Gas-fired hot water boiler based on efficient heat exchange and control method thereof

Abstract

The application discloses a gas hot water boiler based on efficient heat exchange and a control method thereof, relates to the technical field of intelligent control, and aims to solve the technical problem of low long-term operation efficiency and serious transient energy efficiency loss of the gas hot water boiler caused by the fact that a traditional static efficiency model cannot be calibrated online and adapt to dynamic disturbance working conditions, and the application comprises the following steps: an efficient heat exchanger is used for heating cold water flowing through the inside of the efficient heat exchanger; a combustion system is used for providing a heat source for the efficient heat exchanger; a water inlet temperature sensor, a water outlet temperature sensor, a water flow sensor, a gas flow sensor and a fan are used for collecting operation parameters of the system; and a controller is connected with each sensor, the combustion system and the fan. The application fundamentally solves the misalignment problem of the static efficiency model caused by equipment aging and working condition drift by constructing an efficiency model updating mechanism with online self-learning capability, and realizes the continuous and efficient operation of the gas hot water boiler in the whole life cycle.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control, more particularly, to a gas hot water boiler based on efficient heat exchange and a control method thereof. BACKGROUND

[0002] As a widely used commercial and industrial hot water supply device, the energy efficiency of the gas hot water boiler is directly related to the energy consumption cost of the user and the total energy consumption of the society. At present, one of the mainstream technical routes to improve the operating efficiency of the gas hot water boiler is to use a high-efficiency heat exchanger and to optimize the control of the combustion system through a pre-stored efficiency MAP (Modeling and Analysis Platform) diagram, so that the system can achieve high-efficiency operation near a specific working condition point.

[0003] However, this static optimization strategy based on a fixed efficiency model faces fundamental challenges in actual complex application environments. During the long-term use of the gas hot water boiler, the heat exchange performance of the heat exchanger will gradually deteriorate due to reasons such as scale deposition and carbon deposition, and factors such as fluctuations in gas source composition and drift of sensor parameters will also cause the actual operating characteristics of the system to gradually deviate from the ideal model calibrated at the factory.

[0004] This causes the pre-stored static efficiency MAP diagram to gradually "distort", and the optimal path planned by the controller according to this "distorted" MAP diagram cannot guide the system to reach the real high-efficiency point, so the long-term operating efficiency will inevitably decrease. More seriously, under the condition of "multi-point concurrent water use" caused by the simultaneous start and stop of multiple water points, the system will instantaneously deviate from the steady-state high-efficiency zone, and the traditional control strategy, which prioritizes water temperature stability, often uses a method of greatly adjusting the combustion state, which causes the system to run in a low-efficiency region during the dynamic process for a long time, resulting in considerable energy waste.

[0005] Therefore, how to enable the control system of the gas hot water boiler to have self-calibration capability and still intelligently maintain the high-efficiency operation of the system under complex dynamic disturbance conditions, rather than simply pursuing temperature stability, has become a core technical problem that needs to be solved and has a strong pertinence in the field. In view of this, the present application provides a gas hot water boiler based on efficient heat exchange and a control method thereof. SUMMARY

[0006] The present application aims to provide a gas hot water boiler based on efficient heat exchange and a control method thereof, to solve the technical problem that the traditional static efficiency model cannot be calibrated online and adapted to dynamic disturbance conditions, resulting in low long-term operating efficiency and serious transient energy loss of the gas hot water boiler.

[0007] To solve the above technical problems, the present application provides the following technical solution: a gas hot water boiler based on efficient heat exchange, comprising:

[0008] A high-efficiency heat exchanger for heating cold water flowing through its interior;

[0009] A combustion system for providing a heat source for the high-efficiency heat exchanger;

[0010] An inlet water temperature sensor, an outlet water temperature sensor, a water flow sensor, a gas flow sensor, and a fan for collecting operating parameters of the system;

[0011] A controller connected to each sensor, the combustion system, and the fan;

[0012] The controller is configured to perform:

[0013] A baseline efficiency MAP graph based on the thermal characteristics of the high-efficiency heat exchanger is pre-stored, which is used to represent the theoretical thermal efficiency of the system under different combinations of inlet water temperature, water flow, and gas flow;

[0014] During system operation, actual operating parameters are continuously collected, actual instantaneous thermal efficiency is calculated, and it is compared with the theoretical thermal efficiency of the corresponding working condition point in the baseline efficiency MAP graph;

[0015] When the deviation between the actual instantaneous thermal efficiency and the theoretical thermal efficiency continuously exceeds a pre-set range, an online learning correction process is started, the parameters of the local area that has deviated in the baseline efficiency MAP graph are fine-tuned, and an updated current efficiency MAP graph is generated;

[0016] Based on the current efficiency MAP graph and the real-time and predicted load state of the system, an optimal efficiency control trajectory is dynamically generated or updated;

[0017] The combustion system and the fan are controlled to drive the working point of the system to operate along the optimal efficiency control trajectory.

[0018] The present application solves the problem of static efficiency model deviation caused by equipment aging and working condition drift by constructing an efficiency model update mechanism with online self-learning ability, and realizes the continuous and efficient operation of the gas-fired hot water boiler throughout its life cycle. The online learning correction process proposed in the present application can automatically detect and quantify the model deviation by continuously comparing the actual operating efficiency with the model prediction value. Once the deviation continuously exceeds a reasonable range, the system starts the adaptive algorithm to gradually fine-tune the local efficiency MAP parameters, generating a current efficiency MAP graph that matches the current hardware state in real time. This process enables the cognitive model of the control system to evolve synchronously with the performance changes of the physical system, breaking the deadlock in traditional control where the model becomes outdated once it is fixed. As a result, no matter how long the hot water boiler is used, the efficiency MAP graph relied on by the control core always maintains high fidelity, ensuring that the control decisions always target the true optimal efficiency of the system, laying a foundation for long-term and stable energy saving.

