An internet of things energy-saving control system and method for central heating

Abstract

The present application relates to the technical field of heat supply system control, in particular to a kind of internet of things energy-saving control system and method for central heating, comprising: obtaining the action event data of each heating terminal electric regulating valve and the temperature data of temperature sensor in corresponding area, based on the action event data, using sliding window least square inversion method, obtain dynamic time delay-gain spectrum, temperature residual sequence is quantitatively analyzed, to calculate net causal contribution degree, and the correlation between temperature sensor and valve is calculated by combining energy weighting method, identify deviating temperature sensor;Space causal reassociation and virtual heat demand fusion correction are carried out to deviating temperature sensor, to obtain corrected heat demand signal;Dynamic time delay-gain spectrum is embedded into feedforward-feedback composite control framework, and causal adaptive parameter adjustment is carried out based on corrected heat demand signal.The present application improves the energy-saving effect in the process of central heating.

Description

Technical Field

[0001] This invention relates to the field of heating system control technology, and specifically to an Internet of Things (IoT) energy-saving control system and method for centralized heating. Background Technology

[0002] Central heating systems are a crucial infrastructure for the winter operation of large commercial complexes, and their energy efficiency directly impacts the complex's operating costs and indoor thermal comfort levels. In traditional heating systems for large commercial complexes, heat exchange stations typically employ climate compensation or manual, experience-based adjustments, meaning they adjust the water supply temperature based on outdoor temperature changes and only make passive corrections upon receiving user complaints about heating or cooling. With the development of IoT technology, some central heating systems have begun to incorporate indoor temperature sensors, uploading temperature data from various points to a control platform via wireless networks as a reference for heat exchange station adjustments.

[0003] Due to the significant heat transfer delay in the heating network of large commercial complexes, and the substantial differences in the thermal inertia of building envelopes across different areas, there is a temporal asynchrony between the temperature changes sensed by temperature sensors and the actual heat output of the corresponding heating terminals. Furthermore, existing systems often assume that each temperature sensor serves only the heating terminals within its designated area, without verifying the causal relationship between the temperature changes measured by the sensor and the actual heat output of the terminals in that area. When a temperature sensor's temperature change pattern deviates significantly from the actual output characteristics of its associated heating terminal due to spatial location, airflow disturbances, or interference from nearby heat sources, the system still directly uses the sensor's measurement for control decisions at the heat exchange station in that area, resulting in a logical misalignment between the water supply temperature adjustment and the actual heat demand of the area. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an Internet of Things (IoT) energy-saving control system and method for centralized heating.

[0005] One embodiment of the present invention provides an Internet of Things (IoT) energy-saving control method for centralized heating, the method comprising the following steps:

[0006] Acquire action event data of electric regulating valves at each heating terminal and temperature data of temperature sensors in the corresponding area. The action event data includes area number, action time, opening degree before action, and opening degree after action.

[0007] Based on the action event data, a dynamic time delay-gain spectrum is obtained by using the sliding window least squares inversion method. The dynamic time delay-gain spectrum includes unity gain, time delay parameter and thermal time constant.

[0008] The temperature residual sequence is quantitatively analyzed to calculate the net causal contribution, and the correlation between the temperature sensor and the valve is calculated by combining the energy weighting method to identify the temperature sensor deviation.

[0009] Spatial causal recordation and virtual thermal demand fusion correction are performed on the temperature sensor to obtain the corrected thermal demand signal;

[0010] By embedding the dynamic time delay-gain spectrum into the feedforward-feedback composite control framework, and performing causal adaptive parameter adjustment based on the corrected heat demand signal, active time delay compensation and heating control are achieved.

[0011] Optionally, the specific method for obtaining the dynamic delay-gain spectrum is as follows:

[0012] Obtain baseline data, and for any valve action event, calculate the change in valve opening and the net temperature response, wherein the net temperature response is the difference between the average measured temperature and the average baseline data.

[0013] A sliding window with a preset window size is used, and it slides to the right from the moment the valve action is executed with a preset step size until the right end of the window reaches the end of the interval.

[0014] During the sliding window process, for each window position, the least squares optimization problem is solved by combining the valve opening change and the net temperature response to obtain the corresponding unity gain, time delay parameter and thermal time constant; and the unity gain, time delay parameter and thermal time constant obtained for all window positions are arranged according to the window center time to form a corresponding sequence, thereby obtaining the dynamic time delay-gain spectrum.

[0015] Optionally, the quantitative analysis of the temperature residual sequence to calculate the net causal contribution specifically involves:

[0016] An ideal causal prediction curve is constructed using parameters in the dynamic time delay-gain spectrum. The temperature residual between the actual net temperature and the ideal predicted value in the ideal causal prediction curve is calculated, and a temperature residual sequence is formed.

