-dimensional time series data;

[0077] The original multi-dimensional time series data output by the acoustic vibration sensor, strain gauge sensor and infrared temperature sensor has a synchronous sampling frequency of not less than 1Hz

[0078] A data preprocessing and feature extraction module is used for filtering, denoising and principal component analysis of the original multi-dimensional time series data, to extract purified key features;

[0079] A trend prediction module is used for inputting the key features into a prediction model based on long short-term memory network (LSTM), to predict the future pressure-time variation trend;

[0080] An adaptive control decision module is configured to match the future pressure-time trend with a preset theoretical pressure maintaining curve, and generate an execution strategy based on a model predictive control (MPC) algorithm;

[0081] The model predictive control algorithm includes a set cost function, which is used to optimize the execution strategy between maintaining pressure maintaining quality and minimizing energy consumption;

[0082] The formula of the set cost function is as follows:

[0083]

[0084] Wherein, J is the numerical value of the cost function;

[0085] N: prediction time domain length;

[0086] Scenario: the number of future time periods planned for the entire pressure maintaining stage of the vault pouring (e.g. 1 min per step, a total of N=30 periods). Used for the prediction depth of future state.

[0087] y t+i|t : the pressure prediction value of the future t+i moment under the reference of the t moment

[0088] Scenario: the value collected by the tank pressure sensor and calculated by the trend prediction module (LSTM+Transformer). Reflects the future pressure maintaining condition of the tank under the current control.

[0089] y ref (t+i): the theoretical pressure reference value of the t+i moment

[0090] Scenario: the optimal pressure maintaining curve preset according to the concrete solidification curve and the structure safety standard.

[0091] Error weighted two-norm

[0092] Wherein, e=y t+i|t -y ref (t+i)

[0093] Q∈R m×m : the importance weighting matrix (diagonal matrix) of the pressure deviation at different positions or different time periods.

[0094] Meaning: to ensure higher punishment for the pressure deviation of the key position.

[0095] u t+i|t : the control amount prediction value of the t+i moment;

[0096] Scenario: Valve opening, fan speed are executed according to the MPC output.

[0097] Control increment weighted two-norm;

[0098] where Δu = u t+i|t - u t+i-1|t : Control increment weighted two-norm

[0099] R ∈ R p×p : Penalty matrix (diagonal matrix) for different actuator action rate changes.

[0100] Used for smoothing control action, avoiding frequent and large switching of valves or fans.

[0101] γ > 0: Control smoothing weighting coefficient

[0102] Scenario: At night or during sensitive equipment wear period, γ can be appropriately increased to reduce action frequency. Used for balancing pressure maintenance accuracy and equipment life.

[0103] V(x t+N|t ): Terminal stability penalty function

[0104] Form: V(x) = x T Px, where x contains terminal pressure error and key state variables (such as structural strain).

[0105] P ∈ R n×n : Stability weight matrix of terminal state (positive definite matrix).

[0106] Meaning: Ensure that the control sequence can still maintain system stability after the end of prediction, so as not to appear "let go and collapse" phenomenon.

[0107] β > 0: Terminal stability penalty coefficient;

[0108] Scenario: For high-risk areas (such as weak points on the tank top), β can be adjusted to strengthen the terminal stability constraint.

[0109] Meaning: While ensuring the overall pressure maintenance quality, it provides stability guarantee.

[0110] The first term ensures that the pressure closely follows the theoretical curve (pressure maintenance quality);

[0111] The second term smooths the control action (equipment life and energy consumption);

[0112] The third term improves the terminal stability (safety guarantee).

[0113] Scenario-based self-adaptation:

[0114] γ and β can be dynamically adjusted according to the construction stage, temperature gradient or equipment health diagnosis results to meet the optimal control requirements under different working conditions.

[0115] By using the "terminal stability penalty function" V(xt+N|t), which is specially designed for long-term high-risk pressure maintaining scenarios, the double weight coefficients γ and β are cross-regulated to achieve "precision-life-safety" triple optimization.

[0116] The formula of the model predictive control method is:

[0117] x k+1 =Ax k +Bu k ,y k =Cx k

[0118] Where xk∈Rn: the k-th time system state vector (such as pressure deviation, temperature gradient, etc.);

[0119] uk∈Rm: the k-th time control vector (such as valve opening, fan speed);

[0120] yk∈Rp: the k-th time measurable output (corresponding to the pressure reading);

[0121] Matrices A, B, and C are derived from system identification or physical modeling

[0122] An execution and feedback closed loop module is used to drive the fan and valve actuator according to the execution strategy, and real-time execution feedback signals are collected to form feedback data, which is fed back to the data preprocessing and feature extraction module in a closed loop.

