Multi-purpose intelligent gate valve maintenance platform

Abstract

The application relates to the technical field of gate valve equipment, in particular to a multipurpose intelligent gate valve maintenance and repair platform, which comprises a data acquisition module, a data processing module, an intelligent monitoring module and a maintenance and repair module; the data acquisition module is used for acquiring intelligent gate valve data; the data processing module is used for processing the acquired intelligent gate valve data; the intelligent monitoring module is used for monitoring according to the processed intelligent gate valve data; and the maintenance and repair module is used for maintaining and repairing the gate valve according to the intelligent monitoring result. The application can simultaneously maintain and repair various uses of the intelligent gate valve in a working state by constructing an LSTM network to learn different uses of the intelligent gate valve, can judge the running state of the intelligent gate valve, and can generate an optimal maintenance scheme according to the running state, so that the normal work of the intelligent gate valve is ensured, the maintenance of the intelligent gate valve is realized, the maintenance cost is reduced, the maintenance efficiency is improved, and the safety hidden danger of each device is reduced.

Description

Technical Field

[0001] This invention relates to the field of gate valve equipment technology, and in particular to a multi-purpose intelligent gate valve maintenance and repair platform. Background Technology

[0002] A gate valve is a common type of on-off valve. It uses a gate (the opening and closing element, also called a gate or sluice gate) to connect (fully open) or cut off (fully close) the medium in a pipeline. It is not permitted for shut-off purposes; the gate should not be slightly opened during use, as the erosion from the high-speed flowing medium will accelerate damage to the sealing surface. The gate's movement direction is perpendicular to the fluid direction, cutting off the medium in the pipeline like a sluice gate, hence the name gate valve. An intelligent gate valve is an integrated, multi-turn electric actuator that can be directly connected to a computer for control, without the need for a control box. The computer can input a 4-20mA signal to control the valve's opening, closing, and intermediate position. Intelligent gate valves are the latest type of intelligent integrated electric actuators. They are suitable for valves where the opening and closing elements move linearly, such as gate valves, diaphragm valves, piston valves, and sluice gate valves. They are used for opening, closing, or regulating valves. They are essential drive devices for achieving remote, centralized, and automatic control of valves. They are characterized by comprehensive functions, reliable performance, advanced control systems, small size, light weight, and convenient use and maintenance. They are widely used in power, metallurgy, petroleum, chemical, papermaking, and wastewater treatment industries.

[0003] Gate valves are widely used in plant pipeline laying, solving the problem of controlling the on / off state of multiple pipelines. However, existing gate valves are mostly used in various critical locations, making gate valve maintenance essential. Maintenance ensures the normal operation of gate valves and reduces safety hazards in various equipment. However, closing gate valves for maintenance incurs significant losses, and the slow speed of gate valve control reduces the efficiency of related production activities in the plant. Furthermore, directly repairing or replacing faulty gate valves cannot accurately assess the status of intelligent gate valves, resulting in high maintenance costs. Therefore, we propose an intelligent gate valve maintenance platform capable of synchronously maintaining intelligent gate valves in operation. By constructing an LSTM network to intelligently learn the operating status of intelligent gate valves, the platform can determine their operating status and generate optimal maintenance plans based on this status, thus addressing the aforementioned problems. Summary of the Invention

[0004] The purpose of this invention is to address the deficiencies mentioned in the background art by proposing a multi-purpose intelligent gate valve maintenance and repair platform.

[0005] The technical solution adopted in this invention is as follows:

[0006] Provides a multi-purpose intelligent gate valve maintenance and repair platform, including:

[0007] Data acquisition module: used to collect data from intelligent gate valves;

[0008] Data processing module: used to process the collected smart gate valve data;

[0009] Intelligent monitoring module: used for monitoring based on the processed intelligent gate valve data;

[0010] Maintenance and repair module: Used for maintenance and repair of gate valves based on intelligent monitoring results.

[0011] As a preferred embodiment of the present invention, it also includes a database for storing various data of the maintenance platform.

