Thermal power circulating water pump group control method and system, electronic device and storage medium

Abstract

The application provides a thermal power circulating water pump system operation method, system, device and storage medium considering equipment health degree, and belongs to the technical field of intelligent control of circulating water systems of thermal power plants. The method synchronously collects electrical, mechanical and performance data of each candidate water pump, calculates the real-time health degree index, and comprehensively considers energy consumption and equipment health degree attenuation in multi-objective optimization decision, to form an operation combination and frequency control strategy. Further, the decision result is scored by an operation effect prediction model, and when the score is lower than a threshold, the health degree calculation and optimization parameters are adaptively adjusted to realize closed-loop optimization. The application can significantly reduce energy consumption and delay equipment degradation under the premise of guaranteeing the operation requirements of the circulating water system, and improves the overall operation safety and economy of the system.

Description

Technical Field

[0001] This invention belongs to the field of operation and control technology of circulating water system in thermal power plants, and in particular relates to control methods, systems, electronic equipment and storage media for circulating water pump sets in thermal power plants. Background Technology

[0002] The circulating water system in a thermal power plant is a critical component for condenser cooling and typically consists of multiple circulating water pumps. Traditional operating models often employ fixed combinations or manual, experience-based switching. This lack of consideration for the health status of individual equipment leads to accelerated deterioration of high-load equipment while low-health equipment continues to be frequently used. The selection of operating combinations and frequencies is primarily based on experience or single energy consumption indicators, lacking comprehensive optimization that considers both energy consumption and equipment lifespan. Furthermore, the absence of a closed-loop feedback mechanism prevents dynamic adjustments to control strategies based on actual operating results, resulting in long-term energy efficiency decline and increased maintenance costs. These problems directly make it difficult to simultaneously achieve energy conservation and consumption reduction while ensuring equipment reliability in the thermal power plant's circulating water system. Summary of the Invention

[0003] In view of this, the present invention aims to overcome the defects in the prior art and proposes a control method, system, electronic equipment and storage medium for thermal power circulating water pump sets, so as to solve the technical problems in the prior art that it is impossible to ensure the system operation requirements while taking into account the lowest energy consumption and the optimal equipment health degradation, and lacks a closed-loop adaptive optimization mechanism.

[0004] To achieve the above objectives, the technical solution created by this invention is implemented as follows: The control method for thermal power plant circulating water pump sets includes the following steps: S1. Obtain the current operating requirements of the circulating water system and simultaneously collect the electrical, mechanical, and performance operating data of each candidate water pump; S2. Based on the electrical, mechanical and performance operation data, calculate the health factors that reflect the electrical status, mechanical status and operating efficiency status of the water pump, and weight and fuse the health factors into a comprehensive real-time health index for each water pump. S3. Taking the current operating requirements as constraints, and combining the real-time health index of each water pump calculated in step S2, a comprehensive optimization calculation is performed with the goal of minimizing the total energy consumption of the system and optimizing the decline of equipment health, and then a water pump operation combination and the operating frequency of each water pump that meet the requirements are determined. S4. Execute the decision results of step S3 to control the operating status of each water pump; at the same time, based on the feedback of the execution effect of this decision, construct and solve the parameter optimization problem with historical operating effect as the evaluation target, perform closed-loop adaptive adjustment of the health calculation weight in step S2 and / or the optimization calculation weight in step S3, and apply the adjusted new parameter set to subsequent control cycles.

[0005] Furthermore, in step S2, the calculation of health factors and weighted fusion are achieved through the following steps: S2.1: Obtain three-phase electrical signals from the frequency converter of the water pump motor, obtain vibration signals from the water pump body, and obtain the real-time flow rate, head data and input power of the water pump; S2.2: Perform spectrum analysis on the three-phase electrical signals, and calculate the electrical health factor (HE) based on the ratio of the energy of a specific fault frequency component to the fundamental frequency energy. e The calculation formula is: HE e = max(0, 1 - K e * (E fault / E fund )), where K e E is the calibration coefficient. fault For fault characteristic frequency band energy, E fund It is the fundamental wave energy; S2.3: Perform envelope demodulation and spectrum analysis on the vibration signal, and calculate the mechanical health factor (HE) based on the amplitude at the bearing's characteristic frequency. m The calculation formula is: HE m = max(0, 1 - Σ(W n * A n / A threshold_n )), where W n A is the weighting coefficient. n For the characteristic frequency amplitude, A threshold_n For the corresponding alarm threshold; S2.4: Based on the real-time flow rate Q, head H, and input power P in Calculate the current efficiency η actual The calculation formula is: Where: ρ: density of circulating water, generally taken as 1000 kg / m³ 3 g: acceleration due to gravity, taken as 9.81 m / s² 2 Q: Real-time flow rate of the water pump, in meters (m³) 3 / s; H: Real-time head of the water pump, in meters; P in : Input power of the water pump, in watts (W); S2.5: Calculate the current efficiency η obtained in step S1.4. actual Compared with the baseline efficiency η under healthy conditions baseline Comparison, calculation of performance health factor HE p The calculation formula is: HE p = η actual / η baseline The baseline efficiency η under healthy conditions baselineThe efficiency is obtained in advance by: when the pump is in a healthy state (e.g., after installation, commissioning, or overhaul), obtaining its head, flow rate, and input power at different operating frequencies (or flow rates) through frequency conversion testing; calculating and plotting its efficiency-flow rate (or efficiency-frequency) characteristic curve, which serves as the health baseline efficiency curve for the pump and is stored. During online operation, the corresponding η can be queried from the baseline curve based on the real-time flow rate (or operating frequency). baseline Values. For example, when the system is running stably and the equipment is in a "healthy" state, by changing the inverter frequency (e.g., from 30Hz to 50Hz in fixed steps), at each frequency point, after the system stabilizes, synchronously record the flow rate Q, head H, and input power at that frequency. P in Then according to the formula Calculate the efficiency value for each test point under "healthy condition". Then, plot the flow rate Q (or frequency f) on the x-axis and the efficiency η on the y-axis to fit a smooth curve (or create a lookup table). This smooth curve is the healthy baseline efficiency curve η. baseline When used online, the corresponding η can be obtained by interpolation or table lookup from the baseline curve based on the real-time measured flow rate Q (or operating frequency f). baseline ; S2.6: The electrical health factor HE e Mechanical health factor HE m With performance health factor HE p The weighted fusion is performed to obtain the comprehensive real-time health index HI, calculated as follows: HI = w e * HE e + w m * HE m + w p * HE p , where w e It is the electrical health factor HE e The weighting coefficient, w m It is the mechanical health factor HE m The weighting coefficient, w p It is the performance health factor HE p The weighting coefficient, w e +w m +w p =1.

