Incineration heat grading utilization method based on energy quality matching

Abstract

The invention belongs to the technical field of waste incineration, and particularly relates to an incineration heat grading utilization method based on energy and quality matching. According to the method, a real-time dynamic grading decision is made based on thermodynamic calculation of real-time steam parameters, the limitation of a fixed pressure grade is broken through, and accurate mapping of steam quality and an energy consumption scene is established; secondly, a steam demand dynamic response model is constructed through an LSTM load prediction module, supply and demand gaps of steam of different qualities are recognized in advance, and a priority distribution matrix is formed; finally, a fuzzy PID control engine and a three-level path switching mechanism are combined, a multi-level optimization system with the economic benefit maximization and the diverse loss minimization as targets is constructed, and optimal dynamic matching of steam parameters and energy consumption equipment is achieved.

Description

Technical Field

[0001] The present application belongs to the technical field of waste incineration, and more specifically, relates to a method for graded utilization of incineration heat based on energy-quality matching. Background Art

[0002] With the acceleration of urbanization, power generation from incineration of domestic waste has become a core means of urban solid waste treatment and energy recovery. During the incineration process, the heat energy of the waste is converted into steam through the waste heat boiler to drive the steam turbine unit to generate electricity. At the same time, part of the steam is extracted for factory heating or municipal heating, forming a typical energy cascade utilization system.

[0003] Traditional systems generally use a fixed pressure distribution model, dividing the steam network into three fixed pressure levels (e.g., 5MPa, 2.5MPa, and 0.8MPa): high, medium, and low. Mechanical pressure reducing valves are used to adjust steam parameters. However, due to the large fluctuations in the calorific value of garbage (4,200-9,600 kJ / kg), steam production varies dynamically with operating conditions, forcing the use of high-quality steam to be downgraded. Summary of the Invention

[0004] The present invention provides a method for graded utilization of incineration heat based on energy-quality matching, which aims to solve the technical problem that high-grade steam is forced to be downgraded due to the traditional method.

[0005] The incineration heat graded utilization method based on energy-quality matching includes the following steps: Data collection and preprocessing: Collect incineration boiler steam parameters and user demand data, preprocess the collected data to obtain preprocessed incineration boiler steam parameters and user demand data; Enthalpy and heat value calculation and dynamic classification decision-making: Based on the pre-treated incineration boiler steam parameters, the steam enthalpy value is calculated according to the thermodynamic model, and then the steam heat value is calculated based on the steam enthalpy value. The steam is divided into three grades based on the enthalpy value and heat value, and the steam grade is determined based on the calculated enthalpy value and heat value. Load forecasting and demand matching: Based on steam levels, historical load data, and external data, the LSTM load forecasting module outputs future power generation and heat load demands. The forecast results then calculate the supply and demand gap for each steam level, generate priority allocation recommendations, and obtain future and heat demand forecast curves and a steam allocation priority matrix. Multi-objective optimization and control command generation: Based on the steam level, load forecast and demand matching results, as well as real-time equipment status, a fuzzy PID control engine and a three-level path intelligent switching mechanism are used. A dynamically adjusted three-dimensional fuzzy rule library is used to optimize PID parameters in real time. The objective function is to maximize the balance between economic benefits and quality losses, and steam distribution is achieved through a hierarchical control strategy.

[0006] The present invention makes real-time dynamic graded decisions based on thermodynamic calculations of real-time steam parameters, breaks through the limitations of fixed pressure levels, and establishes a precise mapping of steam quality and energy usage scenarios. Secondly, a dynamic response model for steam demand is constructed through the LSTM load forecasting module to identify the supply and demand gaps of steam of different qualities in advance and form a priority allocation matrix. Finally, a fuzzy PID control engine is combined with a three-level path switching mechanism to construct a multi-level optimization system with the goal of maximizing economic benefits and minimizing heat loss, thereby achieving optimal dynamic matching between steam parameters and energy-consuming equipment.

