Multi-objective constrained flue gas collaborative treatment optimization scheduling system

Abstract

The invention relates to the technical field of coal-fired power plant environmental protection, and discloses a multi-objective constrained flue gas collaborative treatment optimization scheduling system, which comprises a data acquisition module, a data analysis module, an optimization scheduling module and a control execution module, wherein the data acquisition module is responsible for acquiring flue gas emission data, treatment equipment operation state data, fire coal quality data and the like of a coal-fired power plant in real time. According to the method, the global search capability of the genetic algorithm is combined with the rapid convergence characteristic of the particle swarm algorithm, three-dimensional balance optimization of pollutant discharge (SO / NOx concentration reduction), treatment cost (annual operation cost reduction) and equipment efficiency (desulfurization efficiency improvement) is achieved, and the method is based on the constraint boundary dynamic adjustment technology of the real-time working condition. The pollutant emission limit value can be automatically tightened according to the peak regulation requirement of the power grid, the treatment cost constraint is dynamically relaxed when the coal quality changes, and the linear coupling of the equipment efficiency constraint and the unit load is realized.

Description

Technical Field

[0001] This invention relates to the field of environmental protection technology for coal-fired power plants, specifically to a multi-objective constrained flue gas collaborative treatment and optimization scheduling system. Background Technology

[0002] With increasingly stringent environmental protection requirements, the flue gas treatment of coal-fired power plants, as one of the major sources of pollution, has received increasing attention. Traditional flue gas treatment methods often focus only on the removal of single pollutants, neglecting the synergistic effects between multiple pollutants and the economic and efficiency issues in the treatment process. Therefore, it is particularly important to develop a multi-objective constrained flue gas synergistic treatment optimization scheduling system that can comprehensively consider the synergistic treatment of multiple pollutants, as well as economic and efficiency considerations.

[0003] Currently, flue gas treatment systems in coal-fired power plants typically employ independent treatment equipment to treat single pollutants, such as desulfurization equipment, denitrification equipment, and dust removal equipment. The lack of effective coordination and scheduling among these devices leads to resource waste and low efficiency during the treatment process. Furthermore, due to the complex and variable operating conditions of coal-fired power plants, traditional fixed-parameter control methods are ill-suited to the treatment needs under different operating conditions.

[0004] To address the aforementioned issues, this invention proposes a multi-objective constraint-based flue gas collaborative treatment and optimization scheduling system. This system integrates advanced sensor technology, data analysis technology, and optimization algorithms to monitor flue gas emissions from coal-fired power plants in real time. Based on different operating conditions and treatment requirements, it dynamically adjusts the operating parameters and scheduling strategies of the treatment equipment to achieve collaborative treatment and optimized scheduling of multiple pollutants. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-objective constrained flue gas collaborative treatment optimization scheduling system to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a multi-objective constrained flue gas collaborative treatment and optimization scheduling system, comprising a data acquisition module, a data analysis module, an optimization scheduling module, and a control execution module; wherein, the data acquisition module is responsible for real-time acquisition of flue gas emission data, treatment equipment operating status data, and coal quality data from coal-fired power plants; the data analysis module preprocesses and analyzes the acquired data to extract key feature parameters related to flue gas treatment; the optimization scheduling module, based on the extracted key feature parameters and preset multi-objective constraints, uses advanced optimization algorithms to perform optimization calculations to obtain the optimal treatment equipment operating parameters and scheduling strategy; the control execution module, based on the optimization results obtained by the optimization scheduling module, performs real-time control and adjustment of the treatment equipment;

[0007] The data acquisition module includes, but is not limited to, flue gas emission sensors, treatment equipment operation status sensors, and coal quality detectors;

[0008] The data analysis module employs data cleaning, data fusion, and feature extraction steps, as detailed below:

[0009] S1 data cleaning steps:

[0010] The data cleaning described in S1.1 uses a combined Z-score and IQR method to remove outliers. The Z-score method can capture Gaussian distribution outliers, while the IQR method can handle asymmetric distribution outliers.

[0011] The data cleaning described in S1.2 uses a KNN-LSTM hybrid model with null correlation to handle missing values;

[0012] S2 data fusion steps:

[0013] S2.1 Time alignment is performed between flue gas online monitoring data (high frequency) and DCS control data (low frequency) using the formula. In the formula, φ represents the alignment path, i and n are the indices of sequences X and Y, respectively, and ||x i -y φ(i) || represents the distance between corresponding points, ensuring that the two are synchronized in the time dimension, thus making subsequent causal analysis more effective;

[0014] S2.2 utilizes spatial fusion technology, through formulas In the formula, m(A) represents the confidence level of event A, B∩C=A means that the intersection of two evidence sources B and C is A, m1(B) and m2(C) are the confidence levels of evidence sources B and C for event A, respectively, and K is the conflict coefficient, which is used to measure the degree of conflict between evidence sources. Through this formula, multi-source data can be effectively integrated, information conflict can be reduced, and the confidence level of the fusion result can be improved.

[0015] S2.3 utilizes data correlation analysis techniques, through formulas In the formula, I(X,Y) represents the mutual information entropy of X and Y, p(x,y) is the joint probability distribution of X and Y, and p(x) and p(y) are the marginal probability distributions of X and Y, respectively. This formula can accurately quantify the correlation strength between variables, reveal the intrinsic relationship of data, and further optimize the fusion effect.

[0016] S3 Feature Extraction Steps:

[0017] S3.1 Extracts short-time energy based on temporal features, using the formula... In the formula, N is the number of data points, x iThis is the value of the i-th data point. The current energy changes significantly during short-term operation when the equipment starts and stops; this characteristic can be used for operating condition identification. Waveform factors are extracted using the formula... In the formula, FF is the average value of the data, and RMS is the root mean square value. When FF>1.5, it indicates that there is abnormal vibration in the equipment, such as a failure of the induced draft fan bearing.