[0006] Furthermore, in step S3, the method for comprehensive optimization calculation with the goal of minimizing total system energy consumption and optimizing equipment health degradation is as follows: Construct an objective function that minimizes both total system energy consumption and the rate of equipment health degradation, and solve for the pump operating combinations and operating frequencies of each pump that meet the requirements: The calculation formula is: J = α * Σ(x i * P i (f i )) + β *Σ[ x i * (1 - HI i ) * (P i (f i ) / P i_rated ) γ ]; where x i f represents the start / stop status of water pump i. i For the operating frequency, P i (f i P represents the corresponding power. i_rated The rated power is α and β, which are weighting coefficients and α+β=1. γ>1 is the load influence coefficient. The constraints include at least the following: the total flow provided by all operating pumps is equal to the total flow value in the current operating demand; the operating frequency of each pump is within its allowable range; and pumps with a health index lower than the safety threshold are not included in the optimization.

[0007] Furthermore, the adaptive adjustment in step S4 is achieved through a pre-trained machine learning model, specifically including the following steps: S4.1: Construct a performance prediction model M eval Its input feature vector F includes at least the following three categories: a) Health-related characteristics F H During the current decision-making cycle, the electrical health factor (HE) of each candidate water pump is... e Mechanical health factor HE m HE (Health Factor) p and its comprehensive health index HI; b) Decision-related characteristics F D The operating frequency f of each pump in the optimal operating combination output by step S3. i Its corresponding power P i (f i The objective function value J* obtained from the decision calculation; c) Operating condition related characteristics F C The current system's total traffic demand, Q demand Circulating water inlet temperature T in ; S4.2: The predicted performance model M eval The training objective is to output the predicted score V. pred The predicted score V predDefined as the expected average unit energy consumption E of the system over the next N control cycles. avg With the average health decay rate ΔHI avg The weighted composite score is calculated using the following formula: V pred = 1 / [λ1* E avg + λ2* ΔHI avg ], where λ1 is the average unit energy consumption E avg The weighting coefficient, λ2 is the average health decay rate ΔHI avg The weighting coefficients, λ1 + λ2 = 1, are preset positive normalized weighting coefficients used to adjust the relative importance of energy consumption targets and equipment health maintenance targets in the comprehensive score. S4.3: When the predicted score V output by the running effect prediction model is... pred The parameter adjustment process is triggered when any of the following conditions are met. a) When the predicted score V pred continuously below the preset threshold V th Reaching M times, b. When the predicted score V pred Below the dynamic threshold V th_dynamic , V th_dynamic = μ * avg(V hist ), where avg(V hist ) represents the average historical predicted score over the most recent K control periods, where K is a preset positive integer and μ is the sensitivity coefficient (0 < μ < 1).

[0008] Furthermore, the parameter adjustment process in step S4.3 includes: First, define the set of parameters to be optimized, θ = {w}, consisting of the health calculation weights and the optimization target weights. e w m w p , α, β}; Secondly, an optimization problem is constructed using historical operational results as feedback. Specifically, this involves using the comprehensive operational performance score V recorded in the historical database over the most recent K control cycles. actual The objective is to maximize the average value, i.e., the objective function is maximize avg(V). actual (θ)) and impose a limit on the magnitude of parameter variation (|θ) new - θ old | ≤ Δ max Constraints; Then, the Bayesian optimization algorithm is used to optimize the parameter set θ under the constraints, and a candidate new parameter set θ is obtained. newThen, based on the system simulation model, it was verified whether it met the core operational constraints of flow balance and equipment safety; Finally, the new parameter set θ will be verified. new It is applied to the next control cycle to complete the closed-loop iteration and adaptive update of the control strategy.

[0009] Furthermore, the performance prediction model M eval Train and update in the following ways: S4.4: Collect historical running data and construct a supervised learning sample set, where each sample consists of an input feature vector F and its corresponding true performance label V. true constitute; The constituent data of the feature vector F correspond to the starting state of a complete decision cycle, including the health-related features of each water pump at that moment. Realistic effect label V true Based on the average unit energy consumption E actually observed in N consecutive control cycles following this decision cycle. avg_true The decrease in average health ΔHI avg_true Through formula V true = 1 / [λ1* E avg_true + λ2* ΔHI avg_true The calculation shows that λ1 + λ2 = 1, where λ1 and λ2 are preset positive normalized weighting coefficients used to adjust the relative importance of energy consumption targets and equipment health maintenance targets in the comprehensive score. Using the above sample set, with feature vector F as input and true effect label V... true To output labels, supervised training is performed on the gradient boosting decision tree or deep neural network model to obtain an initial performance prediction model M. eval ; S4.5: After every L control cycles, the system automatically records the newly generated {F, V} within that time period. true The samples are added to the training set to predict the running effect M. eval Incremental learning is used to track the slow changes in equipment aging and system operating conditions.

