Heat and power load optimization method for combined heat and power generation heat supply unit

Abstract

The invention discloses a combined heat and power generation heat supply unit heat and power load optimization method. The method comprises the following steps that S1, heat and power load data, equipment operation parameters and environment parameters of a combined heat and power generation unit are collected in real time; s2, preprocessing the collected data, including noise filtering, data standardization and feature extraction; s3, constructing a dynamic digital twinborn model based on the preprocessed data, and calculating an optimal thermoelectric load distribution scheme through an improved intelligent optimization algorithm; s4, adjusting thermoelectric load output of the unit according to an optimization result, and verifying a control effect in real time through a digital twin model; and S5, feeding back the actual operation data to the digital twinborn model, and dynamically updating model parameters to improve the prediction precision. The invention belongs to the field of thermoelectric units, and particularly relates to a thermoelectric load optimization method for a combined heat and power generation heat supply unit.

Description

Technical Field

[0001] The present invention belongs to the field of thermal power unit optimization, and specifically refers to a method for optimizing the thermal power load of a cogeneration heating unit. Background Art

[0002] The existing load optimization technology for cogeneration units has the following problems:

[0003] Model static defects:

[0004] Traditional methods rely on offline modeling (e.g., the static thermal model used in patent CN114065487A), which cannot dynamically reflect the actual operating status of the unit. For example, when coal quality fluctuates or equipment ages, the model prediction error can reach 15%-20%, rendering the optimization results ineffective.

[0005] Typical case: A power plant used a linear regression model based on historical data. Under deep peak-shaving conditions, the model did not include the dynamic characteristics of variable load, resulting in an excessive heating pressure accident.

[0006] The algorithm lacks real-time performance:

[0007] Existing optimization methods (such as the multi-objective genetic algorithm in CN113779997A) take up to 5-10 minutes to calculate, which cannot meet the AGC frequency modulation second-level response requirements.

[0008] Comparative data: The traditional particle swarm algorithm (CN113111984A) requires more than 300 iterations to converge in a 50-dimensional optimization problem, while the power grid dispatch instruction update cycle is only 2-4 seconds.

[0009] Weak ability to cope with uncertainty:

[0010] The failure to integrate environmental parameters (such as atmospheric pressure and cooling water temperature) and equipment health data results in poor robustness in the optimization solution. For example, patent CN112906347A fails to account for the decline in condenser efficiency under high summer temperatures, resulting in an 8% increase in coal consumption after optimization. Summary of the Invention

[0011] The technical problems to be solved by the present invention are the static defects of the model, the insufficient real-time performance of the algorithm and the weak ability to cope with uncertainty.

[0012] To solve the above problems, the technical solution adopted by the present invention is as follows: The method for optimizing the thermal power load of a cogeneration heating unit proposed in the present invention comprises the following steps:

[0013] S1: Real-time collection of thermal load data, equipment operating parameters and environmental parameters of the cogeneration unit;

[0014] S2: Preprocessing of collected data, including noise filtering, data standardization and feature extraction;

[0015] S3: Based on the preprocessed data, a dynamic digital twin model is constructed and the optimal thermal and electric load distribution plan is calculated using an improved intelligent optimization algorithm;

[0016] S4: Adjust the thermal and electric load output of the unit based on the optimization results, and verify the control effect in real time through the digital twin model;

[0017] S5: Feedback actual operation data to the digital twin model and dynamically update model parameters to improve prediction accuracy.

[0018] Furthermore, the improved intelligent optimization algorithm is an improved particle swarm optimization algorithm, the inertia weight of which is set to a dynamic adjustment value, and the learning factors are 1.5 and 1.7 respectively.

[0019] Furthermore, the dynamic adjustment formula of the inertia weight is:

[0020] w(t)=w 初始 ×e -λt +w 终值

[0021] Among them, w 初始 =0.9, w 终值 =0.4, λ is the attenuation coefficient.