[0018] S3.2 Power spectral density is extracted based on frequency domain features, using the formula... In the formula, T is the sampling period, f is the frequency, and x is the frequency. n It is the value of the nth sampling point. This formula is used to capture the frequency components during equipment operation and identify energy changes in a specific frequency band.

[0019] S3.3 Extracts wavelet packet energy spectrum based on time-frequency features, using the formula In the formula, d j,k (n) represents the wavelet coefficients of the k-th node in the j-th layer. This formula accurately captures the energy distribution of the equipment in different frequency bands and effectively identifies minor faults under complex working conditions.

[0020] S3.4 Desulfurization efficiency model based on mechanism model-driven feature extraction, through formula In the formula, C in It is the SO2 concentration in the inlet flue gas, C out The SO2 concentration in the flue gas is given by L / G, the liquid-to-gas ratio is given by L / G, pH is the pH value of the slurry, and T is the temperature. The theoretical desulfurization efficiency is calculated by combining the real-time pH value and the liquid-to-gas ratio, and then compared with the measured value to verify and optimize the model.

[0021] Preferably, the specific implementation steps for collecting real-time flue gas emission data from coal-fired power plants using sensors are as follows:

[0022] Step S1: Start all sensors, set the sampling frequency and data transmission protocol, where the sampling frequency is set to f. s The highest frequency of the measured signal is f max The sampling frequency setting formula f is adopted. s ≥2·f max To ensure that the sampling frequency is at least twice the highest frequency of the signal, avoid spectral aliasing. For example, if the SO2 concentration fluctuation frequency is 1Hz, the sampling frequency must be ≥2Hz. As for the data transmission protocol, Modbus TCP or MQTT can be selected. Modbus TCP is suitable for wired transmission and realizes communication between the sensor and the host computer through register mapping. MQTT is suitable for wireless transmission and uses a publish-subscribe model to reduce network load. Choose the protocol according to the transmission distance, bandwidth and real-time requirements.

[0023] Step S2: Collect real-time data on flue gas emissions from coal-fired power plants (such as the concentration of pollutants like SO2, NOx, and dust), operating status data of treatment equipment (such as parameters like operating current, voltage, and temperature), and coal quality data (such as calorific value, ash content, and volatile matter content) through sensors and detectors.

[0024] Specifically, for pollutant concentration calculation (taking SO2 as an example), the sensor output current is set to... The zero-point current is I0, and the sensor sensitivity (μA / ppm) is S. The formula is used... The pollutant concentration is calculated, where K is the calibration coefficient (considering temperature and pressure compensation);

[0025] Among them, for monitoring the operating status data of treatment equipment (taking the desulfurization tower slurry circulation pump as an example), the current, voltage and temperature of the equipment are monitored in real time, and the equipment is judged to be operating normally by combining preset thresholds; it includes single parameter threshold judgment, multi-parameter comprehensive judgment, threshold adaptive adjustment and alarm triggering and hierarchical mechanism;

[0026] Specifically, for the calculation of coal mass (taking calorific value as an example), the formula is used.

[0027] Among them, Q net It is the calorific value of coal (unit: kJ / kg or kcal / kg), and C, H, O, and S are the mass fractions of carbon, hydrogen, oxygen, and sulfur in coal, respectively.

[0028] Step S3: Transmit the collected data to the data center or host computer for processing via wired or wireless means. During the data transmission process, encryption technology is used to ensure data security.

[0029] Step S4: To ensure the stability and reliability of data transmission, a CRC checksum formula is used.

[0030] CRC=(M(x)·x 16 mod G(x)

[0031] G(x)=x 16 +x 12 +x 5 +1, where M(x) represents the polynomial of the data and G(x) is the standard generator polynomial. Before sending data, the sending end calculates the CRC value of the data according to the above formula and appends it to the end of the data before sending it. After receiving the data, the receiving end recalculates the CRC value of the received data. The receiving end compares the calculated CRC value with the CRC value appended by the sending end. If the two CRC values ​​match, the data transmission is considered to be error-free. If they do not match, it is considered that an error has occurred during the data transmission, triggering the retransmission mechanism.

[0032] To more effectively handle packet loss and latency issues in the network, an adaptive retransmission algorithm, RTO = β·RTT, is adopted. 平滑 This allows for dynamic adjustment of the retransmission interval, where β is an adjustment factor used to adjust the RTO (Retransmission Timeout) and RTT (Retransmission Time To Trace) based on network conditions. 平滑 It is a weighted average of round-trip times, used to estimate the average time it takes for a data packet to travel from the sender to the receiver and back to the sender.

[0033] When the sending end sends a data packet and starts timing, the receiving end sends an acknowledgment (ACK) to the sending end upon receiving the data packet. Upon receiving the ACK, the sending end stops timing and records this RTT (Round-Trip Time). The sending end calculates the RTT based on the multiple recorded RTTs. 平滑 The sending end is based on RTT 平滑 The RTO is calculated using the adjustment factor β; if no ACK is received within the RTO time, the data packet is considered lost, triggering the retransmission mechanism and recalculating the RTO.

[0034] Preferably, the single parameter threshold determination is as follows:

[0035] Involving a single current threshold, let the real-time current value be I. 实际 Through formula To determine if the current is abnormal, the threshold is dynamically adjusted based on load changes. For example, the current of the desulfurization tower slurry circulation pump is related to the flue gas flow rate. Among them, I 上限 It is the maximum allowable current limit under certain conditions, I 额定 The rated current of the equipment or system, k is the load factor, Q 烟气 Q represents the actual flue gas flow rate during operation. 额定 The rated flue gas flow rate on which the equipment or system was designed;

[0036] Involving a single voltage threshold, let the rated voltage of the device be U. 额定 The actual voltage of the device is U. 实际 Through formula To determine if the voltage is abnormal, ΔU 允许 Allowable voltage fluctuation range;

[0037] Involving a single temperature threshold, let the actual temperature of the equipment be T. 实际 Through formula To determine if the temperature exceeds the limit, T 安全上限 The temperature resistance limit of the equipment material; simultaneously, the rate of temperature rise is used for judgment. Where dT represents a small change in temperature, and dt represents a small change in time. This is the critical value for the rate of temperature rise. When the rate of temperature rise exceeds the critical value, the system will automatically alarm to indicate that the equipment may be malfunctioning.

