A water flow monitoring method for a water cooling system of a flexible DC converter valve based on a predator algorithm
Through the water flow monitoring method based on the predation algorithm, non-invasive ultrasonic sensors and data analysis are used to solve the problem of misjudgment of the numerical value of water flow in the water cooling system, and the accurate monitoring and misjudgment of the health status of the water cooling system is achieved.
Patent Information
- Application Number
- CN202210576440.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-05-25
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2042-05-25
AI Technical Summary
In high-voltage DC transmission systems, the water flow monitoring of the water cooling system is easily affected by environmental and random factors, resulting in misjudgment of the mutation value and causing system failure.
The water flow monitoring method of the flexible direct converter valve water cooling system based on predation algorithm is adopted, and data is collected using non-invasive ultrasonic flow sensors. Through dirty data cleaning, data statistics and predation model analysis, the causes of mutation values are distinguished to avoid misjudgment.
By quantifying parameters reflect the reasons for the sudden change in water flow signal, the health status of the water-cooled system is monitored to avoid misjudgment, and the reliability and accuracy of the system are improved.
Smart Images

Figure CN115184053B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent power transmission technology, and in particular to a method for monitoring water flow of a flexible DC converter valve water cooling system based on a predator algorithm. Background Art
[0002] With the improvement of power electronic devices, my country's high-voltage direct current transmission technology has developed rapidly. Flexible direct current transmission technology based on voltage source converter (VSC) and modular multilevel converter (MMC) has stood out in direct current transmission technology. The flexible direct current transmission system uses fully controllable insulated gate bipolar transistor (IGBT), and its operation is accompanied by the loss of electric energy, which is converted into heat. Therefore, the monitoring of non-electrical quantity data such as temperature and water flow has important practical significance.
[0003] During the water flow monitoring process, due to the influence of the working environment of the monitoring equipment or other random factors, sudden changes in water flow values may still occur when the water cooling system is operating normally. The sudden changes in values may lead to misjudgment of water cooling system failures. Summary of the invention
[0004] The technical problem to be solved and the technical task proposed by the present invention are to improve and perfect the existing technical solutions, and provide a water flow monitoring method for a water cooling system of a flexible direct current converter valve based on a predatory algorithm, so as to distinguish the causes of sudden changes in the values and avoid misjudgment. To this end, the present invention adopts the following technical solutions.
[0005] A method for monitoring water flow of a water cooling system of a flexible direct current converter valve based on a predator algorithm, wherein the water cooling system of the flexible direct current converter valve comprises a water cooling radiator, a temperature sensor, a flow sensor, a signal transmission device and a host computer; the flexible direct current converter valve is arranged in the water cooling radiator, and the temperature sensor and the flow sensor are connected to the host computer through the signal transmission device; the flow sensor is a non-invasive ultrasonic flow sensor, which is installed at the water inlet and / or outlet of the water cooling radiator in a non-breaking pipe manner and attached to the outer wall of the converter valve branch pipe. The water flow monitoring method comprises the following steps:
[0006] 1) Set the dirty data cleaning scheme, normal water flow data value interval, and abnormal water flow data value interval;
[0007] 2) Obtain the water flow signal of the water cooling system of the flexible DC converter valve through a non-invasive ultrasonic sensor;
[0008] 3) Convert the collected water flow signal into a water flow digital signal through an ADC module;
[0009] 4) The converted digital signal is stored as raw data in the data cache module, and the data cache module communicates with the host computer through the 485 protocol to send the raw data to the raw data database, and inserts the current time into the database through the setDate method;
[0010] 5) Clean the data in the database according to the dirty data cleaning solution, and return the cleaned data to the target database;
[0011] 6) Repeat steps 2) to 5) until the calibration time when the last data is written into the database and the calibration time when the first data is written into the database reach the set time;
[0012] 7) According to the set water flow normal data value interval and water flow abnormal data value interval, count the data volume of normal data and the data volume of mutation data, and calculate the mutation frequency;
[0013] 8) The statistical data and mutation frequency are brought into the predator-prey model to calculate the regular mutation and random mutation, and the relationship between normal and abnormal data in the time domain is obtained. The parameter change trend is displayed through a visualization framework, so as to observe the trend graph to monitor the health status of the converter valve water cooling system.
