Combined heat and power generation whole process control system based on intelligent prediction

Abstract

The invention discloses a combined heat and power generation whole-process control system based on intelligent prediction. The system comprises an intelligent sensing module, a control module and a control module, wherein a vibration sensor, a temperature probe and a gas analyzer are mounted on key equipment such as a boiler and a steam turbine; all sensor data are uploaded in real time through a 5G network; the prediction brain module is a load prediction model which is trained by using historical data and can predict the heat / electricity demand in the next 24 hours; an equipment health degree evaluation model (judging equipment loss by analyzing vibration and temperature); the automatic control module is used for automatically adjusting the fuel supply amount, the steam valve opening degree and the power generation power according to the prediction result; under an abnormal condition, the standby equipment is automatically switched and an alarm is given; the system can judge the electricity consumption in an intelligent measurement and calculation mode, so that the coal feeding amount is accurately controlled, and heat and electricity are fully utilized.

Description

Technical Field

[0001] The present invention relates to the technical field of cogeneration, and in particular to a cogeneration full-process control system based on intelligent prediction, which realizes fully automated optimized operation of the cogeneration process through intelligent perception, predictive analysis and automatic control. Background Art

[0002] Combined Heat and Power (CHP) is a highly efficient energy utilization technology that simultaneously produces electricity and usable heat. Through an integrated system, the energy generated by fuel combustion is utilized in a graded manner, significantly improving energy efficiency (up to over 80%, far exceeding the 40%-50% of traditional power generation).

[0003] Core Principles

[0004] Energy cascade utilization: The high-temperature steam generated by fuel combustion first drives the steam turbine to generate electricity (high-quality energy), and then the waste heat after power generation (low-quality energy) is recovered and used for heating (such as industrial steam, district heating) or cooling.

[0005] System integration: Achieve electricity and heat co-production through equipment such as gas turbines, internal combustion engines or fuel cells.

[0006] Main types

[0007] Industrial type: refineries, chemical plants, etc. use waste heat to supply steam required for production.

[0008] Regional type: Centralized heating for residential areas or commercial buildings (such as the district heating system widely used in Northern Europe).

[0009] Micro CHP: Homes or small units use fuel cells or Stirling engines to generate electricity and heat.

[0010] Advantages

[0011] Energy saving and environmental protection: reduce fuel consumption and lower carbon emissions (compared with electric and thermal production, emissions can be reduced by more than 30%).

[0012] Economical: Reduce the overall energy costs of enterprises / communities.

[0013] Flexibility: Can be combined with renewable energy (biomass, biogas) to adapt to a diversified energy structure.

[0014] Application Scenario

[0015] Industrial parks (such as paper mills that require continuous steam)

[0016] Central heating in cold areas

[0017] Emergency power and heat sources (hospitals, data centers)

[0018] Cogeneration is a key technology being promoted, playing a crucial role in clean heating transformation efforts, particularly in northern China. However, its limitation lies in its reliance on a stable heat load, making it suitable for areas with sustained heat demand.

[0019] Combined Heat and Power (CHP) is a highly efficient energy utilization technology that simultaneously generates electricity and usable heat. Traditional CHP system control relies primarily on manual experience or simple feedback control, which presents the following technical issues:

[0020] Response lag: Traditional control systems respond slowly to load changes and are unable to adapt to rapidly changing energy demands;

[0021] Low energy efficiency: Lack of forward-looking control strategies, unable to adjust operating parameters in advance based on predicted demand;

[0022] High equipment loss: Unable to accurately assess equipment health status in real time, and preventive maintenance is not timely;

[0023] Insufficient security: Handling of abnormal situations relies on manual judgment, resulting in slow response.

[0024] Based on the above, we designed a cogeneration process control system based on intelligent prediction, which can determine the power consumption through intelligent measurement, and then accurately control the coal supply to make full use of thermal power, and can also evaluate the health of equipment. Summary of the Invention

[0025] The technical problem to be solved by the present invention is to provide a cogeneration full-process control system based on intelligent prediction, which can judge the electricity consumption through intelligent measurement, and then accurately control the coal supply to make full use of thermal power, and can evaluate the health of equipment.

