A self-tuning method for soft measurement parameters of cooling towers based on iterative calculation

Abstract

The present invention relates to a self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation, comprising the steps of: constructing a soft measurement model for the air-water ratio of the cooling tower through a BP neural network algorithm; constructing a soft measurement model for the outlet air temperature of the cooling tower through a BP neural network algorithm. The beneficial effects of the present invention are as follows: By using the existing design data or performance test data through the neural network algorithm, soft measurement models for the air-water ratio and the outlet air temperature of the cooling tower are constructed, and the soft measurement value of the air-water ratio is corrected by combining mathematical and mechanism calculations to improve accuracy. In the production environment, by using the known parameters such as atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature, and circulating water flow rate, the operation of the cooling tower system can be monitored in real time and the state optimization analysis can be carried out. It provides a basic technical basis for the operation, maintenance, overhaul and transformation of the cooling tower, and provides reliable data support for calculating the outlet water temperature for cold end optimization.

Description

Technical Field

[0001] The present invention belongs to the field of thermal performance calculation of cooling towers, and particularly relates to a self-tuning method for soft measurement parameters of cooling towers based on iterative calculation. Background Art

[0002] The cooling tower is one of the important cold-end devices in thermal power plants. The quality of its cooling performance directly affects the economy and safety of the entire power plant operation. At present, the thermal performance calculation of large cooling towers all adopts the experimental method, and regular tests are carried out, which have problems such as many installed meters, long time consumption, poor test environment, and high costs.

[0003] In the thermal performance calculation of cooling towers, the air-water ratio is the core parameter for calculating the performance of cooling towers and is a necessary index for calculating the outlet water temperature of cooling towers. Due to problems such as the harsh operating environment of cooling towers, most of the outlet air temperatures are not equipped with real-time online measurement meters, and there are large errors in real-time environmental wind speed measurement data, it is very difficult to use the mechanism calculation model to calculate the cooling number in real time, so as to obtain the air-water ratio and calculate the outlet water temperature.

[0004] At present, there are methods that use algorithms such as neural networks to perform soft measurement of the outlet water temperature of cooling towers. However, since the outlet water temperature and the air-water ratio are closely related in mechanism, the lack of the air-water ratio as a training parameter for the outlet water temperature results in a deviation between the fitted value and the actual value of the outlet water temperature in the production environment, and it is impossible to accurately evaluate the cooling performance of the cooling tower in real time and provide reliable data for cold-end optimization. Therefore, it is very necessary to calculate the air-water ratio and the outlet water temperature using limited parameters in the actual production environment, achieve accurate calculation, and construct a cooling tower performance model. Summary of the Invention

[0005] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a self-tuning method for soft measurement parameters of cooling towers based on iterative calculation.

[0006] This self-tuning method for soft measurement parameters of cooling towers based on iterative calculation includes the following steps:

[0007] Step 1: Adopt the existing cooling tower design data or performance test data, and construct a soft measurement model of the air-water ratio of the cooling tower through the BP neural network algorithm;

[0008] Step 2: Adopt the existing cooling tower design data or performance test data, and construct a soft measurement model of the outlet air temperature of the cooling tower through the BP neural network algorithm;

[0009] Step 3: Use the ambient temperature as the initial value of the outlet water temperature of the cooling tower, and substitute it together with the parameters collected at the current operating condition point into the soft measurement model of the air-water ratio of the cooling tower, and output the soft measurement value of the air-water ratio as the initial iteration value for subsequent soft measurement of the air-water ratio;

[0010] Step 4: Combine the soft measurement model of the air-water ratio of the cooling tower and the soft measurement model of the outlet air temperature of the cooling tower to construct an iterative computer mechanism model of the air-water ratio;

[0011] Step 4.1: Substitute the initial iterative value of the air-water ratio obtained in Step 3, atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature of the tower, and circulating water flow rate into the soft measurement model of the outlet air temperature of the cooling tower to output the soft measurement value of the outlet air temperature;

[0012] Step 4.2: Calculate the constant parameters through the iterative computer mechanism model of the air-water ratio: Query the enthalpy table of dry and wet air and the enthalpy-entropy table of water vapor to obtain seawater density, seawater specific heat, gravitational acceleration constant, specific volume of dry and wet air, and specific enthalpy of dry and wet air; Substitute the queried data and the soft measurement value of the outlet air temperature into the cooling number calculation formula to obtain the cooling number; The calculation formula of the cooling number is:

[0013]

[0014] In the above formula, c w is the specific heat capacity of water, with the unit of kJ / (kg·°C); K a is the volumetric mass transfer coefficient, with the unit of kg / (m 3 ·h); V is the volume of the water distribution packing, with the unit of m 3 ; Q t is the measured inlet water flow rate of the tower, with the unit of kg / h; h" is the specific enthalpy of saturated air corresponding to the water temperature, with the unit of kJ / kg; t is the water temperature, with the unit of °C; h is the specific enthalpy of wet air, with the unit of kJ / kg(DA); Δt is the cooling water temperature difference, with the unit of °C; h m is the average value of the specific enthalpy of wet air at the inlet and outlet of the tower, with the unit of kJ / kg; is the specific enthalpy of saturated air at the inlet water temperature t 1 of the tower, with the unit of kJ / kg; is the specific enthalpy of saturated air at the outlet water temperature t 2 of the tower, with the unit of kJ / kg; is the specific enthalpy of saturated air corresponding to the average water temperature t m at the inlet and outlet of the tower, with the unit of kJ / kg;

[0015] Step 4.3: Adopt the design data of the cooling tower corrected by performance tests, and use the least squares method to fit the functional relationship between the cooling number and the air-water ratio segmented by wind speed, and calculate the air-water ratio cyclically; The functional relationship between different wind speeds and the air-water ratio segmented by wind speed is:

[0016] Ω = Aλ m

[0017] In the above formula, Ω is the cooling number, λ is the air-water ratio, and A and m are constants;

[0018] Step 4.4: Compare the calculated value of the air-water ratio with the soft measurement value of the air-water ratio. If the error between the two is less than the set value, iteratively correct the soft measurement value of the air-water ratio, output the performance parameters of the cooling tower to improve accuracy; otherwise, decrease the outlet water temperature by the set amplitude value, return to execute Step 3, and continue the loop calculation until the error between the calculated value of the air-water ratio and the soft measurement value of the air-water ratio is less than the set value.

[0019] Preferably, Step 1 specifically includes the following steps:

[0020] Step 1.1: Collect the existing cooling tower design data and performance test data of the cooling tower system within the set time period;

[0021] Step 1.2: After removing the abnormal data from the data collected in Step 1, shuffle the remaining data and divide it into two parts: training data and verification data;

[0022] Step 1.3: Establish a BP neural network, use the air-water ratio as the prediction parameter of the BP neural network, and use atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature, outlet water temperature, and circulating water flow rate as training parameters;

[0023] Step 1.4: Input the training data obtained in Step 1.2 into the established BP neural network, verify the trained soft measurement model of the cooling tower air-water ratio with the verification data obtained in Step 1.2, and finally output the weight matrices of each layer of the soft measurement model of the cooling tower air-water ratio and the air-water ratio fitting function.

[0024] Preferably, the existing cooling tower design data and performance test data in Step 1 include: air-water ratio, atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature, and circulating water flow rate, which are used to monitor the operation of the cooling tower system in real time and analyze the status optimization.

[0025] Preferably, Step 2 specifically includes the following steps:

[0026] Step 2.1: Collect the performance test data and cooling tower design data of the cooling tower system during the test time period;

[0027] Step 2.2: After removing the abnormal data from the data collected in Step 2.1, shuffle the remaining data and divide it into two parts: training data and verification data;

[0028] Step 2.3: Establish a BP neural network, use the outlet air temperature as the prediction parameter of the BP neural network, and use atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature, air-water ratio, and circulating water flow rate as training parameters;

[0029] Step 2.4: Input the training data obtained in Step 2.2 into the established neural network, and use the validation data obtained in Step 2.2 to verify the trained soft sensor model of the outlet air temperature of the cooling tower. Finally, output the weight matrices of each layer of the soft sensor model of the outlet air temperature of the cooling tower and the air-water ratio fitting function.

[0030] Preferably, the parameters collected at the current operating condition point in Step 3 include atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature of the tower, and circulating water flow rate.

[0031] Preferably, the performance parameters of the cooling tower in Step 4.4 include outlet water temperature, outlet air temperature, and cooling number.

