System for predicting and diagnosing running state of dry-wet combined cooling tower

Abstract

The invention relates to the technical field of industrial control, and discloses a dry-wet combined cooling tower operation state prediction and diagnosis system comprising a load prediction module used for obtaining production plan data and environment information, constructing a basic heat load prediction model, and outputting a total prediction heat load; the thermal modeling module is used for establishing a heat transfer model and an energy consumption model, the heat transfer model outputs cooling amounts in different operation modes, and the energy consumption model outputs total predicted power; the dry-wet decision module is used for constructing a multi-objective optimization function and solving the multi-objective optimization function to obtain a dry-wet switching strategy; and the control execution module is used for executing the dry-wet switching strategy. According to the method, the basic thermal load prediction model is established, so that the prediction precision of the thermal load is improved, and differentiated cooling strategies are provided for different production stages; on the basis of real-time load requirements and environmental conditions, a dry mode or a wet mode is intelligently selected for operation, and a multi-objective optimization function is constructed, so that the balance between cooling requirement meeting and system energy consumption minimization is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial control, more particularly, it relates to a dry-wet combined cooling tower operation state prediction and diagnosis system. BACKGROUND

[0002] With the rapid development of modern industrial production, especially in the fields of intelligent manufacturing, new energy automobile battery production, precision electronic manufacturing, etc., the production process has higher and higher requirements for temperature control precision. As an important industrial cooling equipment, the dry-wet combined cooling tower plays a key role in these high-precision temperature control scenes. The dry-wet combined cooling tower integrates dry cooling and wet cooling two working modes. When the ambient temperature is low or the heat load is small, the dry cooling mode can be used to save water resources and reduce energy consumption. When the ambient temperature rises or the heat load increases, it automatically switches to the wet cooling mode to improve the cooling efficiency through the evaporation latent heat of the sprayed water.

[0003] The existing Chinese patent with the authorization announcement number CN107203180B discloses a cooling tower monitoring method, device and system. The working parameter values transmitted by the collector are received. According to the received working parameter values and the normal value range of each working parameter in the preset multiple working parameters, it is determined whether there is an abnormal working parameter value in the working parameter values. If there are multiple abnormal working parameter values in the working parameter values, according to the abnormal working parameter values and the mapping relationship between the preset working parameter type and the working processing user, the working processing user matched with each abnormal working parameter value is determined respectively. For each determined working processing user, according to at least one abnormal working parameter value matched with the working processing user, the working task associated with the working processing user is determined, and the determined working task is pushed to the working processing user.

[0004] Although the existing technology can detect abnormal working parameters and push maintenance tasks, it can only respond after the parameter abnormality occurs and cannot predict the impact of production plan changes on cooling demand. In the intelligent manufacturing environment, the heat load of the production line often has certain regularity and predictability. The production plan in the manufacturing execution system (MES) or enterprise resource planning system (ERP) can be used to predict the production capacity arrangement in the next few hours, but the existing system cannot use this information to make forward-looking cooling strategy adjustments. In the rapidly changing production environment, the best adjustment opportunity may be missed, resulting in poor overall energy efficiency. SUMMARY

[0005] The purpose of the present application is to provide a dry-wet combined cooling tower operation state prediction and diagnosis system to solve the above problems.

[0006] The present application provides a dry-wet combined cooling tower operation state prediction and diagnosis system, which comprises:

[0007] a load prediction module, obtaining production plan data and environment information, constructing a basic heat load prediction model based on the production plan data and the environment information, the basic heat load prediction model outputting a total predicted heat load;

[0008] a heat modeling module, establishing a heat transfer model and an energy consumption model based on the total predicted heat load and the environment information, the heat transfer model outputting cooling capacity under different operation modes, and the energy consumption model outputting a total predicted power;

[0009] a dry-wet decision module, constructing a multi-objective optimization function based on the cooling capacity under different operation modes and the total predicted power, and solving the multi-objective optimization function to obtain a dry-wet switching strategy;

[0010] a control execution module, executing the dry-wet switching strategy.

[0011] Further, a value of the multi-objective optimization function is a sum of a preset load deviation weight multiplied by a temperature deviation term, a preset energy consumption weight multiplied by an energy consumption term, a preset switching cost weight multiplied by a switching cost term, and a preset fatigue damage weight multiplied by a fatigue damage term;

[0012] The temperature deviation term is a square of an absolute value of a load deviation, and the load deviation is a difference between the total predicted heat load and the cooling capacity under different operation modes output by the heat transfer model under selected operation modes, fan rotating speeds and spraying flow rates; the operation modes include a dry mode and a wet mode;

[0013] The energy consumption term is the total predicted power under the operation modes;

[0014] The switching cost is a preset basic switching cost multiplied by a switching indication function, the switching indication function being 1 when a mode switching occurs, and otherwise being 0; the mode switching occurring means that the operation mode has changed;

[0015] The fatigue damage term is a preset fatigue coefficient multiplied by a square of an absolute value of a load difference at switching, and the load difference at switching is an absolute value of a difference between the cooling capacity under different operation modes before and after the operation mode switching.

[0016] Further, a dynamic programming algorithm is used to solve the multi-objective optimization function, the dynamic programming algorithm including a value function, an immediate cost function and an action space; a value function at a current time is a sum of the immediate cost function at the current time and a preset discount factor multiplied by a value function at a next time, and the immediate cost function is a single-step cost generated by executing an action in the action space at a state s in a state space and at a time t, and the single-step cost is a value of the multi-objective optimization function;

[0017] Each state in the state space includes an operation mode and an operation parameter combination, and the operation parameter combination includes fan rotating speeds and spraying flow rates;

[0018] Each action in the action space comprises a mode decision action and a parameter adjustment action; the mode decision action comprises keeping the dry mode, keeping the wet mode, switching from the dry mode to the wet mode, and switching from the wet mode to the dry mode; and the parameter adjustment action comprises a fan speed adjustment action and a spraying flow adjustment action.

[0019] Further, the specific steps of obtaining the dry-wet switching strategy comprise:

[0020] A Bellman equation is established, which is that, for each state at each time, an action is selected from multiple actions to minimize the value function;

[0021] The dynamic programming algorithm starts from a terminal time of a preset optimization time domain, sets a boundary condition that the value function at the terminal time is a preset value, and calculates the value function of each state at each time in a reverse order of time;

[0022] The value function of each state at each time is obtained by solving the Bellman equation in a reverse recursion manner, and the solving method of the Bellman equation is that a sequence of actions that minimizes the value function is selected as the dry-wet switching strategy; and the dry-wet switching strategy is output, which comprises a sequence of operation modes and operation parameters; the sequence of operation modes is a time sequence decision result, which guides the selection of the operation mode of the system in a future time period; and the operation parameters comprise time sequence set values of the fan speed and the spraying flow.

[0023] Further, the cooling capacity in different operation modes comprises a dry mode cooling capacity and a wet mode cooling capacity.

[0024] The dry mode cooling capacity is a product of a preset dry mode efficiency coefficient, a pipe heat transfer area, a dry mode heat transfer coefficient, and a dry mode temperature difference; the pipe heat transfer area is an effective heat transfer area of a cooling tower coil, the dry mode heat transfer coefficient is a preset standard heat transfer coefficient multiplied by an environmental temperature correction factor and a fan speed correction factor; the environmental temperature correction factor is 1 plus a preset temperature correction factor multiplied by a difference between a predicted environmental temperature in environmental information and a preset reference temperature, and the fan speed correction factor is a preset exponential power of a ratio of the fan speed to a preset rated fan speed; and the dry mode temperature difference is a difference between a cooling water inlet temperature and a predicted environmental dry-bulb temperature; and the predicted environmental dry-bulb temperature is an environmental air temperature in a future time period.

