Integrated dispatching method for central heating system based on operation data

Abstract

The invention relates to a centralized heating system comprehensive scheduling method based on operation data, and aims to solve the technical problems of excessive heating and uneven cooling and heating of a current heating system. The technical scheme is that the method comprises the steps of data acquisition and transmission, construction of a dynamic thermal index model based on multi-dimensional data, and guidance of accurate distribution of heat; dynamic regulation and control of heat are realized through linkage of an algorithm and equipment; the actual energy consumption and the index energy consumption are compared, the user complaint rate and the room temperature standard reaching rate are combined, a heating station grading assessment result is generated, and the system stability and the prediction precision are continuously improved. Whole-network comprehensive energy consumption indexes with the heating stations as units are established, energy consumption indexes of heat, coal, gas, water, electricity and the like of the whole network are formulated according to the overall comprehensive energy consumption level of a company, the energy consumption indexes of the heating stations are refined through the average room temperature, and a scheduling process and evaluation system according to the indexes of the heating stations based on the intelligent heat supply platform is established.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban centralized heating, and in particular relates to a comprehensive scheduling method for a centralized heating system based on operation data. Background Art

[0002] In recent years, with the annual increase in urban heating area, the energy demand of centralized heating systems has increased, but the overall heat source heating capacity is tight, and the heat source fluctuation phenomenon has become increasingly frequent due to the influence of electricity spot trading and environmental protection. At the same time, a smart heating platform and data middle platform have been built, and a room temperature monitoring platform has been established to realize the unified dispatching and operation management of "source, network, station, and household". After the regional reform of heating, the correlation between the heating network and branches has weakened. The original comprehensive energy management and assessment system based on the heating network can no longer adapt to the existing management model. The phenomena of excessive heating and uneven heating and cooling are more obvious, and the level of refined management of the heating network needs to be further improved.

[0003] Patent CN113251451A uses a PID algorithm to adjust the heating power in real time, eliminating steady-state errors, suppressing overshoot, and predicting dynamic changes through proportional (P), integral (I), and differential (D) terms. However, its PID parameters rely on empirical tuning, making it difficult to adapt to multi-variable coupling scenarios (such as sudden weather changes). It has significant lag and cannot predict load fluctuations in advance. Patent CN112556028A combines meteorological change rate characteristics (such as temperature and humidity change rates) with machine learning models (such as TCN-GRU) to predict future load trends, but it lacks a real-time control mechanism, and the prediction results need to rely on the response of external actuators.

[0004] Under the existing conditions, initiating load forecasting from user terminals and realizing a comprehensive energy management and evaluation system based on thermal power stations is of great significance to ensuring the economic operation of the heating system. It is also a key link in the development of centralized heating from "standard heating" to "comfortable heating" and finally to "heating on demand". Summary of the Invention

[0005] The purpose of the present invention is to solve the above problems and provide a comprehensive scheduling method for a centralized heating system based on operating data.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A method for comprehensive scheduling of a central heating system based on operating data comprises the following steps:

[0008] Step 1) Data Collection and Transmission: Real-time data collection of user room temperature, heat, water, and electricity consumption is performed. Heat meters installed on the return pipes of the heating station zones or the primary network return main monitor heat, water and electricity meters equipped with optical character recognition (OCR) dials monitor water and electricity consumption, and room temperature is monitored using user-side room temperature collectors. Abnormal data is automatically flagged and repairs are triggered. The data is uploaded to the smart heating platform in real time via a dedicated data line and verified.

[0009] Step 2) Dynamic thermal index model establishment:

[0010] Step 2.1) Calculate energy consumption indicators using a phased approach: Divide the heating season into the early cold period, the severe cold period, and the late cold period;

[0011] Eliminate abnormal data: thermal index >40W / m² or <0 in the early cold period, >70W / m² or <20W / m² in the severe cold period, >40W / m² or <0 in the late cold period;

[0012] Step 2.2) Based on the outdoor temperature, the heat load ratio interpolation method is used to complete the thermal index value of the discontinuous temperature interval. The heat load ratio calculation formula is:

[0013]

[0014] Among them, t n is the indoor design temperature, t p is the average outdoor temperature, t wn Design temperature for outdoor use;

[0015] Step 2.3) Calculate the average thermal index of each heating station subsystem at different times according to the following formula:

[0016] Heat load ratio × maximum heat index = actual heat index;

[0017] Establish the index model for each temperature point within the range of -12~12℃;

[0018] Step 2.4) Based on the nonlinear relationship between the shutdown rate and heat consumption per unit, energy consumption is predicted using a third-order regression model:

[0019] Heat consumption per unit = C1 + C2 × stop rate − C3 × stop rate^2 + C4 × stop rate^3

[0020] Among them, C1-C4 are coefficients to be fitted, which are fitted through a high-order regression model. The suspension rate = the total suspension area of ​​the stall heating station / the total contract area of ​​the stall heating station. The unit heat consumption = the total heat consumption of the stall heating station / the total operating area of ​​the stall heating station;

[0021] Step 2.5) Based on the real-time room temperature data, the thermal index is corrected in real time using the correction formula:

[0022]

[0023] Where: q is the corrected thermal index, unit is W / ㎡; T i is the actual room temperature, unit is ℃; T j The unit of outdoor heating temperature is ℃, which is -9.7℃; Q d is the thermal index before correction, unit is W / ㎡, and the suspension rate is the proportion of users' suspension reports;

[0024] Step 3) Dynamic control of thermal indicators:

[0025] PID algorithm control: PID control uses the stop rate prediction value from step 2.4) as the feedforward input. First, adjust the proportional coefficient KP to suppress valve oscillation (KP value is 0.08-0.5), then introduce the integral coefficient KI to eliminate the residual error, and finally optimize the dynamic performance through the differential coefficient KD.