[0026] In order to solve the above problems, the present invention adopts the following technical solutions:

[0027] A whole-process control system for combined heat and power generation based on intelligent prediction, including:

[0028] 1.1 Intelligent perception module:

[0029] Install vibration sensors, temperature probes, and gas analyzers on key equipment such as boilers and steam turbines;

[0030] All sensor data is uploaded in real time via the 5G network;

[0031] 1.2 Prediction Brain Module:

[0032] Load forecasting model trained with historical data (can predict heat / electricity demand for the next 24 hours);

[0033] Equipment health assessment model (determining equipment loss by analyzing vibration and temperature);

[0034] 1.3 Automatic control module:

[0035] Automatically adjust fuel supply, steam valve opening, and power generation based on prediction results;

[0036] Automatically switch to backup equipment and alarm in abnormal situations:

[0037] The control method of the whole process control system of cogeneration based on intelligent prediction includes the following steps: 2.1 Data acquisition stage:

[0038] Key parameters such as boiler combustion temperature and power generation are collected once every minute;

[0039] Use an infrared camera to capture a pipeline thermal map every half hour;

[0040] 2.2 Prediction and decision-making stage:

[0041] When heat demand is predicted to increase, increase the boiler temperature 30 minutes in advance;

[0042] When the equipment loss score exceeds 80 points (out of 100), the load is automatically reduced to protect the equipment; 2.3 Execution protection phase:

[0043] Control instructions operate valves and motors through explosion-proof actuators;

[0044] In an emergency, the fuel supply can be cut off within 0.5 seconds.

[0045] The control method of the whole process control system of cogeneration based on intelligent prediction, equipment life prediction includes: 3.1 Accumulated loss calculation:

[0046] Record the cumulative working time of the equipment under high temperature / high pressure;

[0047] Calculate the loss percentage based on the material fatigue curve;

[0048] 3.2 Visual detection:

[0049] Using AI to analyze microscope photos of metal pipes;

[0050] When cracks or voids are found with an area greater than 5%, it is marked as high risk.

[0051] Preferably, the AI training process includes:

[0052] 4.1 Data Preparation:

[0053] Collect more than 50,000 photos comparing pipeline health and damage;

[0054] Each photo was manually annotated by multiple experts to identify damaged areas of the pipeline;

[0055] 4.2 Intelligent Scoring:

[0056] AI compares the predicted damage area with the actual marking

[0057] Scoring formula:

[0058] Score = (2 × number of overlapping pixels) / (number of AI-predicted pixels + number of manually annotated pixels);

[0059] The model is automatically retrained when the score is less than 0.7.

[0060] Preferably, the vibration sensor is a high-precision vibration sensor (sampling frequency ≥ 10 kHz).

[0061] Preferably, the gas analyzer is a laser gas analyzer (detection accuracy ±0.5% FS).

[0062] Preferably, the prediction brain module is composed of a feedforward channel (based on dynamic differential flatness theory) and a feedback channel (using H∞ robust control algorithm), and the control period is ≤100ms.

[0063] The beneficial effects of the present invention are:

[0064] Advantage 1: This system uses AI training to learn past electricity consumption data and generate a load forecast model to determine the heat / electricity demand for the next 24 hours. It then adjusts the fuel supply, steam valve opening, and power generation capacity based on the forecast data, so that the heat / electricity supply can meet demand and control excess, avoiding energy waste and insufficient heat / electricity supply.

[0065] The second advantage is that this system uses the equipment health assessment model to determine the health of the entire boiler system. When the health exceeds 80, it will automatically reduce the load to protect the equipment.

[0066] Advantage three: According to the heat / electricity supply needs, before the peak arrives, the boiler temperature can be raised 30 minutes in advance to meet the subsequent peak electricity consumption.

[0067] Advantage 4: Through AI learning and judgment, the pipeline life can be judged and a system warning can be issued when the life is insufficient to avoid leakage accidents.

[0068] Advantage five: It reduces 80% of manual intervention and reduces the workload of operators. DETAILED DESCRIPTION

[0069] All features disclosed in this specification, or all steps in the disclosed methods or processes, except mutually exclusive features and / or steps, can be combined in any manner.

[0070] Any feature disclosed in this specification (including any appended claims and abstract), unless otherwise stated, may be replaced by other equivalent or similar features. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

[0071] Example 1

[0072] A whole-process control system for combined heat and power generation based on intelligent prediction, including:

[0073] 1.1 Intelligent perception module:

[0074] Install vibration sensors, temperature probes, and gas analyzers on key equipment such as boilers and steam turbines;

[0075] All sensor data is uploaded in real time via the 5G network;

[0076] The vibration sensor uses a high-precision MEMS accelerometer, which is installed in the key parts of the equipment bearing and housing, with a sampling frequency of ≥10kHz

[0077] The temperature probe uses a PT100 platinum resistance temperature sensor with a measurement range of 0-600℃ and an accuracy of ±0.5℃.