[0019] Preferably, the online learning correction process specifically comprises:

[0020] When the system is running at a quasi-steady state operating point, record the actual operating parameter group and the calculated actual instantaneous thermal efficiency of the point ;

[0021] Calculate the instantaneous efficiency deviation value , wherein is based on the old parameter vector and the theoretical thermal efficiency prediction value obtained from the reference efficiency MAP;

[0022] If the instantaneous efficiency deviation value exceeds the preset deviation threshold value in a plurality of consecutive sampling periods, it is determined that the regional MAP needs to be corrected;

[0023] Based on the actual instantaneous thermal efficiency and the corresponding operating parameter group, an adaptive algorithm is used to gradually adjust the parameters of the corresponding region in the reference efficiency MAP until the instantaneous efficiency deviation value converges within the deviation threshold value, and the update of the current efficiency MAP is completed;

[0024] A confidence evaluation mechanism is established for the current efficiency MAP, and the confidence level is updated by , wherein is a forgetting factor, is a data quality score calculated based on the steady-state efficiency deviation after learning convergence and the cumulative number of effective data points, , wherein is a deviation penalty coefficient, is a data quantity saturation coefficient, represents the steady-state efficiency deviation after learning convergence, represents the cumulative number of effective data points used to correct the region model;

[0025] When generating the optimal efficiency control trajectory, the system operating point is preferentially guided to pass through the region with high confidence level.

[0026] Preferably, dynamically generating or updating the optimal efficiency control trajectory specifically comprises:

[0027] According to the current inlet water temperature, the target outlet water temperature and the water flow variation trend, predict the future load change path of the system;

[0028] starting from the current operating point of the system, and guided by the predicted load change path, a continuous path is sought in the multi-dimensional efficiency space defined by the current efficiency MAP, connecting the starting point and the expected end point, with the path satisfying the condition that the projection on the current efficiency MAP corresponds to a value of thermal efficiency as high as possible, and the path is smooth and satisfies the constraints of combustion stability and water temperature fluctuation;

[0029] The seeking of the continuous path is realized by solving a multi-objective optimization problem, by minimizing the continuous-time objective function , where is the expected dimensionless thermal efficiency at time on the trajectory, is the rate of change of the control variable vector at time , is the characteristic vector of the control variable rate of change, and are the weight coefficients.

[0030] Preferably, when a step change in water flow or multiple changes in flow with different amplitudes are detected, it is determined that the system enters a multi-point concurrent water disturbance working condition;

[0031] In this condition, the controller is further configured to:

[0032] real-time calculation of the normalized water flow disturbance intensity , where and are the instantaneous rate of change and acceleration of water flow, respectively, is the characteristic value, is the characteristic time constant, and are the proportional coefficients;

[0033] According to the normalized water flow disturbance intensity , the weight coefficients and in the objective function are dynamically adjusted;

[0034] In the initial stage of severe disturbance, the restriction on water temperature fluctuation is temporarily relaxed by adjusting the weight coefficients, allowing the operating point to quickly cross the possible low-efficiency region;

[0035] After the flow tends to be stable, the restriction on water temperature fluctuation is tightened, guiding the operating point to the high-efficiency point in the new steady state.

[0036] Preferably, the controller is further configured to perform high-efficiency standby control suitable for frequent start-stop water mode, including:

[0037] identifying short, intermittent water usage patterns based on historical water usage data;

[0038] when a water usage event ends and is identified as the short water usage pattern, instead of immediately shutting down the combustion system, the system operating point is adjusted to a predetermined high-efficiency standby point in the low-load region of the current efficiency MAP;

[0039] at the predetermined high-efficiency standby point, a minimum level of combustion power and circulation water flow is maintained, keeping the heat exchanger core temperature at a level higher than ambient temperature and with a high thermal efficiency;

[0040] when a next water usage request is predicted or detected, the system is quickly loaded from the predetermined high-efficiency standby point to the target power.

[0041] Preferably, the predetermined high-efficiency standby point is not a fixed value, but is selected by solving a constrained optimization problem, whose objective is to find an operating point in the feasible set of low-load operating conditions such that the estimated reheat time from this point to the target power point does not exceed an allowed maximum value while maximizing the normalized thermal efficiency of this point , mathematically described as , with the constraint .

[0042] A control method for a gas-fired hot water boiler, comprising the following steps:

[0043] S1: the system pre-stores a reference efficiency MAP based on the thermal characteristics of the high-efficiency heat exchanger;

[0044] S2: real-time acquisition of system operating parameters and calculation of actual instantaneous thermal efficiency;

[0045] S3: continuous comparison and learning of the actual instantaneous thermal efficiency with the theoretical value of the reference efficiency MAP, when a continuous deviation is detected, online correction and update to the current efficiency MAP;

[0046] S4: based on the current efficiency MAP and the real-time load state of the system, dynamically generating an optimal efficiency control trajectory;

[0047] S5: controlling the combustion system and the fan to drive the system operating point to run along the optimal efficiency control trajectory.

[0048] Preferably, the step S4 of dynamically generating an optimal efficiency control trajectory is achieved by solving a discrete-time model predictive control problem, at each control period , solving a future ​Predict the optimal control sequence in the time domain step by step. ;

[0049] The constraints are: , , ;

[0050] In the formula, For control sequences, This is the predicted state value. To control actions.

[0051] Preferably, when multiple concurrent water usage disturbances are detected, step S4 further includes:

[0052] Based on real-time calculated normalized water flow disturbance intensity The constraints and boundaries in the model predictive control problem are dynamically adjusted.

[0053] Normalized boundary constraining water temperature fluctuations Adjusted to: In the formula, The upper limit of allowable fluctuations in base water temperature. For adjustment coefficients;

[0054] In a strong disturbance transient, by increasing Slight fluctuations in water temperature are allowed to prioritize efficiency and system stability; however, this constraint is tightened in the later stages of the disturbance to quickly stabilize the water temperature.