[0038] Preferably, the multi-parameter comprehensive judgment is as follows:

[0039] Involving current and voltage power factor thresholds, let the actual power be P. 实际 The actual voltage is U 实际 The actual current is I 实际 Through formula To determine if the power factor is abnormal, cosφ 下限 This is the lower limit of the power factor;

[0040] Involving the current and temperature thermal balance threshold, let the ambient temperature T of the equipment be... 环境 The heat dissipation capacity R of the equipment 热阻 The actual current of the equipment during operation The equipment's running time is t, which is calculated using the formula... Perform temperature and thermal balance prediction when T 实际 >T 预测 +ΔT 安全 When this occurs, an overheat alarm is triggered, where ΔT 安全 This is the preset safe temperature change amount;

[0041] This involves multi-parameter fusion fuzzy logic judgment, using the current deviation membership function μ. I Taking (x) as an example, through

[0042] Used to evaluate current deviation;

[0043] Combining the comprehensive anomaly index formula When the index exceeds the preset threshold, the system will activate an emergency mechanism to balance the false alarm and false negative rates.

[0044] Preferably, the threshold adaptive adjustment is based on a dynamic threshold I derived from historical data. 上限 (t)=μ I (t-1)+3σ I (t-1), where μ I (t-1) represents the average current over the past 24 hours, σ I (t-1) represents the standard deviation of the current over the past 24 hours.

[0045] Preferably, the alarm triggering and grading mechanism includes:

[0046] Level 1 Alarm (Early Warning): When a single parameter approaches the threshold (e.g., current ≥ 90% of the upper limit), a work order is automatically generated to warn operators of potential problems in advance so that preventive measures can be taken in a timely manner;

[0047] Level 2 Alarm (Emergency): Multiple parameters exceed limits or critical parameters (such as temperature) exceed limits, indicating a serious immediate danger or malfunction. The alarm is directly pushed to the mobile phone of maintenance personnel, requiring immediate corrective action to prevent equipment damage or safety accidents.

[0048] Using formula Where R is the comprehensive alarm threshold, w i It is the weight of the i-th parameter, δ i R is the deviation between the actual value of the i-th parameter and its set threshold, where n is the number of parameters involved in the calculation. 触发 It is the preset comprehensive alarm trigger threshold.

[0049] Preferably, to more accurately predict the calorific value of coal, adjustments can be made based on industrial analysis data. The following are the adjustment steps:

[0050] S2.1 Calculate the mass fraction of fixed carbon FC = 100 - AVO, where A is the mass fraction of ash, V is the mass fraction of volatile matter, and O is the mass fraction of oxygen;

[0051] S2.2 Adjusting the carbon content in the coal mass calculation formula: Since fixed carbon mainly contains carbon, the mass fraction of fixed carbon can be regarded as part of the mass fraction of carbon. Therefore, the adjusted carbon mass fraction C adj It can be represented as C adj =C+FC;

[0052] S2.3 Recalculate the calorific value: Use the adjusted carbon mass fraction C adj Substituting this into the coal mass calculation formula, we get:

[0053] Preferably, the optimization algorithm used by the optimization scheduling module includes genetic algorithm, particle swarm optimization algorithm, and simulated annealing algorithm; the multi-objective constraints include pollutant emission concentration limit, treatment cost limit, and equipment operating efficiency limit.

[0054] Preferably, the control execution module includes a PLC controller, an actuator, and a human-machine interface, and further includes a PLC digital input module, a PLC analog input module, a PLC digital output module, a PLC analog output module, and a PLC communication module; the PLC controller collects equipment status signals in real time through the PLC digital input module, obtains environmental parameters through the PLC analog input module, controls equipment start and stop through the PLC digital output module, adjusts equipment operating parameters through the PLC analog output module, and realizes data interaction through the PLC communication module. The PLC controller is also electrically connected to field sensors and actuators to realize data acquisition and command issuance.

[0055] This invention provides a multi-objective constrained flue gas collaborative treatment optimization scheduling system. It has the following beneficial effects:

[0056] (1) By adopting a high-frequency sampling strategy combined with a dynamic calibration formula, this invention effectively avoids spectral aliasing and ensures real-time and accurate capture of pollutant concentrations such as SO2 / NOx, equipment operating parameters, and coal quality indicators; the three types of sensors work together to cover emission monitoring, equipment health diagnosis, and fuel characteristic analysis, forming a closed-loop data chain to provide a complete decision-making basis for process optimization.

[0057] (2) This invention adopts a three-level protection system for equipment status monitoring, and constructs an intelligent diagnosis and early warning system with a single-parameter threshold adaptive, multi-parameter fusion decision-making, and hierarchical alarm mechanism. It introduces a dynamic adjustment mechanism, advanced algorithms such as power factor, thermal balance model and fuzzy logic, and dynamically adjusts the thresholds of key parameters such as current, voltage and temperature according to the historical operating data and real-time operating conditions of the equipment. This enables comprehensive analysis and judgment of multiple parameters, which can more comprehensively reflect the health status of the equipment and improve the accuracy of fault identification.

[0058] (3) This invention uses multi-dimensional data processing technologies such as data cleaning, data fusion, and feature extraction to eliminate noise interference, improve data quality, ensure the accuracy and reliability of data processing, further optimize the algorithm model, improve system response speed and decision efficiency, realize the optimal scheduling scheme under multi-objective constraints, reduce governance costs, improve equipment operating efficiency, and ensure that pollutant emissions meet standards.