[0022] Furthermore, in step S2, a low-pass filter with a cutoff frequency of 10 Hz is used for noise filtering, and a Z-score normalization method is used for data normalization.

[0023] Furthermore, the dynamic digital twin model includes the following submodules:

[0024] Thermal system simulation module: simulates the dynamic thermal process of boilers, steam turbines and heating systems;

[0025] Electrical system simulation module: simulates generators, grid interfaces and power regulation processes;

[0026] Control logic simulation module: reproduces the unit DCS control strategy and AGC frequency regulation logic.

[0027] Furthermore, the digital twin model updates its parameters every 10 minutes based on actual operating data.

[0028] Furthermore, the optimization goal in step S3 is:

[0029] min(α·coal consumption rate+β·NO x Emission + γ AGC frequency deviation)

[0030] Among them, α, β, and γ are weighting coefficients, which are determined through historical data training.

[0031] Furthermore, the method integrates a predictive control module to predict the grid load demand in the next 15 minutes through an LSTM neural network and pre-adjust the unit operating status.

[0032] Furthermore, the XX.

[0033] The beneficial effects achieved by the present invention using the above method are as follows:

[0034] 1. This proposal proposes a method for optimizing the thermal load of a cogeneration unit. By constructing a dynamic digital twin model, the method updates parameters every 10 seconds and integrates data such as coal quality correction factors and rotor stress monitoring. This significantly improves model accuracy (main steam pressure prediction error ≤ 1.5%) and allows for adaptation to complex operating conditions.

[0035] 2. The proposed method for optimizing the thermal load of a cogeneration heating unit adopts an improved particle swarm optimization algorithm, introduces dynamic inertia weights and adaptive learning factors, shortens the calculation time to less than 25 seconds, and improves the convergence speed by 40% (the number of iterations is reduced from 300 to 180), meeting real-time control requirements.

[0036] 3. The proposed method for optimizing the thermal load of cogeneration heating units improves system robustness through a multi-objective optimization function (coal consumption, emissions, and frequency regulation performance) and a predictive control module (LSTM network predicts load demand in the next 15 minutes). BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the overall steps of the present invention.

[0038] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0039] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0040] like Figure 1 As shown, the present invention proposes a method for optimizing the thermal load of a cogeneration heating unit, comprising the following steps:

[0041] S1: Real-time collection of thermal load data, equipment operating parameters and environmental parameters of the cogeneration unit;

[0042] S2: Preprocessing of collected data, including noise filtering, data standardization and feature extraction;

[0043] S3: Based on the preprocessed data, a dynamic digital twin model is constructed and the optimal thermal and electric load distribution plan is calculated using an improved intelligent optimization algorithm;

[0044] S4: Adjust the thermal and electric load output of the unit based on the optimization results, and verify the control effect in real time through the digital twin model;

[0045] S5: Feedback actual operation data to the digital twin model and dynamically update model parameters to improve prediction accuracy.

[0046] The improved intelligent optimization algorithm is the improved particle swarm optimization algorithm, whose inertia weight is set to a dynamic adjustment value, and the learning factors are 1.5 and 1.7 respectively.

[0047] The dynamic adjustment formula of inertia weight is:

[0048] w(t)=w 初始 ×e -λt +w 终值

[0049] Among them, w 初始 =0.9, w 终值 =0.4, λ is the attenuation coefficient.

[0050] In step S2, noise filtering was performed using a low-pass filter with a cutoff frequency of 10 Hz, and data normalization was performed using the Z-score normalization method.

[0051] The dynamic digital twin model includes the following submodules:

[0052] Thermal system simulation module: simulates the dynamic thermal process of boilers, steam turbines and heating systems;

[0053] Electrical system simulation module: simulates generators, grid interfaces and power regulation processes;

[0054] Control logic simulation module: reproduces the unit DCS control strategy and AGC frequency regulation logic.

[0055] The digital twin model updates its parameters every 10 minutes based on actual operating data.