[0055] Preferably, the high-efficiency standby point Dynamic selection involves evaluating a set of candidate low-load points in each decision cycle. Selection of comprehensive benefit score To achieve this, select the point with the highest score;

[0056] The formula for calculating the comprehensive benefit score is as follows:

[0057] ;

[0058] In the formula, This represents the normalized thermal efficiency at the candidate point. To estimate the reheat time, The characteristic constant of reheat time, To normalize maintenance energy consumption, , , These are the weighting coefficients.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] 1. This invention fundamentally solves the problem of static efficiency model inaccuracy caused by equipment aging and operating condition drift by constructing an efficiency model update mechanism with online self-learning capabilities, thus achieving continuous and efficient operation of the gas-fired hot water boiler throughout its entire lifecycle. The online learning and correction process proposed in this invention automatically detects and quantifies model deviations by continuously comparing actual operating efficiency with model predictions. Once the deviation exceeds a reasonable range, the system initiates an adaptive algorithm to progressively fine-tune the local efficiency MAP parameters, generating a current efficiency MAP that matches the current hardware state in real time. This process allows the cognitive model of the control system to evolve synchronously with the performance changes of the physical system, breaking the deadlock of traditional control where models gradually become outdated once they are fixed. As a result, regardless of how long the hot water boiler has been in use, the efficiency MAP on which its control core relies always maintains high fidelity, ensuring that control decisions always aim at the true optimal system efficiency, laying the foundation for long-term and stable energy-saving effects.

[0061] 2. This invention also effectively overcomes the industry-wide problem of efficiency drops during transient processes by designing an intelligent disturbance rejection control strategy based on dynamic efficiency trajectory planning and elastic constraints. This invention innovatively constructs dynamic process control as a multi-objective optimization problem. Based on a real-time updated efficiency MAP, it predicts and generates an optimal efficiency control trajectory that balances high thermal efficiency and control stability. Specifically, when strong water flow disturbances are detected, the system can dynamically calculate the disturbance intensity and intelligently adjust the weights of the control strategy accordingly: in the early stages of the disturbance, the restrictions on instantaneous water temperature fluctuations are appropriately relaxed, allowing the operating point to quickly and smoothly traverse potential inefficiency zones, prioritizing the system's avoidance of inefficient or unstable states; after the disturbance subsides, the control is rapidly tightened, guiding the system to accurately return to the steady-state high-efficiency point. This strategy, combining elastic constraints and trajectory planning, transforms the system from a passive, temperature-stabilized reactive response to an active, globally optimal intelligent management approach, significantly improving overall energy efficiency in complex water use scenarios.

[0062] 3. This invention also addresses the high standby point selection strategy based on multi-objective optimization, solving the efficiency problem under dynamic operating conditions while further overcoming the technical pain point of high standby energy consumption under frequent start-stop modes, thus achieving energy efficiency improvement across the entire operating cycle. After achieving efficient operation in steady-state and dynamic processes, energy waste under frequent start-stop modes (such as kitchen washing) becomes the next energy-saving bottleneck. This invention identifies short-term intermittent patterns by analyzing water usage history and innovatively proposes maintaining the system at an optimized, highly efficient standby point during water usage intervals, rather than completely shutting it down. The selection of this standby point is not fixed but dynamically determined by solving a constrained optimization problem. Its core is to maximize thermal efficiency during standby while meeting the user's requirements for reheating speed. This essentially finds the optimal balance between reheating response speed and standby energy consumption. When the next water usage is predicted, the system can quickly load from this highly efficient standby point, greatly reducing energy consumption and waiting time from cold start, thereby extending energy-saving optimization from a single operating process to a complete start-stop cycle, achieving a dual improvement in energy saving effect and user experience. Attached Figure Description

[0063] Figure 1 This is a schematic diagram of the system architecture of the present invention;

[0064] Figure 2 This is the overall control flowchart of the present invention;

[0065] Figure 3 This is a flowchart of the online learning modification sub-process of the present invention;

[0066] Figure 4 This is a flowchart of the dynamic trajectory generation and disturbance rejection control sub-process of the present invention;

[0067] Figure 5 This is a flowchart of the efficient standby control sub-process of the present invention. Detailed Implementation

[0068] Example 1: As Figure 1 As shown, the present invention relates to a gas-fired hot water boiler based on high-efficiency heat exchange, comprising:

[0069] High-efficiency heat exchangers are used to heat cold water flowing through them;

[0070] A combustion system is used to provide a heat source for the high-efficiency heat exchanger;

[0071] Inlet water temperature sensor, outlet water temperature sensor, water flow sensor, gas flow sensor, and fan are used to collect system operating parameters;

[0072] The controller is connected to the various sensors, the combustion system, and the fan.

[0073] It should be noted that, unless otherwise specified, in this invention, the system refers to the entire gas-fired hot water boiler comprising the aforementioned high-efficiency heat exchanger, combustion system, sensor group (inlet water temperature sensor, outlet water temperature sensor, water flow sensor, and gas flow sensor), fan, and controller, and operating the control method disclosed in this invention. This term will refer to this integrated entity in the following description.

[0074] The controller is configured to execute:

[0075] Based on the thermal characteristics of the high-efficiency heat exchanger, its baseline efficiency MAP is pre-stored. The baseline efficiency MAP represents the theoretical thermal efficiency of the system under different combinations of inlet water temperature, water flow rate and gas flow rate.

[0076] During system operation, actual operating parameter sets are continuously collected, and the corresponding actual instantaneous thermal efficiency is calculated.

[0077] The actual instantaneous thermal efficiency is compared with the theoretical thermal efficiency at the corresponding operating point in the benchmark efficiency MAP. When the deviation continues to exceed the allowable range, it is determined that the benchmark efficiency MAP has shifted, and an online learning and correction process is initiated.

[0078] Through the online learning correction process, the parameters of the baseline efficiency MAP of the local area where the offset has occurred are fine-tuned to generate an updated current efficiency MAP.

[0079] Based on the current efficiency MAP, an optimal efficiency control trajectory is dynamically generated or updated for the current and predicted load state of the system.

[0080] The combustion system and the fan are controlled to drive the system's operating point along the optimal efficiency control trajectory.

[0081] In an embodiment of the present invention, the online learning correction process specifically includes:

[0082] When the system is operating at a quasi-steady-state condition point, record the actual operating parameter set at that point and the calculated actual instantaneous thermal efficiency;

[0083] The actual instantaneous thermal efficiency is compared with the corresponding theoretical thermal efficiency obtained from the reference efficiency MAP chart, and the instantaneous efficiency deviation value is calculated.

[0084] If the instantaneous efficiency deviation value exceeds the preset deviation threshold in multiple consecutive sampling periods, it is determined that the MAP map of the region needs to be corrected.

[0085] Based on the actual instantaneous thermal efficiency and the corresponding set of operating parameters, one or more influencing parameters in the corresponding region of the baseline efficiency MAP are progressively adjusted using recursive least squares or neural network algorithms until the instantaneous efficiency deviation value converges to within the deviation threshold, thereby completing the update of the current efficiency MAP.