[0026] Actuator control logic: When increasing the load, the valve is opened first and the pump is adjusted; when decreasing the load, the pump is lowered first and the valve is closed; the electric regulating valve has an operating cycle of 7 minutes / time, the return water booster pump has an operating cycle of 20 minutes / time, the frequency increase amplitude is 0.6Hz / time, and the frequency reduction amplitude is 0.3Hz / time;

[0027] Differentiated grouping strategy: Public heating stations with no nighttime heat demand will have their nighttime heat index temperature increased by 5%-10%. Terminal heating stations in the heating network will have their temperature reduced by 3%-8% based on the heat source index. Old residential communities will maintain their all-day temperature control and reduce it at night.

[0028] Mode switching: The thermal index control mode is adopted in the severe cold period, and the temperature-index mixed control mode is adopted in the early cold period and the late cold period;

[0029] Step 4) Energy consumption assessment and evaluation:

[0030] Taking the heating station as the unit, statistics are collected on heating type, construction year, enclosure structure and secondary pipe network attributes;

[0031] Compare actual energy consumption with the dynamic indicator energy consumption obtained in step 2). Combined with the user complaint rate and room temperature compliance rate, a room temperature between 18°C ​​and 24°C is considered to meet the standard. Room temperatures below 18°C ​​and above 24°C are considered to have poor control effects. The control effects are differentiated by the number of effective complaints per user.

[0032] Energy consumption score calculation: The absolute value of energy consumption ranking of thermal power stations of the same type accounts for X%, and the actual energy consumption and target energy consumption savings and excess scores account for Y%, and the energy consumption year-on-year ratio score accounts for Z%. Where X+Y+Z=100% and the weights are adjusted dynamically, the energy consumption score is obtained, and the performance salary is determined based on the energy consumption score;

[0033] Step 5) System optimization and iteration:

[0034] Heat source fault handling: When the heat source is under-supplied / over-supplied, the nearest heating station will prioritize adjusting the valve opening to maintain the heat network balance and automatically mark abnormal data for interpolation and completion;

[0035] Equipment fault handling: Establish records for subsystems that deviate from indicators, use PID algorithm for stable control, and solve valve oscillation problems through KP, KI, and KD parameter optimization;

[0036] Control failure handling: When PID control times out and fails to meet the standard, it switches to manual control;

[0037] Model iteration: After the end of each heating season, the third-order regression model coefficients are refitted based on the historical data of the past three years;

[0038] Energy consumption optimization: K-means clustering is used to identify high-energy-consuming thermal power stations for priority transformation.

[0039] Furthermore, in step 1), the heat meter is installed on the return pipe of each zone of the thermal power station or the return main of the primary network to monitor the heat; the water meter / electricity meter is equipped with an OCR dial to realize remote data transmission for monitoring the water and electricity volume, and the image recognition error rate of the OCR dial is ≤0.5%.

[0040] Furthermore, the coefficients of the third-order regression model in step 2.4) are fitted by the least squares method, specifically including:

[0041] Energy saving - radiator type: C1=0.3032, C2=0.4557, C3=1.205, C4=2.227;

[0042] Energy saving - floor heating type: C1=0.2670, C2=0.9038, C3=2.695, C4=3.660;

[0043] Non-energy-saving-radiator type: C1=0.3342, C2=0.4492, C3=1.570, C4=3.394.

[0044] Furthermore, the KP value optimization in step 3) reduces the fluctuation range of the thermal index by 60%. When the heat source fluctuates, the thermal index weight coefficient of the terminal thermal power station is modified through manual intervention.

[0045] Furthermore, in step 4), the weight of the heating station with great energy-saving potential is increased by Y%; the weight of the heating station with high energy consumption benchmark level is increased by Z%.

[0046] Furthermore, the valve adjustment priority for heat source fault handling in step 5) is: when the heat source is undersupplied, the proximal heat station prioritizes opening the valve, requiring manual intervention to exit the control mode; when the heat source is oversupplied, the proximal heat station prioritizes reducing the valve opening.

[0047] Furthermore, the method enables the centralized heating system to achieve: control response time ≤ 20 minutes; overall network energy consumption decreased by ≥ 4.27%; and room temperature compliance rate increased to 99%.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. The present invention uses historical data to infer the actual hourly thermal index of each heating station, and conducts personalized regulation and control for each area and each heating station based on the thermal index;

[0050] 2. This invention solves the problems of extensive energy consumption management, low control accuracy, and imperfect assessment system in the existing technology by formulating dynamic energy consumption indicators, intelligent control algorithms, real-time data collection, and a multi-dimensional assessment system, thus achieving a balance between the economy and comfort of the heating system.