[0014] This technical solution uses a non-invasive water flow ultrasonic sensor to measure the flow rate of the coolant in the converter valve without contacting it. The non-invasive water flow ultrasonic sensor sends and receives ultrasonic signals to collect information, transmits analog signals, integrates the conversion between different signals, and simplifies the collection device. This technical solution counts the amount of normal data and the amount of mutation data, quantifies the dynamic law of the converter valve water cooling system, and obtains the comparison relationship between the mutation signal and the normal signal in the time domain through the predator model. This quantitative parameter can reflect the cause of the water flow mutation signal; thereby obtaining the health status of the converter valve water cooling system without the need for hardware equipment to perceive relevant information.
[0015] As a preferred technical means: in step 1), the data type of dirty data is a data type other than double type or empty, and the dirty data cleaning method is missing data, data type error data, duplicate data, and the data is defined as empty and the collection time is retained. The traffic data uses double type. After limiting the data storage type, by processing data of non-limited types, it can effectively save processing time, improve processing efficiency, and help improve the accuracy of subsequent processing.
[0016] As a preferred technical means: in step 6), the set time is 360 hours; the non-invasive ultrasonic sensor collects water flow every 2 hours; and a total of 180 data collected for 360 hours is recorded as a statistical cycle. Using 360 hours can make the interval have mutation data; because the flow change is a continuous process, collecting water flow once every 2 hours can avoid data redundancy and capture the sudden change of flow.
[0017] As a preferred technical means: in step 8), the predator-prey model is expressed as:
[0018]
[0019] Among them, x is the amount of normal data, y is the amount of mutation data, λ is the mutation frequency, ε is the acquisition frequency, dx / dt and dy / dt represent the time-varying rate of normal data and the time-varying rate of mutation data. By solving two differential equations, the contrast relationship between the mutation signal and the normal signal in the time domain is obtained. This quantitative parameter can reflect the cause of the water flow mutation signal to a certain extent.
[0020] By solving two differential equations, the contrast relationship between the mutation signal and the normal signal in the time domain is obtained. The calculation is simple and accurate. This quantitative parameter can reflect the cause of the water flow mutation signal to a certain extent.
[0021] As a preferred technical means: ε is the acquisition frequency of 0.00556.
[0022] As a preferred technical means: the setDate method in step 4) is the setDate method in the PreparedStatement class in the java class library, and the data type of the acquisition time stored in the database as an attribute is long int type.
[0023] As a preferred technical means: in step 7), the statistical data is obtained by using the MapReduce model to obtain each normal data interval and the number of normal data values in each interval, each abnormal data value interval and the number of abnormal data values in each interval; the mutation frequency λ is calculated, which is expressed as the ratio of the number of mutation values to the total number of collected values 180.
[0024] As a preferred technical means: in step 5), cleaning the data in the database includes the following steps:
[0025] 501) Obtaining data from a database;
[0026] 502) Determine whether the data is missing, if so, proceed to step 505); if not, proceed to step 503);
[0027] 503) Determine whether the data type is wrong, if so, proceed to step 505); if not, proceed to step 504);
[0028] 504) Determine whether the data is repeated, if so, proceed to step 505);
[0029] 505) Define the data as empty and retain the time attribute.
[0030] Beneficial effects:
[0031] This technical solution adopts a non-invasive water flow ultrasonic sensor to measure the flow rate of the coolant in the converter valve without contacting it. The non-invasive water flow ultrasonic sensor sends and receives ultrasonic signals to collect information, transmits analog signals, integrates the conversion between different signals, and simplifies the collection device.
[0032] This technical solution counts the amount of normal data and the amount of mutation data, quantifies the dynamic laws of the converter valve water cooling system, and obtains the contrast relationship between the mutation signal and the normal signal in the time domain through the predator model. This quantitative parameter can reflect the cause of the water flow mutation signal; thereby obtaining the health status of the converter valve water cooling system without the need for hardware equipment to perceive relevant information.
[0033] This technical solution displays parameter change trends through a visualization framework and monitors the health status of the converter valve water cooling system by observing the trend graph. The monitoring is intuitive, so the cause of sudden changes in values can be quickly and directly distinguished to avoid misjudgment. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 It is a flow chart of the present invention.
[0035] Figure 2 It is a data cleaning flowchart invented.
[0036] Figure 3 It is a schematic diagram of the structure of the invention.
[0037] In the figure: 1. Water-cooled radiator; 2. Temperature sensor; 3. Flow sensor; 4. Signal transmission device; 5. Host computer; 6. Converter valve; 7. DC energy storage capacitor. DETAILED DESCRIPTION
[0038] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings.