[0078] The gas analyzer uses a laser spectrometer to monitor the concentrations of gases such as O2, CO, and CO2 in combustion products in real time.

[0079] All sensor data is uploaded to the central processing unit in real time through the 5G industrial Internet of Things gateway, with a transmission delay of <10ms and a data update frequency of ≥1Hz.

[0080] 1.2 Prediction Brain Module:

[0081] Load forecasting model trained with historical data (can predict heat / electricity demand for the next 24 hours);

[0082] Input parameters: historical load data (at least 1 year), weather forecast, weekday / holiday marking, production plan;

[0083] Model architecture: time series prediction model based on LSTM neural network;

[0084] Prediction capability: It can predict heat / electricity demand for the next 24 hours with a resolution of 15 minutes and an average error of <3%;

[0085] Training data: Use at least three years of the company's historical operating data for supervised learning.

[0086] Equipment health assessment model (determining equipment loss by analyzing vibration and temperature);

[0087] Input parameters: vibration spectrum characteristics (FFT analysis), temperature gradient, operating time, maintenance records;

[0088] Analysis method: Combine deep learning and physical models to calculate the remaining useful life (RUL) of equipment;

[0089] Output: Health score (0-100%), prediction of potential failures 7 days in advance;

[0090] Adaptability: Model parameters automatically adjust as the device ages.

[0091] 1.3 Automatic control module:

[0092] Automatically adjust fuel supply, steam valve opening, and power generation based on prediction results;

[0093] Fuel supply regulation: Based on the predicted load, the coal feeder / gas valve is precisely controlled by the PID controller;

[0094] Steam valve control: Electric actuator adjusts the main steam valve opening, with a response time of <2s;

[0095] Power generation regulation: Cooperate with the grid dispatching system to automatically adjust the generator excitation current.

[0096] Calculate the optimal operating point in real time, balance the heat / electricity output ratio, and optimize the start-stop sequence to reduce energy consumption based on dynamic programming algorithms;

[0097] Automatically switch to backup equipment and alarm in abnormal situations:

[0098] Anomaly detection: Establish multi-parameter joint alarm thresholds (such as vibration + temperature + efficiency comprehensive judgment);

[0099] Automatic switching: In case of failure, the backup device is started within 50ms, with seamless switching;

[0100] Alarm classification: trigger different levels of alarms (warning / emergency / shutdown) according to the severity.

[0101] Example 2

[0102] The control method of the whole process control system of cogeneration based on intelligent prediction includes the following steps: 2.1 Data acquisition stage:

[0103] Key parameters such as boiler combustion temperature and power generation are collected once every minute;

[0104] Use an infrared camera to capture a pipeline thermal map every half hour;

[0105] High-frequency monitoring: Core data such as boiler combustion temperature, power generation, steam pressure, etc. are collected once every minute to ensure real-time performance.

[0106] Thermal status analysis: Infrared cameras are used to capture pipeline thermal images every half hour to identify abnormal temperature distribution or local overheating risks.

[0107] Data storage: All data is stored in encrypted form, supporting historical backtracking and trend analysis.

[0108] 2.2 Prediction and decision-making stage:

[0109] Based on machine learning algorithms and heat load prediction models, the system dynamically adjusts its operating strategy:

[0110] When heat demand is predicted to increase, increase the boiler temperature 30 minutes in advance;

[0111] When the equipment loss score exceeds 80 points (out of 100), the load will be automatically reduced to protect the equipment;

[0112] Heat demand preconditioning:

[0113] When an increase in heat demand is predicted in the next 30 minutes (such as a sudden change in weather or an increase in user load), the boiler temperature is automatically increased in advance to avoid response delays.

[0114] Equipment health management:

[0115] Calculate equipment loss scores (based on operating hours, temperature fluctuations, mechanical stress, etc.) in real time. If the score exceeds 80 points (out of 100), immediately reduce the load by 10% to 20% and trigger a maintenance alarm.

[0116] Adaptive optimization: Intelligently allocates the thermal and power output ratio based on peak and valley electricity prices and fuel costs.

[0117] 2.3 Execution protection stage:

[0118] Control instructions operate valves and motors through explosion-proof actuators;

[0119] In an emergency, the fuel supply can be cut off within 0.5 seconds.

[0120] Ensure reliable execution of control instructions through industrial-grade hardware:

[0121] Explosion-proof actuator operation:

[0122] Commands such as valve opening and motor speed are implemented through explosion-proof actuators (ATEX certified), suitable for flammable and explosive environments.