[0123] It also includes a mobile application module, which includes:

[0124] Real-time acquisition and display of pressure signals output by the sensor fusion acquisition module;

[0125] Automatic generation of fan start-stop instructions based on the pressure signal and transmission to the execution and feedback closed loop module through the network to trigger automatic start-stop of the fan;

[0126] For manual input of on / off fan instructions by remote operators, the instructions are transmitted and executed through the network within 3 seconds;

[0127] For real-time viewing and playback of historical pressure fluctuation curves.

[0128] It also includes a local operation screen module, which has the same function as the mobile application module, and the two are configured through interlocking logic to ensure that only one control method is effective at the same time.

[0129] For when the local operation screen is in control state, automatically shield the fan and valve control instructions from the mobile terminal;

[0130] For the local operation screen and the mobile terminal can actively switch the automatic / manual control mode of the relief valve.

[0131] It also includes a three-level pressure alarm module, including:

[0132] Class A alarm (L): When the detected pressure is lower than 95% of the standard value, the mobile terminal pushes the alarm information to the responsible person, the on-site buzzer sounds, and the relief valve is fully closed;

[0133] Class B alarm (H): When the detected pressure is within 5% of the standard value, the system only records the alarm log, without buzzer and valve action;

[0134] Class C alarm (HH): When the detected pressure exceeds the standard value by more than 5%, the mobile terminal pushes the alarm information, the on-site buzzer sounds, and automatically adjusts the relief valve opening according to the PID calculation result to maintain the safe pressure;

[0135] After any alarm action is triggered, the responsible person must manually reset on the mobile terminal or local operation screen to restore normal control;

[0136] When the network is normal and the cold / hot / complete start mode, the total start time is not more than 10s, the control instruction response and execution time is not more than 3s, and the corresponding device state is feedback to the mobile terminal and local operation screen within 3s.

[0137] It also includes a device and pressure monitoring and protection module, including:

[0138] Real-time monitoring of the operation state of the fan, electric control cabinet and key actuator, including but not limited to current, voltage, vibration, temperature;

[0139] When extreme conditions (such as power failure, communication interruption, abnormal device vibration) are detected, automatically switch to the backup control scheme to place the system in "manual-on-site" or "safety shutdown" mode;

[0140] Both monitoring data and control data are chained and stored for evidence, ensuring that any abnormality or switching operation can be traced.

[0141] As can be seen from the above, through the double-channel interlocking design of the mobile terminal APP and the local operation screen, a flexible control mode of "remote monitoring + on-site priority" is realized. On the one hand, managers can always monitor the pressure in the tank through the mobile phone and remotely issue start / stop instructions; on the other hand, on-site operators only need to switch to the local screen to take over the control right, ensuring that the construction demand can still be responded quickly in the case of network fluctuations or emergencies.

[0142] Example two:

[0143] Casting and pressure maintaining at room temperature (20℃):

[0144] Sensor fusion acquisition:

[0145] Equipment: pressure sensor (range 0-1 MPa, accuracy ±0.005 MPa), infrared temperature sensor (accuracy ±0.3℃), acoustic vibration and strain gauge sensor;

[0146] Sampling: 1 Hz, synchronous acquisition for 60 min, original data multi-dimensional time series data.

[0147] Data preprocessing and feature extraction:

[0148] Wavelet threshold filtering denoising, linear baseline correction;

[0149] PCA reduced to 2-dimensional principal components, obtained purified key features.

[0150] LSTM + Transformer trend prediction:

[0151] LSTM (hidden unit 64, time domain window 10 min) prediction pressure; Transformer attention correction mutation;

[0152] Predict 30 min future pressure-time trend.

[0153] Adaptive MPC decision:

[0154] Model: xk+1=Axk+Buk, N=20;

[0155] Weight: Q=diag(100, 50), R=diag(10, 10), γ=0.8, β=1.2.

[0156] Cost function calculation (at the 20th min):

[0157] Predicted pressure y21|20=0.88 MPa, reference value yref(21)=0.90 MPa, error e=-0.02;

[0158] Error term: eTQe=0.0004×100=0.04;

[0159] Control increment Δu=[-0.03, +0.02] (valve %, fan %), ΔuTRΔu=0.0004×10+0.0004×10=0.008;

[0160] Terminal state final error 0.01, terminal penalty βxTPx≈1.2×0.0001=0.00012;

[0161] Single step cost is about 0.04+0.008+0.00012=0.04812, 20 steps cumulative J≈0.96.

[0162] Output strategy: ±4% adjustment of fan, ±3% of valve.