[0025] The wet mode cooling capacity is a product of a preset wet mode efficiency coefficient, a sum of a heat transfer term and an evaporation term, and the pipe heat transfer area.

[0026] Further, the total predicted power is a sum of a fan power and a spraying pump power; the fan power is a product of a preset fan power coefficient, a preset fan exponential power of a ratio of the fan speed to a preset rated fan speed, and a preset rated fan power; and the spraying pump power is a product of a preset spraying pump power coefficient, a preset spraying pump exponential power of a ratio of the spraying flow to a preset rated spraying flow, and a preset rated spraying pump power.

[0027] Further, the heat transfer term is a wet mode heat transfer coefficient multiplied by a wet mode temperature difference; the wet mode heat transfer coefficient is a preset standard wet mode heat transfer coefficient multiplied by an environmental temperature correction factor and a fan speed correction factor, and the wet mode temperature difference is a difference between the cooling water inlet temperature and a predicted environmental wet-bulb temperature; the predicted environmental wet-bulb temperature is an environmental temperature considering the humidity influence in a future period;

[0028] The evaporation term is a preset water vaporization latent heat multiplied by an evaporation mass flow density; the evaporation mass flow density is a product of a preset evaporation coefficient, a steam pressure difference and a spray flow correction factor and a fan speed correction factor, the steam pressure difference is a difference between a saturated steam pressure at the predicted wet-bulb temperature and a predicted environmental steam partial pressure, and the spray flow correction factor is a preset spray index power of a ratio of the spray flow to a preset rated spray flow.

[0029] Further, the saturated steam pressure at the predicted wet-bulb temperature is a saturated water vapor pressure corresponding to the predicted environmental wet-bulb temperature, and a calculation method is that the saturated steam pressure at the predicted wet-bulb temperature is a first wet-bulb term power of a preset base number multiplied by a preset pressure conversion coefficient; the first wet-bulb term is a preset Antoine constant A minus a second wet-bulb term, the second wet-bulb term is a preset Antoine constant B divided by a third wet-bulb term, and the third wet-bulb term is the predicted environmental wet-bulb temperature plus a preset Antoine constant C.

[0030] The predicted environmental steam partial pressure is an actual partial pressure of water vapor in the environment air, and the predicted environmental steam partial pressure is a saturated steam pressure at a predicted environmental dry-bulb temperature multiplied by a predicted relative humidity; the saturated steam pressure at the predicted environmental dry-bulb temperature is a first dry-bulb term power of a preset base number multiplied by a preset pressure conversion coefficient; the first dry-bulb term is a preset Antoine constant A minus a second dry-bulb term, the second dry-bulb term is a preset Antoine constant B divided by a third dry-bulb term, and the third dry-bulb term is the predicted environmental dry-bulb temperature plus a preset Antoine constant C.

[0031] Further, the production plan data includes a predicted production power, a production scale, a set temperature and a production stage, wherein the predicted production power is a planned running power of the production line in a future period, the production scale is a rated running capacity of the production line, the set temperature is a cooling water target temperature required by a process, and the production stage is different process stages including a formation stage and a packaging stage; and the environmental information includes a predicted environmental temperature and a predicted relative humidity, wherein the predicted environmental temperature is an environmental temperature in a future period, and the predicted relative humidity is an environmental humidity in a future period.

[0032] Further, the step of outputting the total predicted heat load includes:

[0033] calculating a thermal load output, the thermal load output being a sum of a preset first regression coefficient multiplied by a predicted production power, a preset second regression coefficient multiplied by a production scale, a preset third regression coefficient multiplied by a temperature term, and a preset constant term; the temperature term being a difference between a preset reference temperature and a set temperature;

[0034] calculating a basic predicted thermal load, the basic predicted thermal load being the thermal load output multiplied by a preset correction coefficient corresponding to the process stage;

[0035] calculating an environmental factor correction amount, the environmental factor correction amount being a difference between a predicted environmental temperature and a preset reference temperature multiplied by a preset temperature correction coefficient, plus a difference between a predicted relative humidity and a preset reference humidity multiplied by a preset humidity correction coefficient;

[0036] the total predicted thermal load being a sum of the basic predicted thermal load and the environmental factor correction amount.

[0037] The present application has the beneficial effects that: the present application improves the prediction accuracy of the thermal load by integrating the MES / ERP system data and the meteorological information to establish a basic thermal load prediction model; the prediction accuracy is effectively improved by using a two-stage prediction strategy, first establishing a basic model based on the production plan data, and then combining with the environmental information for correction; the cooling strategy is differentiated for different production stages according to the characteristics of different process stages, and the adaptability is strong.

[0038] Based on real-time load demand and environmental conditions, the dry mode or wet mode operation is intelligently selected to avoid unnecessary energy waste; through the weight configuration of the multi-objective optimization function, the system energy consumption is minimized under the premise of meeting the cooling demand; the equipment fatigue factor is considered in the multi-objective optimization function, and the equipment service life is prolonged by reasonable switching strategy to reduce frequent start-stop of the equipment; the balance between meeting the cooling demand and minimizing the system energy consumption is realized by constructing a multi-objective optimization function that comprehensively considers the temperature control accuracy, system energy consumption, switching cost and equipment fatigue.

[0039] By precise operation parameter control, the equipment is prevented from running in non-optimal conditions to reduce equipment wear and tear; the slow optimization control layer and the fast closed-loop control layer are combined to ensure stable operation of the system; a perfect safety protection mechanism including fan overload protection, spray system leakage detection, temperature overrun alarm, etc. is established; real-time monitoring of equipment state and automatic fault diagnosis are realized to improve system reliability.

[0040] By intelligent optimization control, the energy consumption of the fan and the spray pump is significantly reduced to realize operation cost saving; by precise spray flow control during wet mode operation, water resource waste is avoided; precise temperature control ensures production process requirements to improve product quality and production efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1is a module example diagram of a dry-wet combined cooling tower operation state prediction and diagnosis system of the present application;

[0042] Figure 2 is a thermal modeling module example diagram of a dry-wet combined cooling tower operation state prediction and diagnosis system of the present application;

[0043] Figure 3 is a dry-wet decision module example diagram of a dry-wet combined cooling tower operation state prediction and diagnosis system of the present application;

[0044] Figure 4 is a control execution module example diagram of a dry-wet combined cooling tower operation state prediction and diagnosis system of the present application. DETAILED DESCRIPTION

[0045] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that discussions of these implementations are merely provided to enable those skilled in the art to better understand and thus implement the subject matter described herein, and are not intended to limit the scope of the disclosure. Various procedures or components can be omitted, altered, or supplemented to those described in some examples. In addition, features described in relation to some examples can be combined in other examples.

[0046] Example 1:

[0047] A dry-wet combined cooling tower operation state prediction and diagnosis system, as shown in Figure 1 , comprises:

[0048] The load prediction module 101 acquires production plan data and environmental information, constructs a basic thermal load prediction model based on the production plan data and environmental information, and the basic thermal load prediction model outputs a total predicted thermal load.

[0049] The load prediction module, as the starting module of the system, acquires production plan data through the MES / ERP system interface, combines environmental information through the meteorological data interface, predicts the cooling thermal load demand in the future period, and provides basic data support for the subsequent modules. This module adopts a two-stage prediction strategy, first establishes a basic thermal load prediction model based on production plan data, then combines environmental information for correction, and finally outputs the total predicted thermal load considering the production process characteristics and environmental influence.