[0051] 3. This invention establishes a comprehensive energy consumption index for the entire network with the heating station as the unit. Based on the company's overall comprehensive energy consumption level, it formulates energy consumption indicators for heat, coal, gas, water, electricity, etc. for the entire network. It also uses the average room temperature to refine the energy consumption indicators of the heating station, and establishes a scheduling process and evaluation system based on the heating station indicators based on the smart heating platform;

[0052] 4. The present invention takes into account the characteristics of different users when establishing thermal indicators, fully reflecting the characteristics of each thermal power station;

[0053] 5. The present invention decomposes personalized scheduling into multiple layers to achieve clear responsibility decomposition and coherent connection between the front and the back, which can significantly improve the execution efficiency of personalized regulation and significantly reduce the difficulty of execution. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is the outdoor temperature curve for the 2023-2024 heating season;

[0055] Figure 2 This is the fitted line diagram of the energy-saving radiator type heating station of the present invention;

[0056] Figure 3 This is the main effect diagram of the energy-saving radiator type heat consumption of the present invention;

[0057] Figure 4 A model Pareto chart is obtained for the energy-saving-radiator type fitting model of the present invention;

[0058] Figure 5 The residual diagram of heat consumption per unit after fitting the model of the energy-saving radiator type heating station of the present invention;

[0059] Among them, (a) is the normal probability plot; (b) is the scatter plot with the fitted value; (c) is the histogram; (d) is the curve plot with the order;

[0060] Figure 6 This is the 24H heat consumption curve of the present invention when the KP value is 0.5;

[0061] Figure 7 This is the 24H heat consumption curve of the present invention when the KP value is 0.08;

[0062] Figure 8 This is a diagram of the thermal index grading logic architecture of the present invention;

[0063] Figure 9 is a normal distribution diagram of the maximum thermal index of the floor heating / radiator system of the present invention;

[0064] Figure 10 This is a curve diagram of the control effect of the thermal power station of the present invention. DETAILED DESCRIPTION

[0065] The present invention will be further described below with reference to the accompanying drawings and examples.

[0066] Example

[0067] like Figure 1-10 As shown, this embodiment uses the thermal power stations of seven branches of Xiaodian Company 1, Xiaodian Company 2, Yingze Company, Jiancaoping, Wanbailin, Xinghualing, and Jinyuan Company in Taiyuan City, Shanxi Province during the 2023-2024 heating season as samples for analysis.

[0068] A method for comprehensive scheduling of a central heating system based on operating data comprises the following steps:

[0069] Step 1) Data Collection and Transmission: Real-time data collection of user room temperature, heat, water, and electricity consumption is performed. Heat meters installed on the return pipes of the heating station zones or the primary network return main monitor heat, water and electricity meters equipped with optical character recognition (OCR) dials monitor water and electricity consumption, and room temperature is monitored using user-side room temperature collectors. Abnormal data is automatically flagged and repairs are triggered. Data is uploaded to the smart heating platform in real time via 4G / 5G or fiber optic network data lines, and the data is verified.

[0070] Among them, the remote transmission meter is installed in a pipeline location with sufficient space and no obstruction. For the thermal power station with absorption heat exchanger unit, the heat meter is installed in the return pipe of each partition of the thermal power station or the return main of the primary network to monitor the heat; the water meter / electricity meter is equipped with an OCR dial to realize remote data transmission for monitoring water and electricity volume, and the image recognition error rate of the OCR dial is ≤0.5%.

[0071] Deploy 200,000 room temperature collectors, uploading temperature data every 30 minutes (covering residential / public buildings, radiators / floor heating systems);

[0072] This embodiment installs / replaces 2466 pieces of heat meter, 1534 pieces of remote water meter, 369 pieces of remote electricity meter, and 1556 pieces of OCR meter.

[0073] Mark 483 pieces of data anomaly heat meter (value is 0 or mutation), using adjacent data interpolation repair.

[0074] Step 2) Dynamic heat index model establishment:

[0075] Step 2.1) Adopting a phased approach to energy consumption index calculation: the heating season is divided into early cold period (November), severe cold period (December-February) and late cold period (March);

[0076] The basic data is removed according to the time and space dimension, and the sub-system index less than 0 and greater than 40W / ㎡ in the early cold period is removed as abnormal data, the heat index less than 20W / ㎡ and greater than 70W / ㎡ in the severe cold period is removed as abnormal data, the sub-system index less than 0 and greater than 40W / ㎡ in the late cold period is removed as abnormal data, and the individual index mutation in the whole heating season is corrected and used, and the abnormal data in the whole heating season is removed, which is used as the basis for heat meter debugging.

[0077] The heat index value of the discontinuous temperature interval is completed by using the heat load ratio interpolation method;

[0078] Step 2.2) Because of the large number of heat stations and the large system, it is difficult to accurately calculate each system, so the heat index model is established in the early stage, and the heat load ratio method is used to calculate the heat index under each outdoor temperature. The heat load ratio is one of the commonly used calculation methods for predicting heat load, predicting heat index, and calculating energy consumption in winter heating. Taking the calculation of heat index as an example, its principle is that when the outdoor temperature reaches the indoor calculation temperature, the heat index is 0, when the outdoor temperature reaches the design outdoor temperature, the heat index is maximum, and when the outdoor temperature is between the design outdoor temperature and the indoor calculation temperature, the heat index changes linearly. The method is simple and easy to understand, and convenient for calculation;

[0079] Based on outdoor temperature, the heat index value of the discontinuous temperature interval is completed by using the heat load ratio interpolation method, and the heat load ratio calculation formula is:

[0080]