[0010] Furthermore, the method is executed in an edge-cloud collaborative architecture, wherein step S1 is executed in an edge computing unit deployed on the water pump side, and steps S2 and S4 are executed in a cloud server deployed at the plant level. The edge computing unit and the cloud server interact with each other through an industrial communication protocol.

[0011] A control system for a thermal power plant circulating water pump set, used to implement the above method, includes: A health calculation module is used to perform steps S1 and S2 in claim 1; An optimization decision-making and adaptive update module is used to perform steps S3 and S4 in claim 1; The control execution interface is used to connect to and control the circulating water pump and its frequency converter.

[0012] An electronic device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the above method.

[0013] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0014] Compared with existing technologies, the present invention has the following advantages: This invention acquires the three-phase electrical signals, vibration signals, and flow and head data of the water pump simultaneously during the data acquisition phase. It then uses spectrum analysis and envelope demodulation to calculate electrical health factors, mechanical health factors, and performance health factors, respectively, and performs weighted fusion to obtain a comprehensive real-time health index, thereby quantifying the equipment's operating status at the source. In the multi-objective optimization decision-making phase, this health index is introduced into the mixed-integer optimization problem, forming a weighted objective function together with the target energy consumption. This ensures that the optimization result is not only optimal in terms of energy consumption but also prioritizes equipment with higher health, reducing the rate of degradation.

[0015] Furthermore, a multi-feature-based operational performance prediction model is introduced during execution to comprehensively score energy consumption and health degradation over a future period. Once the score falls below a threshold, an adaptive parameter adjustment process is triggered, performing Bayesian optimization on the health calculation weights and optimization target weights. This complete technology chain, consisting of real-time status perception, health-driven optimization decision-making, and closed-loop feedback adjustment, enables the dynamic self-evolution of the operational strategy. It can simultaneously reduce energy consumption and extend equipment life while ensuring the operational needs of the circulating water system, thereby improving the system's safety, economy, and intelligence. Attached Figure Description

[0016] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments and descriptions of the invention are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This invention provides a flowchart of a control method for a thermal power plant circulating water pump unit. Figure 2 This invention provides a schematic diagram of the structure of a thermal power plant circulating water pump unit control system; Figure 3 This invention provides a structural schematic diagram of an electronic device for a thermal power circulating water pump unit. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0018] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0019] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0020] The invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] The control method for thermal power plant circulating water pump sets includes the following steps: S1. Obtain the current operating requirements of the circulating water system and simultaneously collect the electrical, mechanical, and performance operating data of each candidate water pump; S2. Based on the electrical, mechanical and performance operation data, calculate the health factors that reflect the electrical status, mechanical status and operating efficiency status of the water pump, and weight and fuse the health factors into a comprehensive real-time health index for each water pump. S3. Taking the current operating requirements as constraints, and combining the real-time health index of each water pump calculated in step S2, a comprehensive optimization calculation is performed with the goal of minimizing the total energy consumption of the system and optimizing the decline of equipment health, and then a water pump operation combination and the operating frequency of each water pump that meet the requirements are determined. S4. Execute the decision results of step S3 to control the operating status of each water pump; at the same time, based on the feedback of the execution effect of this decision, construct and solve the parameter optimization problem with historical operating effect as the evaluation target, perform closed-loop adaptive adjustment of the health calculation weight in step S2 and / or the optimization calculation weight in step S3, and apply the adjusted new parameter set to subsequent control cycles.

[0022] In step S2, the calculation of health factors and weighted fusion are achieved through the following steps: S2.1: Obtain three-phase electrical signals from the frequency converter of the water pump motor, obtain vibration signals from the water pump body, and obtain the real-time flow rate, head data and input power of the water pump; S2.2: Perform spectrum analysis on the three-phase electrical signals, and calculate the electrical health factor (HE) based on the ratio of the energy of a specific fault frequency component to the fundamental frequency energy. e The calculation formula is: HE e = max(0, 1 - K e * (E fault / E fund )), where K e E is the calibration coefficient. fault For fault characteristic frequency band energy, E fund It is the fundamental wave energy; "Fundamental frequency energy" is the reference energy for normal operation; "energy of specific fault frequency components" is the abnormal energy characterizing a fault. By comparing the ratio of the two, the current electrical health status of the motor can be quantified, and the electrical health factor HE can be calculated. e The method for obtaining "fundamental wave energy" involves performing a Fast Fourier Transform on the acquired three-phase current signal (usually processing a single phase or a composite vector), finding the amplitude corresponding to 50Hz (or the corresponding operating frequency) on the spectrum, and then calculating its square (or a calculation related to signal length and sampling rate) to represent the fundamental wave energy. The method for obtaining "energy of specific fault frequency components" involves performing spectral analysis on the current signal, locating these expected fault characteristic frequency points or their narrow bands on the spectrum, and then calculating the total energy of the signal components within that band. Under healthy conditions, E... fault The ratio approaches 0 and is very small. e Approaching 1. As the fault progresses, E fault Increase, ratio increases, HE e Decrease.