[0059] (4) This invention combines the global search capability of the genetic algorithm with the fast convergence characteristic of the particle swarm algorithm to achieve a three-dimensional balance optimization of pollutant emissions (reduction of SO2 / NOx concentration), treatment costs (reduction of annual operating costs) and equipment efficiency (improvement of desulfurization efficiency). Based on the real-time operating condition constraint boundary dynamic adjustment technology, the pollutant emission limit can be automatically tightened with the peak shaving demand of the power grid, the treatment cost constraint can be dynamically relaxed when the coal quality changes, and the equipment efficiency constraint can be linearly coupled with the unit load. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of the system modules of the present invention;

[0061] Figure 2 This document presents a flowchart of the method steps for this invention. Detailed Implementation

[0062] 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the invention, and should not be construed as limiting the invention.

[0064] A preferred embodiment of the multi-objective constrained flue gas collaborative treatment optimization scheduling system provided by the present invention is as follows: Figure 1-2 The system, as shown, is a multi-objective constrained flue gas collaborative treatment and optimization scheduling system, comprising a data acquisition module, a data analysis module, an optimization scheduling module, and a control execution module. The data acquisition module is responsible for real-time acquisition of flue gas emission data, treatment equipment operating status data, and coal quality data from coal-fired power plants. The data analysis module preprocesses and analyzes the acquired data, extracting key characteristic parameters related to flue gas treatment. The optimization scheduling module, based on the extracted key characteristic parameters and preset multi-objective constraints, uses advanced optimization algorithms to perform optimization calculations, obtaining the optimal operating parameters for the treatment equipment and the scheduling strategy. The control execution module, based on the optimization results obtained from the optimization scheduling module, performs real-time control and adjustment of the treatment equipment.

[0065] Data acquisition modules include, but are not limited to, flue gas emission sensors, treatment equipment operation status sensors, and coal quality detectors;

[0066] The specific implementation steps for collecting real-time flue gas emission data from coal-fired power plants using sensors are as follows:

[0067] Step S1: Start all sensors, set the sampling frequency and data transmission protocol, where the sampling frequency is set to f. s The highest frequency of the measured signal is f max The sampling frequency setting formula f is adopted. s ≥2·f max To ensure that the sampling frequency is at least twice the highest frequency of the signal, avoid spectral aliasing. For example, if the SO2 concentration fluctuation frequency is 1Hz, the sampling frequency must be ≥2Hz. As for the data transmission protocol, Modbus TCP or MQTT can be selected. Modbus TCP is suitable for wired transmission and realizes communication between the sensor and the host computer through register mapping. MQTT is suitable for wireless transmission and uses a publish-subscribe model to reduce network load. Choose the protocol according to the transmission distance, bandwidth and real-time requirements.

[0068] Step S2: Collect real-time data on flue gas emissions from coal-fired power plants (such as the concentration of pollutants like SO2, NOx, and dust), operating status data of treatment equipment (such as parameters like operating current, voltage, and temperature), and coal quality data (such as calorific value, ash content, and volatile matter content) through sensors and detectors.

[0069] Specifically, for pollutant concentration calculation (taking SO2 as an example), the sensor output current is set to... The zero-point current is I0, and the sensor sensitivity (μA / ppm) is S. The formula is used... The pollutant concentration is calculated, where K is the calibration coefficient (considering temperature and pressure compensation);

[0070] Among them, for monitoring the operating status data of treatment equipment (taking the desulfurization tower slurry circulation pump as an example), the current, voltage and temperature of the equipment are monitored in real time, and the equipment is judged to be operating normally by combining preset thresholds; including single parameter threshold judgment, multi-parameter comprehensive judgment, threshold adaptive adjustment and alarm triggering and classification mechanism;

[0071] Single parameter threshold judgment:

[0072] Involving a single current threshold, let the real-time current value be I. 实际 Through formula To determine if the current is abnormal, the threshold is dynamically adjusted based on load changes. For example, the current of the desulfurization tower slurry circulation pump is related to the flue gas flow rate. Among them, I 上限 It is the maximum allowable current limit under certain conditions, I 额定 The rated current of the equipment or system, k is the load factor, Q 烟气 Q represents the actual flue gas flow rate during operation. 额定 The rated flue gas flow rate on which the equipment or system was designed;

[0073] Involving a single voltage threshold, let the rated voltage of the device be U. 额定 The actual voltage of the device is U. 实际 Through formula To determine if the voltage is abnormal, ΔU 允许 Allowable voltage fluctuation range;

[0074] Involving a single temperature threshold, let the actual temperature of the equipment be T. 实际 Through formula To determine if the temperature exceeds the limit, T 安全上限 The temperature resistance limit of the equipment material; simultaneously, the rate of temperature rise is used for judgment. Where dT represents a small change in temperature, and dt represents a small change in time. This is the critical value for the rate of temperature rise. When the rate of temperature rise exceeds the critical value, the system will automatically alarm to indicate that the equipment may be faulty.

[0075] Multi-parameter comprehensive judgment:

[0076] Involving current and voltage power factor thresholds, let the actual power be P. 实际 The actual voltage is U 实际 The actual current is I 实际 Through formula To determine if the power factor is abnormal, cosφ 下限 This is the lower limit of the power factor;

[0077] Involving the current and temperature thermal balance threshold, let the ambient temperature T of the equipment be... 环境 The heat dissipation capacity R of the equipment 热阻 The actual current of the equipment during operation The equipment's running time is t, which is calculated using the formula... Perform temperature and thermal balance prediction when T 实际 >T 预测 +ΔT 安全 When this occurs, an overheat alarm is triggered, where ΔT 安全 This is the preset safe temperature change amount;

[0078] This involves multi-parameter fusion fuzzy logic judgment, using the current deviation membership function μ. I Taking (x) as an example, through

[0079] Used to evaluate current deviation;

[0080] Combining the comprehensive anomaly index formula When the index exceeds the preset threshold, the system will activate an emergency mechanism to balance the false alarm and false negative rates.