[0081] Among them, tn is the indoor design temperature, tp is the outdoor average temperature, and twn is the outdoor design temperature. Taking 4208 Hengguangyuan floor heating low sub-system as an example, tn=18℃, twn=-11℃, tp=-4.36℃;

[0082] ;

[0083] Step 2.3) Calculate the average heat index of each heat substation system at different times. The substation refers to the independent control of the substation partition circuit or the secondary pipe network subsystem. The formula is as follows:

[0084] Heat load ratio x maximum heat index = actual heat index (measured average of 37.6 W / ㎡ in severe cold period);

[0085] The maximum heat index is calculated as 37.6 / 0.768 = 48.96

[0086] According to the linear interpolation of heat load ratio, the index curve of -12~12℃ is established, as shown in Figure 1 ;

[0087] Step 2.4) Based on the historical operation data of the six districts of Taiyuan City (Xiaodian District, Yingze District, Jiancaoping District, Wanbailin District, Xinghualing District, and Jinyuan District) from 2019 to 2022, according to the nonlinear relationship between the report stop rate and heat consumption, the energy consumption is predicted by a third-order regression model (taking the energy-saving radiator type heat station as an example):

[0088] Energy-saving radiator type: heat consumption = 0.3032 + 0.4557*report stop rate -1.205*report stop rate²+2.227*report stop rate³;

[0089] Table 1 Energy-saving radiator type heat station grading table

[0090]

[0091] Regression analysis is performed on the grading data, and a single-factor high-order (3rd order) interaction model is selected for fitting. The high-order regression model is shown below, and the fitting line graph is shown in Figure 2 . The model has a very high fitting degree with the data. The R-sq value is high, indicating that the model has good prediction performance. The main effect graph of heat consumption is shown in Figure 3 . The fitted model shows the model Pareto graph as shown in Figure 4 , which verifies the correctness of the variance analysis result.

[0092] As shown in Figure 2-4 , there is a significant positive correlation between the average heat index and the report stop rate. The increase in the report stop rate is directly related to the improvement of the heat index. By timely adjusting relevant strategies or measures, we can effectively control the change of heat consumption, thereby ensuring its stability and reliability. The residual error of heat consumption after fitting the model is shown in Figure 5 . The residual error fluctuates within the range of -0.01 to 0.01, indicating that the deviation between the data points and the fitting line is very small, indicating that the model fitting effect is quite good.

[0093] Based on the above results analysis, the energy-saving-heating coil type heat station sub-model between the regression analysis of the stop rate and the heat consumption has good fitting effect, which can provide a theoretical basis for subsequent analysis. Based on the model, the heat station can be preliminarily predicted in the second year of operation combined with the stop rate and the given heat consumption (newly operated heat stations are not allowed to stop), which can provide a basic value for energy saving and consumption reduction of heat stations.

[0094] Energy-saving-floor heating type: heat consumption = 0.2670 + 0.9038*stop rate - 2.695*stop rate²+3.660*stop rate³;

[0095] Non-energy-saving-heating coil type: heat consumption = 0.3342 + 0.4492*stop rate - 1.570*stop rate²+3.394*stop rate³;

[0096] The coefficients of the third-order regression model are fitted by the least square method, which specifically includes:

[0097] Energy-saving-heating coil type: C1=0.3032, C2=0.4557, C3=-1.205, C4=2.227;

[0098] Energy-saving-floor heating type: C1=0.2670, C2=0.9038, C3=-2.695, C4=3.660;

[0099] Non-energy-saving-heating coil type: C1=0.3342, C2=0.4492, C3=-1.570, C4=3.394.

[0100] Step 2.5) Based on real-time room temperature data, the heat index is corrected in real time by the correction formula, and the correction formula is:

[0101]

[0102] In the formula: q is the corrected heat index, unit W / ㎡; T i is the actual room temperature; T j is the outdoor reference temperature for heating, unit ℃, considering -9.7℃; Q d is the heat index before correction, unit W / ㎡, the stop rate is the proportion of user stop;

[0103] According to the above formula, the heat index is corrected, the room temperature is corrected to 20℃, and it can be found that each heat station is reduced by about 0.12 W / ㎡ on the original index, and the average index is reduced by about 0.38%. The actual heating area of 278 heat stations is about 30 million square meters. Combined with the adjusted data of the heat index, the instantaneous load can be reduced by about 3.6 MW, and the heat can be saved by about 47,000 GJ in the whole heating season.

[0104] Step 3) Dynamic regulation of heat index:

[0105] PID algorithm control: PID control uses the stop rate prediction value from step 2.4) as the feedforward input. First, adjust the proportional coefficient KP to suppress valve oscillation (KP value is 0.08-0.5). Optimize the KP value to reduce the fluctuation of the thermal index by 60%. Then introduce the integral coefficient KI to eliminate the residual error. Finally, use the differential coefficient KD to optimize dynamic performance.

[0106] The valve opening and pump frequency are adjusted through the PID algorithm, and differentiated control is achieved by combining the thermal station grouping strategy. The specific contents are as follows:

[0107] Using a trial-and-error method, the PID controller's three parameters, KP, KI, and KD (proportional, integral, and differential), are freely set through the front end to achieve subsystem stability curve control. The proportional coefficient, KP, is first adjusted to suppress valve oscillation, followed by the integral coefficient, KI, to eliminate residual errors. Finally, the differential coefficient, KD, is used to optimize dynamic performance. Generally, if the proportional coefficient, KP, is adjusted first and stable control is not achieved, the integral KI and differential KD controls are then introduced. The valve operation cycle is set to 7 minutes per cycle, and the distributed variable frequency pump frequency is increased by 0.6 Hz per cycle and decreased by 0.3 Hz per cycle. Valve oscillation is resolved by adjusting the KP value, which is set within a range of 0.08-0.5. The response characteristics of each parameter are shown in Table 2 below.

