Power plant boiler waste heat recovery system

Abstract

The invention discloses a power plant boiler waste heat recovery system, and relates to the technical field of boiler waste heat detection. The recovery system comprises a flue gas parameter acquisition and pretreatment module, a waste heat potential evaluation and classification module, a heating power matching and heat exchange path planning module and a dynamic operation optimization and feedback control module which are operated in sequence, wherein the flue gas parameter acquisition and preprocessing module is used for acquiring flue gas related parameters corresponding to a target electric field boiler and performing data preprocessing action; according to the technical key points, a multi-objective optimization model is established, heat recovery is weighted to the maximum degree under the condition that the minimum temperature difference and matching constraints are met, and high-grade waste heat is preferentially supplied to high-value users; meanwhile, the model dynamically generates an optimal heat distribution scheme, and energy depreciation and resource mismatching are avoided, so that the problems of energy level mismatching and low efficiency caused by a traditional fixed connection mode are solved, cascade and economization of waste heat utilization are realized, and the intelligent decision-making capability of system resource allocation is embodied.

Description

Technical Field

[0001] This invention relates to the field of boiler waste heat detection technology, specifically a power plant boiler waste heat recovery system. Background Technology

[0002] Boiler waste heat detection involves deploying sensors in the tail flue to collect parameters such as exhaust gas temperature, flow rate, pressure, and composition in real time, obtaining information on the thermal state of the flue gas. It also uses temperature and flow data to assess the amount of heat carried away by the exhaust gas, and combines this with ambient temperature to determine the potential for waste heat utilization. An online monitoring system summarizes and analyzes the data to reflect the heat loss status during boiler operation. Existing technologies sometimes incorporate thermodynamic analysis methods to distinguish between the heat of waste heat and usable energy, providing a preliminary assessment of its work capacity. The detection results are used to monitor boiler efficiency changes and guide operational adjustments. This technology, based on static parameter measurement, identifies the existence and basic thermodynamic characteristics of waste heat, providing data support for energy-saving diagnostics.

[0003] Existing boiler waste heat detection methods largely rely on single temperature or flow rate monitoring, resulting in relatively simple data processing. However, they lack effective filtering of sensor noise and are prone to large reading errors due to factors such as airflow disturbances. For example, instantaneous jumps in flue gas temperature are often misinterpreted as changes in operating conditions, affecting operational judgment. Waste heat assessment is generally based on heat balance calculations of heat loss, focusing only on quantity and neglecting quality, failing to distinguish the difference in work capacity between high-temperature flue gas and low-temperature hot water. Although some existing systems have introduced the concept of heat loss, the grading thresholds are fixed, such as uniformly using 80 kJ / kg as the high heat loss standard, when the fuel is bituminous coal. When switching to high-moisture lignite, the flue gas humidity increases while the usable energy decreases. If the original standard is still applied, it will cause an overestimation problem. Thermal matching often adopts a fixed heat exchange path. For example, the low-temperature economizer is only connected to a certain stage heater. It cannot be dynamically adjusted according to the waste heat grade, resulting in high-grade waste heat being used for low-grade heating, causing energy devaluation. Traditional boiler waste heat recovery systems may experience a decline in heat exchange performance after long-term operation due to factors such as ash accumulation and corrosion. However, the model parameters are not updated, and the predicted recovered heat remains high. The lack of a feedback correction mechanism ultimately renders the optimization strategy ineffective, and the overall control accuracy gradually decreases. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: A waste heat recovery system for power plant boilers, the system comprising: The flue gas parameter acquisition and preprocessing module acquires relevant flue gas parameters corresponding to the target power plant boiler, performs data preprocessing, and forms a standard input vector. The waste heat potential assessment and classification module, based on standard input vectors, pre-builds a flue gas usable heat assessment model, and divides waste heat into set level ranges according to heat values, outputting waste heat level label information; The thermal matching and heat exchange path planning module establishes a multi-objective matching model based on waste heat level label information and heat user needs, and plans the optimal heat exchange path. The dynamic operation optimization and feedback control module establishes an operation cost function based on real-time operation and economic indicators, and introduces a feedback mechanism. Based on the feedback comparison results, it selects whether to trigger the model recalibration mechanism to update the multi-objective matching model.