[0081] Threshold adaptive adjustment based on historical data dynamic threshold I 上限 (t)=μ I (t-1)+3σ I (t-1), where μ I (t-1) represents the average current over the past 24 hours, σ I (t-1) represents the standard deviation of the current over the past 24 hours;

[0082] Alarm triggering and grading mechanisms include:

[0083] Level 1 Alarm (Early Warning): When a single parameter approaches the threshold (e.g., current ≥ 90% of the upper limit), a work order is automatically generated to warn operators of potential problems in advance so that preventive measures can be taken in a timely manner;

[0084] Level 2 Alarm (Emergency): Multiple parameters exceed limits or critical parameters (such as temperature) exceed limits, indicating a serious immediate danger or malfunction. The alarm is directly pushed to the mobile phone of maintenance personnel, requiring immediate corrective action to prevent equipment damage or safety accidents.

[0085] Using formula Where R is the comprehensive alarm threshold, w i It is the weight of the i-th parameter, δ i R is the deviation between the actual value of the i-th parameter and its set threshold, where n is the number of parameters involved in the calculation. 触发 It is the preset comprehensive alarm trigger threshold;

[0086] Specifically, for the calculation of coal mass (taking calorific value as an example), the formula is used.

[0087] Among them, Q net It is the calorific value of coal (unit: kJ / kg or kcal / kg), and C, H, O, and S are the mass fractions of carbon, hydrogen, oxygen, and sulfur in coal, respectively.

[0088] To more accurately predict the calorific value of coal, adjustments can be made by incorporating industrial analysis data. The following are the adjustment steps:

[0089] S2.1 Calculate the mass fraction of fixed carbon FC = 100 - AVO, where A is the mass fraction of ash, V is the mass fraction of volatile matter, and O is the mass fraction of oxygen;

[0090] S2.2 Adjusting the carbon content in the coal mass calculation formula: Since fixed carbon mainly contains carbon, the mass fraction of fixed carbon can be regarded as part of the mass fraction of carbon. Therefore, the adjusted carbon mass fraction C adj It can be represented as C adj =C+FC;

[0091] S2.3 Recalculate the calorific value: Use the adjusted carbon mass fraction C adj Substituting this into the coal mass calculation formula, we get:

[0092] Step S3: Transmit the collected data to the data center or host computer for processing via wired or wireless means. During the data transmission process, encryption technology is used to ensure data security.

[0093] Step S4: To ensure the stability and reliability of data transmission, a CRC checksum formula is used.

[0094] CRC=(M(x)·x 16 mod G(x)

[0095] G(x)=x16 +x 12 +x 5 +1, where M(x) represents the polynomial of the data and G(x) is the standard generator polynomial. Before sending data, the sending end calculates the CRC value of the data according to the above formula and appends it to the end of the data before sending it. After receiving the data, the receiving end recalculates the CRC value of the received data. The receiving end compares the calculated CRC value with the CRC value appended by the sending end. If the two CRC values ​​match, the data transmission is considered to be error-free. If they do not match, it is considered that an error has occurred during the data transmission, triggering the retransmission mechanism.

[0096] To more effectively handle packet loss and latency issues in the network, an adaptive retransmission algorithm, RTO = β·RTT, is adopted. 平滑 This allows for dynamic adjustment of the retransmission interval, where β is an adjustment factor used to adjust the RTO (Retransmission Timeout) and RTT (Retransmission Time To Trace) based on network conditions. 平滑 It is a weighted average of round-trip times, used to estimate the average time it takes for a data packet to travel from the sender to the receiver and back to the sender.

[0097] When the sending end sends a data packet and starts timing, the receiving end sends an acknowledgment (ACK) to the sending end upon receiving the data packet. Upon receiving the ACK, the sending end stops timing and records this RTT (Round-Trip Time). The sending end calculates the RTT based on the multiple recorded RTTs. 平滑 The sending end is based on RTT 平滑 The RTO is calculated using the adjustment factor β; if no ACK is received within the RTO time, the data packet is considered lost, triggering the retransmission mechanism and recalculating the RTO.

[0098] The data analysis module employs data cleaning, data fusion, and feature extraction methods, as detailed below:

[0099] S1 data cleaning steps:

[0100] S1.1 When cleaning data, the Z-score and IQR combined method is used to remove outliers. The Z-score method can capture Gaussian distribution outliers, and the IQR method can handle asymmetric distribution outliers.

[0101] The Z-score method identifies outliers by calculating the deviation of each data point from the mean, expressed in standard deviation. It uses a formula... Where Z is the standard score, X is the original data point, μ is the mean of the dataset, and σ is the standard deviation of the dataset; the threshold is set to |Z|>3, which is considered an outlier and can be adjusted according to the actual situation.

[0102] The IQR method identifies and removes outliers exceeding 1.5 times the IQR range by calculating the interquartile range of the data. The threshold is set at... Adjustments can be made based on the actual situation;

[0103] S1.2 Data cleaning uses a KNN-LSTM hybrid model with null correlation to handle missing values:

[0104] Based on spatial KNN interpolation, the KNN algorithm is used, and the formula is applied. To estimate missing values, where, It is a normalization coefficient to ensure that the sum of the weights is 1. W is the feature similarity weight. For each missing value position i, its similarity to its neighboring positions j (j∈N) is calculated. k The feature similarity between j (i.e., the set of k nearest neighbors belonging to i) and the weight W. ij This reflects the degree to which position j contributes to the estimation of the missing value at position i;

[0105] Based on time interpolation, a recursive prediction is performed using an LSTM model, with the following steps:

[0106] S1.2.1 Prepare time series data: Collect time series data containing missing values, ensuring that the data is arranged in chronological order;

[0107] S1.2.2 Training the LSTM model: Train the LSTM network using complete historical data to enable it to learn long-term dependencies in the time series; the LSTM model can receive data from the previous n time steps (x... t-1 ,x t-2 ,...,x t-n () as input, and predict the value at the next time step;