[0017] Discretize the temperature residual sequence into a symbol sequence;

[0018] Calculate the Lempel-Ziv complexity of the symbol sequence, denoted as the first parameter;

[0019] Calculate the signal energy of the ideal causal prediction curve and normalize it, denoted as the second parameter;

[0020] Calculate the net causal contribution based on the first and second parameters.

[0021] Optionally, discretizing the temperature residual sequence into a symbol sequence specifically involves:

[0022] Preset a first threshold and a second threshold;

[0023] When the residual value of any element in the temperature residual sequence is greater than the first threshold, 1 is used as the sign value of the corresponding element.

[0024] When the residual value of any element in the temperature residual sequence is less than the second threshold, -1 is used as the sign value of the corresponding element;

[0025] When the residual value of any element in the temperature residual sequence does not meet the conditions of being greater than the first threshold and less than the second threshold, the sign value of the element corresponding to 0 will be changed.

[0026] Thus, the sequence of sign values ​​corresponding to all elements in the temperature residual value sequence is taken as the sign sequence.

[0027] Optionally, the identification of the temperature sensor deviating from the target temperature is specifically as follows:

[0028] Collect statistics on all valve action events between the current temperature sensor and valve pair within a preset time window, and obtain the opening change for each valve action event;

[0029] The net causal contribution of valve action events is used as a weight, combined with the change in valve opening, to calculate the deviation of the temperature sensor-valve pair.

[0030] Based on the deviation distribution of all temperature sensor-valve pairs, a deviation threshold is determined. When the deviation exceeds the deviation threshold, the corresponding temperature sensor is marked as a deviation temperature sensor.

[0031] Optionally, the spatial causal re-association specifically refers to:

[0032] For any off-temperature sensor, calculate the potential causal strength between the off-temperature sensor and each other valve;

[0033] The valve corresponding to the maximum potential causal strength is identified as a candidate valve;

[0034] When the ratio of the potential causal strength of a candidate valve to the correlation strength of the valve originally associated with the temperature sensor is greater than a preset multiple, and the potential causal strength of the candidate valve is greater than the lower bound of the normal correlation strength of the system, the candidate valve is determined to meet the re-association condition.

[0035] The temperature sensor is decoupled from the original associated valve, and a new association relationship is established between the temperature sensor and the candidate valve that meets the re-association condition.

[0036] Optionally, the virtual thermal demand fusion correction specifically includes:

[0037] A potential causal threshold is preset. For temperature sensors that cannot be spatially re-correlated, the relevant valve set of the temperature sensors is selected based on the potential causal strength and the potential causal threshold.

[0038] Obtain the spatial distance and time delay consistency factor between the valve and the temperature sensor. Based on the potential causal strength, spatial distance and time delay consistency factor between the valve and the temperature sensor in the relevant valve set, determine the attention weight of each valve.

[0039] Based on the attention weights of each valve and its historical opening status, the virtual thermal demand deviating from the temperature sensor is calculated.

[0040] Optionally, embedding the dynamic time delay-gain spectrum into the feedforward-feedback composite control framework specifically involves:

[0041] Feedforward control section: Based on the ratio of the target temperature adjustment to the current dynamic gain, the feedforward control signal is determined, and the command transmission time is determined based on the dynamic time delay to achieve active time delay compensation;

[0042] Feedback control section: Based on the deviation between the corrected heat demand signal and the target temperature, a PID control algorithm is used to generate a feedback control signal;

[0043] The proportional gain of the PID controller is dynamically adjusted based on the changing trend of the potential causal strength.

[0044] Optionally, the dynamic adjustment of the proportional gain of the PID controller specifically involves:

[0045] Calculate the average value of the global potential causal strength at the current moment, and determine the adjustment coefficient of the proportional gain by combining the deviation between the current global causal correlation strength and the historical statistical level;

[0046] The initial proportional gain of the PID controller is corrected by adjusting the coefficients to obtain the corrected proportional gain.

[0047] An embodiment of the present invention also provides an Internet of Things (IoT) energy-saving control system for centralized heating, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of any one of the IoT energy-saving control methods for centralized heating.

[0048] The beneficial effects of the technical solution of this invention are as follows: A dynamic time-delay-gain spectrum is constructed through sliding window least squares inversion, achieving accurate characterization of time-varying heat transfer characteristics; the randomness of complex quantified temperature residual sequences is introduced, and the net causal contribution is calculated by combining signal energy analysis, establishing a quantitative evaluation system for the strength of sensor-valve causal correlation; for identified deviating sensors, a dual correction mechanism combining spatial causal re-correlation and virtual heat demand is adopted to ensure the authenticity and reliability of the control input signal; the dynamic time-delay-gain spectrum is embedded into a feedforward-feedback composite control framework, and through active time-delay compensation and causal adaptive parameter adjustment, dynamic matching between the control strategy and the actual thermodynamic state of the system is achieved, thereby improving the accuracy of on-demand heating and the system's energy-saving effect, while also enhancing the control strategy's adaptability to time-varying factors such as seasonal changes and building renovations. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating the steps of an IoT-based energy-saving control method for centralized heating according to the present invention.