[0007] Preferably, before calculating the enthalpy value, a moving average filter is used to perform a moving average filter on the pressure and temperature in the steam parameters; and then, based on the data after the moving average filter, it is identified whether it is in a transition state; wherein the pressure and temperature are unstable during the process of the transition state value transitioning from one level to another, and the determination of the transition state is based on a set threshold, and if it belongs to the transition state, it is marked as a transition state; In the process of determining the steam level, the transition state is introduced and the steam level is compensated using a fuzzy logic method. In the transition state, if the deviation of pressure and temperature is less than a threshold, the current level is maintained; if the deviation is greater than the threshold, the level is allowed to be switched.

[0008] Preferably, the steps of the LSTM load forecasting module for forecasting are as follows: Before prediction, the data is time-series aligned and aggregated. All data are time-series aligned. The steam data is aggregated using the sliding average method to align with the time series of other data. Feature construction is performed based on the time-aligned and aggregated data, including steam grade ratio, meteorological compensation factor, and generation plan weight. The steam grade ratio is the proportion of steam of different grades in the total steam flow; the meteorological compensation factor is a compensation coefficient constructed based on the impact of outdoor temperature on steam demand; the production plan weight is the normalized waste treatment volume, reflecting the impact of production intensity on steam generation; A standardized feature matrix is ​​constructed based on the time-aligned and aggregated data and the data obtained by feature construction, the standardized feature matrix is ​​used as the input of the LSTM load forecasting module, and a prediction is performed based on the LSTM load forecasting module to obtain future power generation demand and heating demand; The LSTM load forecasting module includes an LSTM layer, an Attention layer, and a fully connected layer. The Attention layer dynamically calculates the contribution weight of the time step in the output sequence of the LSTM layer to the current prediction, obtaining a weighted time series. The output is then mapped to the predicted values ​​of power generation and heating demand through the fully connected layer.

[0009] Preferably, the steps of generating the priority allocation suggestion are as follows: Demand-steam conversion: Calculate the required high-pressure steam flow rate based on the turbine efficiency and the enthalpy of high-pressure steam; calculate the required medium-pressure steam flow rate based on the supply and return water temperature difference and specific heat capacity; Gap rate calculation: Calculate the difference between power generation steam demand and actual high-pressure steam flow to reflect the supply and demand gap of high-pressure steam; calculate the difference between heating steam demand and actual medium-pressure steam flow to reflect the supply and demand gap of medium-pressure steam; calculate the difference between actual low-pressure steam flow and leachate treatment demand to reflect the excess of low-pressure steam; Dynamic priority weight allocation: Dynamically adjust power generation weight according to grid frequency deviation; dynamically adjust heating weight according to equipment scaling; dynamically adjust leachate weight according to low-pressure steam flow; The obtained weights are combined with the gap ratio to generate weight coefficients, and then the steam allocation priority matrix is ​​obtained, which includes the weight coefficients of power generation, heat supply and leachate treatment.

[0010] Preferably, the fuzzy PID control engine includes the following processing steps: Fuzzy rule matching: Through the preset fuzzy rule library, the input load deviation, deviation change rate and equipment health index are fuzzified; PID parameter adjustment: adjust the PID parameters according to the PID parameter adjustment amount output by the fuzzy rule to obtain preliminary PID parameters; Objective function optimization: Taking the balance between maximizing comprehensive benefits and quality loss as the objective function, preliminary PID parameters are used and the gradient descent method is used to iterate to find the optimal PID parameters that meet the objective function.

[0011] Preferably, the three-level path intelligent switching mechanism includes a high-pressure steam path, a medium-pressure steam path and a low-pressure steam path; The high-pressure steam path monitors the grid frequency deviation in real time and linearly adjusts the steam inlet of the back-pressure unit to smooth out grid fluctuations. If the frequency deviation exceeds the threshold of ±0.2Hz, the heat accumulator is immediately activated to store heat. Medium-pressure steam path: Feedforward-feedback composite control is used. Feedforward control adjusts the heating valve opening in advance based on predicted heating demand. Feedback control, based on optimized PID parameters, fine-tunes the heating valve opening in a cycle of ≤5 seconds to eliminate real-time heating temperature / flow deviations. Low-pressure steam path: Calculates the evaporator heat transfer coefficient in real time. If the efficiency is less than 85%, it automatically switches to a standby evaporation unit and increases the steam distribution volume to compensate for the efficiency loss. It also dynamically intercepts low-pressure steam resources based on a priority matrix.