[0023] S2.3: Perform envelope demodulation and spectrum analysis on the vibration signal, and calculate the mechanical health factor (HE) based on the amplitude at the bearing's characteristic frequency. m The calculation formula is: HE m = max(0, 1 - Σ(W n * An / A threshold_n )), where W n A is the weighting coefficient. n For the characteristic frequency amplitude, A threshold_n For the corresponding alarm threshold; The weighting coefficient W n This is a preset value, and its magnitude reflects the relative importance of the failure modes corresponding to different bearing characteristic frequencies on the overall mechanical health of the equipment. W n The weights can be determined through one of the following methods: 1. Based on expert experience: Domain experts allocate weights according to the severity of each bearing failure mode; 2. Based on historical data: Statistical analysis of historical vibration data and failure records is performed, and the weights of each characteristic frequency component are determined based on the strength of their correlation with the occurrence of the failure; 3. Based on industry standards: Weights are set according to relevant vibration assessment standards. Each weight coefficient must satisfy a normalization condition (e.g., ΣW). n = 1), to ensure the rationality of the health factor calculation. It should be noted that other weighting coefficients involved in this application are also determined by one of the above methods.

[0024] S2.4: Based on the real-time flow rate Q, head H, and input power P in Calculate the current efficiency η actual The calculation formula is: Where: ρ: density of circulating water, generally taken as 1000 kg / m³ 3 g: acceleration due to gravity, taken as 9.81 m / s² 2 Q: Real-time flow rate of the water pump, in meters (m³) 3 / s; H: Real-time head of the water pump, in meters; P in : Input power of the water pump, in watts (W); S2.5: Calculate the current efficiency η obtained in step S1.4. actual Compared with the baseline efficiency η under healthy conditions baseline Comparison, calculation of performance health factor HE p The calculation formula is: HE p = η actual / η baseline The baseline efficiency η under healthy conditions baseline The efficiency is obtained in advance by: when the pump is in a healthy state (e.g., after installation, commissioning, or overhaul), obtaining its head, flow rate, and input power at different operating frequencies (or flow rates) through frequency conversion testing; calculating and plotting its efficiency-flow rate (or efficiency-frequency) characteristic curve, which serves as the health baseline efficiency curve for the pump and is stored. During online operation, the corresponding η can be queried from the baseline curve based on the real-time flow rate (or operating frequency). baselineValues. For example, when the system is running stably and the equipment is in a "healthy" state, by changing the inverter frequency (e.g., from 30Hz to 50Hz in fixed steps), at each frequency point, after the system stabilizes, synchronously record the flow rate Q, head H, and input power at that frequency. P in Then according to the formula Calculate the efficiency value for each test point under "healthy condition". Then, plot the flow rate Q (or frequency f) on the x-axis and the efficiency η on the y-axis to fit a smooth curve (or create a lookup table). This smooth curve is the healthy baseline efficiency curve η. baseline When used online, the corresponding η can be obtained by interpolation or table lookup from the baseline curve based on the real-time measured flow rate Q (or operating frequency f). baseline .

[0025] S2.6: The electrical health factor HE e Mechanical health factor HE m With performance health factor HE p The weighted fusion is performed to obtain the comprehensive real-time health index HI, calculated as follows: HI = w e * HE e + w m * HE m + w p * HE p , where w e It is the electrical health factor HE e The weighting coefficient, w m It is the mechanical health factor HE m The weighting coefficient, w p It is the performance health factor HE p The weighting coefficient, w e +w m +w p =1.

[0026] In step S3, the method for comprehensive optimization calculation with the goal of minimizing total system energy consumption and optimizing equipment health degradation is as follows: Construct an objective function that minimizes both total system energy consumption and the rate of equipment health degradation, and solve for the pump operating combinations and operating frequencies of each pump that meet the requirements: The calculation formula is: J = α * Σ(x i * P i (f i )) + β *Σ[ x i * (1 - HI i ) * (P i (f i ) / P i_rated )γ ]; where x i f represents the start / stop status of water pump i. i For the operating frequency, P i (f i P represents the corresponding power. i_rated The rated power is α and β, which are weighting coefficients and α+β=1. γ>1 is the load influence coefficient. The constraints include at least the following: the total flow provided by all operating pumps is equal to the total flow value in the current operating demand; the operating frequency of each pump is within its allowable range; and pumps with a health index lower than the safety threshold are not included in the optimization.

[0027] The objective function J is the core evaluation index of the mixed-integer optimization problem constructed in step S2 of this invention, used to comprehensively evaluate the overall cost of different pump operation schemes. Its expression is J = α * Σ(x i * P i (f i )) + β * Σ[ x i * (1 - HI i ) * (P i (f i ) / P i_rated ) γ It consists of two parts: the first part is α * Σ(x) i * P i (f i )) represents the total energy consumption of the system, and the second part is β * Σ[ x i * (1 - HI i ) * (P i (f i ) / P i_rated ) γ The symbol ] represents the penalty for device health degradation. Where x i P represents the start / stop state of water pump i. i (f i Let f be the operating frequency of the i-th water pump. i The actual input power under HI i P is a real-time health index. i_rated Let J be the rated input power of the i-th pump under its nameplate or design conditions, α and β be the weighting coefficients for adjusting the priority of energy saving and equipment health maintenance, and γ>1 be the load influence coefficient for amplifying the impact of high loads. The optimization solution aims to minimize J. Among all feasible solutions that satisfy flow balance, frequency range, and health protection constraints, the pump operation combination and frequency setting with low total energy consumption and slow health degradation are selected. This achieves the dual technical effect of ensuring the operational needs of the circulating water system while simultaneously achieving energy saving and extending equipment lifespan.

[0028] It should be noted that α is the total energy consumption term of the system Σ(x i * P i (f i The weighting coefficient α represents the system's overall energy consumption. A larger value of α indicates that the optimization algorithm prioritizes reducing the system's total energy consumption during decision-making. β is the penalty term for device health degradation Σ[x]. i * (1-HI) i ) * (P i (f i ) / P i_rated ) γ The larger the value of β, the more the optimization algorithm focuses on protecting the health of the equipment and slowing down its degradation rate when making decisions.