[0050] The MES / ERP system interface is a data interface for a manufacturing execution system (Manufacturing Execution System) and an enterprise resource planning system (Enterprise Resource Planning). The interface is connected through an OPC-UA protocol or a RESTful API to obtain real-time production plan, equipment state, process parameters and other data. The data transmission frequency is once every 5 minutes, and the data format adopts a JSON standard format. The weather data interface is connected to a national meteorological bureau or a third-party weather service provider to obtain weather forecast data for the next 2-8 hours, including temperature, humidity, wind speed, air pressure and other parameters. The data update frequency is once every hour.

[0051] The production plan data includes predicted production power, production scale, set temperature, production stage, etc. The predicted production power is the planned running power of the production line in the future period, with a unit of kilowatts, which is obtained through a production scheduling module of the MES system. The data source is a daily production plan and shift arrangement made by a production planning department. The production scale is the rated running capacity of the production line, with a unit of kilowatts. This parameter is a device nameplate parameter, which reflects the power consumption when the production line is running at full load. The set temperature is the target temperature of the cooling water required by the process, with a unit of degrees Celsius. This temperature is determined by a process engineer according to the battery production process requirements. The cooling water temperature required by the formation process is 15-20°C, and the cooling water temperature required by the packaging process is 20-25°C. The production stage is different process stages, including the formation stage, the packaging stage, etc. The formation stage is the battery activation process, which generates a large amount of heat. The packaging stage is the battery packaging process, which generates relatively small amount of heat.

[0052] The environmental information includes predicted environmental temperature and predicted relative humidity. The predicted environmental temperature is the environmental air dry bulb temperature in the future period, with a unit of degrees Celsius, which is obtained through the weather data interface. The predicted relative humidity is the environmental air relative humidity in the future period, with a unit of %.

[0053] The load prediction module first establishes a mapping relationship between the production plan data and the thermal load to construct a basic thermal load prediction model. The model is divided into production parameter mapping and process stage correction.

[0054] In the production parameter mapping part, a multiple linear regression model is adopted, which calculates the corresponding heat load output by taking the predicted production power, production scale and set temperature as input variables. The specific calculation method is that the heat load output is equal to the predicted production power multiplied by a preset first regression coefficient, plus the production scale multiplied by a preset second regression coefficient, plus the temperature term multiplied by a preset third regression coefficient, and finally plus a preset constant term. The role of the preset constant term is to compensate for the fixed heat load and system basic load not considered in the multiple linear regression model, so as to ensure that the system can still maintain the basic running state when the input variables are zero. The temperature term is the difference between the preset reference temperature and the set temperature, and the default value of the preset reference temperature is 25 degrees Celsius. The reference temperature is determined based on the local annual average temperature and the process suitable temperature. The process suitable temperature is the suitable temperature condition required for the cooling water in the production process to ensure product quality, stable equipment operation or process parameter compliance. These preset regression coefficients are obtained by least squares fitting of historical data. The default value of the first regression coefficient is 0.85, which is dimensionless. The default value of the second regression coefficient is 1.2, which is dimensionless. The default value of the third regression coefficient is 2.5, with the unit of kilowatts per degree Celsius. The default value of the constant term is 50, with the unit of kilowatts. The unit of the third regression coefficient is kilowatts per degree Celsius, which ensures the consistency of the units of each term in the formula. The values of the above regression coefficients and constant term are based on linear regression analysis of at least 6 months of historical production data. When the historical data is insufficient, the industry standard value is used.

[0055] In the process stage correction part, the calculation method of the basic predicted heat load is to multiply the heat load output by the preset correction coefficient corresponding to the process stage to obtain the value of the basic predicted heat load.

[0056] The preset correction coefficient of the process stage is determined according to different production stages. The preset correction coefficient of the formation stage is 1.3, corresponding to a high heat load process. The preset correction coefficient of the packaging stage is 0.8, corresponding to a low heat load process. The preset correction coefficient of other stages is 1.0. The default value of the preset correction coefficient of the process stage is determined based on process heat load characteristic analysis and field test data.

[0057] On the basis of obtaining the basic predicted heat load, the basic heat load prediction model further considers the influence of environmental information on cooling demand, calculates the environmental factor correction amount through the environmental correction model, and superimposes the basic predicted heat load to obtain the final total predicted heat load.

[0058] The environmental factor correction amount is calculated by an empirical formula, which considers the influence of predicted environmental temperature and predicted relative humidity on heat load.

[0059] The calculation method of the environmental factor correction amount is that the difference between the predicted ambient temperature and the preset reference temperature is multiplied by a preset temperature correction coefficient, and the difference between the predicted relative humidity and the preset reference humidity is multiplied by a preset humidity correction coefficient. The default value of the preset reference temperature is 25°C, the default value of the preset reference humidity is 60%, the default value of the preset temperature correction coefficient is 2.5 kilowatts per degree Celsius, and the default value of the preset humidity correction coefficient is 1.8 kilowatts per percentage point, which are determined based on statistical analysis of local meteorological conditions and heat transfer theory calculation.

[0060] The calculation method of the total predicted heat load is that the basic predicted heat load is added to the environmental factor correction amount to obtain the total predicted heat load considering the environmental impact. The total predicted heat load is the final output of the load prediction module and provides accurate load demand data for the subsequent thermal modeling module.

[0061] The thermal modeling module 102 establishes a heat transfer model and an energy consumption model based on the total predicted heat load and environmental information, and the heat transfer model outputs the cooling capacity under different operating modes, and the energy consumption model outputs the total predicted power.

[0062] As shown in Figure 2 , it includes:

[0063] The thermal modeling module receives the total predicted heat load and environmental information output by the load prediction module, establishes an accurate heat transfer model based on the dry-wet combined heat transfer theory, calculates the cooling capacity and energy consumption characteristics under different operating modes, and provides a theoretical basis for operating mode decision-making. This module realizes accurate prediction of the heat transfer performance of the cooling tower under different operating modes by establishing a heat transfer model and a system energy consumption model. The heat transfer model includes a dry mode heat transfer model and a wet mode heat transfer model.

[0064] The thermal modeling module establishes a unified heat transfer model framework based on the dry-wet combined heat transfer theory. The operating modes of the dry-wet combined cooling tower include dry mode and wet mode. In dry mode, the cooling water flows in a closed coil, and the heat is indirectly transferred to the air through the pipe wall without the participation of spray water, and the heat transfer mechanism is forced convection heat transfer; in wet mode, both heat transfer and mass transfer (evaporation) mechanisms are used to achieve cooling, in addition to the convective heat transfer between the cooling water and the air, the closed coil surface is also wetted by spraying, and a large amount of latent heat is absorbed by water evaporation, so the cooling efficiency is higher. The heat transfer mechanisms of the two modes are different, and corresponding mathematical models need to be established respectively.

[0065] When running in dry mode, the cooling tower cools down only through air convection heat transfer, without starting the spray system. The calculation of the dry mode cooling capacity in the dry mode heat transfer model is based on Newton's cooling law, which states that the cooling rate of an object is directly proportional to the temperature difference between the object and the environment. The calculation method is to multiply the preset dry mode efficiency coefficient, the pipe heat transfer area, the dry mode heat transfer coefficient, and the dry mode temperature difference. The pipe heat transfer area is the effective heat transfer area of the 304 stainless steel coil pipe, with a default value of 120 square meters, which is determined based on actual equipment parameters. This parameter is a common basis for dry mode and wet mode heat transfer calculations and directly affects the heat transfer capacity of the cooling tower.

[0066] The dry mode heat transfer coefficient represents the intensity of convective heat transfer between air and the surface of the coil pipe in dry mode, and its value is affected by the predicted ambient temperature and fan speed. The calculation method of the dry mode heat transfer coefficient is to multiply the preset standard heat transfer coefficient by the ambient temperature correction factor and the fan speed correction factor. The ambient temperature correction factor is 1 plus the preset temperature correction factor multiplied by the difference between the predicted ambient temperature and the preset reference temperature, and the fan speed correction factor is the preset exponential power of the ratio of the fan speed to the preset rated fan speed.