[0032] The beneficial effects of the present invention are as follows: By using the existing design data or performance test data through the neural network algorithm, the present invention constructs a soft sensor model of the air-water ratio and the outlet air temperature of the cooling tower, and corrects the soft measurement value of the air-water ratio by combining mathematical and mechanism calculations to improve the accuracy. In the production environment, by using the known parameters such as atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature of the tower, and circulating water flow rate, the operation of the cooling tower system can be monitored in real time and the state optimization analysis can be carried out. It provides a basic technical basis for the operation and maintenance, repair and transformation of the cooling tower, and provides reliable data support for the optimized calculation of the outlet water temperature at the cold end. Description of the Drawings

[0033] Figure 1 It is a flowchart of self-tuning of soft sensor parameters for the cooling tower;

[0034] Figure 2 It is a comparison trend chart of the soft measurement value and the corrected value of the outlet water temperature in the example;

[0035] Figure 3 It is a comparison trend chart of the soft measurement value and the corrected value of the air-water ratio in the example. Detailed Embodiments

[0036] The following further describes the present invention with reference to embodiments. The description of the following embodiments is only for helping to understand the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several modifications can still be made to the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

[0037] In view of the limitations of the existing technology and the development trend of intelligent technology, the present invention proposes a self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation. By combining mathematics and mechanism, the present invention constructs a performance model of the cooling tower, and the performance model of the cooling tower is an air-water ratio iterative computer mechanism model; the BP neural network algorithm is used to perform soft measurement on the air-water ratio and the outlet tower air temperature of the cooling tower, and the soft measurement value is corrected by the mechanism calculation method to obtain the outlet tower water temperature, and the operation of the cooling tower system is monitored and the state optimization analysis is carried out in real time.

[0038] Embodiment 1

[0039] Embodiment 1 of the present application provides a self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation as follows: Figure 1 as shown:

[0040] Step 1: Use the existing cooling tower design data or performance test data to construct a soft measurement model of the air-water ratio of the cooling tower through the BP neural network algorithm;

[0041] Step 2: Use the existing cooling tower design data or performance test data to construct a soft measurement model of the outlet tower air temperature of the cooling tower through the BP neural network algorithm;

[0042] Step 3: Take the environmental temperature as the initial value of the outlet tower water temperature of the cooling tower, and substitute it together with the parameters collected at the current working condition point into the soft measurement model of the air-water ratio of the cooling tower, and output the soft measurement value of the air-water ratio as the initial iteration value of the subsequent soft measurement of the air-water ratio;

[0043] Step 4: Combine the soft measurement model of the air-water ratio of the cooling tower and the soft measurement model of the outlet tower air temperature of the cooling tower to construct an air-water ratio iterative computer mechanism model;

[0044] Step 4.1: Substitute the initial iteration value of the soft measurement of the air-water ratio obtained in Step 3, together with the atmospheric pressure, dry bulb temperature, relative humidity, environmental wind speed, unit load, inlet tower water temperature, and circulating water flow rate, into the soft measurement model of the outlet tower air temperature, and output the soft measurement value of the outlet tower air temperature;

[0045] Step 4.2: Calculate the constant parameters through the air-water ratio iterative computer mechanism model: query the enthalpy table of dry and wet air and the enthalpy-entropy table of water vapor to obtain the seawater density, seawater specific heat, gravitational acceleration constant, dry and wet air specific volume, and dry and wet air specific enthalpy; substitute the data obtained by the query and the soft measurement value of the outlet tower air temperature into the cooling number calculation formula to obtain the cooling number; the calculation formula of the cooling number is:

[0046]

[0047] In the above formula, c w is the specific heat capacity of water, with the unit of kJ / (kg·℃); K a is the volumetric mass transfer coefficient, with the unit of kg / (m3 · h); V is the volume of the water distribution packing, in m 3 ; Q t is the measured water flow rate into the tower, in kg / h; h" is the saturated air specific enthalpy corresponding to the water temperature, in kJ / kg; t is the water temperature, in °C; h is the specific enthalpy of humid air, in kJ / kg(DA); Δt is the cooling water temperature difference, in °C; h m is the average value of the specific enthalpy of humid air at the inlet and outlet of the tower, in kJ / kg; is the saturated air specific enthalpy of the water temperature t 1 into the tower, in kJ / kg; is the saturated air specific enthalpy of the water temperature t 2 out of the tower, in kJ / kg; is the saturated air specific enthalpy corresponding to the average water temperature t m at the inlet and outlet of the tower, in kJ / kg;