[0005] Furthermore, the relevant flue gas parameters include: flue gas temperature Tg, flue gas mass flow rate mg, water vapor volume fraction xH2O in the flue gas, flue gas pressure Pg, ambient temperature Ta, and ambient humidity ϕa; the data preprocessing steps for these flue gas parameters include: The original flue gas parameters are processed by sliding window filtering. All filtered flue gas parameters are normalized to [0, 1].

[0006] Furthermore, the operation process of the pre-built flue gas availability assessment model is as follows: calculate the flue gas physical efficiency e_x, based on: e_x=(hg-ha)-Ta(sg-sa); where hg, ha: specific enthalpy and specific entropy of flue gas in state (Tg, Pg); sg, sa: reference enthalpy and entropy in environmental state (Ta, Pa), Pa represents environmental pressure; the calculation basis of the total flue gas flow rate Ex is: Ex=mg×e_x.

[0007] Furthermore, the criteria for dividing the levels into defined ranges are as follows: When e_x > 2M, it is determined to be: High-level Hi; When 2M≥e_x>M, it is determined to be: medium level Mz; When e_x≤M, it is determined as: low level Ld; The output waste heat level label information includes: L∈{Hi, Mz, Ld} and the corresponding total flow rate, where L is the level.

[0008] Furthermore, the value of M is set based on a pre-constructed characteristic dynamic adjustment function; where M represents the reference value, which is a function that depends on the flue gas temperature, ambient temperature and flue gas component characteristics, based on: M=f(Tg, Ta, xH2O); specifically in the form: M=α×(|Tg-Ta|)+β×Ta×ln(1+γ×xH2O).

[0009] Furthermore, the specific process of establishing a multi-objective matching model is as follows: Construct a set of hot users U = {u1, u2, ..., u...} n}; Each hot user u qIt has the lowest heat absorption temperature Tmin q Maximum heat absorption Qmax q And the demand for Ereq q ; Define the matching function M(L, u) q )∈{0,1}, based on the following: ; Where ΔTmin is the minimum heat transfer temperature difference constraint; q = 1, 2, ..., n, and q is the number of each heat user; Establish the objective function for path planning: ; The objective function for this path planning is constrained by: ; Among them, Q q w represents the actual recovered heat allocated to user q. q The value represents the user priority weight, ranging from 0 to 1. ηrec represents the heat exchange system efficiency, cp represents the average specific heat capacity of the flue gas, and Tout,min represents the minimum allowable exhaust gas temperature.

[0010] Further, real-time operating and economic indicators include: boiler fuel cost Cf, boiler efficiency ηb, actual recovered heat Qrec, electricity price Pe, flue gas side pressure drop Δp, and fan power consumption penalty coefficient v.

[0011] Furthermore, the operating cost function is established based on: C = Cf - ηb × Qrec × Pe + v × Δp 2 The introduced feedback mechanism is as follows: a model error feedback term ϵ(t) = Qpred(t) − Qmeas(t) is introduced to correct the heat exchange system efficiency ηrec, based on: ηrec(k+1) = ηrec(k) + Kp × ϵ(t); where Qpred(t) represents the predicted recovered heat, Qmeas(t) represents the measured value of the recovered heat, ηrec(k+1) represents the estimated value of the heat exchange system efficiency in the (k+1)th iteration, ηrec(k) represents the estimated value of the heat exchange system efficiency in the kth iteration, and Kp represents the proportional gain coefficient, with a value range of 0 to 0.1.

[0012] Furthermore, the feedback comparison process is as follows: when ϵ(t) exceeds the set threshold for at least 3 time periods, the model recalibration mechanism is triggered; Conversely, the model recalibration mechanism will not be triggered.