[0108] Table 2 PID parameter adjustment response characteristics

[0109]

[0110] The thermal index adjustment interface is adjusted by adjusting the KP / KI / KD values. If the adjustment range is too large, the KP value can be reduced and the KI value can be increased. If the range is too small, the opposite is true. Figure 6 As shown in the figure, when the KP value is 0.5, the valve action amplitude is large, and the thermal index is prone to fluctuate, such as Figure 7 As shown in the figure, when KP is adjusted to 0.08, the valve adjustment range becomes smaller, and it is easier to adjust to the planned thermal index value. The thermal index fluctuation range is reduced by 60% through KP value optimization.

[0111] The adjustment cycle affects the time it takes for the thermal indicators to be adjusted. If the adjustment cycle is too long, the subsystem will be in an oversupply or undersupply state for a long time. If the adjustment cycle is too short, the secondary network will not show the adjustment effect after the first adjustment, and the second adjustment has already started, resulting in over-adjustment. Based on experience, the valve adjustment cycle is finally determined to be 7 minutes and the return water pressure pump adjustment cycle is 20 minutes.

[0112] The adjustment range affects the effect of thermal index adjustment. If the adjustment range is too large, it will not be adjusted in place. If the adjustment range is too small, the adjustment time will be too long. The valve adjustment range can be controlled by adjusting the PID value.

[0113] Since the regulating effect of the return water booster pump is greater than that of the valve, the frequency of increase and decrease should not be too large. After many tests and optimizations, the regulating amplitude is 0.6Hz for increasing the frequency and 0.3Hz for decreasing the frequency. When increasing the load, open the valve first and then adjust the pump. When decreasing the load, lower the pump first and then close the valve.

[0114] Each system can be controlled according to thermal indicators, with quick response and good results.

[0115] The adjustment cycle affects the time it takes for the thermal indicators to be adjusted. If the adjustment cycle is too long, the subsystem will be in an oversupply or undersupply state for a long time. If the adjustment cycle is too short, the secondary network will not show the adjustment effect after the first adjustment, and the second adjustment has already started, resulting in over-adjustment. Based on experience, the valve adjustment cycle is finally determined to be 7 minutes and the return water pressure pump adjustment cycle is 20 minutes.

[0116] The adjustment range affects the effect of thermal index adjustment. If the adjustment range is too large, it will not be adjusted in place. If the adjustment range is too small, the adjustment time will be too long. The valve adjustment range can be controlled by adjusting the PID value.

[0117] Actuator control logic: When increasing the load, the valve is opened first and the pump is adjusted; when decreasing the load, the pump is lowered first and the valve is closed; the electric regulating valve has an operating cycle of 7 minutes / time, the return water booster pump has an operating cycle of 20 minutes / time, the frequency increase amplitude is 0.6Hz / time, and the frequency reduction amplitude is 0.3Hz / time;

[0118] Since the regulation effect of the return water booster pump is greater than that of the valve, the frequency amplitude of the increase and decrease should not be too large. After many tests and optimizations, the regulation amplitude is 0.6Hz for increasing the frequency and 0.3Hz for decreasing the frequency. When increasing the load, open the valve first and then adjust the pump. When decreasing the load, reduce the pump first and then close the valve. Under the action of the inertia of the heating network, suddenly closing the valve will cause a surge in return water pressure (water hammer effect). First reduce the pump frequency to gradually reduce the flow rate, and then close the valve in steps to avoid pipeline stress shock (the measured pressure fluctuation is from 0.8MPa to 0.3MPa).

[0119] Each system can be controlled according to thermal indicators, with quick response and good results.

[0120] Differentiated grouping strategy: Public heating stations with no nighttime heat demand have their nighttime heat index temperature increased by 8% to offset heat delivery delays during the pump-down process. Terminal heating stations in the heating network have their temperature reduced by 5% based on the heat source index. Older residential communities maintain a full-day temperature control and a 3% nighttime temperature reduction to prevent oversupply caused by residual heat in the pipe network. By simultaneously implementing both the timing and grouping strategies, the temperature fluctuation range at the terminal end of the heating network has narrowed from ±4°C to ±1.2°C.

[0121] The heat meter of the heat station of the absorption heat exchange unit is installed on the return water main to distinguish the heat of each zone when the unit is started or stopped;

[0122] Mode switching: The thermal index control mode is adopted in the severe cold period, and the temperature-index mixed control mode is adopted in the early cold period and the late cold period;

[0123] When the heat source fluctuates, the operating conditions of the terminal thermal power station are guaranteed by manually modifying the thermal index weight. For example, when the heat source output drops by 15%, the thermal index weight coefficient of the terminal thermal power station is increased by 20%. The optimized control logic is as follows: the electric regulating valve action time is 7 minutes / time, the return water pressure pump action time is 20 minutes / time, the frequency of increase is 0.6Hz / time, and the frequency of decrease is 0.3Hz / time. When increasing the load, the valve is opened first, and then the return water pressure pump frequency is adjusted. When decreasing the load, the return water pressure pump frequency is reduced first, and then the valve is closed.