[0108] S1.2.3 Perform imputation prediction: For each missing time point t, input the previous n known data points into the LSTM model; the model outputs the predicted value. Used to fill in missing values ​​at that time point;

[0109] By combining spatial KNN interpolation with temporal LSTM interpolation, a comprehensive interpolation strategy is formed, using the following formula: in, The value is obtained through KNN spatial interpolation. The value is obtained through LSTM time interpolation. α is a weighting parameter used to control the relative importance of the two interpolation results. Usually, α = 0.6, which is adjusted according to the actual data distribution and missing data to ensure that the interpolation result reflects both spatial characteristics and captures the time trend, thereby improving the overall estimation accuracy. Through the comprehensive interpolation strategy, it not only makes up for the shortcomings of a single interpolation method, but also enhances the integrity and reliability of the data, making it suitable for data repair of complex dynamic systems.

[0110] S2 data fusion steps:

[0111] S2.1 Time alignment is performed between flue gas online monitoring data (high frequency) and DCS control data (low frequency) using the formula. In the formula, φ represents the alignment path, i and n are the indices of sequences X and Y, respectively, and ||x i -y φ(i) || represents the distance between corresponding points, ensuring that the two are synchronized in the time dimension, thus making subsequent causal analysis more effective;

[0112] S2.2 utilizes spatial fusion technology, through formulas In the formula, m(A) represents the confidence level of event A, B∩C=A means that the intersection of two evidence sources B and C is A, m1(B) and m2(C) are the confidence levels of evidence sources B and C for event A, respectively, and K is the conflict coefficient, which is used to measure the degree of conflict between evidence sources. Through this formula, multi-source data can be effectively integrated, information conflict can be reduced, and the confidence level of the fusion result can be improved.

[0113] S2.3 utilizes data correlation analysis techniques, through formulas In the formula, I(X,Y) represents the mutual information entropy of X and Y, p(x,y) is the joint probability distribution of X and Y, and p(x) and p(y) are the marginal probability distributions of X and Y, respectively. This formula can accurately quantify the correlation strength between variables, reveal the intrinsic relationship of data, and further optimize the fusion effect.

[0114] S3 Feature Extraction Steps:

[0115] S3.1 Extracts short-time energy based on temporal features, using the formula... In the formula, N is the number of data points, x i This is the value of the i-th data point. The current energy changes significantly during short-term operation when the equipment starts and stops; this characteristic can be used for operating condition identification. Waveform factors are extracted using the formula... In the formula, FF is the average value of the data, and RMS is the root mean square value. When FF>1.5, it indicates that there is abnormal vibration in the equipment, such as a failure of the induced draft fan bearing.

[0116] S3.2 Power spectral density is extracted based on frequency domain features, using the formula... In the formula, T is the sampling period, f is the frequency, and x is the frequency. n It is the value of the nth sampling point. This formula is used to capture the frequency components during equipment operation and identify energy changes in a specific frequency band.

[0117] S3.3 Extracts wavelet packet energy spectrum based on time-frequency features, using the formula In the formula, d j,k(n) represents the wavelet coefficients of the k-th node in the j-th layer. This formula accurately captures the energy distribution of the equipment in different frequency bands and effectively identifies minor faults under complex working conditions.

[0118] S3.4 Desulfurization efficiency model based on mechanism model-driven feature extraction, through formula In the formula, C in It is the SO2 concentration in the inlet flue gas, C out The SO2 concentration in the flue gas is given by L / G, the liquid-to-gas ratio is given by L / G, pH is the pH value of the slurry, and T is the temperature. The theoretical desulfurization efficiency is calculated by combining the real-time pH value and the liquid-to-gas ratio, and then compared with the measured value to verify and optimize the model.

[0119] By extracting features in the time domain, frequency domain, time-frequency domain, and based on mechanistic models, the operational status and treatment effectiveness of flue gas treatment systems can be comprehensively evaluated. These features not only facilitate real-time monitoring and early warning but also provide data support for system optimization and improvement.

[0120] The optimization algorithms used in the optimization scheduling module include genetic algorithm, particle swarm optimization, and simulated annealing algorithm; the multi-objective constraints include pollutant emission concentration limits, treatment cost limits, and equipment operating efficiency limits.

[0121] The control execution module includes a PLC controller, an actuator, and a human-machine interface. It further includes a PLC digital input module, a PLC analog input module, a PLC digital output module, a PLC analog output module, and a PLC communication module. The PLC controller acquires equipment status signals in real time through the PLC digital input module, obtains environmental parameters through the PLC analog input module, controls equipment start-up and shutdown through the PLC digital output module, adjusts equipment operating parameters through the PLC analog output module, and achieves data interaction through the PLC communication module. The PLC controller is also electrically connected to field sensors and actuators to realize data acquisition and command issuance.

[0122] In summary, this invention, through the collaborative work of the data acquisition module, data analysis module, optimization scheduling module, and control execution module, constructs a highly efficient and intelligent flue gas treatment system, achieving precise monitoring and control of flue gas emissions and significantly improving treatment efficiency.