[0012] Preferably, feedback correction is also included, comparing the instruction with the execution result, calculating the cumulative deviation, and triggering PID parameter adjustment based on the gradient descent method if the cumulative deviation is less than or equal to a preset deviation threshold; If the accumulated deviation is greater than the preset deviation threshold, expert diagnosis is performed.

[0013] The beneficial effects of the present invention include: The present invention makes real-time dynamic graded decisions based on thermodynamic calculations of real-time steam parameters, breaks through the limitations of fixed pressure levels, and establishes a precise mapping of steam quality and energy usage scenarios. Secondly, a dynamic response model for steam demand is constructed through the LSTM load forecasting module to identify the supply and demand gaps of steam of different qualities in advance and form a priority allocation matrix. Finally, a fuzzy PID control engine is combined with a three-level path switching mechanism to construct a multi-level optimization system with the goal of maximizing economic benefits and minimizing heat loss, thereby achieving optimal dynamic matching between steam parameters and energy-consuming equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0015] Figure 1 This is a flowchart of the overall steps provided by an embodiment of the present invention.

[0016] Figure 2 This is a specific step block diagram of step S2 provided in an embodiment of the present invention.

[0017] Figure 3 This is a specific step block diagram of step S3 provided in an embodiment of the present invention.

[0018] Figure 4 This is a specific step block diagram of step S4 provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the technical problems, technical solutions and beneficial effects to be solved by this application more clearly understood, this application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0020] See also Figure 1 As shown, the incineration heat graded utilization method based on energy-quality matching includes the following steps: S1. Data collection and preprocessing: Collect incineration boiler steam parameters and user demand data, preprocess the collected data, and obtain preprocessed incineration boiler steam parameters and user demand data; The incineration boiler steam parameters include the pressure, temperature and flow of the incineration boiler outlet steam parameters; the user side demand parameters include the grid dispatch instructions (frequency, power demand), the heating pipe network pressure and the leachate treatment volume; The preprocessing includes outlier detection, time synchronization and other processing to ensure data quality. The preprocessing to ensure data quality is a conventional technical means in this field, so this application will not be described in detail. S2. Enthalpy and RH calculation and dynamic classification decision-making: Based on the pretreated incineration boiler steam parameters, the steam enthalpy is calculated according to a thermodynamic model. The RH is then calculated based on the steam enthalpy. The steam is classified into three grades based on the enthalpy and RH values, resulting in a steam grade. The steam grade is determined based on the calculated enthalpy and RH values. See also Figure 2 As shown, first, in order to eliminate sensor noise and smooth the data, the sliding window technology is used to smooth the real-time steam pressure and temperature, where the window length is 30 seconds and the period is 10 seconds, and the smoothed filtered data is obtained based on the sliding average filter; Based on the smoothed filtered data, the pressure and temperature changes are used to determine whether the state is in a transition state. The definition of the boundary area is determined by the pressure and temperature thresholds: High voltage / medium voltage boundary: ; Medium voltage / low voltage boundary:

[0021] Temperature boundary: ; The identification rules of transition states are as follows: like or , it is marked as a transition state; represents the steam pressure at time t; represents the steam temperature at time t; Calculate the specific enthalpy and specific heat of steam and quantify the energy quality of steam based on the specific enthalpy and specific heat of steam: The specific enthalpy of steam is calculated according to the simplified model of IAPWS-IF97 standard. The specific expression is as follows: ; Where: represents the specific heat capacity at constant pressure; Indicates the density of steam; represents the latent heat of phase change; represents the pressure correction factor, and e represents the base of the natural logarithm; Steam ratio calculation: ; Where: represents the specific enthalpy under reference ambient conditions; represents the specific entropy under reference environmental conditions; h represents the calculated specific enthalpy value; Represents the calculated specific entropy value; the steam specific entropy is calculated based on the following formula: ; Where: represents the specific entropy of steam; Indicates the reference ambient temperature; represents the filtered temperature; R represents the gas constant of water vapor; Indicates the steam pressure after filtering; Indicates the baseline ambient pressure; Dynamic classification decision: steam classification is performed based on filtered pressure, temperature, ratio value and whether it is in a transition state, as follows: high pressure: , ; Medium pressure: , ; Low pressure: , ; During the transition state, deviations in pressure and temperature can affect the grade determination; therefore, fuzzy logic is used to adjust the grade to avoid frequent switching of the transition state. The fuzzy logic rules are as follows: like and ,but ; (Priority is given to retaining the original level); like or ,but ; (allow level switching); in and Represents pressure deviation and temperature deviation respectively. Pressure deviation is based on the real-time pressure and the corresponding threshold value. calculate; Decomposition value based on real-time temperature and response Calculated; The compensated value is: ; By adjusting the compensated value: , avoid frequent level switching.

[0022] In this embodiment, the transition state is defined by pressure and temperature thresholds, and the unstable area during level switching is identified. After introducing the fuzzy logic compensation mechanism, when the deviation is small, the original level is preferentially retained through the compensation factor to avoid frequent level switching due to slight parameter fluctuations, thereby significantly improving the system control stability, reducing unnecessary actions of the actuator, and extending the life of the equipment; secondly, since traditional steam classification only relies on static thresholds and is easily affected by operating condition fluctuations, this embodiment dynamically identifies the transition state and combines it with fuzzy compensation to enable the classification decision to have the ability of "inertia maintenance", maintain the current level when the parameters fluctuate, and trigger only when it deviates significantly from the threshold, thereby avoiding control loop oscillation and improving the operating stability of the incineration boiler steam system.

[0023] S3. Load Forecasting and Demand Matching: Based on steam levels, historical load data, and external data, the LSTM load forecasting module outputs future power generation and heat load demands. The forecast results then calculate the supply and demand gap for each steam level, generate priority allocation recommendations, and produce future and heat demand forecast curves and a steam allocation priority matrix. The steam level is the real-time steam level label (high pressure, medium pressure, low pressure), heat value, and flow rate, with a sampling period of 10 seconds. The historical load data is the power generation and heating demand of the past 72 hours, with a time granularity of 5 minutes and including date types. External data includes weather forecasts and production plans, where the weather forecast is the temperature and wind speed for the next 24 hours with a time granularity of 1 hour; the production plan is the waste treatment volume for the next 24 hours with a time granularity of 1 hour. See also Figure 3 As shown, first, the input data is aligned and aggregated in time series: the 10-second steam data is aggregated into a 5-minute granularity using a sliding average algorithm (window width 30, overlap rate 50%); Then align the timestamps of the external data and convert the 1-hour granularity meteorological data into 5-minute granularity through linear interpolation; Feature construction: including steam grade ratio, meteorological compensation factor and production plan weight; The steam grade ratio is the ratio of high-pressure, medium-pressure, and low-pressure steam in the total flow. For example, the high-pressure steam ratio is calculated as follows: ; Where: Indicates the proportion of high-pressure steam in the total flow; Indicates high-pressure steam flow; Indicates medium pressure steam flow; Indicates low-pressure steam flow; In this embodiment, the steam grade proportion can reflect the current energy quality distribution, provide real-time energy quality state information for the prediction model, and can more comprehensively characterize the steam supply-demand relationship compared with single pressure / temperature parameters.

[0024] The meteorological compensation factor is designed according to the influence of outdoor temperature on steam demand, and the specific expression is as follows: ; In the formula: represents the outdoor temperature; represents the meteorological compensation factor; In this embodiment, the influence of outdoor temperature on heating demand is quantified, enabling the model to adapt to climate changes.

[0025] The production plan is seriously based on the normalized value of the garbage disposal volume, reflecting the influence of the production plan on steam demand: ; In the formula: represents the actual garbage disposal volume; the maximum value is 30.

[0026] In this embodiment, the garbage disposal volume is normalized to directly relate to the steam generation volume, enabling the model to capture production load changes in advance.