[0124] Step 4) Energy consumption assessment and evaluation:

[0125] Comprehensively analyze and understand the basic properties of the heating station subsystem. Taking the heating station as a unit, complete the statistics and confirmation of the heating type, construction year, enclosure structure, and secondary pipe network properties (Table 3). At the same time, it is required to establish a statistical table of thermal indicators of the heating station subsystem and a statistical table of energy consumption of the heating station subsystem by period (week, month, heating season). Compare the actual energy consumption with the dynamic indicator energy consumption obtained in step 2). Combined with the user complaint rate and the room temperature compliance rate, a room temperature of 18-24℃ is considered to meet the standard. Room temperatures below 18℃ and above 24℃ are considered to have poor control effects. The control effect is distinguished by the number of effective complaints per unit user. The grading logic is shown in the figure. Figure 8 shown.

[0126] Table 3 Classification of thermal indicators of heating stations

[0127]

[0128] According to statistics, the average value of the radiator system at -9.7℃ is 40.9W / ㎡, and the average value of the floor heating system is 37.6W / ㎡. The normal distribution diagram of the maximum thermal index is shown in Figure 9 The average temperature of this heating season is -0.07. After conversion, the average value of the radiator system is 25.5W / ㎡, and the average value of the floor heating system is 23.4W / ㎡. Equivalent to the entire heating season of 151 days, the energy consumption of the radiator system is 0.333GJ / ㎡•a, and the energy consumption of the floor heating system is 0.305GJ / ㎡•a.

[0129] Comparing the energy consumption after converting the outdoor temperature during the same period, the comprehensive energy consumption of the group company's thermal power station from February 26 to March 25 in the 2023-2024 heating season was 1.64kgce / ㎡ (corresponding to an average outdoor temperature of 5.43℃), and the comprehensive energy consumption of the thermal power station during the same period in the 2022-2023 heating season was 1.36kgce / ㎡ (corresponding to an average outdoor temperature of 8℃). After converting the meteorological parameters according to the heat load ratio to this heating season for comparison and calculation, the comprehensive energy consumption of the thermal power station decreased by 4.1% compared with the previous heating season.

[0130] Energy consumption score calculation: Establish an energy consumption evaluation system for heat stations as a unit, compare actual energy consumption with index energy consumption, and use differentiated energy consumption cost evaluation methods to ultimately realize online evaluation system; The absolute value ranking score of energy consumption of heat stations of the same type accounts for X%, the energy consumption and index energy consumption of the actual energy consumption accounts for Y%, and the energy consumption same ratio accounts for Z%, wherein X+Y+Z=100% and the weight is dynamically adjusted, the heat station with great energy saving potential increases Y% weight; The heat station with high energy consumption benchmark increases Z% weight, and the energy consumption score is obtained, and the performance salary is determined according to the energy consumption score;

[0131] Energy consumption score real-time feedback to PID control module: When the energy consumption score of the heat station is lower than the threshold (such as 60 points), it is determined that the station control response is lagging or the residual error is accumulated seriously, and the system automatically triggers the PID parameter re-determination mechanism:

[0132] Increase the integral coefficient KI: For the heat station with low Y% score (actual energy consumption and index energy consumption), the residual error elimination ability is insufficient, the KI value is increased by 20%-50% (1.2-1.5 times of the original benchmark value), and the accumulated deviation of room temperature and target value is eliminated;

[0133] Reduce the proportional coefficient KP: For the heat station with X% score continuously at the bottom, it means that the valve oscillation is serious, and the KP value is reduced to 0.08-0.15 interval to suppress overshoot;

[0134] Dynamic adjustment period: For the station with Z% score deterioration, shorten the actuator action period (valve from 7 minutes / time to 5 minutes / time), and improve the response frequency.

[0135] In Taiyuan pilot, the room temperature compliance rate of 21 heat stations with score <60 is improved from 83% to 97% after PID parameter adaptive adjustment, and the control response time is shortened to 12 minutes.

[0136] After the control system ends in a dispatching period (such as 7 days, a month, etc.), the target heat consumption Q x of each heat exchange control system is calculated according to the actual outdoor temperature history value and the heating area heat index of each heat exchange system of the branch company xf , the total target heat consumption Q s of the branch company is calculated, and the actual heat consumption Q sf of each heat exchange system is calculated according to the heat table, and the total actual heat consumption Q xf of the branch company is calculated.

[0137] (1) The total station examines the branch company

[0138] The total station compares the total target heat consumption Q xf of the branch company with the total actual heat consumption Q sf of the branch company to examine the branch company.

[0139] (2) Assessment of branch dispatchers (or station managers) by each branch

[0140] The branch uses the target heat consumption Q of each heat exchange control system x and actual calorie consumption Q s The comparison is made to evaluate the dispatcher (or station manager) responsible for the heat exchange system.

[0141] Step 5) System optimization and iteration:

[0142] Heat source fault handling: When the heat source is under-supplied / over-supplied, the valve adjustment priority is as follows: when under-supplied, the nearest heating station prioritizes opening the valve; when over-supplied, the nearest heating station prioritizes reducing the valve opening. The nearest heating station prioritizes adjusting the valve opening to maintain the heat network balance and automatically marks abnormal data for interpolation and completion.