[0012] As a preferred technical solution of the present invention: the data acquisition module collects smart gate valve data for different uses according to the purpose of the smart gate valve.

[0013] As a preferred technical solution of the present invention: the data processing module processes and converts the collected intelligent gate valve data.

[0014] As a preferred technical solution of the present invention: the intelligent monitoring module monitors the working status of the intelligent gate valve by constructing a neural network for deep learning.

[0015] As a preferred embodiment of the present invention, the deep learning steps are as follows:

[0016] The LSTM unit obtains the input at time t and the LSTM input at time t-1 according to the forget gate, and calculates the obtained input under the action of multiple excitations to obtain the output f of the forget gate. t ;

[0017] The formula for the filtration quantization ratio is:

[0018] f t =σ(w fx x t +w fh h t-1 +b f )

[0019] Where, x t For the input smart gate valve data, h t-1 The hidden information of the LSTM at time t-1; σ is the Sigmoid function;

[0020] The above formula determines the historical input information, and then other information is merged into the LSTM unit, including the input information x at time t that has already been removed. t The hidden information h at LSTMt-1 time t-1 Input gate information i t And new content The formula is:

[0021] i t =σ(w ix x t +w ih h t-1 +b i )

[0022]

[0023] Where tanh is the hyperbolic tangent function;

[0024] By combining the information removed and the newly acquired information from the above formula, we obtain the memory information C of the LSTM at time t. t The formula is:

[0025]

[0026] Among them, C t-1 This represents the memory information of the LSTM at time t-1;

[0027] Based on the current output of the LSTM, further calculations are performed by the output gate to obtain the output to the current state h. t The ratio of o t and the current status information h t :

[0028] o f =σ(w ix x f +w ih h t-1 +b i )

[0029] h t =o t ·tanh(C t )

[0030] Obtain the current state information of the output stream, and then finally calculate the output value y of the LSTM network. t :

[0031] y t =σ(w y h t +b y )

[0032] In the above calculation process, the sigmoid function is used as the activation function to filter information; w fx w fh w ix w ih w cx w ch w ixw ih w y b is the weighting coefficient; f b i b c b i b y This is the bias value;

[0033] The calculation process of the LSTM network is as follows:

[0034] S1: Select multidimensional measurement data from the front-end measurement unit to form multidimensional input data and single-dimensional or multidimensional output for training;

[0035] S2: Using the training data as input, the prediction data is obtained through a multi-unit LSTM;

[0036] S3: After the calculation in S2, the comparison value between the predicted data and the actual data is obtained; this comparison value is used as input, merged into the original input, and then processed by LSTM to learn the weights.

[0037] S4: Use the learned LSTM weights as the basis for subsequent calculations to obtain future predicted values.

[0038] As a preferred technical solution of the present invention: the intelligent monitoring module uses the intelligent gate valve data under different uses as the input of each LSTM unit to construct an LSTM network and monitor the working status of the intelligent gate valve under each use; the intelligent monitoring module uses the intelligent gate valve data under different uses as input, obtains feedback through the LSTM network, and transmits the intelligent gate valve data that detects abnormalities to the maintenance and repair module.

[0039] As a preferred technical solution of the present invention: the maintenance and repair module generates a maintenance strategy based on the abnormal data of the intelligent gate valve, and performs maintenance on the intelligent gate valve according to the generated maintenance strategy.

[0040] As a preferred technical solution of the present invention: In the maintenance strategy, the deterioration state of the intelligent gate valve is predicted based on the intelligent gate valve data predicted by the intelligent monitoring module. Let S0 be the normal operating state of the intelligent gate valve, and S... R To address the fault state of the intelligent gate valve, the two states of the intelligent gate valve are extended to multiple states, that is, a deterioration state is added between the normal operation state and the fault state of the intelligent gate valve. m and S m+1 For two consecutive intelligent gate valve deterioration states, and deterioration state S m m+1 ​A multi-state Markov model of an intelligent gate valve considering maintenance conditions is constructed. The maintenance of the intelligent gate valve is divided into minor and major repairs. Minor repairs cause the intelligent gate valve to revert to its previous deteriorated state; major repairs cause it to return to its normal state. Maintenance is carried out according to the following plan:

[0041] The smart gate valve was detected to be in a deteriorated state S. m During minor repairs, the intelligent gate valve was detected to be in a deteriorated state (S). m+1 Overhaul to ensure the intelligent gate valve is in optimal operating condition;

[0042] Repair Plan 1: The intelligent gate valve is detected to be in a deteriorated state (S). m During minor repairs, the intelligent gate valve was detected to be in a deteriorated state (S). m+1 Overhaul to ensure the intelligent gate valve is in optimal operating condition;

[0043] Maintenance Plan Two: The intelligent gate valve was detected to be in a deteriorated state (S). m and S m+1 Perform minor repairs to restore the smart gate valve to its previous state.

[0044] Maintenance Plan 3: The intelligent gate valve was detected to be in a deteriorated state (S). m No maintenance is required; maintenance is only required until the intelligent gate valve reaches a deteriorated state (S). m+1 After a major overhaul, the intelligent gate valve was restored to its normal state.

[0045] Maintenance Plan 4: The intelligent gate valve was detected to be in a deteriorated state (S). m No maintenance is required; only when the intelligent gate valve is detected to be in a deteriorated state (S) will maintenance be performed. m+1 After minor repairs, the intelligent gate valve returned to its deteriorated state. m .

[0046] As a preferred embodiment of the present invention: the maintenance and repair module (400) constructs an objective function A with the goal of minimizing maintenance costs:

[0047] A = A L +A M

[0048] in:

[0049]

[0050] Among them, A L This indicates the loss due to the failure of the smart gate valve; A M D1 and D2 represent the labor costs incurred during maintenance. mi D ma and D F These represent the durations of routine inspection, minor repair, major repair, and fault repair, respectively; k1, kmi k ma and k F These are the cost coefficients for each time period; based on the objective function and the current state of the intelligent gate valve, the maintenance scheme with the lowest cost that is suitable for the current state of the intelligent gate valve is calculated and selected.

[0051] The multi-purpose intelligent gate valve maintenance and repair platform provided by this invention has the following advantages compared with the prior art:

[0052] This invention constructs an LSTM network to learn the different uses of intelligent gate valves, enabling simultaneous maintenance of intelligent gate valves in operation for various purposes. It judges the operating status of intelligent gate valves and generates the best maintenance plan based on their operating status, ensuring the normal operation of intelligent gate valves, realizing the maintenance of intelligent gate valves, reducing maintenance costs, improving maintenance efficiency, and reducing safety hazards of various equipment. Attached Figure Description

[0053] Figure 1 This is a block diagram of the maintenance platform according to a preferred embodiment of the present invention.

[0054] The meanings of the various markings in the diagram are as follows: 100, data acquisition module; 200, data processing module; 300, intelligent monitoring module; 400, maintenance and repair module; 500, database. Detailed Implementation

[0055] It should be noted that, unless otherwise specified, the embodiments and features described in this embodiment can be combined with each other. The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0056] Reference Figure 1 A preferred embodiment of the present invention provides a multi-purpose intelligent gate valve maintenance and repair platform, comprising:

[0057] Data acquisition module 100: Used to acquire data from intelligent gate valves;

[0058] Data processing module 200: Used to process the collected smart gate valve data;

[0059] Intelligent monitoring module 300: Used for monitoring based on the processed intelligent gate valve data;

[0060] Maintenance and repair module 400: Used for maintenance and repair of gate valves based on intelligent monitoring results.

[0061] It also includes a database 500, which is used to store various data from the maintenance platform.

[0062] The data acquisition module 100 collects data on the intelligent gate valve for different applications according to the intended use of the intelligent gate valve.

[0063] The data processing module 200 processes and converts the collected smart gate valve data.

[0064] The intelligent monitoring module 300 monitors the working status of the intelligent gate valve by constructing a neural network for deep learning.