[0027] Based on the data after feature construction, time alignment, and aggregation, a standardized feature matrix is constructed, including 15-dimensional features (time, steam grade proportion, exergy value, meteorological factor, production weight, etc.), with a time granularity of 5 minutes, and is standardized by Z-score (mean is​​​​​​​​​​​​​​​​​​​​​​​​​The time series output by the second LSTM layer is weighted based on the Attention layer to obtain a weighted time series. This weighted time series is then processed by a fully connected layer, which uses 32 neurons and mapped to the final load forecast value. Based on this, the power generation and heating demand for the next four hours is output with a time granularity of 5 minutes.

[0029] Power generation steam demand: Based on the turbine efficiency and the high-pressure steam value, the required high-pressure steam flow rate is calculated: ; Where: Indicates the efficiency of the steam turbine; Indicates the value of high-pressure steam; represents the power generation demand; Indicates the required high-pressure steam flow rate; Heating steam demand: Calculate the required steam flow based on the supply and return water temperature difference and specific heat capacity: ; Where: Indicates the water supply temperature; Indicates return water temperature; represents specific heat capacity; Indicates heating demand; Indicates the required steam flow rate; The required flow rate is calculated based on the turbine efficiency and high-pressure steam field value (energy quality), rather than the traditional distribution based solely on pressure levels. This allows the power generation system to reduce flow consumption when the high-pressure steam field value is high (such as 300°C, 5MPa) (the higher the field value, the higher the energy utilization rate).

[0030] Calculate the high-pressure, medium-pressure and low-pressure steam gap rates separately: ; ; ; Where: Indicates high-pressure steam flow; Indicates the high-pressure steam gap rate; Indicates the medium pressure steam gap rate; Indicates medium pressure steam flow; Express the low-pressure steam gap ratio; Indicates leachate treatment needs; Indicates low-pressure steam flow; The supply-demand gap ratio calculated 、 、 , positive values ​​indicate shortages, and negative values ​​indicate surpluses; Power generation weight calculation: Dynamically adjusted according to grid frequency deviation. The greater the frequency deviation, the higher the power generation priority: ; Where: Indicates the real-time frequency deviation of the power grid; represents the power generation weight; Heating weight calculation: ; Where: Indicates the scaling coefficient; represents the heating weight; Leachate weight: ; Where: Indicates low-pressure steam flow; represents the leachate weight; The Softmax function is used to combine the weights and the gap rate to generate the steam allocation priority matrix: ; in is the weight coefficient of high-pressure steam; is the weight coefficient of medium-pressure steam; is the weight coefficient of low-pressure steam; The weight and gap rate data are normalized based on the above Softmax function to obtain the final steam allocation priority matrix.

[0031] S4. Multi-Objective Optimization and Control Command Generation: Based on steam level, load forecast and demand matching results, and real-time equipment status, a fuzzy PID control engine and a three-level intelligent path switching mechanism are used to optimize PID parameters in real time using a dynamically adjusted three-dimensional fuzzy rule base. The objective function is to maximize the balance between economic benefits and quality losses, and steam distribution is achieved through a hierarchical control strategy. See also Figure 4 As shown, the fuzzy PID control engine includes the following processing steps: Fuzzy rule matching: Through the preset fuzzy rule library, the input load deviation, deviation change rate and equipment health index are fuzzified; The load deviation is the output value of the current power generation / heating demand and the predicted value; the variation rate is the change trend of the load deviation over time; the equipment health index includes the real-time health score of key equipment such as steam turbines and heat accumulators; The fuzzy processing is performed by matching through the fuzzy rule base, as follows: According to the magnitude of the load deviation (such as "positive and large" or "negative and small") and the rate of change (such as "rapid increase" or "slow decrease"), combined with the equipment health status (such as "good" or "warning"), a matching control strategy is selected from the predefined fuzzy rule library; For example, if the power generation demand deviation is "largely negative" (severely insufficient) and the equipment health status is "good", the steam turbine output is prioritized and the weight of heat distribution is reduced; PID parameter adjustment: adjust the PID parameters according to the PID parameter adjustment amount output by the fuzzy rule to obtain preliminary PID parameters; The adjustment amount output by the fuzzy rule (such as "increase the proportional coefficient Kp") acts on the PID controller to optimize the control response speed and stability in real time. For example: When the load deviation increases rapidly, increase the proportional coefficient (Kp) to speed up the response.