[0051] Figure 2 This is a structural block diagram of an Internet of Things (IoT) energy-saving control system for centralized heating according to the present invention. Detailed Implementation

[0052] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an Internet of Things (IoT) energy-saving control system and method for centralized heating proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0054] The following description, in conjunction with the accompanying drawings, details a specific scheme for an IoT-based energy-saving control system and method for centralized heating provided by the present invention.

[0055] Please see Figure 1The diagram illustrates a flowchart of an IoT-based energy-saving control method for centralized heating, according to an embodiment of the present invention. The method includes the following steps:

[0056] Step S001: Obtain the action event data of the electric regulating valves at each heating terminal and the temperature data of the temperature sensors in the corresponding area.

[0057] It should be noted that in the heating systems of large commercial complexes, due to the heat transfer delay of the heating network and the differences in thermal inertia of the building envelope, there is a time lag between the temperature changes sensed by the temperature sensors and the actual output at the heating terminals. In order to accurately quantify this time delay characteristic and identify the true causal relationship between the temperature sensors and the valves, it is necessary to first obtain the corresponding data of valve action events and temperature responses.

[0058] Specifically, to implement the IoT-based energy-saving control method for centralized heating proposed in this embodiment, it requires action event data of the electric regulating valves at each heating terminal and monitoring data from temperature sensors in the corresponding areas. The specific process is as follows:

[0059] First, deploy an action recorder at the electrically operated regulating valve at each heating terminal. Set the trigger condition to a valve opening change exceeding a preset dead zone (e.g., 2%). When the trigger condition is met, record the zone number, action time, pre-action opening, and post-action opening. Combine the zone number, action time, pre-action opening, and post-action opening into a single valve action event and record it as follows: ,in Indicates the area code; Indicates the moment of the action; Indicates the opening degree before the action. This indicates the opening degree after the action; all valve action events are stored in chronological order of the action time, forming an event stream.

[0060] Then, temperature sensors are installed in each area to monitor the temperature. The time interval for collecting monitoring data is the time interval between two adjacent valve actions. Whenever a valve action event occurs, all temperature readings of the temperature sensor in that area within the corresponding time interval are automatically recorded to form temperature data.

[0061] Finally, within the set time window, the time periods during which no valves operate are extracted. The temperature data for these time periods are then linearly fitted using the least squares method, and the fitting results are used as the baseline data for the corresponding temperature data.

[0062] In addition, since the natural temperature drift is approximately linear in the absence of active heating intervention, baseline data can be used to calculate the net temperature response.

[0063] Thus, the valve action event data of the electric regulating valves at each heating terminal and the monitoring data of the temperature sensors in the corresponding areas are obtained through the above method, i.e., temperature data.

[0064] Step S002: Based on the action event data, the dynamic time delay-gain spectrum is obtained by using the sliding window least squares inversion method.

[0065] It should be noted that after the heating terminal changes its opening degree, there is a transmission delay before the distant temperature sensor begins to respond. This transmission delay is called the time delay, and the temperature change caused by each 1% change in valve opening degree is called the gain. Due to the different thermal inertia of different areas, the time delay and gain are not fixed values. If a fixed time delay is used for control, premature or delayed heating may occur, leading to energy waste or poor temperature control. Therefore, it is necessary to analyze the dynamic changes in time delay and gain throughout the entire response process for each valve action.

[0066] As a preferred embodiment, the method for obtaining the dynamic time delay-gain spectrum based on action event data using the sliding window least squares inversion method includes the following specific steps:

[0067] Obtain baseline data, and for any valve action event, calculate the change in valve opening and the net temperature response, wherein the net temperature response is the difference between the average measured temperature and the average baseline data.

[0068] A sliding window with a preset window size is used, and it slides to the right from the moment the valve action is executed with a preset step size until the right end of the window reaches the end of the interval.

[0069] During the sliding window process, for each window position, the least squares optimization problem is solved by combining the valve opening change and the net temperature response to obtain the corresponding unity gain, time delay parameter and thermal time constant; and the unity gain, time delay parameter and thermal time constant obtained for all window positions are arranged according to the window center time to form a corresponding sequence, thereby obtaining the dynamic time delay-gain spectrum.

[0070] As an optional embodiment, the specific formula of the least squares optimization function can be expressed as:

[0071]

[0072] in, This represents the unit gain, which is the temperature rise caused by a 1% change in valve opening (unit: ℃ / %). This indicates the change in valve opening (unit: %). This parameter represents the time delay from the start of valve operation to the start of temperature rise (unit: minutes). The thermal time constant, expressed as the time (in minutes) required for the temperature to rise from its initial rise to 63.2% of its final steady-state value, reflects the thermal inertia of the system. The larger the value, the slower the temperature rises. Represents the time in the sliding window; Indicates the start time of the sliding window; This indicates the window size of the sliding window.