[0065] The deep learning steps are as follows:

[0066] The LSTM unit obtains the input at time t and the LSTM input at time t-1 according to the forget gate, and calculates the obtained input under the action of multiple excitations to obtain the output f of the forget gate. t ,

[0067] The formula for the filtration quantization ratio is:

[0068] f t =σ(w fx x t +w fh h t-1 +b f )

[0069] Where, x t For the input smart gate valve data, h t-1 The hidden information of the LSTM at time t-1; σ is the Sigmoid function;

[0070] The above formula determines the historical input information, and then other information is merged into the LSTM unit, including the input information x at time t that has already been removed. t The hidden information h at LSTMt-1 time t-1 Input gate information i t And new content The formula is:

[0071] i t =σ(w ix x t +w ih h t-1 +b i )

[0072]

[0073] Where tanh is the hyperbolic tangent function;

[0074] By combining the information removed and the newly acquired information from the above formula, we obtain the memory information C of the LSTM at time t.t The formula is:

[0075]

[0076] Among them, C t-1 This represents the memory information of the LSTM at time t-1;

[0077] Based on the current output of the LSTM, further calculations are performed by the output gate to obtain the output to the current state h. t The ratio of o t and the current status information h t :

[0078] o f =σ(w ix x f +w ih h t-1 +b i )

[0079] h t =o t ·tanh(C t )

[0080] Obtain the current state information of the output stream, and then finally calculate the output value y of the LSTM network. t :

[0081] y t =σ(w y h t +b y )

[0082] In the above calculation process, the sigmoid function is used as the activation function to filter information; w fx w fh w ix w ih w cx w ch w ix w ih w y b is the weighting coefficient; f b i b c b i b y This is the bias value;

[0083] The calculation process of the LSTM network is as follows:

[0084] S1: Select multidimensional measurement data from the front-end measurement unit to form multidimensional input data and single-dimensional or multidimensional output for training;

[0085] S2: Using the training data as input, the prediction data is obtained through a multi-unit LSTM;

[0086] S3: After the calculation in S2, the comparison value between the predicted data and the actual data is obtained; this comparison value is used as input, merged into the original input, and then processed by LSTM to learn the weights.

[0087] S4: Use the learned LSTM weights as the basis for subsequent calculations to obtain future predicted values.

[0088] The intelligent monitoring module 300 uses intelligent gate valve data for different applications as input to each LSTM unit to construct an LSTM network and monitor the working status of intelligent gate valves for each application. The intelligent monitoring module 300 uses intelligent gate valve data for different applications as input, obtains feedback through the LSTM network, and transmits the intelligent gate valve data that detects abnormalities to the maintenance module.

[0089] The maintenance and repair module 400 generates a repair strategy based on abnormal data of the intelligent gate valve, and performs repairs on the intelligent gate valve according to the generated repair strategy.

[0090] In the maintenance strategy, the deterioration state of the intelligent gate valve is predicted based on the intelligent monitoring module 300's predicted intelligent gate valve data. Let S0 be the normal operating state of the intelligent gate valve, and S... R To address the fault state of the intelligent gate valve, the two states of the intelligent gate valve are extended to multiple states, that is, a deterioration state is added between the normal operation state and the fault state of the intelligent gate valve. m and S m+1 For two consecutive intelligent gate valve deterioration states, and deterioration state S m m+1 A multi-state Markov model of an intelligent gate valve considering maintenance conditions is constructed. The maintenance of the intelligent gate valve is divided into minor and major repairs. Minor repairs cause the intelligent gate valve to revert to its previous deteriorated state; major repairs cause it to return to its normal state. Maintenance is carried out according to the following plan:

[0091] Repair Plan 1: The intelligent gate valve is detected to be in a deteriorated state (S). m During minor repairs, the intelligent gate valve was detected to be in a deteriorated state (S). m+1 Overhaul to ensure the intelligent gate valve is in optimal operating condition;

[0092] Maintenance Plan Two: The intelligent gate valve was detected to be in a deteriorated state (S). m and S m+1 Perform minor repairs to restore the smart gate valve to its previous state.