[0032] When the equipment health index is low, reduce the integral factor (Ki) to avoid equipment overload.

[0033] Based on this, preliminary PID parameters are obtained, and then multi-objective optimization calculations are performed, with the objective function of maximizing the balance between comprehensive benefits and quality loss. The preliminary PID parameters are used and the gradient descent method is iterated to find the optimal PID parameters that meet the objective function. The expression of the objective function is as follows: ; Where: and represents the quality loss function; Indicates actual steam pressure; Indicates the actual temperature; Indicates the base pressure; Indicates the reference temperature; In this embodiment, the equipment health index (such as turbine bearing temperature and heat accumulator leakage rate) is introduced as a fuzzy input. When the health status is "warning", the integral coefficient Ki is automatically reduced (for example, from 0.5 to 0.3) to avoid equipment overload due to excessive integral adjustment.

[0034] The three-level path intelligent switching mechanism includes a high-pressure steam path, a medium-pressure steam path, and a low-pressure steam path; The high-pressure steam path monitors the grid frequency deviation in real time and linearly adjusts the steam intake of the back-pressure unit to smooth out grid fluctuations. The linear adjustment expression for the back-pressure unit intake is as follows: ; Where: represents the frequency adjustment coefficient; Indicates grid frequency deviation; Indicates the steam inlet volume of the back pressure unit; Indicates the designed steam inlet volume; If the frequency deviation exceeds the threshold of ±0.2Hz, the heat storage device is immediately activated to store heat. The heat storage capacity of the heat storage device is calculated based on the following formula: ; Where: Indicates heat storage efficiency; Indicates the heat storage capacity of the heat accumulator; Indicates the capacity of the heat storage tank; In this embodiment, the grid frequency deviation is monitored in real time, and the fluctuation is quickly smoothed out by linearly adjusting the steam inlet volume of the back-pressure unit.

[0035] Medium-pressure steam path: Feedforward-feedback composite control is adopted, in which feedforward control adjusts the heating valve opening in advance based on the predicted heating demand; ; Where: Indicates the opening of the heating valve; Indicates the basic opening; Indicates the heating regulation coefficient; represents the predicted building heat load; Feedback control is based on optimized PID parameters, fine-tuning the heating valve opening in a cycle of ≤5 seconds to eliminate real-time heating temperature / flow deviation; ; Where: Indicates heating deviation; 、 、 Represents the proportional, integral, and differential parameters of the PID controller; Indicates the last heating valve opening; Integral term representing heating deviation; rate of change of heating deviation; In this embodiment, feedback control is combined with optimized PID parameters (5-second fine-tuning cycle) to eliminate flow deviation in real time, so that the pressure fluctuation of the heating network is ≤0.05MPa.

[0036] Low-pressure steam path: Continuously monitor the evaporator's heat transfer coefficient. When the heat transfer coefficient is less than 85%, it indicates that the evaporator is fouled, affecting the heat transfer efficiency. Therefore, it triggers the switch to the backup evaporation unit to maintain the efficiency of steam generation. The evaporation capacity after the switch is calculated using the following formula: ; Where: represents the scaling influence coefficient; represents the heat transfer coefficient; Indicates the current evaporation capacity of the evaporation unit after switching; Indicates the adjusted evaporation capacity of the evaporation unit after switching; wherein the scaling influence coefficient is calculated based on the current heat transfer efficiency and the benchmark heat transfer efficiency; that is, it is obtained by subtracting the ratio of the current heat transfer efficiency to the benchmark heat transfer efficiency from 1; When distributing steam, the following formula is used according to the priority of leachate treatment: ; Where: represents the priority weight of leachate treatment; Indicates the total distribution of low-pressure steam; Indicates the amount of steam used for leachate treatment; In this embodiment, when the heat transfer coefficient is lower than 85%, the standby evaporation unit is automatically switched, and the switching time is ≤30 seconds, to avoid the decrease in steam production due to scaling (for example, when scaling reduces the heat transfer efficiency by 20%, the standby unit can compensate for the 15% loss in evaporation), thus ensuring the continuous operation of leachate treatment (traditional manual switching takes 2-3 hours and easily leads to interruption of sewage treatment); secondly, the priority weight ensures that low-pressure steam is supplied first for leachate treatment (an industry necessity). Even if the total amount of low-pressure steam is insufficient, it can still guarantee more than 70% of the treatment demand, avoiding the risk of leachate overflow.