[0048] Step 4.3: Using the design data of the cooling tower corrected by performance tests, the functional relationship between the cooling number and the air-water ratio segmented by wind speed is obtained by least squares fitting, and the air-water ratio is calculated iteratively; the functional relationship between different wind speeds and the air-water ratio segmented by wind speed is:

[0049] Ω = Aλ m

[0050] In the above formula, Ω is the cooling number, λ is the air-water ratio, and A and m are constants;

[0051] Step 4.4: Compare the calculated value of the air-water ratio with the soft measurement value of the air-water ratio. If the error between the two is less than the set value, the soft measurement value of the air-water ratio is iteratively corrected, and the performance parameters of the cooling tower are output to improve accuracy; otherwise, the water temperature out of the tower is reduced by the set amplitude value, and return to execute Step 3 to continue the iterative calculation until the error between the calculated value of the air-water ratio and the soft measurement value of the air-water ratio is less than the set value.

[0052] Example 2

[0053] Based on Example 1, Example 2 of this application provides the application of the self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation in a cooling tower system of a certain power plant:

[0054] The effectiveness of the method proposed in the present invention is verified by using the data of a cooling tower in a certain power plant. The soft measurement models of the air-water ratio and the outlet air temperature of the cooling tower are constructed using the data during the test period from September 3, 2020 to September 17, 2020 of the cooling tower system and the design data of the cooling tower. The data of 16 operating points from October 10, 2020 to October 20, 2020 of the cooling tower system are used to test the model effect, and the data sample length is 112.

[0055] In this embodiment, the soft measurement of the water temperature out of the tower for comparison uses the BP neural network algorithm. The water temperature out of the tower is used as the target parameter of the neural network, and the atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, water temperature into the tower, and circulating water flow rate are used as training parameters to compare the correction effect of the comparative example.

[0056] Step 1 specifically includes the following steps:

[0057] Step 1.1: Collect the data of the cooling tower system during the test period from September 3, 2020 to September 17, 2020 and the design data of the cooling tower. The sampling frequency is 1 minute, the data length is 5400, and the collected parameters include: air-water ratio, atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, water temperature into the tower, water temperature out of the tower, and circulating water flow rate.

[0058] Step 1.2: After removing the abnormal data, shuffle the data and divide it into two parts, among which 4320 pieces are used for training data and 1080 pieces are used for verification data.

[0059] Step 1.3: Establish a BP neural network, use the air-water ratio as the prediction parameter of the neural network, and use the atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, water temperature into the tower, water temperature out of the tower, and circulating water flow rate as training parameters.

[0060] Step 1.4: Put 4320 pieces of training data into the established neural network, and verify the trained model with the verification data, and output the weight matrix of each layer of the final model and the air-water ratio fitting function. The air-water ratio weight and normalization parameters are shown in Table 1 below.

[0061] Table 1 Air-water ratio weight and normalization parameter table

[0062] Parameter Weight min max Constant 0.589987 Atmospheric pressure -0.01082 1010 1016 Ambient temperature 0.006083 24.4 25.9 Ambient humidity 0.014734 52.73 64.09 Ambient wind speed 0.000288 0 11.5 Unit load 0.094944 650 1000 Inlet tower water temperature 0.133434 32.56 38.31 Water flow -0.00974 79020 80040 Outlet tower water temperature -0.06228 25.26 26.55

[0063] Step 2 specifically includes the following steps:

[0064] Step 2.1: Collect the data of the cooling tower system during the test period from September 3, 2020 to September 17, 2020 and the design data of the cooling tower. The sampling frequency is 1 minute, the data length is 5400, and the collected parameters include: air-water ratio, atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, water temperature into the tower, air temperature out of the tower, and circulating water flow rate.

[0065] Step 2.2: After removing the abnormal data, shuffle the data and divide it into two parts, among which 4320 pieces are used for training data and 1080 pieces are used for verification data.

[0066] Step 2.3: Establish a BP neural network, using the outlet tower air temperature as the prediction parameter of the neural network, and the atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet tower water temperature, air-water ratio, and circulating water flow rate as the training parameters.