[0073] It should be noted that for the least squares optimization function, the physical meaning of this optimization problem is: by adjusting the values ​​of unity gain, time delay parameter and thermal time constant, calculate the difference between the model prediction value and the measured value at each time step, and then sum the squared differences to find a set of parameters that minimizes the sum of squares. This is the least squares fitting, which can give the optimal parameter estimation result within the sliding window.

[0074] It should be noted that, in a specific embodiment of the present invention, the window size of the sliding window is set to 30 minutes, and the corresponding step size during the sliding traversal is preset to 2 minutes. The reason for setting the window size to 30 minutes is that the thermal time constant is usually 30-120 minutes, and the window width is comparable to the minimum time constant, which can capture the complete temperature rise segment; if the window size is too small (e.g., 10 minutes), the complete exponential change shape may not be observed; conversely, if the window size is too large (e.g., 60 minutes), the change characteristics of the parameter over time will be smoothed out. In addition, the reason for setting the step size to 2 minutes is that the sampling interval of the temperature sensor is usually 1-2 minutes, and the step size is consistent with the sampling interval to ensure sufficient time resolution and moderate computational load; in other embodiments, the window size and sliding step size of the sliding window can be adjusted according to the actual situation, and the embodiments of the present invention do not impose specific limitations.

[0075] As the window slides, each window position receives a set of... Arranging these parameters according to the window center time yields three sequences that change over time: , , ,Will , , The sequence set formed by the sequence is used as a dynamic delay-gain spectrum, which characterizes the delay and gain characteristics of the system at different times after the valve is actuated.

[0076] Thus, the dynamic delay-gain spectrum is obtained using the above method.

[0077] Step S003: Perform quantitative analysis on the temperature residual sequence to calculate the net causal contribution, and combine the energy weighting method to calculate the correlation between the temperature sensor and the valve, and identify the temperature sensor deviating from the value.

[0078] It's important to note that while the unity gain, time delay parameters, and thermal time constant obtained through least-squares fitting are optimal, they force the model to account for all temperature changes, including those that might not be caused by valve action (e.g., external disturbances such as solar radiation, foot traffic, and heating effects from neighboring areas). This can lead to an overestimation of the true effect of valve action. Therefore, a metric is needed to quantify what proportion of the observed temperature changes are truly "caused" by the valve action itself, rather than being influenced by external disturbances.

[0079] As a preferred embodiment, the temperature residual sequence is quantitatively analyzed using Lempel-Ziv complexity to obtain the net causal contribution, including the following specific methods:

[0080] An ideal causal prediction curve is constructed using parameters in the dynamic time delay-gain spectrum. The temperature residual between the actual net temperature and the ideal predicted value in the ideal causal prediction curve is calculated, and a temperature residual sequence is formed.

[0081] Discretize the temperature residual sequence into a symbol sequence; calculate the Lempel-Ziv complexity of the symbol sequence, denoted as the first parameter; calculate the signal energy of the ideal causal prediction curve and normalize it, denoted as the second parameter;

[0082] Calculate the net causal contribution based on the first and second parameters.

[0083] The temperature residual between the actual net temperature and the ideal predicted value in the ideal causal prediction curve represents the difference between the actual net temperature and the ideal predicted value in the ideal causal prediction curve. The ideal causal prediction curve shows how the temperature should change solely due to valve operation without external disturbances.

[0084] As a preferred embodiment, the specific method for discretizing the temperature residual sequence into a symbol sequence includes:

[0085] A first threshold and a second threshold are preset. When the residual value of any element in the temperature residual sequence is greater than the first threshold, 1 is used as the sign value of the corresponding element. When the residual value of any element in the temperature residual sequence is less than the second threshold, -1 is used as the sign value of the corresponding element. When the residual value of any element in the temperature residual sequence does not meet the conditions of being greater than the first threshold and less than the second threshold, 0 is used as the sign value of the corresponding element. Thus, the sequence of sign values ​​corresponding to all elements in the temperature residual sequence is used as the sign sequence.

[0086] It should be noted that, in a specific embodiment of the present invention, the first threshold can be set to 0.05℃ and the second threshold can be -0.05℃. When the residual value of any element in the temperature residual sequence is greater than the first threshold, it indicates an abnormal temperature rise; when the residual value of any element in the temperature residual sequence is less than the second threshold, it indicates an abnormal temperature drop; and when the sign value of an element in the temperature residual sequence is 0, it indicates a stable temperature change. In other embodiments, the first and second thresholds can be adjusted according to the actual situation, and the embodiments of the present invention do not impose specific limitations.

[0087] It should be noted that the first parameter represents the number of different substrings contained in the symbol sequence. The more random the sequence (i.e., the greater the perturbation), the larger the first parameter; conversely, the more regular the sequence (i.e., the more it conforms to the ideal prediction), the smaller the first parameter.

[0088] As an optional embodiment, the specific method for calculating the net causal contribution can be: ,in Indicates net causal contribution. Indicates the second parameter; This indicates the first parameter.