[0013] Furthermore, the process of triggering the model recalibration mechanism is as follows: the average specific heat capacity cp of the flue gas is fitted online using the least squares method; at the same time, the minimum allowable flue gas temperature Tout and min are adjusted and updated at intervals of +15K.

[0014] This invention provides a waste heat recovery system for power plant boilers, which has the following beneficial effects: This solution effectively suppresses instantaneous noise and dimensional differences in sensor measurements by performing sliding window filtering and normalization on the data, thereby improving data reliability. At the same time, the preprocessing provides a stable and standardized input vector for the subsequent model, avoiding misjudgments caused by abnormal fluctuations. This scheme uses the second law of thermodynamics to construct a physical model, scientifically quantifies the work capacity of flue gas waste heat, and achieves accurate identification of waste heat quality. By introducing a pre-constructed characteristic dynamic adjustment function, the classification threshold is adaptively adjusted according to the operating conditions, thereby solving the problem of misclassification under different climate and fuel conditions by the traditional fixed threshold, and realizing the universality and accuracy of waste heat assessment. This solution establishes a multi-objective optimization model to maximize the weighted recovery of heat while satisfying the constraints of minimum temperature difference and temperature matching, so as to prioritize the supply of high-grade waste heat to high-value users. At the same time, the model dynamically generates the optimal heat allocation scheme to avoid energy devaluation and resource misallocation, thereby solving the problems of energy level mismatch and low efficiency caused by traditional fixed connection methods, realizing the tiered and economical utilization of waste heat, and demonstrating the intelligent decision-making capability of system resource allocation. This scheme coordinates fuel savings and plant power penalties through a cost function, achieving a balance between economic efficiency and operational effectiveness. It also introduces error feedback and model recalibration mechanisms to continuously perform correction feedback actions, compensate for equipment aging and operating condition drift, thereby solving the problem of long-term accuracy decay of traditional static models, achieving adaptive maintenance of the system, and demonstrating the unity of closed-loop control and long-term stability. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the modular operation of a power plant boiler waste heat recovery system according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1This embodiment provides a waste heat recovery system for power plant boilers, the system comprising: I. Flue Gas Parameter Acquisition and Preprocessing Module: Obtain the relevant flue gas parameters corresponding to the target power plant boiler, perform data preprocessing, and form a standard input vector; In this embodiment, the target electric field boiler can be any actual electric field boiler that is selected. The relevant flue gas parameters include: exhaust gas temperature Tg (K), flue gas mass flow rate mg (kg / s), water vapor volume fraction xH2O in the flue gas (%), flue gas pressure Pg (Pa), ambient temperature Ta (K), and ambient humidity ϕa (%). The data preprocessing steps for these flue gas parameters include: The original flue gas parameters are subjected to sliding window filtering to eliminate instantaneous fluctuation noise, based on the following: Where x_avg(t) represents the smoothed value at time t, i is an integer index representing the i-th sampling point back from the current time t, N is the sliding window length, which takes a value of 5~10 and is set according to the flue gas disturbance frequency, Δt represents the sampling period, and x(t-iΔt) represents the original measurement value of the i-th historical sampling point. It should be noted that the flue gas-related parameters are collected through a pre-configured distributed sensor network. The above formula uses a linear low-pass filtering method to smooth the time series signal and suppress high-frequency noise or instantaneous fluctuations. In industrial settings, such as power plant boiler exhaust ducts, parameters such as temperature and flow rate collected by the distributed sensor network are often affected by airflow disturbances, electromagnetic interference, and measurement delays, generating random noise. This noise manifests as data fluctuating rapidly around the true value. Direct input into subsequent models may lead to misjudgments. Therefore, the moving average uses the arithmetic mean of the most recent N sampling points as the output at the current moment to weaken the influence of individual outliers and retain trend information. Essentially, it performs convolution operations on the signal in the time domain, with a rectangular window kernel function, offering advantages such as simplicity and strong real-time performance. The selected i ranges from 0 to N-1 to ensure that the window contains exactly N consecutive data points, ending at the current time t, i.e., the latest time point. If the traditional method is used, starting from i=1, the current value will be missed; if it ends at N, one point will be extra, disrupting the fixed-length characteristic. Normalize all filtered flue gas parameters to [0, 1]; It should be noted that the data preprocessing steps described above can effectively suppress measurement noise, avoid subsequent model misjudgments, and at the same time, normalization eliminates dimensional differences, improving the model's convergence speed and stability.