[0037] S5. Feedback correction: Compare the command and the execution result, calculate the cumulative deviation, and trigger PID parameter adjustment based on the gradient descent method if the cumulative deviation is less than or equal to the preset deviation threshold; If the accumulated deviation is greater than the preset deviation threshold, expert diagnosis is performed.

[0038] The present invention makes real-time dynamic graded decisions based on thermodynamic calculations of real-time steam parameters, breaks through the limitations of fixed pressure levels, and establishes a precise mapping of steam quality and energy usage scenarios. Secondly, a dynamic response model for steam demand is constructed through the LSTM load forecasting module to identify the supply and demand gaps of steam of different qualities in advance and form a priority allocation matrix. Finally, a fuzzy PID control engine is combined with a three-level path switching mechanism to construct a multi-level optimization system with the goal of maximizing economic benefits and minimizing heat loss, thereby achieving optimal dynamic matching between steam parameters and energy-consuming equipment.

[0039] The above is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A method for hierarchical utilization of incineration heat based on energy-quality matching, characterized in that: The following steps are involved: Data collection and preprocessing: Collect incineration boiler steam parameters and user demand data, preprocess the collected data to obtain preprocessed incineration boiler steam parameters and user demand data; Enthalpy and heat value calculation and dynamic classification decision-making: Based on the pre-treated incineration boiler steam parameters, the steam enthalpy value is calculated according to the thermodynamic model, and then the steam heat value is calculated based on the steam enthalpy value. The steam is divided into three grades based on the enthalpy value and heat value, and the steam grade is determined based on the calculated enthalpy value and heat value. Load forecasting and demand matching: Based on steam levels, historical load data, and external data, the LSTM load forecasting module outputs future power generation and heat load demands. The forecast results then calculate the supply and demand gap for each steam level, generate priority allocation recommendations, and obtain future and heat demand forecast curves and a steam allocation priority matrix. Multi-objective optimization and control command generation: Based on the steam level, load forecast and demand matching results, as well as real-time equipment status, a fuzzy PID control engine and a three-level path intelligent switching mechanism are used. A dynamically adjusted three-dimensional fuzzy rule library is used to optimize PID parameters in real time. The objective function is to maximize the balance between economic benefits and quality losses, and steam distribution is achieved through a hierarchical control strategy.

2. The method for graded utilization of incineration heat based on energy-quality matching according to claim 1 is characterized in that: Before calculating the enthalpy value, a moving average filter is used to perform moving average filtering on the pressure and temperature in the steam parameters; then, based on the data after the moving average filtering, whether it is in a transition state is identified; wherein the pressure and temperature are unstable during the process of the transition state value transitioning from one level to another, wherein the transition state is determined based on a set threshold, and if it is a transition state, it is marked as a transition state; In the process of determining the steam level, the transition state is introduced and the steam level is compensated using a fuzzy logic method. In the transition state, if the deviation of pressure and temperature is less than a threshold, the current level is maintained; if the deviation is greater than the threshold, the level is allowed to be switched.