[0143] Because it takes on-demand heating as the starting point, thermal index control does not perform heat network balance adjustment. Therefore, during the initial cold period of operation, the temperature control mode is first used to complete the balance adjustment of the entire network, and the severe cold period is switched to thermal index control. Due to the property of the thermal index control method that the heat network balance adjustment is autonomously performed without following the fluctuation of the heat source, when the heat source is undersupplied, the near-end thermal power station will open the valve first compared with the far-end thermal power station, the heating effect of the far-end thermal power station will deteriorate, and the balance of the heat network will be disrupted. Therefore, this control method has high requirements for the timely processing efficiency of the dispatching heat source. When the heat source is undersupplied or the operating conditions change significantly, it is necessary to manually exit the index control mode, retain the original valve state of the thermal power station, and perform manual intervention in time according to the changes in the operating conditions of the thermal station. Compared with temperature control, when the heat source is oversupplied, the return water temperature of the thermal power station controlled by thermal index control is generally lower than that of the thermal power station controlled by temperature control; when the heat source is undersupplied, the return water temperature of the thermal power station controlled by thermal index control is generally higher than that of the thermal power station controlled by temperature control; when the heat source is stable, the difference between the two is not much.

[0144] Equipment fault handling: For subsystems that significantly deviate from the indicator concentration area, identify the cause and take corresponding measures, establish a record, use PID algorithm for stable control, and solve valve oscillation problems through KP, KI, and KD parameter optimization;

[0145] Control failure handling: If the PID adjustment fails to meet the standard within the time limit, it will switch to not participating in the balance control, and the operator will manually adjust the pump and valve parameters.

[0146] Model iteration: After the end of each heating season, the third-order regression model coefficients are refitted based on the historical data of the past three years;

[0147] Energy consumption optimization: By analyzing the energy consumption characteristics of thermal power stations, we can identify high-energy-consuming groups and give priority to energy-saving transformation.

[0148] Identify building groups with high unit heat consumption (>0.4125GJ / (㎡·a)), comprehensively analyze and determine the reasons for the high energy consumption from the aspects of basic conditions, operating conditions, user feedback, etc. For reasons such as pipeline network design, age of use, pipeline blockage, etc., give priority to the transformation of heating pipelines to reduce overheating losses.

[0149] This method enables the centralized heating system to achieve: control response time ≤ 20 minutes; overall network energy consumption reduced by ≥ 4.27%; and room temperature compliance rate increased to 99%.

[0150] When this method is applied to a heating network consisting of 1,635 heating stations and 2,998 systems, it achieves the following results:

[0151] During the severe cold season (December-February), 83 pilot heating stations were tested for large-scale heating network, with control response time ≤ 20 minutes;

[0152] The fluctuation range of thermal index was reduced by 60% through PID parameter optimization;

[0153] Ultimately, the overall energy consumption of the entire network decreased by 4.27%, and the room temperature compliance rate increased to 99%.

[0154] Table 4 shows the result data table of some thermal power stations obtained by the method of the present invention. The results show that after the implementation of the present invention, the energy consumption of thermal power stations in the entire network is reduced, the refined management level of the group company is improved, the scheduling process and assessment and evaluation system are optimized, and the comprehensive energy consumption of the entire network is effectively reduced, with significant energy-saving and environmental protection benefits.

[0155] Table 4 Correction results of some thermal power stations

[0156]