[0029] The adaptive adjustment in step S4 is achieved through a pre-trained machine learning model, specifically including the following steps: S4.1: Construct a performance prediction model M eval Its input feature vector F includes at least the following three categories: a) Health-related characteristics F H During the current decision-making cycle, the electrical health factor (HE) of each candidate water pump is... e Mechanical health factor HE m HE (Health Factor) p and its comprehensive health index HI; b) Decision-related characteristics F D The operating frequency f of each pump in the optimal operating combination output by step S3. i Its corresponding power P i (f i The objective function value J* obtained from the decision calculation; c) Operating condition related characteristics F C The current system's total traffic demand, Q demand Circulating water inlet temperature T in ; S4.2: The predicted performance model M eval The training objective is to output the predicted score V. pred The predicted score V pred Defined as the expected average unit energy consumption E of the system over the next N control cycles. avg With the average health decay rate ΔHI avg The weighted composite score is calculated using the following formula: V pred = 1 / [λ1* E avg + λ2* ΔHI avg ], where λ1 is the average unit energy consumption E avgThe weighting coefficient, λ2 is the average health decay rate ΔHI avg The weighting coefficients, λ1 + λ2 = 1, are preset positive normalized weighting coefficients used to adjust the relative importance of energy consumption targets and equipment health maintenance targets in the comprehensive score. S4.3: When the predicted score V output by the running effect prediction model is... pred The parameter adjustment process is triggered when any of the following conditions are met. a) When the predicted score V pred continuously below the preset threshold V th Reaching M times, b. When the predicted score V pred Below the dynamic threshold V th_dynamic , V th_dynamic = μ * avg(V hist ), where avg(V hist The value of μ is the average of the historical predicted scores over the most recent K control periods, where K is a preset positive integer and μ is the sensitivity coefficient (0 < μ < 1). For example, if K = 24 and the control period is 5 minutes, it corresponds to the data from the most recent 2 hours.

[0030] The parameter adjustment process in step S4.3 includes: First, define the set of parameters to be optimized, θ = {w}, consisting of the health calculation weights and the optimization target weights. e w m w p , α, β}; Secondly, an optimization problem is constructed using historical operational results as feedback. Specifically, this involves using the comprehensive operational performance score V recorded in the historical database over the most recent K control cycles. actual The objective is to maximize the average value, i.e., the objective function is maximize avg(V). actual (θ)) and impose a limit on the magnitude of parameter variation (|θ) new - θ old | ≤ Δ max Constraints; Then, the Bayesian optimization algorithm is used to optimize the parameter set θ under the constraints, and a candidate new parameter set θ is obtained. new Then, based on the system simulation model, it was verified whether it met the core operational constraints of flow balance and equipment safety; Finally, the new parameter set θ will be verified. new It is applied to the next control cycle to complete the closed-loop iteration and adaptive update of the control strategy.

[0031] Performance Prediction Model M eval Train and update in the following ways: S4.4: Collect historical running data and construct a supervised learning sample set, where each sample consists of an input feature vector F and its corresponding true performance label V. true constitute; The constituent data of the feature vector F correspond to the initial state of a complete decision-making cycle, including the health-related features of each pump at that moment, such as the HE of each pump. e HE m HE p , HI; and related characteristics of the decision output for this cycle, such as the decision execution frequency f. i Power P i (f i The objective function value J; and the current operating characteristics of the system, such as the total system demand Q. demand ; Realistic effect label V true Based on the average unit energy consumption E actually observed in N consecutive control cycles following this decision cycle. avg_true The decrease in average health ΔHI avg_true Through formula V true = 1 / [λ1* E avg_true + λ2* ΔHI avg_true The calculation shows that λ1 + λ2 = 1, where λ1 and λ2 are preset positive normalized weighting coefficients used to adjust the relative importance of energy consumption targets and equipment health maintenance targets in the comprehensive score. Using the above sample set, with feature vector F as input and true effect label V... true To output labels, supervised training is performed on the gradient boosting decision tree or deep neural network model to obtain an initial performance prediction model M. eval Specifically, the constructed supervised learning sample set is used to minimize the predicted score V. pred Compared to the actual rating V true Using the mean squared error between the two sides as the objective, supervised training is performed on either a Gradient Boosting Decision Tree (GBDT) or a Deep Neural Network (DNN) model to obtain an initial performance prediction model M. eval .

[0032] S4.5: After every L control cycles, the system automatically records the newly generated {F, V} within that time period. true The samples are added to the training set to predict the running effect M. eval Incremental learning is performed to track the slow changes in equipment aging and system operating conditions. The "training set" is a subset of the sample set, specifically partitioned off for training (fitting) the machine learning model. Each sample contains an input feature vector F and a corresponding true performance label V. trueThe model adjusts its parameters (such as the weights of the neural network and the structure of the decision tree) by learning patterns from this data. Specifically, in the initial stage of system deployment, historical records are extracted from the historical database and converted into {F, V}. true The system uses samples in the format {F, V} to create an initial training set, ensuring that the model has the ability to make predictions based on historical experience before being deployed. During the online operation phase, after every L control cycles, the system automatically processes the newly generated operation records during this period into a new {F, V} format. true The samples are added to the existing training set, allowing the training set to continue to grow and cover the latest operational experience.

[0033] As an example, the method is executed in an edge-cloud collaborative architecture, wherein step S1 is executed in an edge computing unit deployed on the water pump side, and steps S2 and S4 are executed in a cloud server deployed at the plant level. The edge computing unit and the cloud server interact with each other through an industrial communication protocol.

[0034] A control system for a thermal power plant circulating water pump set, used to implement the above method, includes: A health calculation module is used to perform steps S1 and S2 in claim 1; An optimization decision-making and adaptive update module is used to perform steps S3 and S4 in claim 1; The control execution interface is used to connect to and control the circulating water pump and its frequency converter.