[0086] Wherein, the instantaneous efficiency deviation value The calculation formula is as follows:

[0087] ;

[0088] In the formula:

[0089] For the first The instantaneous efficiency deviation at each sampling moment is a dimensionless relative error index.

[0090] For the first The actual thermal efficiency at each sampling moment is calculated based on the temperature and flow data measured by the sensor in real time.

[0091] For the old parameter vector The theoretical thermal efficiency prediction value is obtained by interpolation or calculation from the baseline efficiency MAP diagram and the current operating conditions (such as inlet water temperature and flow rate);

[0092] This represents the parameter vector used in the efficiency MAP diagram before this round of comparison and correction.

[0093] Operational Logic: This formula calculates the relative deviation between the system's measured thermal efficiency and the theoretical thermal efficiency predicted based on the old parameter model at a specific sampling time. By comparing the measured value with the model's predicted value and normalizing by dividing by the predicted value, a standardized instantaneous error metric is obtained. This metric is used to determine the model's prediction accuracy under the current operating conditions, providing immediate quantitative feedback on the model's prediction accuracy. By continuously monitoring this standardized deviation, the system can sensitively detect subtle deviations between the model's predicted value and the actual performance, thus providing a reliable and objective criterion for triggering the self-learning correction process and ensuring that the model upon which control depends always maintains high fidelity.

[0094] In an embodiment of the present invention, the controller is further configured to:

[0095] Establish a confidence evaluation mechanism for the current efficiency MAP plot;

[0096] For regions that have been updated through the online learning correction process, a higher confidence level is assigned based on the amount of data used for correction and data consistency.

[0097] For areas that are unverified or have sparse data, maintain a low confidence level;

[0098] When generating the optimal efficiency control trajectory, the system operating point is preferentially guided through regions with high confidence levels.

[0099] The confidence level The following formula is used for evaluation and updating:

[0100] ;

[0101] In the formula:

[0102] For the first The confidence level is updated after each calculation cycle. The higher the value, the more reliable the model is for that region.

[0103] For the first The confidence level for each calculation period (i.e., the previous period);

[0104] Forgetting factor This determines the extent to which historical confidence information is retained during updates. The closer it is to 1, the greater the influence of historical information and the smoother the update;

[0105] For the first Data quality score for each calculation period, ,in This is the deviation penalty coefficient, used to adjust the intensity of the impact of steady-state efficiency deviation on the score. The larger the value, the stronger the penalty for the same deviation on the score, and the more cautious the confidence increase. The value range is [value range missing]. , This is the data saturation coefficient, used to adjust the rate at which the effective data volume affects the saturation of the scoring. The larger the value, the more data is needed to improve the score. Avoid prematurely relying on partial data; its value should be... , The steady-state efficiency deviation after learning convergence reflects the final accuracy of the model after correction. This represents the cumulative number of valid data points used to correct the model for this region;

[0106] Operational Logic: This formula uses an exponentially weighted moving average to dynamically update the confidence levels of each region in the efficiency MAP. It weights and fuses the confidence level of the previous period with the data quality score of the current period. A forgetting factor controls the decay rate of historical information, ensuring that the confidence assessment tracks the latest data quality while maintaining a certain historical inertia, avoiding drastic fluctuations caused by single anomalous data. This dynamic confidence update mechanism endows the efficiency MAP with the attributes of a "confidence MAP." This allows the control system to intelligently distinguish between reliable and uncertain regions of the model. When planning efficient trajectories, it prioritizes and utilizes information from high-confidence regions, significantly improving the robustness and safety of control decisions and preventing control performance degradation due to local model distortion.

[0107] In an embodiment of the present invention, the dynamic generation or updating of an optimal efficiency control trajectory specifically includes:

[0108] Based on the current inlet water temperature, the target outlet water temperature, and the real-time water flow rate or flow rate change trend detected by the water flow sensor, the load change path of the system in the near future is predicted.

[0109] In the multidimensional efficiency space defined by the current efficiency MAP, starting from the current operating point of the system and guided by the predicted load change path, a continuous path connecting the starting point and the expected endpoint is found.

[0110] The continuous path satisfies the following constraints: the thermal efficiency value corresponding to the projection on the current efficiency MAP is as high as possible, and the path is smooth, satisfying the constraints of combustion stability and water temperature fluctuation.

[0111] The found continuous path is defined as the optimal efficiency control trajectory.

[0112] The generation of the optimal efficiency control trajectory is constructed as a multi-objective optimization problem, which is achieved by minimizing the continuous-time objective function defined below. Please provide a solution:

[0113] ;

[0114] In the formula:

[0115] The objective function value is a continuous-time objective, which is the comprehensive performance index that needs to be minimized.

[0116] and These are the start and end times for trajectory optimization, defining the time range for optimization.

[0117] Indicates at time The expected dimensionless thermal efficiency (using characteristic efficiency) along the candidate trajectory, obtained from the current efficiency MAP mapping. (Normalization)

[0118] Indicates at time The rate of change of the control variable vector (which typically includes the rate of change of gas valve opening and the rate of change of fan speed);

[0119] The eigenvector of the rate of change of the control variable is used to... Perform dimensional normalization to make the second term dimensionless;

[0120] and These are dimensionless weighting coefficients, used to balance efficiency optimization and control stationarity. The larger the value, the more the system tends to pursue high efficiency, potentially sacrificing some control stability. The larger the value, the smoother the control action, but the response may be slower and the efficiency slightly reduced. The possible values ​​are as follows: , , and It can be dynamically adjusted according to the operating conditions (steady-state / transient). During commissioning, it is recommended to keep the ratio within a certain range. Avoid extreme strategies;

[0121] Operational Logic: This formula defines an objective function for evaluating the overall performance of a candidate control trajectory in the continuous time domain. It includes two costs: the first penalizes deviations between the expected and ideal efficiencies of the trajectory, encouraging high efficiency; the second penalizes the rate of change of the control variables, encouraging smooth operation. By integrating these two terms from the start time to the end time and applying appropriate weights, the trajectory optimization problem is transformed into a mathematical problem of finding the minimum integral value. This objective function is the mathematical core for achieving multi-objective collaborative optimization. It cleverly unifies the sometimes conflicting objectives of maximizing efficiency and achieving operational smoothness within a single framework. The optimal trajectory obtained by minimizing this objective function is essentially the best trade-off path that the system can achieve between efficiency and smoothness within a given prediction period, thus realizing global optimization of system performance.