[0018] By adopting the above technical solution, sliding window filtering and normalization of the data effectively suppress instantaneous noise and dimensional differences in sensor measurements, improving data reliability. On the other hand, this preprocessing provides a stable and standardized input vector for the subsequent model, avoiding misjudgments caused by abnormal fluctuations, thereby solving the problem of control instability caused by poor data quality in traditional systems and ensuring the robustness of model input.

[0019] II. Waste Heat Potential Assessment and Classification Module: Based on the standard input vector, the second law of thermodynamics is used to pre-build a flue gas availability assessment model, and the waste heat is divided into a set level range according to the waste heat value to guide subsequent thermal matching and output waste heat level label information. The calculation of flue gas physical properties e_x is based on: e_x = (hg - ha) - Ta(sg - sa); where hg and ha are the specific enthalpy and specific entropy of the flue gas in state (Tg, Pg); Ta is mentioned above, i.e., ambient temperature; sg and sa represent the reference enthalpy and entropy in ambient state (Ta, Pa); and Pa represents the ambient pressure, i.e., standard atmospheric pressure. The calculation basis for the total efflux rate Ex is: Ex = mg × e_x; the flue gas physical efflux e_x here is the efflux value mentioned above: the waste heat is divided into efflux values ​​in the set grade range based on the efflux value; The criteria for defining the grade ranges are as follows: When e_x > 2M, it is determined to be: High-level Hi; When 2M≥e_x>M, it is determined to be: medium level Mz; When e_x≤M, it is determined as: low level Ld; The output waste heat level label information includes: L∈{Hi, Mz, Ld} and the corresponding total flow rate, where L is the level; The value of M is set based on a pre-constructed dynamic adjustment function; where M represents the baseline value, which is a function dependent on flue gas temperature, ambient temperature, and flue gas component characteristics, based on: M = f(Tg, Ta, xH2O); specifically, M = α × (|Tg - Ta|) + β × Ta × ln(1 + γ × xH2O); It should be noted that α×(|Tg-Ta|) reflects the temperature difference driving force between flue gas and the environment, and is the main source of physical heat. This item dominates the heat assessment of dry flue gas; β×Ta×ln(1+γ×xH2O) is the latent heat heat that compensates for the condensation and release of water vapor in wet flue gas. Wet flue gas heat includes two parts: physical heat and phase change heat. The latter is related to the dew point temperature and the ambient temperature difference. Using the form of ln(1+γ×xH2O) can simulate the nonlinear characteristic that the heat growth slows down when humidity increases; the values ​​of α and β are both in the range of 0~1, while the value of γ is greater than 0. In this embodiment, it is usually taken as 2 to reasonably amplify the effect of humidity; If the final value of M in this embodiment is 40, and an integer example is used here for easy comparison, then when e_x > 80, it is determined as: high-grade Hi; when 80 ≥ e_x > 40, it is determined as: medium-grade Mz; when e_x ≤ 40, it is determined as: low-grade Ld. The working condition adaptability of the waste heat level classification is realized. The function takes temperature difference as the main driving force and humidity as the correction term, which conforms to the essential composition of thermodynamics. The classification interval [M, 2M] dynamically expands and contracts with the operating conditions, which improves the accuracy of waste heat assessment on the one hand, and avoids the failure problem of fixed threshold under different geographical climate and fuel conditions on the other hand, and has significant engineering practicality.