[0089] It should be noted that in the specific calculation method of net causal contribution, the denominator... The total uncertainty is represented by the sum of the causal signal energy and the perturbation complexity; the molecule... The net causal contribution represents the proportion of the causal signal in the total uncertainty. The larger the net causal contribution, the more likely the temperature change is caused by the valve action. Conversely, the smaller the net causal contribution, the more likely the temperature change is dominated by external disturbances and the weak causal relationship between the temperature sensor and the valve.

[0090] As a preferred embodiment, the specific method for calculating the divergence between the temperature sensor and the valve based on an energy-weighted approach is as follows:

[0091] Collect statistics on all valve action events between the current temperature sensor and valve pair within a preset time window, and obtain the opening change for each valve action event;

[0092] The net causal contribution of valve action events is used as a weight, combined with the change in valve opening, to calculate the deviation of the temperature sensor-valve pair.

[0093] Based on the deviation distribution of all temperature sensor-valve pairs, a deviation threshold is determined. When the deviation exceeds the deviation threshold, the corresponding temperature sensor is marked as a deviation temperature sensor.

[0094] As an optional embodiment, for any deviation between the temperature sensor and the valve, the specific calculation method for the deviation can be:

[0095]

[0096] in, This indicates the degree of deviation between the temperature sensor and the valve. Indicates the first The square of the change in valve opening for each valve action event; Indicates the first Net causal contribution of each valve action event; This indicates the number of all valve action events between the current temperature sensor and valve pair within a preset time window.

[0097] It should be noted that changes in valve opening determine changes in input heat. The greater the heat change, the greater the resulting temperature change, and the higher the signal-to-noise ratio, thus making the net causal contribution C more reliable. Weighting using the square of the valve opening change further amplifies the weight differences in large-amplitude movements, emphasizing the reliability of large-amplitude signals. This represents the energy-weighted average causal strength. The smaller the deviation, the higher the reliability of the temperature sensor's output; while the larger the deviation, the lower the correlation between the temperature sensor reading and the valve action.

[0098] Then, calculate the average value and standard deviation of the energy-weighted average causal intensity of all temperature sensor-valve pairs in the entire commercial complex, and set a deviation threshold; the temperature sensor corresponding to the deviation degree is greater than the deviation threshold is regarded as the deviation temperature sensor.

[0099] As an optional embodiment, the specific calculation method for the deviation threshold can be: ,in, Indicates deviation from the threshold; This represents the average value; It represents the standard deviation.

[0100] It should be noted that, This represents the lower limit of normal causal strength. Assuming the energy-weighted average causal strength approximates a normal distribution, the probability of it being below 1.5 standard deviations is approximately 6.7%. Furthermore, for any temperature sensor-valve pair, if its energy-weighted average causal strength is below this deviation threshold (i.e., the deviation is greater than the deviation threshold), its causal strength is considered significantly lower than the system average, and the temperature sensor is marked as a deviation temperature sensor. Here, 1.5 is a preset multiple, which can be adjusted according to actual conditions in other embodiments. This invention does not impose specific limitations on this embodiment.

[0101] Thus, the net causal contribution, deviation, and deviation temperature sensor identification results are obtained through the above methods.

[0102] Step S004: Perform spatial causal recordation and virtual thermal demand fusion correction on the temperature sensor to obtain the corrected thermal demand signal.

[0103] It's important to note that a temperature sensor's discrepancy with its corresponding valve might be because it's actually more affected by another valve. For example, a temperature sensor installed near an atrium might be more sensitive to valve activity downstairs than to valve activity in its own room due to rising hot air currents in the atrium. Therefore, it's necessary to find the true physical cause for this temperature sensor's behavior or reconstruct its thermal demand signal through multi-source fusion.

[0104] As a preferred embodiment, the specific method for spatial causal re-association is as follows:

[0105] For any off-temperature sensor, calculate the potential causal strength between the off-temperature sensor and each other valve;

[0106] The valve corresponding to the maximum potential causal strength is identified as a candidate valve;

[0107] When the ratio of the potential causal strength of a candidate valve to the correlation strength of the valve originally associated with the temperature sensor is greater than a preset multiple, and the potential causal strength of the candidate valve is greater than the lower bound of the normal correlation strength of the system, the candidate valve is determined to meet the re-association condition.

[0108] The temperature sensor is decoupled from the original associated valve, and a new association relationship is established between the temperature sensor and the candidate valve that meets the re-association condition.

[0109] As an optional embodiment, the specific formula for the reassociation condition can be expressed as:

[0110]

[0111] in, Indicates temperature sensor With candidate valves The potential causal strength between them; Indicates temperature sensor Corresponding valve The potential causal strength; Indicates temperature sensor The average of the potential causal strength among all candidate valves; Indicates temperature sensor The standard deviation of the potential causal strength among all candidate valves.