[0035] The health calculation module collects multi-source data such as electrical signals, vibration signals, flow rate, and head, and calculates a comprehensive real-time health index based on specific algorithms across three dimensions: electrical, mechanical, and performance. This enables precise quantitative assessment of the individual status of each water pump, providing a crucial foundation of equipment "health" information—missing in traditional methods—for subsequent optimization decisions. This module serves as the system's "sensing and assessment center," specifically executing steps S1 (data acquisition) and S2 (health calculation) in claim 1. Physically, the module employs an edge-cloud collaborative architecture to balance computational efficiency and resource concentration. The edge sensing unit, deployed on each water pump, forms the module's data front-end. Its hardware includes data acquisition cards and communication gateways, used to execute step S1: real-time acquisition of three-phase current, voltage, and power signals via communication with the high-voltage frequency converter of the pump motor; and acquisition of vibration acceleration signals from the pump body and flow rate and head data in the pipeline via sensor interfaces. These electrical, mechanical, and performance operating data are initially processed and then uploaded. The cloud-based health calculation engine, deployed on a plant-level server, constitutes the module's core analysis end. Its core is a processor that runs a dedicated algorithm, used to execute step S2, namely: Electrical Health Factor (HE) e Calculation: Perform frequency spectrum analysis (FFT) on the uploaded three-phase current signal, and calculate the energy (E) of the specific fault frequency component. fault ) and fundamental energy (E) fund The ratio of ) is calculated according to the formula HE e = max(0, 1 - K e * (E fault / E fund Perform the calculation.

[0036] Mechanical health factor (HE) m ) Calculation: Perform envelope demodulation and spectrum analysis on the uploaded vibration signal to extract the amplitude (A) of each characteristic frequency of the bearing. n According to the formula HE m = max(0, 1 - Σ(W n * A n / A threshold_n Perform the calculation.

[0037] Performance Health Factor (HE) p Calculation: Based on real-time flow rate Q, head H, and input power P in Calculate the current efficiency η actual and compared with the pre-stored health baseline efficiency η baseline Compare according to formula HE p = η actual / η baseline calculate.

[0038] Health Index (HI) Fusion: Finally, the three factors mentioned above are combined according to the formula HI = w e * HE e + w m * HE m + w p *HE p The weighted fusion is performed to output the comprehensive real-time health index (HI) of each water pump, providing direct input for subsequent optimization decisions.

[0039] The optimization decision-making and adaptive update module uses the real-time health index provided by the health calculation module as one of its core inputs. It constructs a multi-objective optimization problem that simultaneously minimizes the total system energy consumption and the rate of equipment health decay. By solving this problem, it determines the optimal pump combination and operating frequency of each pump under the current flow demand constraint. This module not only introduces health as a decision variable into the optimization model, enabling the system to automatically avoid or reduce high-load use on low-health equipment when pursuing energy conservation, thereby delaying its deterioration, but also embeds an adaptive update mechanism based on historical operating effect feedback. This mechanism can perform closed-loop adjustments to key weight parameters in health calculation and optimization calculation, allowing the entire decision-making strategy to self-evolve as the operating status of the equipment group changes, maintaining optimal performance in the long term. This module is the system's "intelligent decision-making and learning brain," specifically executing steps S3 (optimization decision) and S4 (adaptive update) in claim 1. It is a software service set integrated on a cloud server, containing two closely cooperating sub-modules: Multi-objective optimization decision submodule: Used to execute step S3. This submodule receives the current operating requirements (total flow Q) from the upper-level system (such as DCS). demand It starts immediately after [the process is complete]. Using the HI values ​​of each pump provided by the health calculation module as key inputs, it constructs and solves a mixed integer optimization problem with the objectives of minimizing total system energy consumption and optimizing equipment health degradation. Its objective function is: J = α * Σ(x... i * P i (f i )) + β * Σ[ x i * (1 - HI i ) * (P i (f i ) / P i_rated ) γ Under the constraints of flow balance and equipment capacity, the optimal pump start-stop state x is solved. i and operating frequency f i This process transforms equipment health status from an "assessment indicator" into a "decision variable," enabling an online dynamic trade-off between energy saving and equipment lifespan maintenance.

[0040] The parameter adaptive learning submodule is used to execute step S4. This submodule ensures the long-term optimality of the system, and its core is a performance prediction model M. eval The model takes current health characteristics, decision variables, and operating condition characteristics as inputs to predict a score V for the overall operational performance over a future period. pred When V predWhen the values ​​continuously fall below a threshold or a dynamic threshold, this submodule triggers a closed-loop adjustment: it uses historical performance as feedback to construct a parameter optimization problem and employs algorithms such as Bayesian optimization to adjust the weights (w) in the health calculation module. e , w m , w p The system optimizes the weights (α, β) in the decision-making submodule and applies the optimized parameter set to the next control cycle. This forms a self-evolving closed loop of "decision -> execution -> evaluation -> learning -> re-decision," enabling the system to adapt to equipment aging and changes in operating conditions.

[0041] The control execution interface ensures that the aforementioned intelligent decisions can be translated into actual control actions for the field frequency converters and water pumps without delay or error, completing the crucial link from "digital optimization" to "physical execution." This module is the "execution terminal" connecting the system to the physical world, used to translate the digital instructions generated by the optimization decision and adaptive update module into precise control of the field equipment. It typically includes: Command issuance interface: Located in the cloud, responsible for transmitting the {x} generated by the optimization decision submodule. i , f i The optimal instruction sequence is encapsulated into standard industrial protocol messages (such as Modbus TCP commands).