[0093] Maintenance Plan 3: The intelligent gate valve was detected to be in a deteriorated state (S).​m No maintenance is required; maintenance is only required until the intelligent gate valve reaches a deteriorated state (S). m+1 After a major overhaul, the intelligent gate valve was restored to its normal state.

[0094] Maintenance Plan 4: The intelligent gate valve was detected to be in a deteriorated state (S). m No maintenance is required; only when the intelligent gate valve is detected to be in a deteriorated state (S) will maintenance be performed. m+1 After minor repairs, the intelligent gate valve returned to its deteriorated state. m .

[0095] The maintenance and repair module 400 constructs an objective function A with the goal of minimizing maintenance costs:

[0096] A = A L +A M

[0097] in:

[0098]

[0099] Among them, A L This indicates the loss due to the failure of the smart gate valve; A M D1 and D2 represent the labor costs incurred during maintenance. mi D ma and D F These represent the durations of routine inspection, minor repair, major repair, and fault repair, respectively; k1, k mi k ma and k F These are the cost coefficients for each time period; based on the objective function and the current state of the intelligent gate valve, the maintenance scheme with the lowest cost that is suitable for the current state of the intelligent gate valve is calculated and selected.

[0100] In this embodiment, the data acquisition module 100 collects data from smart gate valves under different applications as input to the LSTM network. The data processing module 200 processes and transforms the collected smart gate valve data, converting it into a data format suitable for LSTM network input. The intelligent monitoring module 300 learns the operating status of the smart gate valves under different applications by constructing an LSTM network.

[0101] The LSTM unit obtains the input at time t and the LSTM input at time t-1 according to the forget gate, and calculates the obtained input under the action of multiple excitations to obtain the output f of the forget gate. t ,

[0102] The formula for the filtration quantization ratio is:

[0103] f t =σ(w fx x t +w fhh t-1 +b f )

[0104] Where, x t For the input smart gate valve data, h t-1 The hidden information of the LSTM at time t-1; σ is the Sigmoid function;

[0105] The above formula determines the historical input information. Furthermore, the amount of information obtained must also be determined to allow other information to be incorporated into the LSTM unit, including the input information x at time t that has already been removed. t The hidden information h at LSTMt-1 time t-1 Input gate information i t And new content The formula is:

[0106] i t =σ(w ix x t +w ih h t-1 +b i )

[0107]

[0108] Where tanh is the hyperbolic tangent function;

[0109] By combining the information removed and the newly acquired information from the above formula, we obtain the memory information C of the LSTM at time t. t The formula is:

[0110]

[0111] Among them, C t-1 This represents the memory information of the LSTM at time t-1;

[0112] Based on the current output of the LSTM, further calculations are performed by the output gate to obtain the output to the current state h. t The ratio of o t and the current status information h t :

[0113] o t =σ(w ix x t +w ih h t-1 +b i )

[0114] h t =o t ·tanh(C t )

[0115] Obtain the current status information of the output stream, and finally calculate the output value y of the LSTM network t :

[0116] y t = σ(w y h t + b y )

[0117] In the above calculation process, the activation function uses the curvature function sigmoid to filter information; w fx 、w fh 、w ix 、w ih 、w cx 、w ch 、w ix 、w ih 、w y are weight coefficients; b f 、b i 、b c 、b i 、b y are bias quantities.

[0118] Select multi-dimensional measurement data from the pre-measurement unit to form multi-dimensional input data and single-dimensional or multi-dimensional output for training; use the training data as input to obtain calculated prediction data through a multi-unit LSTM; after the previous calculation, obtain the comparison value between the prediction data and the actual data; use this comparison value as input, merge it into the original input, and perform the weight learning process through LSTM calculation; use the learned LSTM weights as the basis for subsequent calculations to obtain future prediction values.