Claims

1. A comprehensive scheduling method for a central heating system based on operating data, characterized in that: The steps include: Step 1) Data Collection and Transmission: Real-time data collection of user room temperature, heat, water, and electricity consumption is performed. Heat meters installed on the return pipes of the heating station zones or the primary network return main monitor heat, water and electricity meters equipped with optical character recognition (OCR) dials monitor water and electricity consumption, and room temperature is monitored using user-side room temperature collectors. Abnormal data is automatically flagged and repairs are triggered. The data is uploaded to the smart heating platform in real time via a dedicated data line and verified. Step 2) Dynamic thermal index model establishment: Step 2.1) Calculate energy consumption indicators using a phased approach: Divide the heating season into the early cold period, the severe cold period, and the late cold period; Eliminate abnormal data: thermal index >40W / m² or <0 in the early cold period, >70W / m² or <20W / m² in the severe cold period, >40W / m² or <0 in the late cold period; Step 2.2) Based on the outdoor temperature, the heat load ratio interpolation method is used to complete the thermal index value of the discontinuous temperature interval. The heat load ratio calculation formula is: ， Among them, t n is the indoor design temperature, t p is the average outdoor temperature, t wn Design temperature for outdoor use; Step 2.3) Calculate the average thermal index of each heating station subsystem at different times according to the following formula: Heat load ratio × maximum heat index = actual heat index; Establish the index model for each temperature point within the range of -12~12℃; Step 2.4) Based on the nonlinear relationship between the shutdown rate and heat consumption per unit, energy consumption is predicted using a third-order regression model: Heat consumption per unit = C1 + C2 × stop rate − C3 × stop rate^2 + C4 × stop rate^3 Among them, C1-C4 are coefficients to be fitted, which are fitted through a high-order regression model. The suspension rate = the total suspension area of ​​the stall heating station / the total contract area of ​​the stall heating station. The unit heat consumption = the total heat consumption of the stall heating station / the total operating area of ​​the stall heating station; Step 2.5) Based on the real-time room temperature data, the thermal index is corrected in real time using the correction formula: ， Where: q is the corrected thermal index, unit is W / ㎡; T i is the actual room temperature, unit is ℃; T j The unit of outdoor heating temperature is ℃, which is -9.7℃; Q d is the thermal index before correction, unit is W / ㎡, and the suspension rate is the proportion of users' suspension reports; Step 3) Dynamic control of thermal indicators: PID algorithm control: PID control uses the stop rate prediction value from step 2.4) as the feedforward input. First, adjust the proportional coefficient KP to suppress valve oscillation (KP value is 0.08-0.5), then introduce the integral coefficient KI to eliminate the residual error, and finally optimize the dynamic performance through the differential coefficient KD. Actuator control logic: When increasing the load, the valve is opened first and the pump is adjusted; when decreasing the load, the pump is lowered first and the valve is closed; the electric regulating valve has an operating cycle of 7 minutes / time, the return water booster pump has an operating cycle of 20 minutes / time, the frequency increase amplitude is 0.6Hz / time, and the frequency reduction amplitude is 0.3Hz / time; Differentiated grouping strategy: Public heating stations with no nighttime heat demand will have their nighttime heat index temperature increased by 5%-10%. Terminal heating stations in the heating network will have their temperature reduced by 3%-8% based on the heat source index. Old residential communities will maintain their all-day temperature control and reduce it at night. Mode switching: The thermal index control mode is adopted in the severe cold period, and the temperature-index mixed control mode is adopted in the early cold period and the late cold period; Step 4) Energy consumption assessment and evaluation: Taking the heating station as the unit, statistics are collected on heating type, construction year, enclosure structure and secondary pipe network attributes; Compare actual energy consumption with the dynamic indicator energy consumption obtained in step 2). Combined with the user complaint rate and room temperature compliance rate, a room temperature between 18°C ​​and 24°C is considered to meet the standard. Room temperatures below 18°C ​​and above 24°C are considered to have poor control effects. The control effects are differentiated by the number of effective complaints per user. Energy consumption score calculation: The absolute value of energy consumption ranking of thermal power stations of the same type accounts for X%, and the actual energy consumption and target energy consumption savings and excess scores account for Y%, and the energy consumption year-on-year ratio score accounts for Z%. Where X+Y+Z=100% and the weights are adjusted dynamically, the energy consumption score is obtained, and the performance salary is determined based on the energy consumption score; Step 5) System optimization and iteration: Heat source fault handling: When the heat source is under-supplied / over-supplied, the nearest heating station will prioritize adjusting the valve opening to maintain the heat network balance and automatically mark abnormal data for interpolation and completion; Equipment fault handling: Establish records for subsystems that deviate from indicators, use PID algorithm for stable control, and solve valve oscillation problems through KP, KI, and KD parameter optimization; Control failure handling: When PID control times out and fails to meet the standard, it switches to manual control; Model iteration: After the end of each heating season, the third-order regression model coefficients are refitted based on the historical data of the past three years; Energy consumption optimization: Identify high-energy-consuming thermal power stations, analyze the causes, and prioritize their transformation.

2. A method for comprehensive scheduling of a centralized heating system based on operation data according to claim 1, characterized in that: In step 1), the heat meter is installed on the return pipe of each zone of the thermal power station or the return main of the primary network to monitor the heat; the water meter / electricity meter is equipped with an OCR dial to realize remote data transmission for monitoring the water and electricity volume, and the image recognition error rate of the OCR dial is ≤0.5%.

3. A method for comprehensive scheduling of a centralized heating system based on operation data according to claim 1, characterized in that: The coefficients of the third-order regression model in step 2.4) are fitted using the least squares method, specifically including: Energy saving - radiator type: C1=0.3032, C2=0.4557, C3=1.205, C4=2.227; Energy saving - floor heating type: C1=0.2670, C2=0.9038, C3=2.695, C4=3.660; Non-energy-saving-radiator type: C1=0.3342, C2=0.4492, C3=1.570, C4=3.

394.

4. A method for comprehensive scheduling of a centralized heating system based on operation data according to claim 1, characterized in that: When the KP value in step 3) is optimized to 0.08, the fluctuation range of the thermal index is reduced by 60%. When the heat source fluctuates, the thermal index weight coefficient of the terminal thermal power station is modified through manual intervention. The energy consumption score is fed back to the PID control module in real time. When the energy consumption score is lower than the threshold, the KP / KI value and the actuator period are automatically adjusted.

5. The method for comprehensive scheduling of a centralized heating system based on operation data according to claim 1, characterized in that: In step 4), a Y% weight is added to the heating station with great energy-saving potential; a Z% weight is added to the heating station with a high energy consumption benchmark. The clustering input vector includes the relationship between the shutdown rate and the unit heat consumption, the aging degree of the enclosure structure, and the leakage rate of the pipe network.

6. A method for comprehensive scheduling of a centralized heating system based on operation data according to claim 1, characterized in that: The valve adjustment priority for heat source fault handling in step 5) is: when the heat source is undersupplied, the proximal heat station prioritizes opening the valve, requiring manual intervention to exit the control mode; when the heat source is oversupplied, the proximal heat station prioritizes reducing the valve opening.

7. A method for comprehensive scheduling of a centralized heating system based on operation data according to claim 1, characterized in that: The method enables the centralized heating system to achieve: control response time ≤ 20 minutes; overall network energy consumption reduced by ≥ 4.27%; and room temperature compliance rate increased to 99%.

Images