[0067] The dry mode temperature difference is defined as the difference between the cooling water inlet temperature and the predicted ambient dry-bulb temperature, which is the temperature potential that drives dry mode heat transfer. When the predicted ambient temperature is low or the fan speed is high, the dry mode heat transfer coefficient increases, and the heat transfer effect is enhanced. The cooling water inlet temperature is the temperature of the cooling water flowing into the cooling tower from the production equipment, with a unit of degrees Celsius, which reflects the intensity of the heat load generated during the production process and is an important benchmark parameter for calculating the heat transfer temperature difference. The cooling water inlet temperature is obtained by real-time measurement through a temperature sensor installed on the cooling tower inlet pipe, with a measurement accuracy of ±0.1°C and a data acquisition frequency of once per minute. This parameter is a common basis for dry mode and wet mode temperature difference calculations and directly affects the driving force of the cooling tower's heat transfer and cooling effect.

[0068] The predicted ambient dry-bulb temperature is the ambient air temperature in the future period, with a unit of degrees Celsius, which is an important parameter for determining the temperature difference in dry mode heat transfer calculation. The predicted ambient dry-bulb temperature is obtained through a meteorological data interface, which connects local weather stations or third-party weather service providers to obtain 2-8 hour hourly temperature forecast data. The data update frequency is once per hour, and the prediction accuracy is ±1°C. When the meteorological data interface is not available, the system uses the statistical average of historical same-period temperature data as a backup prediction value.

[0069] The preset dry mode efficiency coefficient default value is 0.85, which is determined based on the dry cooling tower heat transfer efficiency test data and theoretical analysis; the preset standard heat transfer coefficient default value is 45 watts per square meter Kelvin, the preset temperature correction factor default value is 0.02, the preset rated fan speed default value is 1450 revolutions per minute, and the preset index default value is 0.8, which is determined based on heat transfer theory and equipment technical parameters.

[0070] During wet mode operation, the cooling tower simultaneously activates the fan and sprinkler system, achieving cooling through both heat and mass transfer mechanisms. The wet mode cooling capacity in the wet mode heat transfer model is calculated by multiplying the preset wet mode efficiency coefficient, the tube heat transfer area, and the sum of the heat transfer term and the evaporation term.

[0071] The heat transfer term characterizes the sensible heat transfer in wet mode and is calculated as the wet mode heat transfer coefficient multiplied by the wet mode temperature difference. The wet mode heat transfer coefficient characterizes the intensity of convective heat transfer between the wetted surface and the air and its value is affected by the predicted ambient temperature and fan speed. The wet mode heat transfer coefficient is calculated by multiplying the preset standard wet mode heat transfer coefficient by the ambient temperature correction factor and the fan speed correction factor. The ambient temperature correction factor is 1 plus the preset temperature correction factor multiplied by the difference between the predicted ambient temperature and the preset reference temperature. The fan speed correction factor is the ratio of the fan speed to the preset rated fan speed raised to the power of a preset exponential. The wet mode temperature difference is defined as the difference between the cooling water inlet temperature and the predicted ambient wet bulb temperature. This temperature difference is the temperature potential difference that drives the sensible heat transfer in wet mode.

[0072] The predicted ambient wet-bulb temperature (WPBT) is the ambient temperature for the future time period, accounting for humidity. It is expressed in degrees Celsius and is a key parameter for determining temperature differences and evaporation mass transfer in wet-mode heat transfer calculations. The WPBT reflects the humidity level of the air and is always lower than or equal to the dry-bulb temperature. The WPBT calculation method is based on the predicted ambient dry-bulb temperature and predicted relative humidity, using the Stull wet-bulb temperature empirical formula. This ensures that the WPBT remains lower than or equal to the dry-bulb temperature, consistent with physical laws.

[0073] The evaporation term characterizes the latent heat transfer in wet mode and is calculated by multiplying the preset latent heat of vaporization of water by the evaporation mass flow density. The evaporation mass flow density reflects the intensity of water evaporation per unit area and its value is affected by the spray flow rate and fan speed. The evaporation mass flow density is calculated by multiplying the preset evaporation coefficient, the vapor pressure difference, the spray flow correction factor, and the fan speed correction factor. The vapor pressure difference is the difference between the saturated vapor pressure at the predicted wet-bulb temperature and the predicted ambient vapor partial pressure. This pressure difference is the driving force for water evaporation. The spray flow correction factor is the ratio of the spray flow rate to the preset rated spray flow rate raised to the power of the preset spray exponent. The fan speed correction factor is the ratio of the fan speed to the preset rated fan speed raised to the power of the preset exponent.

[0074] The saturated vapor pressure at the predicted wet-bulb temperature is the saturated water vapor pressure corresponding to the predicted ambient wet-bulb temperature, in units of pascal, and is an important component in determining the vapor pressure difference in the evaporation mass transfer calculation. The saturated vapor pressure is calculated using the Antoine Equation, and the calculation method is that the saturated vapor pressure at the predicted wet-bulb temperature is the first wet-bulb term raised to the power of a preset base number multiplied by a preset pressure conversion coefficient; the first wet-bulb term is the preset Antoine constant A minus the second wet-bulb term, the second wet-bulb term is the preset Antoine constant B divided by the third wet-bulb term, and the third wet-bulb term is the predicted ambient wet-bulb temperature plus the preset Antoine constant C. The default value of the preset base number is 10, the default value of the preset Antoine constant A is 8.07131, the default value of the preset Antoine constant B is 1730.63, and the default value of the preset Antoine constant C is 233.426, which is determined based on the thermodynamic properties of water. The default value of the preset pressure conversion coefficient is 133.322, which is used to convert millimeters of mercury to pascal. The formula has high accuracy in the temperature range of 0-100℃, and is suitable for the operating conditions of the cooling tower.

[0075] The predicted ambient vapor partial pressure is the actual partial pressure of water vapor in the ambient air, in units of pascal, and reflects the absolute humidity level of the ambient air, which is another important component in determining the vapor pressure difference in the evaporation mass transfer calculation. The calculation method of the predicted ambient vapor partial pressure is to multiply the saturated vapor pressure at the predicted ambient dry-bulb temperature by the predicted relative humidity percentage value. The saturated vapor pressure at the predicted ambient dry-bulb temperature is calculated using the same Antoine equation calculation method as the saturated vapor pressure at the predicted wet-bulb temperature, and the calculation method is that the saturated vapor pressure at the predicted ambient dry-bulb temperature is the first dry-bulb term raised to the power of a preset base number multiplied by a preset pressure conversion coefficient; the first dry-bulb term is the preset Antoine constant A minus the second dry-bulb term, the second dry-bulb term is the preset Antoine constant B divided by the third dry-bulb term, and the third dry-bulb term is the predicted ambient dry-bulb temperature plus the preset Antoine constant C. This calculation method is based on the definition of relative humidity, which is the ratio of actual vapor partial pressure to saturated vapor pressure at the same temperature.

[0076] The default value of the preset wet mode efficiency coefficient is 0.92, the default value of the preset standard wet mode heat transfer coefficient is 65 watts per square meter kelvin, and the default value of the preset latent heat of water vaporization is 2260 kilojoules per kilogram, which is determined based on the heat and mass transfer theory and experimental data of wet cooling towers; the default value of the preset evaporation coefficient is 1.2x10-8 kilograms per second square meter pascal, the default value of the preset rated spray flow is 45 cubic meters per hour, and the default value of the preset spray index is 0.6, which is determined based on mass transfer theory, wet air property parameters and spray system characteristics.