[0020] By adopting the above technical solutions, a physical quality model is constructed based on the second law of thermodynamics to scientifically quantify the work capacity of flue gas waste heat and achieve accurate identification of waste heat quality. On the other hand, by introducing a pre-constructed characteristic dynamic adjustment function, the classification threshold is adaptively adjusted according to the operating conditions, thereby solving the problem of misclassification under different climate and fuel conditions by the traditional fixed threshold, realizing the universality and accuracy of waste heat assessment, and reflecting the deep integration of quality analysis and operating conditions.

[0021] III. Thermal Matching and Heat Exchange Path Planning Module: Based on the waste heat level label information and combined with the needs of heat users, a multi-objective matching model is established to plan the optimal heat exchange path; The specific steps are as follows: Construct a set of hot users U = {u1, u2, ..., u...} n}; Each hot user u q It has the lowest heat absorption temperature Tmin q Maximum heat absorption Qmax q And the demand for Ereq q ; Define the matching function M(L, u) q )∈{0,1}, based on the following: ; Where ΔTmin is the minimum heat transfer temperature difference constraint, which can be taken as 15K in this embodiment; q=1, 2, ..., n, and q is the number of each heat user; the basis for establishing the path planning objective function is as follows: The objective function for this path planning is constrained by: ; Among them, Q q w represents the actual recovered heat allocated to user q; q The user priority weight is set based on thermodynamic economics, with a value range of 0 to 1; ηrec represents the heat exchange system efficiency, with a value range of 0.85 to 0.95 in this embodiment; cp represents the average specific heat capacity of the flue gas, in kJ / (kg·K); Tout,min represents the minimum allowable flue gas temperature to prevent low-temperature corrosion, which can be selected according to actual needs, and in this embodiment, it is taken as dew point +15K; it should be noted that the above model operates under the premise of satisfying thermodynamic feasibility, maximizing the utilization of high-value thermal energy and avoiding energy devaluation.

[0022] Specifically, this module constructs a heat transfer path planning model based on heat matching and multi-objective optimization to achieve efficient cascade utilization of waste heat resources. First, a matching function is used to screen feasible users that meet the minimum heat transfer temperature difference and heat requirements, ensuring thermodynamic feasibility. Then, a linear programming model is established with the goal of maximizing weighted recovered heat, with weights reflecting the thermoeconomic value of each user. Constraints cover the user's heat absorption capacity, total heat conservation, and the lower limit of flue gas temperature to prevent low-temperature corrosion, ensuring the safety and feasibility of the scheme. This method not only avoids energy devaluation but also automatically adjusts the allocation strategy under varying operating conditions through mathematical optimization, improving the overall energy efficiency and economy of the system. It solves the problems of poor adaptability and low efficiency of traditional fixed allocation methods, realizing intelligent and dynamic decision-making for waste heat utilization paths.

[0023] By adopting the above technical solutions, a multi-objective optimization model is established to maximize the weighted recovery of heat under the constraints of minimum temperature difference and temperature matching, so as to prioritize the supply of high-grade waste heat to high-value users. At the same time, the model dynamically generates the optimal heat allocation scheme to avoid energy devaluation and resource misallocation, thereby solving the problems of energy level mismatch and low efficiency caused by traditional fixed connection methods, realizing the tiered and economical utilization of waste heat, and demonstrating the intelligent decision-making capability of system resource allocation.