[0042] Drive and communication unit: Located on the equipment side or in the control cabinet, it receives commands via an industrial network. Its hardware includes communication modules, relay output boards, analog output modules, etc., used to directly connect to and control the circulating water pump and its frequency converter; that is, according to command x... i The circuit breaker controlling the water pump starts and stops the pump, and operates according to command f. i Write the frequency setting value to the frequency converter to precisely drive the water pump set to operate according to the optimized scheme.

[0043] The system's workflow embodies a logical closed loop of three modules: the health calculation module continuously provides "health check reports" on equipment status; the optimization decision-making and adaptive update module formulates the optimal "scheduling and maintenance plan" based on the health check report and current tasks, and has the ability to learn and improve the plan from historical results; the control execution interface is responsible for ensuring the plan is executed flawlessly. This entire process repeats continuously, achieving safe, efficient, reliable, and adaptive intelligent operation of the thermal power plant's circulating water pump unit.

[0044] In summary, this system, through the closed-loop collaboration of these three modules, deeply integrates real-time equipment health status perception, forward-looking multi-objective optimization decision-making, and continuous strategy self-learning capabilities. Ultimately, it achieves outstanding technical results, including significantly reducing the overall system operating costs (including power consumption and maintenance costs), improving equipment reliability and service life, and possessing long-term adaptive capabilities under complex operating conditions.

[0045] An electronic device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the above method.

[0046] Figure 3 A block diagram of an exemplary electronic device terminal suitable for implementing embodiments of the present invention is shown. It is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of the present invention.

[0047] like Figure 3 As shown, terminal 12 is presented in the form of a general-purpose computing device. The components of terminal 12 may include, but are not limited to: one or more processors or processing units 16, system memory 28, and bus 18 connecting different system components (including system memory 28 and processing unit 16).

[0048] Bus 18 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0049] Terminal 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by terminal 12, including volatile and non-volatile media, removable and non-removable media.

[0050] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. Terminal 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be used to read and write non-removable, non-volatile magnetic media (… Figure 3 Not shown; usually referred to as a "hard drive"). Although Figure 3 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0051] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 42 typically perform the functions and / or methods described in the embodiments of the present invention.

[0052] Terminal 12 can also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), and with one or more devices that enable a user to interact with terminal 12, and / or with any device that enables terminal 12 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed via input / output (I / O) interface 22. Furthermore, terminal 12 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 20. As shown, network adapter 20 communicates with other modules of terminal 12 via bus 18. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with terminal 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0053] The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, such as implementing the thermal power circulating water pump group control method provided in the embodiments of the present invention.

[0054] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0055] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0056] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0057] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including—but not limited to—wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0058] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0059] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A control method for a thermal power plant circulating water pump set, characterized in that, Includes the following steps: S1. Obtain the current operating requirements of the circulating water system and simultaneously collect the electrical, mechanical, and performance operating data of each candidate water pump; S2. Based on the electrical, mechanical and performance operation data, calculate the health factors that reflect the electrical status, mechanical status and operating efficiency status of the water pump, and weight and fuse the health factors into a comprehensive real-time health index for each water pump. S3. Taking the current operating requirements as constraints, and combining the real-time health index of each water pump calculated in step S2, a comprehensive optimization calculation is performed with the goal of minimizing the total energy consumption of the system and optimizing the decline of equipment health, and then a water pump operation combination and the operating frequency of each water pump that meet the requirements are determined. S4. Execute the decision results of step S3 to control the operating status of each water pump; at the same time, based on the feedback of the execution effect of this decision, construct and solve the parameter optimization problem with historical operating effect as the evaluation target, perform closed-loop adaptive adjustment of the health calculation weight in step S2 and / or the optimization calculation weight in step S3, and apply the adjusted new parameter set to subsequent control cycles.

2. The control method for a thermal power plant circulating water pump set according to claim 1, characterized in that, In step S2, the calculation of health factors and weighted fusion are achieved through the following steps: S2.1: Obtain three-phase electrical signals from the frequency converter of the water pump motor, obtain vibration signals from the water pump body, and obtain the real-time flow rate, head data and input power of the water pump; S2.2: Perform spectrum analysis on the three-phase electrical signals, and calculate the electrical health factor (HE) based on the ratio of the energy of a specific fault frequency component to the fundamental frequency energy. e The calculation formula is: HE e = max(0, 1 - K e * (E fault / E fund )), where K e E is the calibration coefficient. fault For fault characteristic frequency band energy, E fund It is the fundamental wave energy; S2.3: Perform envelope demodulation and spectrum analysis on the vibration signal, and calculate the mechanical health factor (HE) based on the amplitude at the bearing's characteristic frequency. m The calculation formula is: HE m = max(0, 1 - Σ(W n * A n / A threshold_n )), where W n A is the weighting coefficient. n For the characteristic frequency amplitude, A threshold_n For the corresponding alarm threshold; S2.4: Based on the real-time flow rate Q, head H, and input power P in Calculate the current efficiency η actual The calculation formula is: ; Where: ρ: density of circulating water; g: gravitational acceleration; Q: real-time flow rate of the water pump; H: real-time head of the water pump; P in : Input power of the water pump; S2.5: Calculate the current efficiency η obtained in step S1.

4. actual Compared with the baseline efficiency η under healthy conditions baseline Comparison, calculation of performance health factor HE p The calculation formula is: HE p = η actual / η baseline ; S2.6: The electrical health factor HE e Mechanical health factor HE m With performance health factor HE p The weighted fusion is performed to obtain the comprehensive real-time health index HI, calculated as follows: HI = w e * HE e + w m * HE m + w p * HE p , where w e It is the electrical health factor HE e The weighting coefficient, w m It is the mechanical health factor HE m The weighting coefficient, w p It is the performance health factor HE p The weighting coefficient, w e +w m +w p =1.