[0077] In order to accurately evaluate the energy consumption characteristics of different operating modes, the thermal modeling module establishes an energy consumption model of the system. The total predicted power is the sum of the fan power and the spray pump power, which respectively corresponds to the main energy consumption equipment in the dry mode and the wet mode.

[0078] The calculation of the fan power is based on the fan similarity law, which is a theory describing the performance parameter relationship of geometrically similar fans under different operating conditions. The calculation method is to multiply the preset fan power coefficient, the preset fan index power of the preset rated fan speed ratio, and the preset rated fan power. The default value of the preset fan index is 3, and the fan power changes with the cube of the speed, which is a common energy consumption component of the dry mode and the wet mode.

[0079] The calculation of the spray pump power is based on the pump similarity law, which is similar to the fan similarity law and describes the performance relationship of geometrically similar pumps under different operating conditions. The calculation method is to multiply the preset spray pump power coefficient, the preset spray pump index power of the preset rated spray flow ratio, and the preset rated spray pump power. The default value of the preset spray pump index is 2, and the spray pump power changes with the square of the flow.

[0080] The preset fan power coefficient has a default value of 0.85, the preset fan index has a default value of 3, and the preset rated fan power has a default value of 15 kW, which is determined based on the fan similarity law and the technical parameters of the equipment. The preset spray pump power coefficient has a default value of 0.75, and the preset rated spray pump power has a default value of 8 kW, which is determined based on the pump similarity law and the technical parameters of the equipment.

[0081] Based on the above heat transfer model and energy consumption model, the thermal modeling module calculates the cooling efficiency under different operating modes to provide a basis for subsequent operating mode decision-making. The cooling efficiency prediction method under different operating modes is to divide the cooling capacity corresponding to the operating mode by the total predicted power to obtain the cooling efficiency. The cooling efficiency includes the dry mode cooling efficiency and the wet mode cooling efficiency.

[0082] The dry mode cooling efficiency is equal to the dry mode cooling capacity divided by the fan power, and the wet mode cooling efficiency is equal to the wet mode cooling capacity divided by the sum of the fan power and the spray pump power. By comparing the cooling efficiencies of different modes, the dry-wet decision-making module can provide an energy efficiency evaluation index.

[0083] The dry-wet decision-making module 103 constructs a multi-objective optimization function based on the cooling capacity and the total predicted power under different operating modes, and solves the multi-objective optimization function to obtain a dry-wet switching strategy.

[0084] As shown in Figure 3 , it includes:

[0085] As the core decision-making unit of the system, the dry-wet decision module receives the total predicted heat load output by the load prediction module and the dry mode predicted cooling capacity and the wet mode predicted cooling capacity output by the heat modeling module. By constructing a multi-objective optimization function and using a dynamic programming algorithm to solve the optimal operation mode sequence and corresponding operation parameters, the balance between cooling demand satisfaction and system energy minimization is achieved. Multi-objective optimization is a mathematical optimization problem that optimizes multiple conflicting objectives simultaneously. The optimal solution set is found through weight allocation or Pareto optimality. The dynamic programming algorithm is a mathematical method for solving multi-stage decision problems. By decomposing complex problems into subproblems, the global optimal solution is found using the optimal substructure property. The decision-making process of this module is divided into three logically progressive stages: data preprocessing and state definition, multi-objective optimization function construction, and dynamic programming algorithm solution.

[0086] The dry-wet decision module first preprocesses the received data and establishes the state space and key parameter definitions required for decision-making.

[0087] The system operation state space uses a hierarchical definition method, including the operation mode layer and the operation parameter layer. The operation mode layer includes dry mode and wet mode. In dry mode, only the fan system is started, and cooling is achieved through air convection heat transfer. In wet mode, both the fan and the sprinkler systems are started, and cooling is achieved through both heat and mass transfer mechanisms. The operation parameter layer defines the specific operation parameter combination under each operation mode, including fan speed and sprinkler flow. Each state s in the state space includes the operation mode and the operation parameter combination.

[0088] The fan speed range is 500-1500 revolutions per minute, discretized into a preset number of fan levels, with a default value of 11. Each level interval is 100 revolutions per minute. The purpose of discretization is to convert continuous control variables into discrete choices, facilitating dynamic programming algorithm solution. The sprinkler flow range is 0-60 cubic meters per hour, discretized into a preset number of sprinkler levels, with a default value of 7. Each level interval is 10 cubic meters per hour. In dry mode, the sprinkler flow is fixed at 0, and in wet mode, the sprinkler flow is selected from 1-7 levels according to cooling demand.

[0089] Each action in the action space includes mode decision action and parameter adjustment action. Mode decision action includes maintaining dry mode, maintaining wet mode, switching from dry mode to wet mode, and switching from wet mode to dry mode. Parameter adjustment action includes fan speed adjustment action and sprinkler flow adjustment action. A complete action a includes mode decision, fan speed adjustment, and sprinkler flow adjustment.

[0090] Based on the preprocessed data and state definition, the dry-wet decision module constructs a multi-objective optimization function that considers temperature control accuracy, system energy consumption, switching cost, and equipment fatigue.

[0091] The multi-objective optimization function is calculated within a preset optimization time domain, the length of which is determined based on a balance between a production planning period and a prediction accuracy, to ensure both the foresight of the decision and the reliability of the prediction data. The default value of the preset optimization time domain is 8 hours, and the time resolution is 15 minutes, i.e., the optimization time domain contains 32 decision times, and the selection of the time resolution is based on the system response time and the control accuracy requirement.

[0092] The multi-objective optimization function adopts a minimization strategy, and a comprehensive optimization function is constructed by weighted summation of the four objective items. The mathematical expression of the multi-objective optimization function is: the value of the multi-objective optimization function is equal to the sum of the preset load deviation weight multiplied by the temperature deviation item, the preset energy consumption weight multiplied by the energy consumption item, the preset switching cost weight multiplied by the switching cost item, and the preset fatigue damage weight multiplied by the fatigue damage item.

[0093] The default value of the preset load deviation weight is 100, the default value of the preset energy consumption weight is 0.5, the default value of the preset switching cost weight is 20, and the default value of the preset fatigue damage weight is 80. The default values of the preset weight coefficients are determined based on the multi-objective optimization theory and engineering experience, through the analytic hierarchy process (AHP) and field debugging. The analytic hierarchy process (AHP) is a multi-criteria decision analysis method, which constructs a hierarchical structure model, makes pairwise comparison and judgment, and calculates the relative importance weight of each factor.

[0094] The temperature deviation item quantifies the matching accuracy of the system cooling capacity and the load demand, and its calculation method is the square of the absolute value of the load deviation. The load deviation is the difference between the total predicted heat load and the cooling capacity under the selected operation mode, fan speed and spray flow rate, under the different operation modes output by the heat transfer model, and the unit is kilowatt. The actual cooling capacity is calculated by the heat transfer model of the heat modeling module according to the selected operation mode and operation parameters, and the dry mode cooling capacity calculation formula is used in the dry mode, and the wet mode cooling capacity calculation formula is used in the wet mode. This item ensures that the system prioritizes meeting the cooling load demand, and the closer the load deviation is to 0, the more accurate the matching of the cooling capacity and the demand.

[0095] The energy consumption item quantifies the energy consumption of the system operation, and its calculation method is the total predicted power under the corresponding operation mode. For the dry mode, the energy consumption item is equal to the fan power; for the wet mode, the energy consumption item is equal to the sum of the fan power and the spray pump power.

[0096] The switching cost term quantifies the additional cost caused by the operation mode switching, and the calculation method is to multiply the preset basic switching cost by the switching indicator function. The switching indicator function is 1 when the mode switching occurs, and 0 otherwise. The default value of the preset basic switching cost is 50, which is determined based on the analysis of the device start-stop energy consumption and operating cost. The occurrence of mode switching indicates that the system operation mode has changed, and additional cost is needed to realize the mode switching.