[0112] It should be noted that, regarding the first condition The requirement is that the causal strength of the new candidate valve must be significantly higher than that of the original corresponding valve (at least twice) for it to be worthwhile to change the association, thus avoiding erroneous reassociation due to random fluctuations. Regarding the second condition... This represents the lower bound of the normal correlation strength of the system, which requires the temperature sensor under the new correlation. The potential causal strength between the temperature sensor and the corresponding candidate valve must be within a normal range. When both conditions are met, reassociation is performed, and the system records the new temperature sensor-valve pair association, i.e., the temperature sensor association will be used in subsequent control processes. The readings will be used to control the candidate valves. Instead of the original corresponding valve .

[0113] As a preferred embodiment, the specific method for performing virtual thermal demand fusion correction on the divergent temperature sensor includes:

[0114] A potential causal threshold is preset. For temperature sensors that cannot be spatially re-correlated, the relevant valve set of the temperature sensors is selected based on the potential causal strength and the potential causal threshold.

[0115] Obtain the spatial distance and time delay consistency factor between the valve and the temperature sensor. Based on the potential causal strength, spatial distance and time delay consistency factor between the valve and the temperature sensor in the relevant valve set, determine the attention weight of each valve.

[0116] Based on the attention weights of each valve and its historical opening status, the virtual thermal demand deviating from the temperature sensor is calculated.

[0117] It should be noted that in one specific embodiment of the present invention, the potential causal threshold is set to one-third of the average value of all potential causal intensities, thereby indicating that valves with potential causal intensities higher than the potential causal threshold have at least some influence on the temperature sensor. In other embodiments, the potential causal threshold can be adjusted according to the actual situation, and the embodiments of the present invention do not impose specific limitations.

[0118] As an optional embodiment, the specific method for calculating the attention weights can be: ,in, Indicates valve Attention weights For valves The causal strength factor, i.e., the temperature sensor With valve The potential causal strength between them Indicates temperature sensor With valve Spatial distance between them; Indicates valve The latency consistency factor; For valves The causal strength factor, i.e., the temperature sensor With valve The potential causal strength between them Indicates temperature sensor With valve Spatial distance between them; Indicates valve The latency consistency factor; Indicates temperature sensor The number of valves in the relevant valve set.

[0119] It should be noted that the temperature sensor With valve The higher the value of the potential causal strength between them, the stronger the association, and the greater its weight should be when merging them.

[0120] As an optional embodiment, the specific calculation method for the valve's time delay consistency factor can be as follows: ,in Indicates temperature sensor With valve The time delay parameters between; Indicates temperature sensor The average value of the time delay parameters of all valves in the corresponding valve set; This indicates the preset allowable delay value.

[0121] It should be noted that, in a specific embodiment of the present invention, the specific value of the preset time delay allowable value can be that of the degree sensor. The median of the time delay parameters of all valves in the corresponding valve set can be adjusted according to the actual situation in other embodiments, and the embodiments of the present invention do not impose specific limitations.

[0122] As an optional embodiment, for any temperature sensor with a deviation from its operating temperature, the specific method for calculating the virtual thermal demand of the temperature sensor with a deviation from its operating temperature is as follows:

[0123]

[0124] in, Indicates the deviation from the temperature sensor at time The value of virtual heat demand under the current conditions. Indicates valve Attention weights This indicates the number of valves in the set of valves that are deviating from the temperature sensor. This indicates the valves in the relevant valve set. Unity gain; Indicates valve At any moment Previous moments The opening below, Indicates separation from temperature sensor and valve The time delay parameters between; Indicates valve The initial valve opening.

[0125] It should be noted that the physical meaning of virtual heat demand is as follows: It is assumed that the temperature change from the temperature sensor is a linear superposition of contributions from multiple valves. The contribution of each valve is proportional to its gain multiplied by the amount of its change beyond the base opening, and historical openings are used to account for time delay effects. The attention weight reflects the relative influence of each valve. Therefore, the value of virtual heat demand is the best estimate of what equivalent value the temperature sensor should measure without disturbance.

[0126] Additionally, for each deviating temperature sensor, either a re-associated valve is obtained to continue using its measured temperature, or a real-time updated virtual heat demand is obtained and used as the input signal for the controller.

[0127] Thus, the corrected heat demand signal is obtained through the above method.

[0128] Step S005: Embed the dynamic time delay-gain spectrum into the feedforward-feedback composite control framework, and perform causal adaptive parameter adjustment based on the corrected heat demand signal to realize active time delay compensation and heating control.

[0129] It should be noted that traditional PID control cannot adapt to the large time delay characteristics of heating systems, easily leading to overshoot or oscillation. By embedding the dynamic time delay-gain spectrum into a feedforward-feedback composite control framework, active compensation for time delay can be achieved, and control parameters can be dynamically adjusted according to changes in the strength of causal correlation, thereby improving the adaptability and robustness of the control.