[0122] In an embodiment of the present invention, when the water flow sensor detects a step change in water flow or multiple flow changes of different amplitudes at the same time, the controller determines that the system has entered a multi-point concurrent water use disturbance condition.

[0123] Under this operating condition, when generating the optimal efficiency control trajectory, the controller is further configured as follows:

[0124] The weight of the "water temperature fluctuation limit" in the constraints is dynamically adjusted according to the magnitude and speed of the flow rate change.

[0125] In the early stages of severe disturbances, the restrictions on water temperature fluctuations are temporarily relaxed to increase the weight of thermal efficiency as much as possible and combustion stability, allowing the operating point to quickly pass through the possible inefficiency zone.

[0126] After the flow rate stabilizes, the restrictions on water temperature fluctuations are quickly tightened to guide the operating point to the high-efficiency point under the new steady state.

[0127] Under the multi-point concurrent water disturbance condition, the objective function Weighting coefficients in and Based on the real-time detected normalized water flow disturbance intensity Make dynamic adjustments, that is , Specifically, this involves using piecewise linear functions or table lookup methods, based on... The interval mapping is adjusted. and .

[0128] Among them, the normalized water flow disturbance intensity The instantaneous rate of change of water flow and its changing acceleration The formula, calculated together, is as follows:

[0129] ;

[0130] In the formula:

[0131] Indicates at time The normalized flow disturbance intensity is a dimensionless scalar.

[0132] This is the instantaneous rate of change of water flow, i.e., the first time derivative of the flow rate;

[0133] This is the instantaneous acceleration of the water flow rate, i.e., the second time derivative of the flow rate;

[0134] The characteristic value of the rate of change of water flow is used to... Perform normalization;

[0135] The characteristic time constant of the water flow disturbance is used for coordination. Its dimensions, making it consistent with Item compatibility;

[0136] and These are dimensionless proportionality coefficients, used to adjust the contribution weights of the rate of change and acceleration terms in the calculation of the total disturbance intensity, respectively. Their values ​​are as follows: , , The larger the value, the more sensitive the system is to mutations;

[0137] Operational Logic: This formula quantifies the severity of water flow disturbances. It combines two dynamic characteristics—the instantaneous rate of change of water flow (first derivative) and the acceleration (second derivative)—and normalizes them. The first term directly reflects the speed of flow change, while the second term, by introducing a characteristic time constant, converts the dimension of acceleration to be consistent with the rate of change, allowing the two to be added together. The final result is a dimensionless index characterizing the overall disturbance intensity. This formula achieves accurate perception and quantification of complex water use disturbance patterns. It not only captures the magnitude of flow change (velocity term) but is also sensitive to the abruptness of change (acceleration term), thus enabling a more accurate distinction between gradual load changes and severe, impactful disturbances. This provides precise operating condition identification input for downstream control strategies (such as dynamically adjusting the weights of the objective function), a key prerequisite for the system to achieve intelligent adaptive control.

[0138] In an embodiment of the present invention, the controller is further configured to perform efficient standby control suitable for frequent start-stop water usage patterns, specifically including:

[0139] Based on historical water usage data, identify short-term and intermittent water usage patterns;

[0140] When a water usage session ends and is identified as a short-term water usage mode, the combustion system is not immediately shut down. Instead, the system operating point is adjusted to a predetermined high-efficiency standby point in the low-load region of the current efficiency MAP.

[0141] At the predetermined high-efficiency standby point, a minimum amount of combustion power and circulating water flow are maintained to keep the core temperature of the heat exchanger at a level higher than the ambient temperature and with high thermal efficiency.

[0142] When the next water demand is predicted or detected, the system is rapidly loaded from the predetermined high-efficiency standby point to the target power.

[0143] In an embodiment of the present invention, the predetermined high-efficiency standby point is not a fixed value, but is dynamically selected based on the ambient temperature, the predicted value of the downtime, and the actual shape of the low-load area in the current efficiency MAP. The selection principle is that the instantaneous thermal efficiency of the point is the highest under the premise of satisfying the rapid reheat capability.

[0144] Among them, the predetermined high-efficiency standby point The selection is achieved by solving a constrained optimization problem, the objective of which is to find the feasible set under low load conditions. Find a work point This enables the power point to be reached from that point. Estimated reheat time Not exceeding the maximum allowed value At the same time, maximize the normalized thermal efficiency at that point. The mathematical description is as follows:

[0145] ;

[0146] ;

[0147] in, , This is estimated based on the system dynamic model, the current efficiency MAP, and the current ambient temperature.

[0148] In the formula:

[0149] The predetermined high-efficiency standby point to be selected (decision variable);

[0150] The set of all feasible low-load operating points, i.e., the search space of the optimization problem;

[0151] Indicates the candidate standby point Normalized thermal efficiency at runtime (objective function);

[0152] Indicates from candidate standby point Start and heat to the target power point Estimated time required;

[0153] This is the maximum allowable reheat time for the system (upper bound of the constraint).

[0154] The target power point set for the user, which is the working state that needs to be achieved the next time water is used;

[0155] Operational Logic: This formula describes a constrained optimization problem to find the standby operating point with the highest thermal efficiency while meeting reheat response time requirements. Its objective function is to maximize the normalized efficiency of the standby point, with the constraint that the time required to heat up from the standby point to the target power point must not exceed the maximum allowable value. Solving this problem means finding the most efficient point within the feasible region (low-load operating condition set) that satisfies the time constraint. This optimization problem elevates the selection of the standby point from experience or fixed rules to model-based scientific decision-making. It explicitly addresses the contradiction between "standby energy efficiency" and "reheat speed," ensuring that the selected standby point is not simply a high-efficiency point, but rather the most efficient point under the constraint of rapid response capability, thereby maximizing energy savings during intermittent periods while ensuring user experience (waiting time).