[0163] Execution and feedback loop:

[0164] Drive execution according to execution strategy, and return feedback data in real time;

[0165] Diagnose slight valve jam, and push warning.

[0166] Mobile APP and on-site operation:

[0167] APP automatic / manual mode triggers fan control; response delay <3s; real-time curve display;

[0168] On-site screen and APP interlocking (on-site priority), both can switch automatic control of pressure relief valve.

[0169] Three-level alarm action:

[0170] When the pressure is <0.855MPa (95%), trigger L-level alarm: buzzer + full valve closing + APP push;

[0171] When the pressure is >0.945MPa (105%), trigger HH-level alarm: buzzer + PID regulating valve + APP push;

[0172] H-level (0.90-0.945MPa) only logs record; manual reset is required to clear the alarm.

[0173] Equipment and pressure monitoring guarantee:

[0174] Real-time monitoring of fan current and cabinet temperature, automatic switching to manual on-site in case of communication interruption;

[0175] All data are chained and stored.

[0176] As can be seen from the above, the system has built-in three-level pressure alarm logic, which can automatically trigger buzzer, valve action and remote push and other multiple linkage measures at different overrun boundary points. Graded alarm not only makes the warning more accurate, but also ensures that there will be no over-intervention when there is a slight deviation, and when there is a major risk, the most stringent safety guarantee action will be automatically taken, and manual reset is required to avoid repeated false actions.

[0177] Example three:

[0178] Low temperature (5℃) pouring and pressure maintaining:

[0179] Data acquisition and preprocessing:

[0180] Increase temperature compensation: -0.002 MPa / ℃ zero drift correction to pressure sensor;

[0181] PCA after removing temperature acclimation noise for the first 10 min, to obtain purified key features.

[0182] Trend prediction:

[0183] LSTM as above, Transformer adds low-temperature factor, volatility prediction range +6%; output future pressure-time trend.

[0184] MPC decision:

[0185] Adjust the weight: γ = 1.0, β = 1.5;

[0186] Single-step cost calculation (20th min):

[0187] y21|20 = 0.87 MPa, reference 0.90 MPa, e = -0.03, error term = 0.09;

[0188] Δu = [-0.02, +0.015], incremental term ≈ 0.0065;

[0189] Terminal penalty ≈ 1.5 x 0.000144 = 0.000216;

[0190] Single-step cost ≈ 0.0967, 20-step cumulative ≈ 1.93.

[0191] Strategy execution strategy: ±3% for fan, ±2% for valve.

[0192] Execution and alarm:

[0193] System detects valve response lag 3s, automatically feeds forward +1% compensation;

[0194] Same level alarm and interlock logic as in Example Two.

[0195] APP / just-in-place and monitoring

[0196] Cold start ≤10s, response ≤3s;

[0197] Extreme power outage cut manual, log and data on-chain.

[0198] As can be seen from the above, the mobile APP supports real-time curve playback, manual / automatic switching and work order generation functions, combined with the simple and intuitive interface of the just-in-place touch screen, providing a unified and smooth interactive experience for operators of different roles, whether in remote dispatching or on-site operation. It can quickly locate problems and respond.

[0199] Comparative example:

[0200] Pressure set value 0.90 MPa, valve / fan preset ±10% fixed frequency action, no prediction, no MPC, no APP / interlock, no health diagnosis, no alarm grading.

[0201] The comprehensive comparison of Example 2, Example 3 and Comparative Example is shown in Table 1 below:

[0202] Table 1

[0203]

[0204]

[0205] As can be seen from the above, the trend prediction based on LSTM+Transformer and the decision algorithm of MPC+terminal stability penalty constitute an end-to-end intelligent closed loop. The system not only can accurately predict the pressure dynamics, but also can adaptively balance the pressure maintaining accuracy, energy consumption and equipment wear and tear, realize online self-correction in the whole process, and improve the system reliability and construction efficiency;

[0206] Through multi-dimensional state monitoring of current, vibration and temperature of key equipment such as fan, electric control cabinet and valve, the system can automatically switch to standby scheme or safe shutdown mode under extreme working conditions, ensuring that it can still maintain basic safety under various emergencies and guarantee the continuity of construction;

[0207] All key monitoring data, control instructions and alarm records are stored on the chain and cannot be tampered with and are fully traceable, providing a credible credential for quality acceptance, accident investigation and subsequent operation and maintenance, significantly enhancing the compliance and transparency of the system;

[0208] Through the automatic upgrading of the existing traditional full-capacity low-temperature storage tank dome concrete pouring pressure maintaining system, the combination of PLC technology, 5G transmission technology, ITO Internet of Things technology, etc. is used to realize full-automatic monitoring and control of pressure during the pressure maintaining process, reduce the labor hour input, reduce the defects caused by the large internal pressure fluctuation of the storage tank, and improve the pouring quality.