[0024] Dynamic operation optimization and feedback control module: Based on real-time operation and economic indicators, an operation cost function is established, and a feedback mechanism is introduced. Based on the feedback comparison results, it is selected whether to trigger the model recalibration mechanism to update the multi-objective matching model. The real-time operation and economic indicators include: boiler fuel cost Cf, boiler efficiency ηb, actual recovered heat Qrec, electricity price Pe, flue gas side pressure drop Δp, and fan power consumption penalty coefficient v, which can be taken as 1×10 in this embodiment. -8 ; The operating cost function is based on the following: C = Cf - ηb × Qrec × Pe + v × Δp 2 ; Where Cf is the positive cost, but increasing Qrec can reduce flue gas losses, indirectly reducing Cf; -ηb×Qrec×Pe represents the benefit of generating more electricity due to heat recovery replacing steam extraction, reflecting the energy-saving economy; v×Δp 2 It approximately reflects the additional power consumption of the induced draft fan, which is proportional to the square of the pressure drop, thus avoiding excessive heat exchange that leads to an increase in plant power consumption. This function guides the system to achieve a balance between energy-saving benefits and power consumption penalties, thereby achieving economical operation. The introduced feedback mechanism is as follows: A model error feedback term ϵ(t) = Qpred(t) − Qmeas(t) is introduced to correct the heat exchange system efficiency ηrec, based on: ηrec(k+1) = ηrec(k) + Kp × ϵ(t); where Qpred(t) represents the predicted recovered heat; Qmeas(t) represents the measured value of the recovered heat; ηrec(k+1) represents the estimated value of the heat exchange system efficiency in the (k+1)th iteration, and ηrec(k) represents the estimated value of the heat exchange system efficiency in the kth iteration; Kp represents the proportional gain coefficient, with a value ranging from 0 to 0.1, and in this embodiment, it can be set to 0.05 to control the feedback correction speed; the feedback comparison process is as follows: when ϵ(t) exceeds the set threshold for at least 3 time periods, the model recalibration mechanism is triggered; otherwise, it is not triggered; the set threshold can be set according to actual needs, which will not be elaborated here, and the length of each time period in the 3 time periods is also set according to needs; The process of triggering the model recalibration mechanism is as follows: Based on the latest measured data, the average specific heat capacity cp of the flue gas is fitted online using the least squares method. At the same time, according to the relationship between flue gas humidity and dew point, the minimum allowable flue gas temperature Tout,min is dynamically adjusted and updated at intervals of +15K. This mechanism is used to update cp and Tout,min, and can also compensate for the property deviations caused by equipment ash accumulation, fuel changes, etc., to ensure the long-term accuracy of the heat transfer model.

[0025] By adopting the above technical solutions, the cost function coordinates fuel savings and plant power penalties, achieving a balance between economic efficiency and operational effectiveness. Simultaneously, an error feedback and model recalibration mechanism is introduced to continuously perform corrective feedback actions, compensating for equipment aging and operating condition drift. This solves the problem of long-term accuracy decay in traditional static models, enabling adaptive maintenance of the system and demonstrating the unity of closed-loop control and long-term stability. The aforementioned modules operate sequentially and interconnectedly, achieving a complete closed loop from data to evaluation, then to decision-making and optimization.

[0026] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0027] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0028] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A waste heat recovery system for power plant boilers, characterized in that, The system includes: The flue gas parameter acquisition and preprocessing module acquires relevant flue gas parameters corresponding to the target power plant boiler, performs data preprocessing, and forms a standard input vector. The waste heat potential assessment and classification module, based on standard input vectors, pre-builds a flue gas usable heat assessment model, and divides waste heat into set level ranges according to heat values, outputting waste heat level label information; The thermal matching and heat exchange path planning module establishes a multi-objective matching model based on waste heat level label information and heat user needs, and plans the optimal heat exchange path. The dynamic operation optimization and feedback control module establishes an operation cost function based on real-time operation and economic indicators, and introduces a feedback mechanism. Based on the feedback comparison results, it selects whether to trigger the model recalibration mechanism to update the multi-objective matching model.

2. The waste heat recovery system for power plant boilers according to claim 1, characterized in that: Flue gas related parameters include: exhaust gas temperature Tg, flue gas mass flow rate mg, water vapor volume fraction xH2O in flue gas, flue gas pressure Pg, ambient temperature Ta, and ambient humidity ϕa; the data preprocessing steps for flue gas related parameters include: The original flue gas parameters are processed by sliding window filtering. All filtered flue gas parameters are normalized to [0, 1].