[0097] The fatigue damage term quantifies the fatigue damage caused by the mode switching to the device, and the calculation method is to multiply the preset fatigue coefficient by the square of the absolute value of the load difference at the time of switching. The default value of the preset fatigue coefficient is 0.1, which is determined based on the material fatigue theory and device life evaluation. The material fatigue theory studies the damage accumulation and failure mechanism of materials under cyclic loading. The load difference at the time of switching is the absolute value of the difference in cooling capacity between the different operation modes before and after the operation mode switching, with the unit of kilowatts. This parameter reflects the degree of influence of mode switching on the stability of the system cooling capacity, and is an important input for calculating the fatigue damage term.

[0098] Based on the constructed multi-objective optimization function, the dynamic programming algorithm is used to solve the optimal operation mode sequence, and the system operation constraints are handled during the solving process to ensure the feasibility and safety of the decision results.

[0099] The dynamic programming algorithm includes the value function, the immediate cost function, and the action space. The value function represents the minimum cumulative cost expectation that the system can obtain by adopting the dry-wet switching strategy at state s and time t, which quantifies the optimal decision value from the current state to the end of the optimization time domain. The calculation formula is that the value function at the current time is equal to the sum of the immediate cost function at the current time and the preset discount factor multiplied by the value function at the next time. The default value of the preset discount factor is 0.95, which is determined based on the dynamic programming theory and the future income discounting principle. The immediate cost function represents the single-step cost generated by the system executing action a at state s and time t, and the single-step cost is equal to the value of the multi-objective optimization function.

[0100] The core of the dynamic programming algorithm is the establishment and solution of the Bellman Equation. The solution method of the Bellman Equation is to select the action that minimizes the value function among all possible actions for each time state. The mathematical expression is:

[0101] V(s,t)=min[C(s,a,t)+γ×V(s′,t+1)]

[0102] V(s, t) = mina€A(C(s, a, t) + γV(s', t + 1)) where V(s, t) is the value function of state s at time t, C(s, a, t) is the immediate cost function of performing action a in state s, γ is a preset discount factor, s' is the next state after performing action a, and min denotes the action that makes C(s, a, t) + γV(s', t + 1) attain the minimum value among all possible actions.

[0103] The dynamic programming algorithm starts from the terminal time T of the optimization time horizon, sets the boundary condition V(s, T) = 0, and then calculates the value functions of states at each time in reverse order of time. The next state value function V(s', t + 1) is obtained through the backward recursion process of dynamic programming and has been calculated and stored in the value function table in the previous iteration.

[0104] To improve the prediction accuracy, the system uses a rolling horizon prediction strategy, which updates the prediction parameters based on the latest sensor data and environmental information at each decision-making time. The prediction parameters include environmental temperature, relative humidity, cooling water inlet temperature, etc. The predicted values of these parameters are determined based on historical data trend analysis and weather forecast information. The prediction of environmental temperature and relative humidity uses short-term forecast data provided by the weather data interface, and the prediction of cooling water inlet temperature is based on the production load trend and heat transfer inertia characteristics.

[0105] In the dynamic programming solution process, the system operation constraints are embedded as hard constraints in the feasibility judgment of the action space, ensuring that all candidate actions meet the constraint requirements.

[0106] The cooling capacity constraint ensures that the selected operating mode and operating parameters can meet the total predicted heat load demand, i.e., the actual cooling capacity under the selected mode and parameter combination is not less than the total predicted heat load.

[0107] The equipment operation constraint considers the operation limits of the equipment, including the fan speed range and the spray flow range. The fan speed constraint is 500-1500 revolutions per minute, and the spray flow constraint is 0-60 cubic meters per hour.

[0108] The switching frequency constraint is to avoid frequent switching that may cause damage to the equipment. The minimum switching interval constraint is set so that the time interval between adjacent mode switches cannot be less than the preset minimum switching interval of 15 minutes. The default value of the preset minimum switching interval is 15 minutes, which is determined based on the characteristics of the equipment and the requirements of operation stability.

[0109] The optimal value function of each state at each time is obtained by solving the Bellman equation by reverse recursion. The optimal operation mode sequence and operation parameters are extracted by forward search, and the action sequence that minimizes the value function is selected as the dry-wet switching strategy. After the dynamic programming algorithm is solved, the dry-wet switching strategy is output, including the optimal operation mode sequence, corresponding operation parameters and predicted switching time. The optimal operation mode sequence is the time sequence decision result, which guides the selection of the operation mode of the system in the future period; the corresponding operation parameters include the time sequence set value of the fan speed and the spraying flow, and the operation parameters are dynamically adjusted according to the load change and environmental conditions within the same operation mode sequence, and are updated once at each decision time; and the predicted switching time is the time node of mode switching, which is used for equipment fatigue prediction and control execution.

[0110] The control execution module 104 executes the dry-wet switching strategy. As shown in the figure, it includes: Figure 4

[0111] The control execution module is the execution terminal of the system, receives the dry-wet switching strategy output by the dry-wet decision module, and realizes the coordinated control of the execution equipment such as the fan frequency converter, the spraying system control valve, the cooling water bypass regulating valve and the like through accurate control algorithm and device interface, so as to ensure that the system stably operates according to the optimal operation mode sequence. The module adopts a hierarchical control architecture, including a slow optimization control layer and a fast closed-loop control layer, the cycle of the slow optimization control layer is 15 minutes, and the cycle of the fast closed-loop control layer is 1 second, realizing the organic combination of operation parameter setting and temperature accurate control. The execution process of the module is divided into three logically progressive stages of strategy analysis and parameter extraction, control instruction generation, and device state monitoring and feedback.

[0112] The control execution module first analyzes the received dry-wet switching strategy, extracts the key control parameters, and establishes the control instruction sequence required for execution.

[0113] The dry-wet switching strategy analysis includes operation mode sequence analysis, operation parameter extraction and switching time determination. The operation mode sequence analysis arranges the optimal operation mode sequence in the dry-wet switching strategy in time sequence, identifies the target operation mode at each decision time, including two basic modes of dry mode and wet mode. The operation parameter extraction extracts the corresponding operation parameters from the dry-wet switching strategy, including the time sequence set value of the fan speed and the spraying flow. The fan speed set value ranges from 500 to 1500 revolutions per minute, and the spraying flow set value ranges from 0 to 60 cubic meters per hour, and the spraying flow is fixed at 0 in the dry mode. The switching time determination is based on the predicted switching time in the dry-wet switching strategy, establishes an accurate time control sequence, and ensures that the mode switching is accurately executed at the predetermined time.

[0114] ​Based on the policy analysis results, the control execution module establishes a control instruction generation mechanism to convert the abstract operation mode and parameters into specific device control signals.

[0115] The fan control instruction generation adopts a frequency conversion control strategy to achieve precise control of the axial flow fan speed through a fan frequency converter. The calculation method of the fan control instruction is to divide the target fan speed by the preset rated fan speed, and then multiply it by the preset rated frequency to obtain the frequency converter output frequency set value. The default value of the preset rated fan speed is 1450 revolutions per minute, and the default value of the preset rated frequency is 50 Hz, which is determined based on the fan technical parameters and the frequency converter specifications. The frequency converter output frequency set value is transmitted to the fan frequency converter through a 4-20 milliamp analog signal or a Modbus communication protocol to achieve continuous adjustable control of the speed.

[0116] The spray system control instruction generation adopts a flow proportional control strategy to achieve precise regulation of the spray flow through the spray system control valve. The calculation method of the spray flow control instruction is to divide the target spray flow by the preset rated spray flow, and then multiply it by the preset control valve full opening to obtain the control valve opening set value. The default value of the preset rated spray flow is 45 cubic meters per hour, and the default value of the preset control valve full opening is 100%, which is determined based on the spray system design parameters and the control valve technical specifications. The control valve opening set value is transmitted to the spray system control valve through a 4-20 milliamp analog signal to achieve proportional regulation control of the flow.