[0130] As a preferred embodiment, the specific method for embedding the dynamic time delay-gain spectrum into the feedforward-feedback composite control framework includes:

[0131] Feedforward control section: Based on the ratio of the target temperature adjustment to the current dynamic gain, the feedforward control signal is determined, and the command transmission time is determined based on the dynamic time delay to achieve active time delay compensation;

[0132] Feedback control section: Based on the deviation between the corrected heat demand signal and the target temperature, a PID control algorithm is used to generate a feedback control signal; the proportional gain of the PID controller is dynamically adjusted according to the changing trend of the potential causal strength.

[0133] Using a standard PID controller, the input error is: ,in Indicates time The error value below, Indicates time The set target temperature Indicates the temperature sensor at time The temperature value below.

[0134] It should be noted that, for When the temperature sensor is not diverging from the temperature sensor, The value is the actual temperature value collected by the temperature sensor; however, when the temperature sensor is a non-temperature sensor and there is a re-associated valve, although the corresponding value is the actual temperature value, the control target during PID control is the valve corresponding to the re-associated valve. Additionally, when the temperature sensor is a non-temperature sensor and there is no re-associated valve... The value is the virtual heat demand at the corresponding time.

[0135] As an optional embodiment, the specific method for calculating the corrected proportional gain of the PID controller can be as follows:

[0136]

[0137] in, This represents the corrected proportional gain of the PID controller. This represents the initial proportional gain of the PID controller; This represents the average potential causal strength of all temperature sensor-valve pairs within a preset time range, and represents the typical causal strength of the system under normal operating conditions. The standard deviation of the potential causal strength of all temperature sensor-valve pairs within a preset time range; This represents the mean potential causal strength of all temperature sensor-valve pairs at the current moment.

[0138] It should be noted that the physical meaning of this correction strategy is: when At that time, the adjustment item is 0. This indicates that the system is in a normal state and uses baseline parameters; when When this occurs, it indicates that the current causal relationship is stronger than the historical average, the temperature sensor readings are more reliable, the adjustment term is positive, Kp increases, and the feedback control is more aggressive; when When the value is negative, it indicates that the current causal relationship is weaker than the historical average (there are more disturbances or deviations), the adjustment term is negative, Kp decreases, and the feedback control is more conservative.

[0139] In addition, the system runs in a preset cycle (e.g., 1 hour), and at the end of each cycle, it performs the following updates: recalculates all parameters based on new event data accumulated in the past 24 hours; recalculates the association mapping table and virtual fusion weights based on the update results; as the seasons change (e.g. from autumn to winter), changes in the pipeline network base temperature cause changes in time delay, and changes in building usage patterns (e.g., changes in thermal inertia caused by shop renovations) cause changes in building usage patterns. The system will continuously and automatically learn and adapt to these changes, so that the control strategy is always based on the latest and most realistic causal relationship state.

[0140] This concludes the embodiment.

[0141] The above steps complete the IoT-based energy-saving control method for centralized heating.

[0142] Please see Figure 2 The diagram illustrates a structural block diagram of an IoT energy-saving control system for centralized heating according to an embodiment of the present invention. The system includes a memory 202, a processor 201, and a computer program 2021 stored in the memory 202 and executable on the processor. When the processor 201 executes the computer program 2021, it implements steps S001 to S005 of the IoT energy-saving control method for centralized heating.

[0143] Furthermore, in an optional embodiment, the memory 202 described above may include read-only memory and random access memory, and provide instructions and data to the processor. The memory 202 may also include non-volatile random access memory. For example, the memory may also store device type information.

[0144] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An IoT-based energy-saving control method for centralized heating, characterized in that, The method includes the following steps: Acquire action event data of electric regulating valves at each heating terminal and temperature data of temperature sensors in the corresponding area. The action event data includes area number, action time, opening degree before action, and opening degree after action. Based on the action event data, a dynamic time delay-gain spectrum is obtained by using the sliding window least squares inversion method. The dynamic time delay-gain spectrum includes unity gain, time delay parameter and thermal time constant. The temperature residual sequence is quantitatively analyzed to calculate the net causal contribution, and the correlation between the temperature sensor and the valve is calculated by combining the energy weighting method to identify the temperature sensor deviation. Spatial causal recordation and virtual thermal demand fusion correction are performed on the temperature sensor to obtain the corrected thermal demand signal; The dynamic time delay-gain spectrum is embedded into the feedforward-feedback composite control framework, and causal adaptive parameter adjustment is performed based on the corrected heat demand signal to realize active time delay compensation and heating control.

2. The IoT-based energy-saving control method for centralized heating according to claim 1, characterized in that, The specific method for obtaining the dynamic delay-gain spectrum is as follows: Obtain baseline data, and for any valve action event, calculate the change in valve opening and the net temperature response, wherein the net temperature response is the difference between the average measured temperature and the average baseline data. A sliding window with a preset window size is used, and it slides to the right from the moment the valve action is executed with a preset step size until the right end of the window reaches the end of the interval. During the sliding process of the sliding window, for each window position, the least squares optimization problem is solved by combining the valve opening change and the net temperature response, so as to obtain the corresponding unity gain, time delay parameters and thermal time constant. The unity gain, time delay parameters, and thermal time constants obtained at all window positions are arranged according to the center time of the window to form a corresponding sequence, thereby obtaining the dynamic time delay-gain spectrum.