[0039] like Figure 3As shown, the water cooling system of the flexible direct current converter valve includes a water cooling radiator 1, a temperature sensor 2, a flow sensor 3, a signal transmission device 4 and a host computer 5; the flexible direct current converter valve 6 is arranged in the water cooling radiator 1, and the temperature sensor 2 and the flow sensor 3 are connected to the host computer 5 through the signal transmission device 4; the flow sensor 3 is a non-invasive ultrasonic flow sensor 3, which is installed at the water inlet and / or outlet of the water cooling radiator 1 in a non-breaking pipe manner and attached to the outer wall of the branch pipe of the converter valve 6; the signal transmission device 4 is fixed to the equipotential point on the upper surface of the shell of the DC energy storage capacitor 7; the signal transmission device is fixed to the equipotential point on the upper surface of the shell of the DC energy storage capacitor, which can reduce the influence of the voltage difference on the signal transmission device, which is conducive to improving the stability of its work. The non-invasive water flow ultrasonic sensor is used to measure the flow rate of the coolant in the converter valve without contacting it. The non-invasive water flow ultrasonic sensor sends and receives ultrasonic signals to collect information, transmits analog signals, integrates the conversion between different signals, and simplifies the collection device.
[0040] like Figure 1 As shown, the present invention includes the steps of:
[0041] S1: Set the dirty data cleaning scheme, normal water flow data value interval, and abnormal water flow data value interval;
[0042] In step S1, the data type of dirty data is a data type other than double type or empty. The dirty data cleaning method is missing data, data type error data, duplicate data, and the data is defined as empty and the collection time is retained. The traffic data uses double type. After limiting the data storage type, by processing data of non-limited types, it can effectively save processing time and improve processing efficiency, and at the same time help improve the accuracy of subsequent processing;
[0043] S2: Obtain the water flow signal of the water cooling system of the flexible direct current converter valve through a non-invasive ultrasonic sensor;
[0044] S3: converting the collected water flow signal into a water flow digital signal through an ADC module;
[0045] S4: The converted digital signal is stored as raw data in the data cache module. The data cache module communicates with the host computer through the 485 protocol to send the raw data to the raw data database, and inserts the current time into the database through the setDate method; the setDate method is the setDate method in the PreparedStatement class in the java class library, and the data type of the acquisition time stored in the database as an attribute is long int.
[0046] S5: Clean the data in the database according to the dirty data cleaning solution, and return the cleaned data to the target database; Figure 2 As shown in the figure, dirty data cleaning includes the following steps:
[0047] S501: Acquire data from a database;
[0048] S502: Determine whether data is missing, if so, proceed to step S505; if not, proceed to step S503;
[0049] S503: Determine whether the data type is wrong, if so, proceed to step S505; if not, proceed to step S504;
[0050] S504: Determine whether the data is repeated, if so, proceed to step S505;
[0051] S505: Define the data as empty and retain the time attribute.
[0052] S6: Repeat steps S2 to S5 until the length of the calibration time when the last data is written into the database and the calibration time when the first data is written into the database reaches 360 hours;
[0053] The set time is 360 hours; the non-invasive ultrasonic sensor collects water flow every 2 hours; a total of 180 data points collected for 360 hours are recorded as a statistical cycle. Using 360 hours can make the data mutation exist in this interval; because the flow change is a continuous process, collecting water flow every 2 hours can avoid data redundancy and capture the sudden change of flow at the same time;
[0054] S7: According to the set water flow normal data value interval and water flow abnormal data value interval, count the data volume of normal data and the data volume of mutation data, and calculate the mutation frequency λ, which is expressed as: the ratio of the number of mutation values to the total number of collected values 180; the statistical data is obtained by counting each normal data interval and the number of normal data values in each interval, each abnormal data value interval and the number of abnormal data values in each interval through the MapReduce model;
[0055] S8: Bring statistical data and mutation frequency into the predator-prey model, calculate regular mutations and random mutations, obtain the relationship between normal and abnormal data in the time domain, and display the parameter change trend through a visualization framework, so as to observe the trend graph to monitor the health status of the converter valve water cooling system;
[0056] The predator-prey model is expressed as:
[0057]
[0058] Among them, x is the amount of normal data, y is the amount of mutation data, λ is the mutation frequency, ε is the acquisition frequency. In this example, ε is the acquisition frequency of 0.00556. dx / dt and dy / dt represent the time-varying rate of normal data and the time-varying rate of mutation data. By solving two differential equations, the contrast relationship between the mutation signal and the normal signal in the time domain is obtained. This quantitative parameter can reflect the cause of the water flow mutation signal to a certain extent.