[0067] Step 2.4: Input 4320 pieces of training data into the established neural network, and verify the trained model with the verification data, and output the weight matrix of each layer of the final model and the air-water ratio fitting function. The weights and normalization parameters of the outlet tower air temperature are shown in Table 2 below.

[0068] Table 2 Weights and normalization parameter table of the outlet tower air temperature

[0069] Parameter Weight min max Constant 13.11629 Atmospheric pressure 3.362424 1010 1016 Dry bulb temperature 6.762348 24.4 25.9 Relative humidity 1.679241 52.73 64.09 Ambient wind speed 11.73644 0 11.5 Load 1.397504 650 1000 Inlet tower water temperature 4.003807 32.56 38.3 Water flow 1.725612 79020 80040 Air-water ratio 3.305476 0.552 0.811

[0070] Step 3: Use the data of 16 operating condition points from October 10, 2020 to October 20, 2020 of the cooling tower system to test the model effect, and the data sample length is 112. The collected parameters include atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet tower water temperature, and circulating water flow rate. Using the ambient temperature as the initial value of the outlet water temperature of the cooling tower, and substituting it into the soft sensor model of the air-water ratio of the cooling tower together with other parameters to output the soft measurement value of the air-water ratio, which is used as the initial iteration value of the subsequent soft measurement of the air-water ratio.

[0071] Step 4 specifically includes the following steps:

[0072] Step 4.1: Substitute the initial iteration value of the air-water ratio soft measurement and the atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet tower water temperature, and circulating water flow rate into the soft measurement model of the outlet tower air temperature to output the soft measurement value of the outlet tower air temperature.

[0073] Step 4.2: Perform the calculation of constant parameters, query the enthalpy table of dry and wet air and the enthalpy-entropy table of water vapor to obtain the seawater density, seawater specific heat, gravitational acceleration constant, specific volume of dry and wet air, and enthalpy of dry and wet air. Substitute them into the cooling number calculation formula together with the soft measurement value of the outlet tower air temperature to obtain the cooling number.

[0074] Step 4.3: Obtain the calculated value of the air-water ratio through the functional relationship between the cooling number and the air-water ratio segmented by wind speed. The expression of the segmented function of the cooling number and the air-water ratio is shown in Table 3 below:

[0075] Table 3 Expression table of the segmented function of the cooling number and the air-water ratio

[0076]

[0077] Step 4.4: Compare the calculated value of the air-water ratio with the soft measurement value. If the error between the two is less than 1×10 -10, complete the iterative calculation and correction of the air-water ratio, and output the water temperature out of the tower, the air temperature out of the tower, the cooling number, and the air-water ratio. Otherwise, subtract 0.1 degree Celsius from the water temperature out of the tower and return to step 3 to continue the loop calculation. See the comparison trend chart of the soft measurement value and the corrected value of the water temperature out of the tower in the example in Figure 2 , see the comparison trend chart of the soft measurement value and the corrected value of the air-water ratio in the example in Figure 3 .