[0156] Example 2: Figures 2 to 5 As shown, a control method for a gas-fired hot water boiler includes the following steps:

[0157] S1: The system pre-stores a baseline efficiency MAP diagram based on the thermal characteristics of high-efficiency heat exchangers;

[0158] S2: Real-time acquisition of system operating parameters and calculation of actual instantaneous thermal efficiency;

[0159] S3: Continuously compare and learn from the actual instantaneous thermal efficiency with the theoretical value of the baseline efficiency MAP. When a continuous deviation is detected, correct it online and update it to the current efficiency MAP.

[0160] S4: Based on the current efficiency MAP and the real-time load status of the system, dynamically generate an optimal efficiency control trajectory;

[0161] S5: Control the combustion system and fan, and drive the system operating point to run along the optimal efficiency control trajectory.

[0162] In another embodiment of the present invention, the online correction and update in step S3 specifically includes:

[0163] Actual efficiency data were collected at the quasi-steady-state operating point;

[0164] Calculate the deviation between its theoretical efficiency value and the corresponding point in the baseline efficiency MAP plot;

[0165] If the deviation continues to exceed the limit, an adaptive filtering algorithm is used to progressively adjust the local parameters of the baseline efficiency MAP to generate a current efficiency MAP that better reflects the current hardware status.

[0166] The adaptive filtering algorithm updates the parameter vector of the efficiency MAP map in a recursive manner. Its discrete-time update formula is as follows:

[0167] ;

[0168] In the formula:

[0169] For the first The estimated values ​​of the model parameter vector updated after each iteration (or at each time step);

[0170] For the first The estimated values ​​of the model parameter vector for each iteration (or the previous time step);

[0171] For the first The adaptive gain matrix of the step determines the degree of influence of the prediction error on the parameter correction amount, and is usually calculated online based on the error covariance and measurement noise;

[0172] Based on parameters from the previous time step and the current input regression vector The model predicts the thermal efficiency based on the calculated parameters (including inlet water temperature, flow rate, etc.).

[0173] Operational Logic: This formula is the core recursive update step of adaptive filtering algorithms (such as recursive least squares). Based on the current model prediction error (the difference between the measured and predicted values), it corrects the model parameter vector using a dynamically calculated gain matrix. The new parameter estimate equals the old parameter estimate plus the correction amount after adjusting the prediction error with the gain matrix, thus allowing the model's predicted output to gradually approach the actual measured value. This recursive update algorithm endows the efficiency model with the ability to learn throughout its life. It enables the model to continuously and gradually correct its parameters using online operating data, thereby adaptively compensating for gradual changes in system performance caused by heat exchanger fouling, gas quality changes, sensor drift, etc. This ensures that the brain (efficiency model) upon which the control algorithm relies is always synchronized with the actual state of the body (physical hardware), which is the cornerstone for achieving long-term stable and efficient operation.

[0174] In another embodiment of the present invention, the step S4 of dynamically generating the optimal efficiency control trajectory specifically includes:

[0175] Based on real-time inlet water temperature, target outlet water temperature, and water flow information, predict load changes;

[0176] In the multidimensional efficiency space formed by the current efficiency MAP diagram, under the premise of satisfying combustion stability and water temperature fluctuation constraints, and with the goal of optimizing the overall thermal efficiency of the system, the optimal path from the current operating point to the predicted endpoint is solved, and this path is the optimal efficiency control trajectory.

[0177] The optimal efficiency control trajectory is generated by solving a discrete-time model predictive control problem in each control cycle. Seeking solutions for the future Predict the optimal control sequence in the time domain step by step. :

[0178] ;

[0179] ;

[0180] ;

[0181] ;

[0182] In the formula:

[0183] For at any time The optimal control sequence obtained by solving;

[0184] The future control input sequence to be optimized. , Indicates the future number Step control actions;

[0185] To predict the length of the time domain, i.e., the number of steps to look forward in the optimization problem;

[0186] Represents the discretized instantaneous cost function, evaluating the value in the predicted state. Apply control The cost;

[0187] For at any time For the future The predicted value of the system state at any given time;

[0188] A discrete-time system dynamic model that describes how the state changes with control and the previous state.

[0189] and These represent discretized inequality constraints (such as temperature and pressure limits) and equality constraints (such as mass conservation), respectively.

[0190] Operational logic: This formula describes the standard form of a discrete-time Model Predictive Control (MPC) problem. At each current time step... The controller solves a finite-time domain ( Open-loop optimization problem (step 1): finding a set of future control inputs This allows us to start from the current predicted state and proceed along the system model. The trajectory of evolution, and its cumulative costs Minimum, while satisfying all process constraints and The first element of the optimal control sequence obtained from the solution is applied to the system, and this process is repeated in the next cycle. The Model Predictive Control Framework transforms optimal control from an offline trajectory-tracking problem into an online, iteratively solved rolling optimization problem. This iterative planning and step-by-step strategy enables the system to continuously revise its predictions of future dynamics and re-optimize the control law based on the latest measurement information, thus naturally possessing the ability of feedforward compensation and feedback correction, and exhibiting extremely strong robustness to model errors and external disturbances.

[0191] In another embodiment of the present invention, when a combined disturbance caused by concurrent water use at multiple points is detected, step S4 further includes:

[0192] Based on the intensity of the disturbance, the weight allocation of the constraints is dynamically optimized; in the transient state of strong disturbance, priority is given to ensuring that the efficiency trajectory does not enter the inefficient or dangerous area, and slight fluctuations in water temperature are allowed to a certain extent; in the later stage of the disturbance, priority is given to ensuring that the water temperature quickly stabilizes to the set value.

[0193] In the model predictive control problem, the discretized instantaneous cost function and constraints The boundary will be determined based on the normalized flow disturbance intensity calculated in real time. Adjustments will be made.

[0194] Specifically, the normalized boundary of the water temperature fluctuation constraint Adjusted to:

[0195] ;

[0196] in:

[0197] Indicates at discrete time Dynamically adjusted, normalized upper limit for water temperature fluctuations;

[0198] This is the normalized upper limit of the allowable fluctuation of the base water temperature, which is a constraint that must be strictly observed under stable operating conditions.

[0199] The dimensionless adjustment coefficient determines the intensity of the water flow disturbance. The extent of the impact on the amount of constraint relaxation The larger the value, the greater the allowable water temperature fluctuation during disturbances, and the more aggressive the system response. Its value is... ;

[0200] Indicates at discrete time The normalized flow disturbance intensity was calculated.