[0119] The intelligent monitoring module 300 takes the intelligent gate valve data under different uses as input, obtains feedback through the LSTM network, and transmits the intelligent gate valve data detected to be abnormal to the maintenance and repair module. The maintenance and repair module 400 generates a maintenance strategy based on the abnormal situation of the intelligent gate valve data. Taking S0 as the normal operation state of the intelligent gate valve, S R as the fault state of the intelligent gate valve, and S1, S2, S3, and S4 as four consecutive deterioration states of the intelligent gate valve, and taking the deterioration state S1 < S2 < S3 < S4 as an example, the maintenance is carried out according to the following maintenance plan:

[0120] If the detected state of the intelligent gate valve is the deterioration state S1, then perform minor repairs. If the detected state of the intelligent gate valve is the deterioration state S2, then perform major repairs to keep the intelligent gate valve in the best normal operation state;

[0121] If the detected states of the intelligent gate valve are both the deterioration states S1 and S2, then perform minor repairs to make the state of the intelligent gate valve return to the previous state;

[0122] If the intelligent gate valve is detected to be in a deteriorated state (S1), no maintenance is required. Only when the intelligent gate valve is detected to be in a deteriorated state (S2) should a major overhaul be performed to restore the intelligent gate valve to its normal state.

[0123] If the smart gate valve is detected to be in a deteriorated state S1, no maintenance is performed. Only if the smart gate valve is detected to be in a deteriorated state S2 should minor maintenance be performed to restore the smart gate valve to the deteriorated state S1.

[0124] Based on objective function A:

[0125] A = A L +A M

[0126] in:

[0127]

[0128] Among them, A L This indicates the loss due to the failure of the smart gate valve; A M D1 and D2 represent the labor costs incurred during maintenance. mi D ma and D F These represent the durations of routine inspection, minor repair, major repair, and fault repair, respectively; k1, k mi k ma and k F These are the cost coefficients for each duration;

[0129] The lowest-cost maintenance strategy is determined when the deterioration states are S1, S2, S3, and S4. For example, when the deterioration states are S1 and S2, maintenance strategy four has the lowest cost, so maintenance strategy four is implemented in real time; when the deterioration states are S3 and S4, maintenance strategy three has the lowest cost, so maintenance strategy four is implemented in real time, thus ensuring that the maintenance cost of the intelligent gate valve is minimized.

[0130] The intelligent gate valve is maintained according to the generated maintenance strategy. All data from the monitoring and maintenance process of the intelligent gate valve is recorded in database 500 for easy retrieval and traceability later.

[0131] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0132] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A multipurpose smart gate valve maintenance platform, characterized in that: Include: Data acquisition module (100): for collecting intelligent gate valve data; Data processing module (200): for processing the collected intelligent gate valve data; Intelligent monitoring module (300): for monitoring according to the processed intelligent gate valve data; Maintenance and repair module (400): for maintaining and repairing the gate valve according to the intelligent monitoring result; The maintenance and repair module (400) generates a repair strategy according to the abnormal condition of the intelligent gate valve data, and repairs the intelligent gate valve according to the generated repair strategy; In the overhaul strategy, the deterioration state of the intelligent gate valve is predicted according to the intelligent gate valve data predicted by the intelligent monitoring module (300), and the intelligent gate valve is set For the normal operation state of the intelligent gate valve, For the fault state of the intelligent gate valve, the two states of the intelligent gate valve are expanded into multiple states, that is, the deterioration state of the intelligent gate valve is added between the normal operation state and the fault state of the intelligent gate valve, And For two consecutive deterioration states of the intelligent gate valve, and the deterioration state , a multi-state Markov model of the intelligent gate valve considering the overhaul state is constructed, the overhaul of the intelligent gate valve is divided into small repair and large repair, the state of the intelligent gate valve is returned to the previous deterioration state after small repair; the state of the intelligent gate valve is returned to the normal state after large repair, and the overhaul is carried out according to the following overhaul scheme: Detecting the smart gate valve state as a deterioration state Then minor repair, detecting the smart gate valve state as a deterioration state Major repair, keeping the smart gate valve in the best normal operation state Detecting that the smart gate valve status is in a deteriorated state and Both minor repairs, the smart gate valve status back to the previous state; Detecting the smart gate valve state as a deteriorated state No repair, only overhaul to the smart gate valve state as a deteriorated state After overhaul, the smart gate valve state returns to normal state Detecting that the smart gate valve status is in a deteriorated state No repair, only detecting that the smart gate valve status is in a deteriorated state After minor repair, the smart gate valve status is returned to a deteriorated state ; The maintenance module (400) is constructed to build an objective function with the lowest maintenance cost as the target : Wherein: wherein, represents the loss due to the failure of the smart gate valve; represents the manual cost of the maintenance; , , and respectively represent the time length of the inspection, the minor repair, the major repair and the breakdown repair; , , and are respectively the cost coefficients of the time lengths; according to the target function and the current state of the smart gate valve, a maintenance scheme that is most suitable for the current state of the smart gate valve and has the lowest cost is selected.