[0209] Although embodiments of the present application have been shown and described, it will be understood by those having ordinary skill in the art that various changes, modifications, substitutions and alterations can be made thereto without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A kind of intelligent pressure maintaining system for tank vault concrete pouring, it is characterized in that, The application relates to a pressure maintenance system for a storage tank, which comprises the following modules: a sensor fusion acquisition module for synchronously collecting pressure, temperature and vibration signals inside a tank dome to obtain original multi-dimensional time sequence data; a data preprocessing and feature extraction module for filtering, denoising and principal component analysis of the original multi-dimensional time sequence data to extract purified key features; a trend prediction module for inputting the key features into a prediction model to predict a future pressure-time variation trend; an adaptive control decision module for matching the future pressure-time variation trend with a preset theoretical pressure maintenance curve to generate an execution strategy based on a model predictive control algorithm; an execution and feedback closed loop module for driving a fan and a valve actuator according to the execution strategy and collecting execution feedback signals in real time to form feedback data which is fed back to the data preprocessing and feature extraction module in a closed loop.

2. The intelligent pressure maintaining system for concrete pouring of tank dome according to claim 1, characterized in that, In the adaptive control decision module, the model predictive control algorithm comprises a set cost function which is used to optimize the execution strategy between maintaining pressure maintenance quality and minimizing energy consumption.

3. The intelligent pressure maintaining system for concrete pouring of a tank dome according to claim 2, characterized in that, The formula of the set cost function is as follows: J = (1 / N) * ((yk-yref)TQ(yk-yref) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T D (xk-xopt) + (uk-uopt)T R (uk-uopt) + (xk-xopt)T S (xk-xopt) + (uk-uopt)T ​ ​ y t+i|t : reference to the pressure prediction value at time t+i in the future at time t y ref (t+i): theoretical pressure maintaining reference value at t+i moment Error-weighted L2 norm; where e = y t+i|t - y ref (t+i); Q ∈ R m×m : a matrix (diagonal matrix) that weights the importance of the pressure deviation at different locations or at different time periods; u t+i|t : control amount prediction value at the t+i time Controlled incremental weighted L2 norm; where Δu = u t+i|t -u t+i-1|t : control delta-weighted two norm; R ∈ R p×p : penalty matrix for different actuator action rate changes; ​ V(x t+N|t ) : terminal stability penalty function; P ∈ R n×n : stability weight matrix for end state ​ 4. The intelligent pressure maintaining system for concrete pouring of tank dome according to claim 1, characterized in that, ​ ​ x k+1 = Ax k + Bu k , y k = Cx k ​ ​ ​ ​ 5. The intelligent pressure maintaining system for concrete pouring of tank dome according to claim 1, characterized in that, ​ 6. The intelligent pressure maintaining system for tank dome concrete pouring according to claim 1, characterized in that, ​ 7. The intelligent pressure maintaining system for concrete pouring of tank dome according to claim 1, characterized in that, ​ ​ ​ ​ ​ 8. The intelligent pressure maintaining system for tank dome concrete pouring according to claim 1, characterized in that, ​ ​ ​ 9. The intelligent pressure maintaining system for concrete pouring of tank dome according to claim 1, characterized in that, ​ ​ ​ Class C alarm HH: When the detected pressure exceeds the standard value by more than 5%, the mobile terminal pushes an alarm message, the on-site buzzer sounds, and the opening degree of the pressure relief valve is automatically adjusted according to the PID calculation result to maintain the safety pressure; After any alarm action is triggered, the responsible person needs to manually reset on the mobile terminal or on-site operation screen to restore normal control; Under normal network, cold, hot, and full start-up modes, the total start-up time of the system does not exceed 10s, the control command response and execution time does not exceed 3s, and the corresponding device state is fed back to the mobile terminal and on-site operation screen within 3s.

10. The intelligent pressure maintaining system for concrete pouring of tank dome according to claim 1, characterized in that, It also includes a device and pressure monitoring and protection module, including: Real-time monitoring of the running state of the fan, electric control cabinet, and key actuator, including but not limited to current, voltage, vibration, and temperature; When extreme working conditions are detected, automatically switch to a backup control scheme to place the system in a safe manual-on-site or safe shutdown mode; Both monitoring data and control data are stored on the chain for evidence, ensuring that any abnormality or switching operation can be traced.

Images