[0201] Operational Logic: This formula defines a dynamic adjustment strategy for the upper limit of water temperature fluctuation constraints. The basic idea is that when a strong water flow disturbance is detected, to prioritize preventing the system from entering an inefficient or unstable state, a larger instantaneous fluctuation in water temperature can be temporarily allowed. Specifically, the basic constraint upper limit is added to an increment proportional to the real-time disturbance intensity, resulting in a dynamically relaxed constraint boundary. The stronger the disturbance, the larger the allowed fluctuation range. This dynamic constraint adjustment mechanism is key to the intelligent flexibility of the control system in the face of extreme conditions. It breaks the limitation of rigid and unchanging constraints in traditional control, allowing the control system to maintain efficiency and stability during brief critical moments when disturbances occur, by gaining higher degrees of operational freedom. Once the disturbance subsides, the constraints are quickly tightened to ensure comfort. This flexible constraint concept enables the system to achieve a high-order unity of stability, comfort, and efficiency in complex dynamic environments.

[0202] In another embodiment of the present invention, the method further includes an efficient standby control step:

[0203] Identify water usage patterns that frequently start and stop;

[0204] During short-term water usage breaks, the control point is moved to the high-efficiency standby point in the low-load area of ​​the current efficiency MAP for heat preservation.

[0205] At the start of the next water usage cycle, it will quickly start from the high-efficiency standby point.

[0206] In another embodiment of the present invention, the location of the high-efficiency standby point is dynamically optimized based on the ambient temperature, expected downtime and current efficiency MAP.

[0207] The high-efficiency standby point The dynamic selection is achieved by solving a simplified real-time evaluation problem in each decision cycle: from a set of candidate low-load points In the middle, select the comprehensive benefit score. The highest point.

[0208] ;

[0209] In the formula:

[0210] Indicates candidate standby point Overall benefit score;

[0211] Indicate candidate points Normalized thermal efficiency;

[0212] To start from candidate points Heat to target point The estimated reheat time;

[0213] This is the characteristic constant of reheat time, used to normalize the reheat time (represented in the formula as dividing by). );

[0214] For candidate points Normalized fuel consumption per unit time;

[0215] , , These are dimensionless weighting coefficients, representing the degree of importance attached to the three dimensions of efficiency, response speed, and operating cost, respectively, with values ​​of [values ​​to be filled in]. , , , , , It can learn and update based on user habits (energy-saving preferences / instant heating preferences).

[0216] Calculation Logic: This formula is used to evaluate and compare the overall merits of different candidate standby points. It uses a weighted scoring function to combine three key standby point attributes—static thermal efficiency, dynamic reheat time, and steady-state sustaining energy consumption—into a total score. Each attribute is normalized to eliminate dimensions and weighted by an appropriate coefficient. , , The score reflects the relative importance of each attribute in the decision-making process. The candidate point with the highest total score is selected as the final standby point. This comprehensive scoring function quantifies and automates multi-attribute decision-making. It transforms the complex trade-off process of selecting a standby point into a computable optimization problem, avoiding the one-sidedness of single-index decision-making. By reasonably setting weights, it can flexibly adapt to different user preferences (such as extreme energy-saving mode or fast response mode), enabling the standby strategy to achieve the optimal balance between energy efficiency, experience, and cost as expected by users, thereby improving the system's intelligence and user satisfaction.

[0217] All weight parameters in this application are dimensionless, and their specific numerical values ​​reflect the relative importance between different targets, rather than absolute physical quantities.

[0218] The embodiments disclosed in this invention are preferred embodiments, but are not limited thereto. Those skilled in the art can easily understand the spirit of this invention based on the above embodiments and make different extensions and variations, but as long as they do not depart from the spirit of this invention, they are all within the protection scope of this invention.

Claims

1. A gas-fired hot water boiler based on high-efficiency heat exchange, characterized in that, include: High-efficiency heat exchangers are used to heat cold water flowing through them; A combustion system is used to provide a heat source for the high-efficiency heat exchanger; Inlet water temperature sensor, outlet water temperature sensor, water flow sensor, gas flow sensor, and fan are used to collect the system's operating parameters; The controller is connected to each sensor, the combustion system, and the fan via signals. The controller is configured to perform: A baseline efficiency MAP diagram based on the thermal characteristics of the high-efficiency heat exchanger is pre-stored. This diagram is used to characterize the theoretical thermal efficiency of the system under different combinations of inlet water temperature, water flow rate, and gas flow rate. During system operation, actual operating parameters are continuously collected, the actual instantaneous thermal efficiency is calculated, and it is compared with the theoretical thermal efficiency of the corresponding operating point in the reference efficiency MAP diagram. When the deviation between the actual instantaneous thermal efficiency and the theoretical thermal efficiency continues to exceed the preset range, an online learning correction process is initiated to fine-tune the parameters of the local area in the benchmark efficiency MAP that has shifted, and generate an updated current efficiency MAP. Based on the current efficiency MAP and the real-time and predicted load status of the system, an optimal efficiency control trajectory is dynamically generated or updated. The combustion system and the fan are controlled to drive the system's operating point along the optimal efficiency control trajectory.

2. The gas-fired hot water boiler based on high-efficiency heat exchange according to claim 1, characterized in that, The online learning correction process specifically includes: When the system is operating at the quasi-steady-state point, record the actual operating parameter set at that point and the calculated actual instantaneous thermal efficiency. ; Calculate the instantaneous efficiency deviation value In the formula, For the old parameter vector And the theoretical thermal efficiency prediction obtained from the baseline efficiency MAP plot; If the instantaneous efficiency deviation value exceeds the preset deviation threshold in multiple consecutive sampling periods, it is determined that the MAP map of that region needs to be corrected. Based on the actual instantaneous thermal efficiency and the corresponding set of operating parameters, an adaptive algorithm is used to progressively adjust the parameters of the corresponding region in the baseline efficiency MAP until the instantaneous efficiency deviation value converges to within the deviation threshold, thus completing the update of the current efficiency MAP. A confidence evaluation mechanism is established for the current efficiency MAP plot, and the confidence level is determined by... To update, in the formula, For the first The confidence level updated after each calculation cycle; Forgetting factor, This is a data quality score calculated based on the steady-state efficiency deviation after learning convergence and the cumulative number of valid data points. ,in This is the deviation penalty coefficient. This is the data saturation coefficient. This represents the steady-state efficiency deviation after learning convergence. This represents the cumulative number of valid data points used to correct the model for this region; When generating the optimal efficiency control trajectory, the system operating point is preferentially guided through regions with high confidence levels.