[0123] Emergency protection mechanism:

[0124] When dangerous signals such as boiler overpressure and pipeline leakage are detected, the system cuts off the fuel supply within 0.5 seconds and activates the pressure relief valve.

[0125] The backup power supply ensures that the protection action can still be performed during power outage.

[0126] Example 3

[0127] The control method of the whole process control system of cogeneration based on intelligent prediction, equipment life prediction includes: 3.1 Accumulated loss calculation:

[0128] Record the cumulative working time of the equipment under high temperature / high pressure;

[0129] Calculate the loss percentage based on the material fatigue curve;

[0130] 3.2 Visual detection:

[0131] Using AI to analyze microscope photos of metal pipes;

[0132] When cracks or voids are found with an area greater than 5%, it is marked as high risk.

[0133] Identify potential equipment defects through microscopic image analysis and deep learning:

[0134] Testing process:

[0135] Electron microscope photographs of metal pipes / welds are collected once a week (resolution ≥ 1 μm / pixel).

[0136] The AI model (based on the ResNet50 architecture) automatically analyzes images to detect defects such as cracks, voids, and corrosion.

[0137] Risk grading standards:

[0138]

[0139] System output and linkage control

[0140] Lifespan prediction panel:

[0141] Displays an estimate of remaining life (based on percent loss + defect growth rate) accurate to ±5 days.

[0142] Example: The boiler superheater is currently at 63% loss, and its estimated remaining life is 82 days (2025-07-10).

[0143] Linked with the control system:

[0144] When loss is greater than 70% or the defect risk is high:

[0145] Automatically reduce equipment load to a safe range (such as power generation -15%).

[0146] Lock the opening of the thermostatic valve to avoid sudden stress changes.

[0147] The AI training process includes:

[0148] 4.1 Data Preparation:

[0149] Collect more than 50,000 photos comparing pipeline health and damage;

[0150] Each photo was manually annotated by multiple experts to identify damaged areas of the pipeline;

[0151] 4.2 Intelligent Scoring:

[0152] AI compares the predicted damage area with the actual marking

[0153] Scoring formula:

[0154] Score = (2 × number of overlapping pixels) / (number of AI-predicted pixels + number of manually annotated pixels);

[0155] The model is automatically retrained when the score is less than 0.7.

[0156] The vibration sensor is a high-precision vibration sensor (sampling frequency ≥ 10 kHz).

[0157] A high-precision vibration sensor is a precision device used to detect, measure, and analyze the vibration parameters of an object (such as displacement, velocity, acceleration, frequency, etc.). It is widely used in industrial monitoring, scientific research, aerospace, and other fields. The following is a detailed introduction:

[0158] 1. Core Features

[0159] High sensitivity: Can detect vibration signals at the micron or even nanometer level.

[0160] Wide frequency response range: covering low frequency (such as earthquake monitoring) to high frequency (such as mechanical fault diagnosis).

[0161] Low noise: Optimized signal processing technology reduces environmental interference.

[0162] High resolution: even tiny vibration changes can be accurately captured.

[0163] Strong stability: small performance fluctuation and temperature drift under long-term operation.

[0164] 2. Main types

[0165] Piezoelectric sensors

[0166] Principle: Use the piezoelectric effect of piezoelectric materials (such as quartz and ceramics) to convert mechanical vibrations into electrical signals.

[0167] Application: Suitable for high frequency vibration measurement (such as engine and shock test).

[0168] MEMS (Micro-Electro-Mechanical Systems) Sensors

[0169] Principle: Based on silicon micromachining technology, it has small size, low power consumption and high integration.

[0170] Applications: Consumer electronics (such as mobile phone gyroscopes), industrial equipment status monitoring.

[0171] Fiber optic vibration sensor

[0172] Principle: Detect vibrations through changes in the phase or intensity of optical signals, and resist electromagnetic interference.

[0173] Applications: oil pipeline monitoring, power facility safety.

[0174] Capacitive sensors

[0175] Principle: Vibration is measured by the change in the distance between the capacitor plates with extremely high accuracy.

[0176] Application: Precision instruments (such as semiconductor manufacturing equipment).

[0177] Laser Doppler vibrometer

[0178] Principle: Utilizes the laser Doppler effect, non-contact measurement, suitable for high-frequency micro-vibration.

[0179] Applications: Material properties analysis, aerospace structural testing.

[0180] The gas analyzer is a laser gas analyzer (detection accuracy ±0.5% FS).

[0181] The laser gas analyzer (detection accuracy ±0.5% FS) is a high-precision gas concentration detection device based on laser absorption spectroscopy technology. It can measure the composition and concentration of specific gases in real time and online, with a detection accuracy of ±0.5% of the full scale (FS).