[0056] The optimization goal in step S3 is:

[0057] min(α·coal consumption rate+β·NO x Emission + γ AGC frequency deviation)

[0058] Among them, α, β, and γ are weighting coefficients, which are determined through historical data training.

[0059] The method integrates a predictive control module, uses an LSTM neural network to predict the grid load demand in the next 15 minutes, and pre-adjusts the unit operating status.

[0060] Example 1: Application case of a 350MW supercritical cogeneration unit

[0061] Data acquisition: 128 parameters including main steam pressure, turbine power, and heating flow are collected at a frequency of 1 Hz;

[0062] Preprocessing: Use Butterworth low-pass filter (cut-off frequency 10 Hz) to eliminate noise;

[0063] Optimization calculation: set the particle swarm size to 50, the maximum number of iterations to 200, and the inertia weight to be dynamically adjusted (initial 0.9 → final value 0.4);

[0064] Control execution: After optimization, the unit's coal consumption was reduced by 2.1g / kWh, and the AGC frequency regulation performance indicator Kp value increased by 12%.

[0065] Example 2: Multi-algorithm fusion optimization

[0066] A simulated annealing algorithm is introduced into the optimization layer to perform global search, preventing the particle swarm from falling into local optima. Tests show that this method improves optimization stability by 18% under load mutation conditions.

[0067] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual method is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, devises methods, approaches, and embodiments similar to the technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. A method for optimizing the thermal load of a cogeneration heating unit, characterized in that: The following steps are involved: S1: Real-time collection of thermal load data, equipment operating parameters and environmental parameters of the cogeneration unit; S2: Preprocessing of collected data, including noise filtering, data standardization and feature extraction; S3: Based on the preprocessed data, a dynamic digital twin model is constructed and the optimal thermal and electric load distribution plan is calculated using an improved intelligent optimization algorithm; S4: Adjust the thermal and electric load output of the unit based on the optimization results, and verify the control effect in real time through the digital twin model; S5: Feedback actual operation data to the digital twin model and dynamically update model parameters to improve prediction accuracy.

2. The method for optimizing thermal load of a cogeneration heating unit according to claim 1, characterized in that: The improved intelligent optimization algorithm is an improved particle swarm optimization algorithm, in which the inertia weight is set to a dynamic adjustment value, and the learning factors are 1.5 and 1.7 respectively.

3. The method for optimizing thermal load of a cogeneration heating unit according to claim 2, characterized in that: The dynamic adjustment formula of the inertia weight is: w(t)=w 初始 ×e -λt +w 终值 Among them, w 初始 =0.9, w 终值 =0.4, λ is the attenuation coefficient.

4. The method for optimizing thermal load of a cogeneration heating unit according to claim 1, characterized in that: In step S2, a low-pass filter with a cutoff frequency of 10 Hz is used for noise filtering, and a Z-score normalization method is used for data normalization.

5. The method for optimizing thermal load of a cogeneration heating unit according to claim 1, characterized in that: The dynamic digital twin model includes the following submodules: Thermal system simulation module: simulates the dynamic thermal process of boilers, steam turbines and heating systems; Electrical system simulation module: simulates generators, grid interfaces and power regulation processes; Control logic simulation module: reproduces the unit DCS control strategy and AGC frequency regulation logic.

6. The method for optimizing thermal load of a cogeneration heating unit according to claim 5, characterized in that: The digital twin model updates its parameters every 10 minutes based on actual operating data.

7. The method for optimizing thermal power load of a cogeneration heating unit according to claim 1, characterized in that: The optimization goal in step S3 is: min(α·coal consumption rate+β·NO x Emission + γ AGC frequency deviation) Among them, α, β, and γ are weighting coefficients, which are determined through historical data training.

8. The method for optimizing thermal power load of a cogeneration heating unit according to claim 1, characterized in that: The method integrates a predictive control module, uses an LSTM neural network to predict the grid load demand in the next 15 minutes, and pre-adjusts the unit operating status.

Images