[0117] The cooling water bypass regulating valve control instruction generation adopts a PID control algorithm to achieve precise control of the outlet water temperature through adjusting the bypass valve opening based on the cooling water outlet temperature feedback signal. The cooling water bypass regulating valve is installed on the cooling tower outlet pipeline to achieve rapid regulation of the outlet water temperature by adjusting the proportion of the bypass flow to mix part of the high-temperature return water with the cooled water. The output of the PID controller is the bypass valve opening set value, and the calculation method is to add the proportional term, the integral term, and the differential term to obtain the controller output. The proportional term is the preset proportional gain multiplied by the temperature deviation, the integral term is the preset integral gain multiplied by the time integral of the temperature deviation, and the differential term is the preset differential gain multiplied by the time differential of the temperature deviation. The temperature deviation is the difference between the process set temperature and the cooling water outlet temperature, with the unit of Celsius. The default value of the preset proportional gain is 2.5, the default value of the preset integral gain is 0.8 per second, and the default value of the preset differential gain is 0.3 seconds, which are determined based on the PID control theory and the system dynamic characteristics. The bypass valve opening set value ranges from 0% to 100%, and is transmitted to the electric regulating valve actuator through a 4-20 milliamp analog signal.

[0118] To ensure the reliability and safety of control execution, the control execution module establishes a perfect device state monitoring and feedback mechanism.

[0119] Real-time status monitoring acquires the running status information of key equipment through a sensor network, including fan speed feedback, spray flow feedback, cooling water inlet and outlet temperature, system pressure and other parameters. Fan speed feedback is obtained through a speed sensor installed on the fan shaft, with a measurement accuracy of ±5 revolutions per minute and a data acquisition frequency of once per second. Spray flow feedback is obtained through an electromagnetic flowmeter installed on the spray pipeline, with a measurement accuracy of ±2% and a data acquisition frequency of once per second. Cooling water inlet and outlet temperature is measured by a Pt100 platinum resistance temperature sensor, with a measurement accuracy of ±0.1°C and a data acquisition frequency of once per minute. System pressure is monitored by a pressure transmitter, with a measurement accuracy of ±0.25% and a data acquisition frequency of once per minute.

[0120] The control deviation detection and correction mechanism ensures the consistency of the actual running parameters and the target set value. The calculation method of control deviation is the difference between the target set value and the actual feedback value. When the control deviation exceeds the preset deviation threshold value, the system automatically starts the correction program. The default value of the fan speed deviation threshold is 50 revolutions per minute, the default value of the spray flow deviation threshold is 5 cubic meters per hour, the default value of the temperature deviation threshold is 1°C, and the default value of the bypass valve opening deviation threshold is 5%. The above values are determined based on the control accuracy requirements and system stability analysis. The correction program adopts an incremental control strategy, calculates a small incremental adjustment based on the current deviation value, and superimposes it on the original control command. By adjusting the final execution command sent to the frequency converter and control valve, the control deviation is gradually eliminated. The incremental adjustment is set with a limiting mechanism, with a fan speed incremental limit of ±100 revolutions per minute, a spray flow incremental limit of ±10 cubic meters per hour, and a bypass valve opening incremental limit of ±10%, to prevent excessive adjustment from affecting system stability.

[0121] The safety protection and fault handling mechanism provides multiple safety protection for the system. When detecting equipment failure or abnormal state, the system automatically executes the corresponding protection measures. Equipment failure detection includes fan overload protection, sprinkler system leakage detection, temperature overrun alarm, low flow protection, pressure anomaly protection and bypass valve failure protection, etc. Fan overload protection is achieved by monitoring the fan current and power. When the current exceeds the preset overload current threshold of 1.2 times the rated current, the system automatically reduces the fan speed or switches to the standby fan. Sprinkler system leakage detection is achieved by comparing the inlet flow and return flow. If the flow difference exceeds the preset leakage threshold of 5 cubic meters per hour, the system automatically closes the sprinkler system and sends an alarm signal. Temperature overrun alarm is triggered when the cooling water outlet temperature exceeds the preset temperature upper limit of 35°C or is lower than the preset temperature lower limit of 15°C. The system automatically adjusts the operation mode or starts the emergency cooling program. Low flow protection is achieved by monitoring the circulating pump inlet flow. If the flow is lower than the preset minimum flow threshold of 10 cubic meters per hour, the system automatically stops the circulating pump from running to prevent damage from dry running. Pressure anomaly protection is achieved by monitoring the system pressure. When the pressure exceeds the preset pressure upper limit of 0.8 megapascals or is lower than the preset pressure lower limit of 0.1 megapascals, the system sends an alarm and automatically adjusts the operating parameters. Bypass valve failure protection is achieved by monitoring the valve opening feedback and setting value deviation. When the deviation continuously exceeds 15% and the duration exceeds 30 seconds, it is determined that the valve is stuck or the actuator fails, and the system switches to manual control mode and sends a fault alarm.

[0122] The output of the control execution module includes device control signals, operation state feedback and fault alarm information. Device control signals are transmitted to each execution device through an industrial communication network to achieve precise control of the fan frequency converter, sprinkler system control valve and cooling water bypass regulating valve. Among them, the fan speed set value and the sprinkler flow set value come from the optimization output of the dry-wet decision module 103, which remains stable within 15 minutes; the bypass valve opening set value is dynamically adjusted by the PID controller within seconds to achieve precise control of the outlet water temperature. The operation state feedback information includes actual operating parameters, control deviation, device status, etc., providing feedback data for closed-loop control for the upper module. Fault alarm information is obtained through a sensor network for real-time state monitoring, including fault type, fault location, fault time, fault severity, etc., which comes from the equipment failure detection system in the safety protection and fault handling mechanism, covering fan overload protection, sprinkler system leakage detection, temperature overrun alarm, low flow protection, pressure anomaly protection and bypass valve failure protection, etc. Detection results are timely notified to the operator through the human-machine interface and remote monitoring system to ensure safe and reliable operation of the system.

[0123] Through the cooperative work of the above four modules, the embodiment 1 constructs a complete dry-wet combined cooling tower intelligent predictive control system, and realizes the intelligent management of the whole process from load prediction to control execution. The system is especially suitable for industrial scenes such as new energy battery production lines which have high requirements on temperature control accuracy and large changes in production load, and has good engineering application value.

[0124] The above describes the embodiments of the present application, but the embodiments are not limited to the specific embodiments described above, and the specific embodiments described above are only illustrative but not restrictive, and those skilled in the art can make more forms of equivalent embodiments under the inspiration of the embodiments, which are all within the protection of the embodiments.

Claims

1. A dry-wet combined cooling tower operating status prediction and diagnosis system, characterized in that: include: A load forecasting module acquires production plan data and environmental information, and constructs a basic heat load forecasting model based on the production plan data and environmental information, wherein the basic heat load forecasting model outputs a total predicted heat load; a thermal modeling module, which establishes a heat transfer model and an energy consumption model based on the total predicted heat load and environmental information, wherein the heat transfer model outputs the cooling amount under different operating modes, and the energy consumption model outputs the total predicted power; A dry-wet decision module constructs a multi-objective optimization function based on the cooling amount and the total predicted power under the different operating modes, and solves the multi-objective optimization function to obtain a dry-wet switching strategy; Control execution module to execute dry-wet switching strategy.