3. The method for graded utilization of incineration heat based on energy-quality matching according to claim 1 is characterized in that: The steps for the LSTM load forecasting module to perform forecasting are as follows: Before prediction, the data is time-series aligned and aggregated. All data are time-series aligned. The steam data is aggregated using the sliding average method to align with the time series of other data. Feature construction is performed based on the time-aligned and aggregated data, including steam grade ratio, meteorological compensation factor, and generation plan weight. The steam grade ratio is the proportion of steam of different grades in the total steam flow; the meteorological compensation factor is a compensation coefficient constructed based on the impact of outdoor temperature on steam demand; the production plan weight is the normalized waste treatment volume, reflecting the impact of production intensity on steam generation; A standardized feature matrix is ​​constructed based on the time-aligned and aggregated data and the data obtained by feature construction, the standardized feature matrix is ​​used as the input of the LSTM load forecasting module, and a prediction is performed based on the LSTM load forecasting module to obtain future power generation demand and heating demand; The LSTM load forecasting module includes an LSTM layer, an Attention layer, and a fully connected layer. The Attention layer dynamically calculates the contribution weight of the time step in the output sequence of the LSTM layer to the current prediction, obtaining a weighted time series. The output is then mapped to the predicted values ​​of power generation and heating demand through the fully connected layer.

4. The method for graded utilization of incineration heat based on energy-quality matching according to claim 1 is characterized in that: The steps for generating the priority allocation suggestion are as follows: Demand-steam conversion: Calculate the required high-pressure steam flow rate based on the turbine efficiency and the enthalpy of high-pressure steam; calculate the required medium-pressure steam flow rate based on the supply and return water temperature difference and specific heat capacity; Gap rate calculation: Calculate the difference between power generation steam demand and actual high-pressure steam flow to reflect the supply and demand gap of high-pressure steam; calculate the difference between heating steam demand and actual medium-pressure steam flow to reflect the supply and demand gap of medium-pressure steam; calculate the difference between actual low-pressure steam flow and leachate treatment demand to reflect the excess of low-pressure steam; Dynamic priority weight allocation: Dynamically adjust power generation weight according to grid frequency deviation; dynamically adjust heating weight according to equipment scaling; dynamically adjust leachate weight according to low-pressure steam flow; The obtained weights are combined with the gap ratio to generate weight coefficients, and then the steam allocation priority matrix is ​​obtained, which includes the weight coefficients of power generation, heat supply and leachate treatment.

5. The method for hierarchical utilization of incineration heat based on energy-quality matching according to claim 1 is characterized in that: The fuzzy PID control engine includes the following processing steps: Fuzzy rule matching: Through the preset fuzzy rule library, the input load deviation, deviation change rate and equipment health index are fuzzified; PID parameter adjustment: adjust the PID parameters according to the PID parameter adjustment amount output by the fuzzy rule to obtain preliminary PID parameters; Objective function optimization: Taking the balance between maximizing comprehensive benefits and quality loss as the objective function, preliminary PID parameters are used and the gradient descent method is used to iterate to find the optimal PID parameters that meet the objective function.

6. The method for graded utilization of incineration heat based on energy-quality matching according to claim 5 is characterized in that: The three-level path intelligent switching mechanism includes a high-pressure steam path, a medium-pressure steam path, and a low-pressure steam path; The high-pressure steam path monitors the grid frequency deviation in real time and linearly adjusts the steam inlet of the back-pressure unit to smooth out grid fluctuations. If the frequency deviation exceeds the threshold of ±0.2Hz, the heat accumulator is immediately activated to store heat. Medium-pressure steam path: Feedforward-feedback composite control is adopted, in which feedforward control adjusts the heating valve opening in advance based on the predicted heating demand; Feedback control is based on optimized PID parameters, fine-tuning the heating valve opening in a cycle of ≤5 seconds to eliminate real-time heating temperature / flow deviation; Low-pressure steam path: Calculates the evaporator heat transfer coefficient in real time. If the efficiency is less than 85%, it automatically switches to a standby evaporation unit and increases the steam distribution volume to compensate for the efficiency loss. It also dynamically intercepts low-pressure steam resources based on a priority matrix.

7. The method for graded utilization of incineration heat based on energy-quality matching according to claim 1 is characterized in that: It also includes feedback correction, comparing instructions with execution results, calculating the cumulative deviation, and triggering PID parameter adjustment based on the gradient descent method if the cumulative deviation is less than or equal to the preset deviation threshold; If the accumulated deviation is greater than the preset deviation threshold, expert diagnosis is performed.

Images