[0059] By solving two differential equations, the contrast relationship between the mutation signal and the normal signal in the time domain is obtained. The calculation is simple and accurate. This quantitative parameter can reflect the cause of the water flow mutation signal to a certain extent.
[0060] By counting the amount of normal data and the amount of mutation data, the dynamic law of the converter valve water cooling system is quantified, and the contrast relationship between the mutation signal and the normal signal in the time domain is obtained through the predator model. This quantitative parameter can reflect the cause of the water flow mutation signal; thus, the health status of the converter valve water cooling system can be obtained without the need for hardware equipment to perceive relevant information.
[0061] The normal data obtained can also be combined with the temperature sensor to calculate the heat carried away by the cooling water, so as to further monitor the flexible DC converter valve water cooling system and the flexible DC converter valve.
[0062] The heat dissipation of cooling water is calculated as follows:
[0063] Q 1 =ρ w ·v in ·A in ·c pw ·(T out -T in ) (1)
[0064] In the formula, Q 1 Indicates the heat taken away by water, in W; ρ w Indicates the density of cooling water, kg / m 3 ;v in Indicates the water inlet flow rate in m / s, measured by an ultrasonic flow sensor; A in
[0065] Indicates the cross-sectional area of the water pipe inlet, in m 2 ;c pw Represents the specific heat capacity of water, in J / (kg·K);
[0066] T in Indicates the inlet water temperature, in K; T outrepresents the outlet water temperature, in K. The density and specific heat capacity of cooling water are affected by temperature. To improve the calculation accuracy, ρ in formula (2) is replaced by w is replaced by the calculated ρ w (T); replace c in formula (1) pw is replaced by the calculated c pw (T);
[0067] ρ w (T)=838.466135+1.40050603T-0.003011237T 2 +3.71822313×10 -7 T 3
[0068] c pw (T)=12010.1471-80.4072879T+0.309866854T 2 -5.38186884×10 -4 T 3 +3.62536437×10 -7 T 4
[0069] Where T is the average value of cooling water inlet and outlet temperatures;
[0070] When the fiber optic temperature sensor probe measures the wall temperature of the inlet and outlet pipes, since the average temperature of the water flow in the pipe cannot be directly obtained, in order to increase the inlet and outlet temperature T of the water cooling radiator in and T out The measurement accuracy is improved by solving the established pipeline heat transfer model to monitor the average temperature of the water flow in the pipe.
[0071] In the pipeline heat transfer model, r i 、r o Respectively represent the inner wall radius and outer wall radius of the tube; r m Surface wall midpoint radius; T f represents the average temperature of the water in the pipe, T a represents the air temperature outside the tube, T i 、T m and T o They represent the temperature of the inner wall node, the center node and the outer wall node of the tube respectively; the density of the tube wall ρ t Specific heat capacity c t , thermal conductivity k t are constants, h i 、h o are the convection heat transfer coefficients inside and outside the tube, respectively;
[0072] Define Δr=(ro -r i ) / 4, then r m =r i +2Δr,r o =r i +4Δr, using Fourier's law of heat conduction and the principle of energy conservation, the inner wall node T i Establishing the heat balance equation
[0073]
[0074] Correspondingly, the middle node T of the pipe wall m
[0075]
[0076] Outer wall node T o
[0077]
[0078] Since only the outer wall temperature T is obtained by real-time monitoring o , and its rate of change with time dT / dt, Therefore, by inverting the above heat balance equation, the relationship between the temperature of the fluid in the tube and the temperature of the outer wall of the tube is established;
[0079] First, establish the node temperature T in the middle of the pipe wall m With T o and T a Direct relationship:
[0080]
[0081] Then establish the inner wall temperature T i With T m 、T o Relationship:
[0082]
[0083] Finally, the temperature of the fluid in the tube and the adjacent nodes T i 、T m Relationship:
[0084]
[0085] At this point, based on the real-time monitored tube outer wall temperature data, equations (5), (6), and (7) are recursively used to obtain the average temperature of the cooling medium fluid in the tube, that is, the inlet and outlet temperatures of the cooling medium in equation (1). The heat Q taken away by the cooling water is obtained 1 , the working status of the flexible DC converter valve water cooling system and the flexible DC converter valve can be obtained.