3. A gas-fired hot water boiler based on high-efficiency heat exchange according to claim 2, characterized in that, Dynamically generating or updating the optimal efficiency control trajectory specifically includes: Based on the current inlet water temperature, target outlet water temperature, and water flow rate trends, predict the future load change path of the system; Starting from the current operating point of the system and guided by the predicted load change path, find a continuous path connecting the starting point and the expected endpoint in the multidimensional efficiency space defined by the current efficiency MAP. This path must satisfy the requirement that the thermal efficiency value corresponding to the projection on the current efficiency MAP is as high as possible, and that the path is smooth and meets the constraints of combustion stability and water temperature fluctuation. The search for the continuous path is achieved by solving a multi-objective optimization problem, which minimizes the continuous-time objective function. To solve, in the formula, For the moment on the trajectory The expected dimensionless thermal efficiency. To control the vector of variables at time 10:00 rate of change, The eigenvector representing the rate of change of the control variable. and These are the weighting coefficients.

4. A gas-fired hot water boiler based on high-efficiency heat exchange according to claim 3, characterized in that, When a step change or multiple flow rate changes of different magnitudes are detected in the water flow rate, the system is determined to have entered a multi-point concurrent water use disturbance condition. Under this operating condition, the controller is further configured as follows: Real-time calculation of normalized water flow disturbance intensity In the formula, and These represent the instantaneous rate of change and acceleration of the water flow rate, respectively. For eigenvalues, The characteristic time constant, and This is the proportionality coefficient; According to the normalized water flow disturbance intensity Dynamically adjust the objective function Weighting coefficients in and ; In the early stages of severe disturbances, the restrictions on water temperature fluctuations are temporarily relaxed by adjusting the weighting coefficients, allowing the operating point to quickly traverse the potential inefficiency zone. After the flow rate stabilizes, the restrictions on water temperature fluctuations are tightened to guide the operating point to a new, efficient point under steady-state conditions.

5. A gas-fired hot water boiler based on high-efficiency heat exchange according to claim 4, characterized in that, The controller is also configured to perform efficient standby control suitable for frequent start-stop water usage patterns, including: Based on historical water usage data, identify short-term and intermittent water usage patterns; When a water usage session ends and is identified as a short-term water usage mode, the combustion system is not immediately shut down. Instead, the system operating point is adjusted to a predetermined high-efficiency standby point in the low-load region of the current efficiency MAP. At the predetermined high-efficiency standby point, a minimum amount of combustion power and circulating water flow are maintained to keep the core temperature of the heat exchanger at a level higher than the ambient temperature and with high thermal efficiency. When the next water demand is predicted or detected, the system is rapidly loaded from the predetermined high-efficiency standby point to the target power.

6. A gas-fired hot water boiler based on high-efficiency heat exchange according to claim 5, characterized in that, The predetermined high-efficiency standby point It is not a fixed value; its selection is achieved by solving a constrained optimization problem whose objective is to find the feasible set under low load conditions. Find a work point This enables the power point to be reached from that point. Estimated reheat time Not exceeding the maximum allowed value At the same time, maximize the normalized thermal efficiency at that point. Mathematically described The constraints are .

7. A control method for a gas-fired hot water boiler, applicable to the gas-fired hot water boiler based on high-efficiency heat exchange as described in claim 6, characterized in that, Includes the following steps: S1: The system pre-stores a baseline efficiency MAP diagram based on the thermal characteristics of high-efficiency heat exchangers; S2: Real-time acquisition of system operating parameters and calculation of actual instantaneous thermal efficiency; S3: Continuously compare and learn from the actual instantaneous thermal efficiency with the theoretical value of the baseline efficiency MAP. When a continuous deviation is detected, correct it online and update it to the current efficiency MAP. S4: Based on the current efficiency MAP and the real-time load status of the system, dynamically generate an optimal efficiency control trajectory; S5: Control the combustion system and fan, and drive the system operating point to run along the optimal efficiency control trajectory.

8. The control method for a gas-fired hot water boiler according to claim 7, characterized in that, In step S4, the dynamic generation of the optimal efficiency control trajectory is achieved by solving a discrete-time model predictive control problem, in each control cycle. Seeking solutions for the future Predict the optimal control sequence in the time domain step by step. ; The constraints are: , , ; In the formula, For at any time The optimal control sequence obtained by solving; The future control input sequence to be optimized. , Indicates the future number Step control actions; To predict the length of the time domain; , Represents the instantaneous cost function of discretization; In order to be in Always looking towards the future The predicted value of the system state at any given time; Representing the dynamic model of a discrete-time system; and These represent the discretized inequality constraints and equality constraints, respectively.

9. The control method for a gas-fired hot water boiler according to claim 8, characterized in that, When multiple concurrent water usage disturbances are detected, step S4 further includes: Based on real-time calculated normalized water flow disturbance intensity The constraints and boundaries in the model predictive control problem are dynamically adjusted. Normalized boundary constraining water temperature fluctuations Adjusted to: In the formula, The upper limit of allowable fluctuations in base water temperature. For adjustment coefficients; In a strong disturbance transient, by increasing Slight fluctuations in water temperature are allowed to prioritize efficiency and system stability; however, this constraint is tightened in the later stages of the disturbance to quickly stabilize the water temperature.

10. The control method for a gas-fired hot water boiler according to claim 9, characterized in that, The high-efficiency standby point Dynamic selection involves evaluating a set of candidate low-load points in each decision cycle. Selection of comprehensive benefit score To achieve this, select the point with the highest score; The formula for calculating the comprehensive benefit score is as follows: ； In the formula, This represents the normalized thermal efficiency at the candidate point. To estimate the reheat time, The characteristic constant of reheat time, To normalize maintenance energy consumption, , , These are the weighting coefficients.

Images