3. The control method for a thermal power plant circulating water pump set according to claim 2, characterized in that, In step S3, the method for comprehensive optimization calculation with the goal of minimizing total system energy consumption and optimizing equipment health degradation is as follows: Construct an objective function that minimizes both total system energy consumption and the rate of equipment health degradation, and solve for the pump operating combinations and operating frequencies of each pump that meet the requirements: The calculation formula is: J = α * Σ(x i * P i (f i )) + β * Σ[x i * (1 - HI i ) * (P i (f i ) / P i_rated ) γ ]; where x i f represents the start / stop status of water pump i. i For the operating frequency, P i (f i P represents the corresponding power. i_rated The rated power is α and β, which are weighting coefficients and α+β=1. γ>1 is the load influence coefficient. The constraints include at least the following: the total flow provided by all operating pumps is equal to the total flow value in the current operating demand; the operating frequency of each pump is within its allowable range; and pumps with a health index lower than the safety threshold are not included in the optimization.

4. The control method for a thermal power plant circulating water pump set according to claim 3, characterized in that: The adaptive adjustment in step S4 is achieved through a pre-trained machine learning model, specifically including the following steps: S4.1: Construct a performance prediction model M eval Its input feature vector F includes at least the following three categories: a) Health-related characteristics F H During the current decision-making cycle, the electrical health factor (HE) of each candidate water pump is... e Mechanical health factor HE m HE (Health Factor) p and its comprehensive health index HI; b) Decision-related characteristics F D The operating frequency f of each pump in the optimal operating combination output by step S3. i Its corresponding power P i (f i The objective function value J* obtained from the decision calculation; c) Operating condition related characteristics F C The current system's total traffic demand, Q demand Circulating water inlet temperature T in ; S4.2: The predicted performance model M eval The training objective is to output the predicted score V. pred The predicted score V pred Defined as the expected average unit energy consumption E of the system over the next N control cycles. avg With the average health decay rate ΔHI avg The weighted composite score is calculated using the following formula: V pred = 1 / [λ1* E avg + λ2* ΔHI avg ], where λ1 is the average unit energy consumption E avg The weighting coefficient, λ2 is the average health decay rate ΔHI avg The weighting coefficients, λ1 + λ2 = 1, are preset positive normalized weighting coefficients used to adjust the relative importance of energy consumption targets and equipment health maintenance targets in the comprehensive score. S4.3: When the predicted score V output by the running effect prediction model is... pred The parameter adjustment process is triggered when any of the following conditions are met. a) When the predicted score V pred continuously below the preset threshold V th Reaching M times, b. When the predicted score V pred Below the dynamic threshold V th_dynamic , V th_dynamic = μ * avg(V hist ), where avg(V hist ) represents the average historical predicted score over the most recent K control periods, where K is a preset positive integer and μ is the sensitivity coefficient (0 < μ < 1).

5. The control method for a thermal power plant circulating water pump set according to claim 4, characterized in that, The parameter adjustment process in step S4.3 includes: First, define the set of parameters to be optimized, θ = {w}, consisting of the health calculation weights and the optimization target weights. e w m w p , α, β}; Secondly, an optimization problem is constructed using historical operational results as feedback. Specifically, this involves using the comprehensive operational performance score V recorded in the historical database over the most recent K control cycles. actual The objective is to maximize the average value, i.e., the objective function is maximizeavg(V). actual (θ)) and impose a limit on the magnitude of parameter variation (|θ) new - θ old | ≤ Δ max Constraints; Then, the Bayesian optimization algorithm is used to optimize the parameter set θ under the constraints, and a candidate new parameter set θ is obtained. new Then, based on the system simulation model, it was verified whether it met the core operational constraints of flow balance and equipment safety; Finally, the new parameter set θ will be verified. new It is applied to the next control cycle to complete the closed-loop iteration and adaptive update of the control strategy.

6. The control method for a thermal power plant circulating water pump set according to claim 4, characterized in that, Performance Prediction Model M eval Train and update in the following ways: S4.4: Collect historical running data and construct a supervised learning sample set, where each sample consists of an input feature vector F and its corresponding true performance label V. true constitute; The constituent data of the feature vector F correspond to the starting state of a complete decision cycle, including the health-related features of each water pump at that moment. Realistic effect label V true Based on the average unit energy consumption E actually observed in N consecutive control cycles following this decision cycle. avg_true The decrease in average health ΔHI avg_true Through formula V true = 1 / [λ1* E avg_true + λ2* ΔHI avg_true The calculation shows that λ1 + λ2 = 1, where λ1 and λ2 are preset positive normalized weighting coefficients used to adjust the relative importance of energy consumption targets and equipment health maintenance targets in the comprehensive score. Using the above sample set, with feature vector F as input and true effect label V... true To output labels, supervised training is performed on the gradient boosting decision tree or deep neural network model to obtain an initial performance prediction model M. eval ; S4.5: After every L control cycles, the system automatically records the newly generated {F, V} within that time period. true The samples are added to the training set to predict the running effect M. eval Incremental learning is used to track the slow changes in equipment aging and system operating conditions.

7. The control method for a thermal power plant circulating water pump set according to claim 1, characterized in that: The method is executed in an edge-cloud collaborative architecture, wherein step S1 is executed in an edge computing unit deployed on the water pump side, and steps S2 and S4 are executed in a cloud server deployed at the plant level. The edge computing unit and the cloud server interact with each other through an industrial communication protocol.

8. A control system for a thermal power plant circulating water pump unit, used to implement the method according to any one of claims 1 to 7, characterized in that, include: A health calculation module is used to perform steps S1 and S2 in claim 1; An optimization decision-making and adaptive update module is used to perform steps S3 and S4 in claim 1; The control execution interface is used to connect to and control the circulating water pump and its frequency converter.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images