[0086] above Figure 1 , 2 The water flow monitoring method of a flexible direct current converter valve water cooling system based on a predation algorithm shown is a specific embodiment of the present invention, which has embodied the essential characteristics and progress of the present invention. According to actual use needs and under the guidance of the present invention, equivalent modifications in shape, structure, etc. can be made to the method, which are all within the protection scope of this scheme.
Claims
1. A method for monitoring water flow of a flexible direct current converter valve water cooling system based on a predator algorithm, wherein the flexible direct current converter valve water cooling system comprises a water cooling radiator (1), a temperature sensor (2), a flow sensor (3), a signal transmission device (4) and a host computer (5); the flexible direct current converter valve (6) is arranged in the water cooling radiator (1), and the temperature sensor (2) and the flow sensor (3) are connected to the host computer (5) through the signal transmission device (4); the flow sensor (3) is a non-invasive ultrasonic flow sensor (3), which is installed at the water inlet and / or outlet of the water cooling radiator (1) in a non-breaking pipe manner and is attached to the outer wall of the branch pipe of the converter valve (6), characterized in that: The water flow monitoring method comprises the following steps: 1) Set the dirty data cleaning plan, normal water flow data value interval, and abnormal water flow data value interval; 2) Obtain the water flow signal of the water cooling system of the flexible DC converter valve through a non-invasive ultrasonic sensor; 3) Convert the collected water flow signal into a water flow digital signal through an ADC module; 4) The converted digital signal is stored as raw data in the data cache module, which communicates with the host computer through the 485 protocol to send the raw data to the database, and inserts the current time into the database through the setDate method; 5) Clean the data in the database according to the dirty data cleaning solution, and return the cleaned data to the target database; 6) Repeat steps 2) to 5) until the calibration time when the last data is written into the database and the calibration time when the first data is written into the database reach the set time; 7) According to the set water flow normal data value interval and water flow abnormal data value interval, count the data volume of normal data and the data volume of mutation data, and calculate the mutation frequency; 8) The statistical data and mutation frequency are brought into the predator-prey model to calculate the regular mutation and random mutation, and obtain the relationship between normal and abnormal data in the time domain. The parameter change trend is displayed through a visualization framework, so as to observe the trend graph to monitor the health status of the converter valve water cooling system.
2. According to claim 1, a method for monitoring water flow of a flexible DC converter valve water cooling system based on a predator algorithm is characterized by: In step 1), the data type of dirty data is a data type other than double type or is empty, and the dirty data cleaning method is missing data, data type error data, duplicate data, and the data is defined as empty and the collection time is retained.
3. The method for monitoring water flow of a flexible DC converter valve water cooling system based on a predator algorithm according to claim 2 is characterized by: In step 6), the set time is 360 hours; the non-invasive ultrasonic sensor collects water flow every 2 hours; and a total of 180 data collected continuously for 360 hours is recorded as a statistical cycle.
4. The method for monitoring water flow of a flexible DC converter valve water cooling system based on a predator algorithm according to claim 1, characterized in that: In step 8), the predator-prey model is expressed as: Where x is the amount of normal data, y is the amount of mutation data, λ is the mutation frequency, ε is the acquisition frequency, dx / dt and dy / dt represent the change rate of normal data and mutation data over time.
5. The method for monitoring water flow of a flexible DC converter valve water cooling system based on a predator algorithm according to claim 4 is characterized in that: ε is the acquisition frequency which is 0.00556.
6. The method for monitoring water flow of a flexible DC converter valve water cooling system based on a predator algorithm according to claim 5, characterized in that: The setDate method in step 4) is the setDate method in the PreparedStatement class in the java class library, and the data type of the acquisition time stored in the database as an attribute is long int type.
7. The method for monitoring water flow of a flexible DC converter valve water cooling system based on a predator algorithm according to claim 6 is characterized by: In step 7), the statistical data is obtained by using the MapReduce model to obtain each normal data interval and the number of normal data values in each interval, each abnormal data value interval and the number of abnormal data values in each interval; the mutation frequency λ is calculated, which is expressed as the ratio of the number of mutation values to the total number of collected values 180.
8. The method for monitoring water flow of a flexible DC converter valve water cooling system based on a predator algorithm according to claim 6, characterized in that: In step 5), cleaning the data in the database includes the following steps: 501) Obtaining data from a database; 502) Determine whether the data is missing, if so, proceed to step 505); if not, proceed to step 503); 503) Determine whether the data type is wrong, if so, proceed to step 505); if not, proceed to step 504); 504) Determine whether the data is repeated, if so, proceed to step 505); 505) Define the data as empty and retain the time attribute.