3. The IoT-based energy-saving control method for centralized heating according to claim 2, characterized in that, The quantitative analysis of the temperature residual sequence to calculate the net causal contribution is specifically as follows: An ideal causal prediction curve is constructed using parameters in the dynamic time delay-gain spectrum. The temperature residual between the actual net temperature and the ideal predicted value in the ideal causal prediction curve is calculated, and a temperature residual sequence is formed. Discretize the temperature residual sequence into a symbol sequence; Calculate the Lempel-Ziv complexity of the symbol sequence, denoted as the first parameter; Calculate the signal energy of the ideal causal prediction curve and normalize it, denoted as the second parameter; Calculate the net causal contribution based on the first and second parameters.

4. The IoT-based energy-saving control method for centralized heating according to claim 3, characterized in that, The discretization of the temperature residual sequence into a symbol sequence specifically involves: Preset a first threshold and a second threshold; When the residual value of any element in the temperature residual sequence is greater than the first threshold, 1 is used as the sign value of the corresponding element. When the residual value of any element in the temperature residual sequence is less than the second threshold, -1 is used as the sign value of the corresponding element; When the residual value of any element in the temperature residual sequence does not meet the conditions of being greater than the first threshold and less than the second threshold, the sign value of the element corresponding to 0 will be changed. Thus, the sequence of sign values ​​corresponding to all elements in the temperature residual value sequence is taken as the sign sequence.

5. The IoT-based energy-saving control method for centralized heating according to claim 1, characterized in that, The specific identification of the temperature sensor is as follows: Collect statistics on all valve action events between the current temperature sensor and valve pair within a preset time window, and obtain the opening change for each valve action event; The net causal contribution of valve action events is used as a weight, combined with the change in valve opening, to calculate the deviation of the temperature sensor-valve pair. Based on the deviation distribution of all temperature sensor-valve pairs, a deviation threshold is determined. When the deviation exceeds the deviation threshold, the corresponding temperature sensor is marked as a deviation temperature sensor.

6. The IoT-based energy-saving control method for centralized heating according to claim 1, characterized in that, The aforementioned spatial causal re-association specifically refers to: For any off-temperature sensor, calculate the potential causal strength between the off-temperature sensor and each other valve; The valve corresponding to the maximum potential causal strength is identified as a candidate valve; When the ratio of the potential causal strength of a candidate valve to the correlation strength of the valve originally associated with the temperature sensor is greater than a preset multiple, and the potential causal strength of the candidate valve is greater than the lower bound of the normal correlation strength of the system, the candidate valve is determined to meet the re-association condition. The temperature sensor is decoupled from the original associated valve, and a new association relationship is established between the temperature sensor and the candidate valve that meets the re-association condition.

7. The IoT-based energy-saving control method for centralized heating according to claim 1 or 6, characterized in that, The virtual thermal demand fusion correction specifically refers to: A potential causal threshold is preset. For temperature sensors that cannot be spatially re-correlated, the relevant valve set of the temperature sensors is selected based on the potential causal strength and the potential causal threshold. Obtain the spatial distance and time delay consistency factor between the valve and the temperature sensor. Based on the potential causal strength, spatial distance and time delay consistency factor between the valve and the temperature sensor in the relevant valve set, determine the attention weight of each valve. Based on the attention weights of each valve and its historical opening status, the virtual thermal demand deviating from the temperature sensor is calculated.

8. The IoT-based energy-saving control method for centralized heating according to claim 1, characterized in that, The embedding of the dynamic time delay-gain spectrum into the feedforward-feedback composite control framework specifically refers to: Feedforward control section: Based on the ratio of the target temperature adjustment to the current dynamic gain, the feedforward control signal is determined, and the command transmission time is determined based on the dynamic time delay to achieve active time delay compensation; Feedback control section: Based on the deviation between the corrected heat demand signal and the target temperature, a PID control algorithm is used to generate a feedback control signal; The proportional gain of the PID controller is dynamically adjusted based on the changing trend of the potential causal strength.

9. The IoT-based energy-saving control method for centralized heating according to claim 8, characterized in that, The dynamic adjustment of the proportional gain of the PID controller is specifically as follows: Calculate the average value of the global potential causal strength at the current moment, and determine the adjustment coefficient of the proportional gain by combining the deviation between the current global causal correlation strength and the historical statistical level; The initial proportional gain of the PID controller is corrected by adjusting the coefficients to obtain the corrected proportional gain.

10. An Internet of Things (IoT) energy-saving control system for centralized heating, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the Internet of Things energy-saving control method for centralized heating as described in any one of claims 1 to 9.

Images