[0123] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0124] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-objective constrained flue gas collaborative treatment and optimization scheduling system, characterized in that, It includes a data acquisition module, a data analysis module, an optimization scheduling module, and a control execution module. The data acquisition module is responsible for collecting real-time data on flue gas emissions from coal-fired power plants, the operating status of treatment equipment, and coal quality. The data analysis module preprocesses and analyzes the collected data, extracting key characteristic parameters related to flue gas treatment. The optimization scheduling module uses advanced optimization algorithms to perform optimization calculations based on the extracted key characteristic parameters and preset multi-objective constraints, obtaining the optimal operating parameters for the treatment equipment and the scheduling strategy. The control execution module performs real-time control and adjustment of the treatment equipment based on the optimization results obtained from the optimization scheduling module. The data acquisition module includes, but is not limited to, flue gas emission sensors, treatment equipment operation status sensors, and coal quality detectors; The data analysis module employs data cleaning, data fusion, and feature extraction steps, as detailed below: S1 data cleaning steps: The data cleaning described in S1.1 uses a combined Z-score and IQR method to remove outliers. The Z-score method can capture Gaussian distribution outliers, while the IQR method can handle asymmetric distribution outliers. The data cleaning described in S1.2 uses a KNN-LSTM hybrid model with null correlation to handle missing values; S2 data fusion steps: S2.1 Time alignment is performed between flue gas online monitoring data (high frequency) and DCS control data (low frequency) using the formula. In the formula, φ represents the alignment path, i and n are the indices of sequences X and Y, respectively, and ||x i -y φ(i) || represents the distance between corresponding points, ensuring that the two are synchronized in the time dimension, thus making subsequent causal analysis more effective; S2.2 utilizes spatial fusion technology, through formulas In the formula, m(A) represents the degree of trust in event A, B∩C=A means that the intersection of two sources of evidence B and C is A, m1(B) and m2(C) are the degrees of trust in event A by sources of evidence B and C, respectively, and K is the conflict coefficient, which is used to measure the degree of conflict between sources of evidence. S2.3 utilizes data correlation analysis techniques, through formulas In the formula, I(X,Y) represents the mutual information entropy of X and Y, p(x,y) is the joint probability distribution of X and Y, and p(x) and p(y) are the marginal probability distributions of X and Y, respectively. S3 Feature Extraction Steps: S3.1 Extracts short-time energy based on temporal features, using the formula... In the formula, N is the number of data points, x i This is the value of the i-th data point. The current energy changes significantly during short-term operation when the equipment starts and stops; this characteristic can be used for operating condition identification. Waveform factors are extracted using the formula... In the formula, FF is the average value of the data, and RMS is the root mean square value. When FF>1.5, it indicates that there is abnormal vibration in the equipment, such as a failure of the induced draft fan bearing. S3.2 Power spectral density is extracted based on frequency domain features, using the formula... In the formula, T is the sampling period, f is the frequency, and x is the frequency. n It is the value of the nth sampling point; S3.3 Extracts wavelet packet energy spectrum based on time-frequency features, using the formula In the formula, d j,k (n) represents the wavelet coefficients of the k-th node in the j-th layer; S3.4 Desulfurization efficiency model based on mechanism model-driven feature extraction, through formula In the formula, C in It is the SO2 concentration in the inlet flue gas, C out The SO2 concentration in the flue gas is given by L / G, the liquid-to-gas ratio is given by L / G, pH is the pH value of the slurry, and T is the temperature. The theoretical desulfurization efficiency is calculated by combining the real-time pH value and the liquid-to-gas ratio, and then compared with the measured value to verify and optimize the model.

2. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 1, characterized in that: The specific implementation steps for collecting real-time flue gas emission data from coal-fired power plants using sensors are as follows: Step S1: Start all sensors, set the sampling frequency and data transmission protocol, where the sampling frequency is set to f. s The highest frequency of the measured signal is f max The sampling frequency setting formula f is adopted. s ≥2·f max To ensure that the sampling frequency is at least twice the highest frequency of the signal, avoid spectral aliasing. For example, if the SO2 concentration fluctuation frequency is 1Hz, the sampling frequency must be ≥2Hz. As for the data transmission protocol, Modbus TCP or MQTT can be selected. Modbus TCP is suitable for wired transmission and realizes communication between the sensor and the host computer through register mapping. MQTT is suitable for wireless transmission and uses a publish-subscribe model to reduce network load. Choose the protocol according to the transmission distance, bandwidth and real-time requirements. Step S2: Collect real-time data on flue gas emissions from coal-fired power plants (such as the concentration of pollutants like SO2, NOx, and dust), operating status data of treatment equipment (such as parameters like operating current, voltage, and temperature), and coal quality data (such as calorific value, ash content, and volatile matter content) through sensors and detectors. Specifically, for pollutant concentration calculation (taking SO2 as an example), the sensor output current is set to... The zero-point current is I0, and the sensor sensitivity (μA / ppm) is S. The formula is used... The pollutant concentration is calculated, where K is the calibration coefficient (considering temperature and pressure compensation); Among them, for monitoring the operating status data of treatment equipment (taking the desulfurization tower slurry circulation pump as an example), the current, voltage and temperature of the equipment are monitored in real time, and the equipment is judged to be operating normally by combining preset thresholds; it includes single parameter threshold judgment, multi-parameter comprehensive judgment, threshold adaptive adjustment and alarm triggering and hierarchical mechanism; Specifically, for the calculation of coal mass (taking calorific value as an example), the formula is used. Among them, Q net It is the calorific value of coal (unit: kJ / kg or kcal / kg), and C, H, O, and S are the mass fractions of carbon, hydrogen, oxygen, and sulfur in coal, respectively. Step S3: Transmit the collected data to the data center or host computer for processing via wired or wireless means. During the data transmission process, encryption technology is used to ensure data security. Step S4: To ensure the stability and reliability of data transmission, a CRC checksum formula is used. CRC=(M(x)·x 16 )modG(x) G(x)=x 16 +x 12 +x 5 +1, where M(x) represents the polynomial of the data and G(x) is the standard generator polynomial. Before sending data, the sending end calculates the CRC value of the data according to the above formula and appends it to the end of the data before sending it. After receiving the data, the receiving end recalculates the CRC value of the received data. The receiving end compares the calculated CRC value with the CRC value appended by the sending end. If the two CRC values ​​match, the data transmission is considered to be error-free. If they do not match, it is considered that an error has occurred in the data transmission process, triggering the retransmission mechanism.

3. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 2, characterized in that: To more effectively handle packet loss and latency issues in the network, an adaptive retransmission algorithm, RTO = β·RTT, is adopted. 平滑 This allows for dynamic adjustment of the retransmission interval, where β is an adjustment factor used to adjust the RTO (Retransmission Timeout) and RTT (Retransmission Time To Trace) based on network conditions. 平滑 It is a weighted average of round-trip times, used to estimate the average time it takes for a data packet to travel from the sender to the receiver and back to the sender. When the sending end sends a data packet and starts timing, the receiving end sends an acknowledgment (ACK) to the sending end upon receiving the data packet. Upon receiving the ACK, the sending end stops timing and records this RTT (Round-Trip Time). The sending end calculates the RTT based on the multiple recorded RTTs. 平滑 The sending end is based on RTT 平滑 The RTO is calculated using the adjustment factor β; if no ACK is received within the RTO time, the data packet is considered lost, triggering the retransmission mechanism and recalculating the RTO.

4. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 2, characterized in that: The single parameter threshold determination: Involving a single current threshold, let the real-time current value be I. 实际 Through formula To determine if the current is abnormal, the threshold is dynamically adjusted based on load changes. For example, the current of the desulfurization tower slurry circulation pump is related to the flue gas flow rate. Among them, I 上限 It is the maximum allowable current limit under certain conditions, I 额定 The rated current of the equipment or system, k is the load factor, Q 烟气 Q represents the actual flue gas flow rate during operation. 额定 The rated flue gas flow rate on which the equipment or system was designed; Involving a single voltage threshold, let the rated voltage of the device be U. 额定 The actual voltage of the device is U. 实际 Through formula To determine if the voltage is abnormal, ΔU 允许 Allowable voltage fluctuation range; Involving a single temperature threshold, let the actual temperature of the equipment be T. 实际 Through formula To determine if the temperature exceeds the limit, T 安全上限 The temperature resistance limit of the equipment material; simultaneously, the rate of temperature rise is used for judgment. Where dT represents a small change in temperature, and dt represents a small change in time. This is the critical value for the rate of temperature rise. When the rate of temperature rise exceeds the critical value, the system will automatically alarm to indicate that the equipment may be malfunctioning.

5. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 2, characterized in that: The comprehensive judgment based on multiple parameters: Involving current and voltage power factor thresholds, let the actual power be P. 实际 The actual voltage is U 实际 The actual current is I 实际 Through formula To determine if the power factor is abnormal, cosφ 下限 This is the lower limit of the power factor; Involving the current and temperature thermal balance threshold, let the ambient temperature T of the equipment be... 环境 The heat dissipation capacity R of the equipment 热阻 The actual current of the equipment during operation The equipment's running time is t, which is calculated using the formula... Perform temperature and thermal balance prediction when T 实际 >T 预测 +ΔT 安全 When this occurs, an overheat alarm is triggered, where ΔT 安全 This is the preset safe temperature change amount; This involves multi-parameter fusion fuzzy logic judgment, using the current deviation membership function μ. I Taking (x) as an example, through Used to evaluate current deviation; Combining the comprehensive anomaly index formula When the index exceeds the preset threshold, the system will activate an emergency mechanism to balance the false alarm and false negative rates.

6. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 2, characterized in that: The threshold adaptive adjustment is based on a dynamic threshold I using historical data. 上限 (t)=μ I (t-1)+3σ I (t-1), where μ I (t-1) represents the average current over the past 24 hours, σ I (t-1) represents the standard deviation of the current over the past 24 hours.

7. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 2, characterized in that: The alarm triggering and grading mechanism includes: Level 1 Alarm (Early Warning): When a single parameter approaches the threshold (e.g., current ≥ 90% of the upper limit), a work order is automatically generated to warn operators of potential problems in advance so that preventive measures can be taken in a timely manner; Level 2 Alarm (Emergency): Multiple parameters exceed limits or critical parameters (such as temperature) exceed limits, indicating a serious immediate danger or malfunction. The alarm is directly pushed to the mobile phone of maintenance personnel, requiring immediate corrective action to prevent equipment damage or safety accidents. Using formula Where R is the comprehensive alarm threshold, w i It is the weight of the i-th parameter, δ i R is the deviation between the actual value of the i-th parameter and its set threshold, where n is the number of parameters involved in the calculation. 触发 It is the preset comprehensive alarm trigger threshold.

8. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 2, characterized in that: To more accurately predict the calorific value of coal, adjustments can be made by incorporating industrial analysis data. The following are the adjustment steps: S2.1 Calculate the mass fraction of fixed carbon FC = 100 - AVO, where A is the mass fraction of ash, V is the mass fraction of volatile matter, and O is the mass fraction of oxygen; S2.2 Adjusting the carbon content in the coal mass calculation formula: Since fixed carbon mainly contains carbon, the mass fraction of fixed carbon can be regarded as part of the mass fraction of carbon. Therefore, the adjusted carbon mass fraction C adj It can be represented as C adj =C+FC; S2.3 Recalculate the calorific value: Use the adjusted carbon mass fraction C adj Substituting this into the coal mass calculation formula, we get:

9. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 1, characterized in that: The optimization algorithms used in the optimization scheduling module include genetic algorithm, particle swarm optimization, and simulated annealing algorithm; the multi-objective constraints include pollutant emission concentration limits, treatment cost limits, and equipment operating efficiency limits.

10. The multi-objective constrained flue gas collaborative treatment optimization scheduling system according to claim 1, characterized in that: The control execution module includes a PLC controller, an actuator, and a human-machine interface. It further includes a PLC digital input module, a PLC analog input module, a PLC digital output module, a PLC analog output module, and a PLC communication module. The PLC controller acquires equipment status signals in real time through the PLC digital input module, obtains environmental parameters through the PLC analog input module, controls equipment start-up and shutdown through the PLC digital output module, adjusts equipment operating parameters through the PLC analog output module, and achieves data interaction through the PLC communication module. The PLC controller is also electrically connected to field sensors and actuators to realize data acquisition and command issuance.

Images