[0182] 1. Core Features

[0183] High precision: Detection accuracy is ±0.5% FS, suitable for scenarios with low concentration or high precision requirements (such as environmental emission monitoring).

[0184] High selectivity: Based on the characteristic absorption spectrum of gas molecules, it has strong anti-cross interference ability.

[0185] Fast response: Real-time monitoring, with response time as low as milliseconds.

[0186] Non-contact measurement: No sample pretreatment is required, avoiding gas contamination or loss.

[0187] Wide measurement range: supports concentration measurements from ppm (parts per million) to percentage (%) levels.

[0188] 2. Working Principle

[0189] Laser Absorption Spectroscopy (TDLAS / TDLAS-OA)

[0190] Tunable Diode Laser Absorption Spectroscopy (TDLAS):

[0191] The laser emits a laser of a specific wavelength (matching the absorption spectrum of the gas being measured) and calculates the gas concentration by measuring the intensity attenuation of the laser after it passes through the gas (Beer-Lambert law).

[0192] Wavelength Modulation Spectroscopy (WMS):

[0193] The laser wavelength is modulated at high frequency to suppress background noise and improve the signal-to-noise ratio.

[0194] Open Optical Path (OA) or In-Situ:

[0195] Open optical path: suitable for monitoring large areas (such as chimney emissions).

[0196] In-situ type: directly inserted into the pipe or container for measurement.

[0197] The prediction brain module consists of a feedforward channel (based on dynamic differential flatness theory) and a feedback channel (using H∞ robust control algorithm), with a control period of ≤100ms.

[0198] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A heat and power cogeneration full process control system based on intelligent prediction, characterized by: include: 1.1 Intelligent perception module: Install vibration sensors, temperature probes, and gas analyzers on key equipment such as boilers and steam turbines; All sensor data is uploaded in real time via the 5G network; 1.2 Prediction Brain Module: Load forecasting model trained with historical data (can predict heat / electricity demand for the next 24 hours); Equipment health assessment model (determining equipment loss by analyzing vibration and temperature); 1.3 Automatic control module: Automatically adjust fuel supply, steam valve opening, and power generation based on prediction results; Automatically switch to backup equipment and alarm in abnormal situations.

2. The control method of the combined heat and power whole process control system based on intelligent prediction according to claim 1 is characterized by: The following steps are involved: 2.1 Data collection stage: Key parameters such as boiler combustion temperature and power generation are collected once every minute; Use an infrared camera to capture a pipeline thermal map every half hour; 2.2 Prediction and decision-making stage: When heat demand is predicted to increase, increase the boiler temperature 30 minutes in advance; When the equipment loss score exceeds 80 points (out of 100), the load will be automatically reduced to protect the equipment; 2.3 Execution protection stage: Control instructions operate valves and motors through explosion-proof actuators; In an emergency, the fuel supply can be cut off within 0.5 seconds.

3. The control method of the combined heat and power whole process control system based on intelligent prediction according to claim 2 is characterized by: Equipment life prediction includes: 3.1 Calculation of cumulative losses: Record the cumulative working time of the equipment under high temperature / high pressure; Calculate the loss percentage based on the material fatigue curve; 3.2 Visual detection: Using AI to analyze microscope photos of metal pipes; When cracks or voids are found with an area greater than 5%, it is marked as high risk.

4. The whole process control system of cogeneration based on intelligent prediction according to claim 3 is characterized by: The AI training process includes: 4.1 Data Preparation: Collect more than 50,000 photos comparing pipeline health and damage; Each photo was manually annotated by multiple experts to identify damaged areas of the pipeline; 4.2 Intelligent Scoring: AI compares the predicted damage area with the actual marking Scoring formula: Score = (2 × number of overlapping pixels) / (number of AI-predicted pixels + number of manually annotated pixels); The model is automatically retrained when the score is less than 0.

7.

5. The whole process control system of cogeneration based on intelligent prediction according to claim 1 is characterized in that: The vibration sensor is a high-precision vibration sensor (sampling frequency ≥ 10 kHz).

6. The whole process control system of cogeneration based on intelligent prediction according to claim 1 is characterized in that: The gas analyzer is a laser gas analyzer (detection accuracy ±0.5% FS).

7. The whole process control system of cogeneration based on intelligent prediction according to claim 1 is characterized in that: The prediction brain module consists of a feedforward channel (based on dynamic differential flatness theory) and a feedback channel (using H∞ robust control algorithm), with a control period of ≤100ms.

Images