2. The multipurpose smart gate valve maintenance platform in accordance with claim 1, wherein: It also includes a database (500) for storing various data of the repair platform.

3. The multipurpose smart gate valve maintenance platform in accordance with claim 1, wherein: The data acquisition module (100) collects intelligent gate valve data under different purposes according to the purpose of the intelligent gate valve.

4. The multipurpose smart gate valve maintenance platform in accordance with claim 1, wherein: The data processing module (200) processes and converts the collected intelligent gate valve data.

5. The multipurpose smart gate valve maintenance platform in accordance with claim 1, wherein: The intelligent monitoring module (300) monitors the working state of the intelligent gate valve by constructing a neural network for deep learning.

6. The multipurpose smart gate valve maintenance platform in accordance with claim 5, wherein: The deep learning steps are as follows: The LSTM unit obtains the forget gate according to the input quantity at the moment The input quantity at the moment and the output of the forget gate are obtained The input quantity at the moment, the output of the forget gate, and the output of the input gate are obtained ; The filtering and quantization ratio formula is: wherein, is the inputted smart gate valve data, is the hidden information of the time LSTM; is a Sigmoid function; The above formula determines the historical input information, and other information is combined into the LSTM unit, including the input information at time t which has been removed , LSTM , the hidden information at time t , the input gate information of the hidden information at time t , and the new content , and the formula is: wherein is the hyperbolic tangent function; The information removed by the above formula, the newly obtained information, and the time point LSTM memory information are combined to obtain Memory information of the time point LSTM The formula is: wherein, is memory information of the time LSTM; According to the current LSTM obtained quantity, further calculated by the output gate, thereby obtaining the output to the current state of the ratio and the current state information : Obtain the current state information of the output stream, and then finally calculate the output value of the LSTM network : In the above calculation process, the activation function is selected as the curvature function sigmoid to realize the filtering information. 、 、 、 、 、 、 、 、 is a weight coefficient; 、 、 、 、 is a bias quantity; The calculation process of the LSTM network is as follows: S1: Select multi-dimensional measurement data from the front measurement unit to form multi-dimensional input data and single-dimensional or multi-dimensional output for training; S2: Take the training data as input, and obtain the calculation prediction data through the multi-unit LSTM; S3: After S2 calculation, the comparison value of the prediction data and the actual data is obtained; take this comparison value as input, merge it into the original input, and perform the weight learning process through LSTM calculation; S4: Take the learned LSTM weight as the basis for subsequent calculation to get the future prediction value.

7. The multipurpose smart gate valve maintenance platform in accordance with claim 6, wherein: The intelligent monitoring module (300) takes the intelligent gate valve data under different purposes as the input of each LSTM unit, constructs an LSTM network, and monitors the working state of the intelligent gate valve under different purposes; The intelligent monitoring module (300) takes the intelligent gate valve data under different purposes as input, obtains feedback through the LSTM network, and transmits the detected abnormal intelligent gate valve data to the maintenance and repair module.

Images