2. A dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 1, characterized in that: The value of the multi-objective optimization function is the sum of a preset load deviation weight multiplied by a temperature deviation term, a preset energy consumption weight multiplied by an energy consumption term, a preset switching cost weight multiplied by a switching cost term, and a preset fatigue damage weight multiplied by a fatigue damage term; The temperature deviation term is the square of the absolute value of the load deviation. The load deviation is the difference between the total predicted heat load and the cooling capacity output by the heat transfer model under the selected operating mode, fan speed, and spray flow rate under different operating modes. The operating modes include dry mode and wet mode. The energy consumption item is the total predicted power under the operating mode; The switching cost is the preset basic switching cost multiplied by the switching indicator function. The switching indicator function is 1 when a mode switch occurs and 0 otherwise. The occurrence of a mode switch indicates that the operating mode has changed. The fatigue damage term is the preset fatigue coefficient multiplied by the square of the absolute value of the load difference during switching. The load difference during switching is the absolute value of the difference in cooling amount under different operating modes before and after the operating mode is switched.

3. A dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 2, characterized in that: A dynamic programming algorithm is used to solve the multi-objective optimization function. The dynamic programming algorithm includes a value function, an immediate cost function, and an action space. The value function at the current moment is the sum of the immediate cost function at the current moment and the preset discount factor multiplied by the value function at the next moment. The immediate cost function is the single-step cost generated by executing action a in the action space at state s in the state space and time t. The single-step cost is the value of the multi-objective optimization function. Each state in the state space includes an operating mode and an operating parameter combination, and the operating parameter combination includes a fan speed and a spray flow rate; Each action in the action space includes a mode decision action and a parameter adjustment action; the mode decision action includes maintaining dry mode, maintaining wet mode, switching from dry mode to wet mode, and switching from wet mode to dry mode; the parameter adjustment action includes fan speed adjustment action and spray flow adjustment action.

4. A dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 3, characterized in that: The specific steps to obtain the dry-wet switching strategy include: Establish the Bellman equation, which is to select the action that minimizes the value function among multiple actions for the state at each moment; The dynamic programming algorithm starts from the terminal moment of the preset optimization time domain, sets the boundary conditions, and the boundary conditions are that the value function at the terminal moment is the preset value; the value function of each state at each moment is calculated step by step in reverse time order; By solving the Bellman equation through reverse recursion, the value function of each state at each moment is obtained. The solution method of the Bellman equation is: select the action sequence that minimizes the value function as the dry-wet switching strategy; output the dry-wet switching strategy, which includes the operating mode sequence and operating parameters; the operating mode sequence is the timing decision result, which guides the system's operating mode selection in the future period; the operating parameters include the timing setting values ​​of the fan speed and the spray flow rate.

5. The dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 1, characterized in that: The cooling capacity under different operating modes includes dry mode cooling capacity and wet mode cooling capacity; The dry mode cooling capacity is the product of the preset dry mode efficiency coefficient, the tube heat transfer area, the dry mode heat transfer coefficient, and the dry mode temperature difference; the tube heat transfer area is the effective heat transfer area of ​​the coil in the cooling tower; the dry mode heat transfer coefficient is the preset standard heat transfer coefficient multiplied by the ambient temperature correction factor and the fan speed correction factor; the ambient temperature correction factor is 1 plus the preset temperature correction factor multiplied by the difference between the predicted ambient temperature in the environmental information and the preset reference temperature; the fan speed correction factor is the ratio of the fan speed to the preset rated fan speed raised to the power of the preset exponential; the dry mode temperature difference is the difference between the cooling water inlet temperature and the predicted ambient dry-bulb temperature; the predicted ambient dry-bulb temperature is the ambient air temperature in the future time period; The wet mode cooling capacity is the product of the preset wet mode efficiency coefficient, the sum of the heat transfer term and the evaporation term, and the heat transfer area of ​​the tube.

6. A dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 5, characterized in that: The total predicted power is the sum of the fan power and the spray pump power; the fan power is the preset fan power coefficient, the ratio of the fan speed to the preset rated fan speed multiplied by the preset fan exponent power and the preset rated fan power; the spray pump power is the preset spray pump power coefficient, the ratio of the spray flow rate to the preset rated spray flow rate multiplied by the preset spray pump exponent power and the preset rated spray pump power.

7. The dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 5, characterized in that: The heat transfer term is the wet mode heat transfer coefficient multiplied by the wet mode temperature difference. The wet mode heat transfer coefficient is the preset standard wet mode heat transfer coefficient multiplied by the ambient temperature correction factor and the fan speed correction factor. The wet mode temperature difference is the difference between the cooling water inlet temperature and the predicted ambient wet bulb temperature. The predicted ambient wet bulb temperature is the ambient temperature in the future period taking into account the influence of humidity. The evaporation term is the preset latent heat of vaporization of water multiplied by the evaporation mass flow density; the evaporation mass flow density is the product of the preset evaporation coefficient, the vapor pressure difference, the spray flow correction factor and the fan speed correction factor. The vapor pressure difference is the difference between the saturated vapor pressure at the predicted wet-bulb temperature and the predicted ambient vapor partial pressure. The spray flow correction factor is the ratio of the spray flow to the preset rated spray flow raised to the power of the preset spray index.

8. The dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 7, characterized in that: The saturated vapor pressure at the predicted wet-bulb temperature is the saturated water vapor pressure corresponding to the predicted ambient wet-bulb temperature, and is calculated as follows: the saturated vapor pressure at the predicted wet-bulb temperature is the first wet-bulb term of a preset base raised to the power of the first wet-bulb term multiplied by a preset pressure conversion coefficient; the first wet-bulb term is the preset Antoine constant A minus the second wet-bulb term, the second wet-bulb term is the preset Antoine constant B divided by the third wet-bulb term, and the third wet-bulb term is the predicted ambient wet-bulb temperature plus the preset Antoine constant C; The predicted ambient vapor partial pressure is the actual partial pressure of water vapor in the ambient air. The predicted ambient vapor partial pressure is the saturated vapor pressure at the predicted ambient dry-bulb temperature multiplied by the predicted relative humidity. The saturated vapor pressure at the predicted ambient dry-bulb temperature is the first dry-bulb term of a preset base raised to the power of the preset pressure conversion coefficient. The first dry-bulb term is the preset Antoine constant A minus the second dry-bulb term, the second dry-bulb term is the preset Antoine constant B divided by the third dry-bulb term, and the third dry-bulb term is the predicted ambient dry-bulb temperature plus the preset Antoine constant C.

9. The dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 1, characterized in that: Production planning data includes predicted production power, production scale, set temperature, and production stage. The predicted production power is the planned operating power of the production line in the future period, the production scale is the rated operating capacity of the production line, the set temperature is the target cooling water temperature required by the process, and the production stage is different process stages, including the formation stage and the packaging stage. Environmental information includes predicted ambient temperature and predicted relative humidity. The predicted ambient temperature is the ambient temperature in the future period, and the predicted relative humidity is the ambient humidity in the future period.

10. The dry-wet combined cooling tower operating status prediction and diagnosis system according to claim 8, characterized in that: The steps to output the total predicted heating load include: Calculate the heat load output, which is the sum of the predicted production power multiplied by the preset first regression coefficient, the production scale multiplied by the preset second regression coefficient, the temperature term multiplied by the preset third regression coefficient, and the preset constant term; the temperature term is the difference between the preset reference temperature and the set temperature; Calculate the basic predicted heat load, which is the heat load output multiplied by the preset correction factor of the corresponding process stage; Calculating an environmental factor correction amount, where the environmental factor correction amount is the difference between the predicted ambient temperature and the preset reference temperature multiplied by the preset temperature correction coefficient, plus the difference between the predicted relative humidity and the preset reference humidity multiplied by the preset humidity correction coefficient; The total predicted heat load is the sum of the basic predicted heat load and the environmental factor correction.

Images