Claims

1. A self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation, characterized in that, it includes the following steps: Step 1: Use the existing cooling tower design data or performance test data to construct a soft measurement model of the air-water ratio of the cooling tower through the BP neural network algorithm; The existing cooling tower design data and performance test data include: air-water ratio, atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet tower water temperature, and circulating water flow; Step 2: Use the existing cooling tower design data or performance test data to construct a soft measurement model of the outlet tower air temperature of the cooling tower through the BP neural network algorithm; Step 3: Use the ambient temperature as the initial value of the outlet tower water temperature of the cooling tower, and substitute it into the soft measurement model of the air-water ratio of the cooling tower together with the parameters collected at the current working condition point to output the soft measurement value of the air-water ratio as the initial iteration value of the subsequent soft measurement of the air-water ratio; Step 4: Combine the soft measurement model of the air-water ratio of the cooling tower and the soft measurement model of the outlet tower air temperature of the cooling tower to construct an air-water ratio iterative calculation mechanism model; Step 4.1: Substitute the initial iteration value of the soft measurement of the air-water ratio obtained in Step 3, atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet tower water temperature, and circulating water flow into the soft measurement model of the outlet tower air temperature to output the soft measurement value of the outlet tower air temperature; Step 4.2: Calculate the constant parameters through the air-water ratio iterative calculation mechanism model: query the psychrometric chart and the enthalpy-entropy chart of water vapor to obtain seawater density, seawater specific heat, gravitational acceleration constant, specific volume of dry and wet air, and specific enthalpy of dry and wet air; substitute the data obtained from the query and the soft measurement value of the outlet tower air temperature into the cooling number calculation formula to obtain the cooling number; the calculation formula of the cooling number is: In the above formula, c w is the specific heat capacity of water, with the unit of kJ / (kg·℃); K a is the volumetric mass transfer coefficient, with the unit of kg / (m 3 ·h); V is the volume of the water spray packing, with the unit of m 3 ; Q t is the measured water flow rate into the tower, with the unit of kg / h; h" is the specific enthalpy of saturated air corresponding to the water temperature, with the unit of kJ / kg; t is the water temperature, with the unit of ℃; h is the specific enthalpy of humid air, with the unit of kJ / kg(DA); Δt is the cooling water temperature difference, with the unit of ℃; h m is the average value of the specific enthalpies of the inlet and outlet humid air, with the unit of kJ / kg; is the specific enthalpy of saturated air of the inlet tower water temperature t 1 , with the unit of kJ / kg; is the specific enthalpy of saturated air of the outlet tower water temperature t 2 , with the unit of kJ / kg; is the specific enthalpy of saturated air corresponding to the average water temperature t m of the inlet and outlet towers, with the unit of kJ / kg; Step 4.3: Use the cooling tower design data corrected by performance tests, and use the least squares method to fit the functional relationship between the cooling number and the air-water ratio segmented by wind speed to calculate the air-water ratio; the functional relationship between different wind speeds and the air-water ratio segmented by wind speed is: Ω = Aλ m In the above formula, Ω is the cooling number, λ is the air-water ratio, and A and m are constants; Step 4.4: Compare the calculated value of the air-water ratio with the soft measurement value of the air-water ratio. If the error between the two is less than the set value, perform iterative correction on the soft measurement value of the air-water ratio and output the performance parameters of the cooling tower; otherwise, reduce the outlet tower water temperature by the set amplitude value and return to execute Step 3 until the error between the calculated value of the air-water ratio and the soft measurement value of the air-water ratio is less than the set value.

2. The self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation according to claim 1, characterized in that, Step 1 specifically includes the following steps: Step 1.1: Collect the existing cooling tower design data and performance test data of the cooling tower system within a set time period; Step 1.2: After removing the abnormal data in the data collected in Step 1, shuffle the remaining data and divide it into two parts: training data and verification data; Step 1.3: Establish a BP neural network, use the air-water ratio as the prediction parameter of the BP neural network, and use atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet tower water temperature, outlet tower water temperature, and circulating water flow as training parameters; Step 1.4: Input the training data obtained in Step 1.2 into the established BP neural network, and verify the trained soft measurement model of the cooling tower air-water ratio with the verification data obtained in Step 1.

2. Finally, output the weight matrices of each layer of the soft measurement model of the cooling tower air-water ratio and the air-water ratio fitting function.

3. The self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation according to claim 1 or 2, characterized in that, Step 2 specifically includes the following steps: Step 2.1: Collect the performance test data and design data of the cooling tower system during the test period; Step 2.2: After removing the abnormal data from the data collected in Step 2.1, shuffle the remaining data and divide it into two parts: training data and verification data; Step 2.3: Establish a BP neural network, use the outlet air temperature of the cooling tower as the prediction parameter of the BP neural network, and use the atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature, air-water ratio, and circulating water flow rate as training parameters; Step 2.4: Throw the training data obtained in Step 2.2 into the established neural network, and verify the trained soft measurement model of the outlet air temperature of the cooling tower with the verification data obtained in Step 2.

2. Finally, output the weight matrices of each layer of the soft measurement model of the outlet air temperature of the cooling tower and the air-water ratio fitting function.

4. The self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation according to claim 1, characterized in that: The parameters collected at the current operating condition point in Step 3 include atmospheric pressure, dry bulb temperature, relative humidity, ambient wind speed, unit load, inlet water temperature, and circulating water flow rate.

5. The self-tuning method for soft measurement parameters of a cooling tower based on iterative calculation according to claim 1, characterized in that: The performance parameters of the cooling tower in Step 4.4 include outlet water temperature, outlet air temperature, and cooling number.

Images