3. The waste heat recovery system for power plant boilers according to claim 2, characterized in that: The operation process of the pre-built flue gas available evaluation model is as follows: calculate the flue gas physical properties e_x, based on: e_x=(hg-ha)-Ta(sg-sa); where hg, ha: specific enthalpy and specific entropy of flue gas in state (Tg, Pg); sg, sa: reference enthalpy and entropy in environmental state (Ta, Pa), Pa represents environmental pressure; the calculation basis of the total flue gas flow rate Ex is: Ex=mg×e_x.

4. The waste heat recovery system for power plant boilers according to claim 2, characterized in that: The criteria for dividing the levels into defined ranges are as follows: When e_x > 2M, it is determined to be: High-level Hi; When 2M≥e_x>M, it is determined to be: medium level Mz; When e_x≤M, it is determined as: low level Ld; The output waste heat level label information includes: L∈{Hi, Mz, Ld} and the corresponding total flow rate, where L is the level.

5. A power plant boiler waste heat recovery system according to claim 4, characterized in that: The value of M is set based on a pre-constructed characteristic dynamic adjustment function; where M represents the reference value, which is a function that depends on the flue gas temperature, ambient temperature and flue gas component characteristics, based on: M=f(Tg, Ta, xH2O); specifically in the form: M=α×(|Tg-Ta|)+β×Ta×ln(1+γ×xH2O).

6. The waste heat recovery system for power plant boilers according to claim 1, characterized in that: The specific process of establishing a multi-objective matching model is as follows: Construct a set of hot users U = {u1, u2, ..., u...} n }; Each hot user u q It has the lowest heat absorption temperature Tmin q Maximum heat absorption Qmax q And the demand for Ereq q ; Define the matching function M(L, u) q )∈{0,1}, based on the following: ； Where ΔTmin is the minimum heat transfer temperature difference constraint; q = 1, 2, ..., n, and q is the number of each heat user; Establish the objective function for path planning: ; The objective function for this path planning is constrained by: ； Among them, Q q w represents the actual recovered heat allocated to user q. q The value represents the user priority weight, ranging from 0 to 1. ηrec represents the heat exchange system efficiency, cp represents the average specific heat capacity of the flue gas, and Tout,min represents the minimum allowable exhaust gas temperature.

7. The waste heat recovery system for power plant boilers according to claim 1, characterized in that: Real-time operating and economic indicators include: boiler fuel cost Cf, boiler efficiency ηb, actual recovered heat Qrec, electricity price Pe, flue gas side pressure drop Δp, and fan power consumption penalty coefficient v.

8. A waste heat recovery system for power plant boilers according to claim 7, characterized in that: The operating cost function is based on: C = Cf - ηb × Qrec × Pe + v × Δp 2 The introduced feedback mechanism is as follows: a model error feedback term ϵ(t) = Qpred(t) − Qmeas(t) is introduced to correct the heat exchange system efficiency ηrec, based on: ηrec(k+1) = ηrec(k) + Kp × ϵ(t); where Qpred(t) represents the predicted recovered heat, Qmeas(t) represents the measured value of the recovered heat, ηrec(k+1) represents the estimated value of the heat exchange system efficiency in the (k+1)th iteration, ηrec(k) represents the estimated value of the heat exchange system efficiency in the kth iteration, and Kp represents the proportional gain coefficient, with a value range of 0 to 0.

1.

9. A waste heat recovery system for power plant boilers according to claim 8, characterized in that: The feedback comparison process is as follows: when ϵ(t) exceeds the set threshold for at least 3 time periods, the model recalibration mechanism is triggered. Conversely, the model recalibration mechanism will not be triggered.

10. A power plant boiler waste heat recovery system according to claim 9, characterized in that: The process of triggering the model recalibration mechanism is as follows: the average specific heat capacity cp of the flue gas is fitted online using the least squares method; at the same time, the minimum allowable flue gas temperature Tout and min are adjusted and updated at intervals of +15K.

Images