Energy-saving high-voltage electrode vacuum boiler and intelligent control system thereof

Abstract

The invention relates to the technical field of intelligent energy-saving control, in particular to an energy-saving high-voltage electrode vacuum boiler and an intelligent control system of the energy-saving high-voltage electrode vacuum boiler. The control system is connected with the temperature sensor, the flow sensor, the pressure sensor and the vacuum pump; the water supply pipe is connected with the vacuum boiler; the pressure sensor is connected with the vacuum boiler and the control system; the vacuum boiler is connected with the pressure sensor, the water supply pipe, the flow sensor and the vacuum pump; the vacuum pump is connected with the vacuum boiler and the control system; and the flow sensor is connected with the vacuum boiler and the control system. Closed-loop feedback is formed through the control system, real-time optimization of operation parameters of the vacuum boiler is achieved, the comprehensive benefits of improving heat efficiency, reducing energy consumption, adapting to load fluctuation and prolonging the service life of equipment are achieved, and the automation level, reliability and safety of the vacuum boiler are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent energy-saving control, and in particular to an energy-saving high-voltage electrode vacuum boiler and an intelligent control system thereof. Background Art

[0002] Traditional high-voltage electrode vacuum boilers often cannot accurately match actual load demands. Due to the lack of an effective load forecasting mechanism, boilers usually operate at a relatively fixed power.

[0003] Chinese Patent Publication No. CN109184812A discloses a system and method for nuclear energy coupled with chemical energy power generation based on a two-loop boiler, comprising: a boiler connected to a gas-molten salt heat exchanger to form a boiler primary loop; the gas-molten salt heat exchanger is respectively connected to a cold molten salt tank and a hot molten salt tank; the cold molten salt tank and the hot molten salt tank are respectively connected to a molten salt-superheated steam heat exchanger to form a boiler secondary loop; the steam outlet of the nuclear island is divided into two paths, one of which is connected to a base load steam turbine and the other is connected to a molten salt-superheated steam heat exchanger; the steam outlet of the molten salt-superheated steam heat exchanger is connected to a coupling steam turbine; the base load steam turbine drives a first generator to generate electricity, and the coupling steam turbine drives a second generator to generate electricity; the present invention fully utilizes the high specific heat capacity of special gases and the process time shifting characteristics of molten salt heat storage, while improving the cycle thermal efficiency and power generation efficiency, while improving the safety of the coupled system operation and adaptability to variable operating conditions. However, this solution cannot adjust the relationship between vacuum degree and evaporation rate in real time according to actual operating conditions. At the same time, since the vacuum degree cannot be adjusted in time, it is difficult to adapt the appropriate vacuum degree according to actual production conditions. As a result, excessive vacuum degree leads to excessive energy consumption of the vacuum boiler. Summary of the Invention

[0004] To this end, the present invention provides an energy-saving high-voltage electrode vacuum boiler and its intelligent control system to overcome the problem in the prior art that it is difficult to adapt the appropriate vacuum degree according to actual production conditions, thereby excessive vacuum degree leads to excessive energy consumption of the vacuum boiler.

[0005] To achieve the above objectives, the present invention provides an energy-saving high-voltage electrode vacuum boiler, comprising:

[0006] A temperature sensor is connected to the vacuum boiler and the control system to collect the ambient temperature;

[0007] A control system, which is connected to a temperature sensor, a flow sensor, a pressure sensor, and a vacuum pump, and is used to control the parameters of the vacuum boiler;

[0008] a water supply pipe connected to the vacuum boiler and used for supplying water to the vacuum boiler;

[0009] A pressure sensor is connected to the vacuum boiler and the control system to collect steam pressure;

[0010] A vacuum boiler connected to a pressure sensor, a water supply pipe, a flow sensor, and a vacuum pump;

[0011] A vacuum pump, which is connected to the vacuum boiler and the control system and is used to pump the vacuum boiler into a vacuum state;

[0012] The flow sensor is connected to the vacuum boiler and the control system and is used to collect steam flow and steam flow change.

[0013] On the other hand, the present invention also provides an intelligent control system, the system comprising:

[0014] Data acquisition module, used to acquire boiler data;

[0015] a load prediction and control module for predicting the predicted load using a load prediction method and based on boiler data to obtain the predicted load, for controlling the vacuum boiler based on the predicted load using a vacuum boiler control method, and for calibrating the vacuum boiler control method;

[0016] An energy efficiency balancing module is used to construct a DQN algorithm model, obtain an optimal evaporation rate based on the DQN algorithm model, perform temperature adjustment on the DQN algorithm model, and perform load correction on the temperature adjustment process;

[0017] The coordination control module is used to calculate the adjustment of the vacuum pump speed according to the optimal evaporation rate, to control the vacuum pump, to perform a first correction on the calculation process of adjusting the vacuum pump speed, and to fine-tune the correction process of the vacuum boiler control method.

[0018] Furthermore, the load prediction and control module predicts the predicted load using a load prediction method and based on boiler data to obtain the predicted load. The load prediction method includes:

[0019] Step A1, constructing a boiler load prediction model to obtain a boiler load prediction model;

[0020] Step A2: Input the load data at time t in the boiler data into the boiler load prediction model to obtain the predicted load yc at time Z output by the boiler load prediction model. n , the predicted load at Z moments is output as the predicted load.

[0021] Furthermore, the load prediction and control module controls the vacuum boiler according to the predicted load using a vacuum boiler control method, wherein the vacuum boiler control method includes:

[0022] Step B1: Based on the predicted load yc n , expected load yq, weight coefficient λj, time sequence n, total time sequence Z, steam flow change △q, electrode power change △P and change sequence j are used to construct the defined objective function J2, and set Get the objective function J2;

[0023] Step B2: solving the optimal control sequence using a quadratic programming algorithm according to the objective function J2 to obtain the optimal control sequence, and controlling the vacuum boiler according to the optimal control sequence.

[0024] Furthermore, the load prediction and control module will predict the load yc when calibrating the vacuum boiler control method. n Compare with the preset load yy, 50MW≤yy≤100MW, judge the predicted load demand based on the comparison result, and calibrate the vacuum boiler control method based on the judgment result, where:

[0025] When yc n When ≤yy, the load prediction and control module determines that the predicted load demand is small and does not calibrate the vacuum boiler control method;

[0026] When yc n When yy is higher, the load prediction and control module determines that the load demand is large, corrects the vacuum boiler control method, corrects the weight coefficient λj according to the demand coefficient xq, obtains the corrected weight coefficient λjz, replaces the weight coefficient λj with the corrected weight coefficient λjz, and reconstructs the objective function J2, setting e is the base of natural logarithm, λjz=λj×xq.

[0027] Furthermore, when constructing the DQN algorithm model, the energy efficiency balancing module constructs the DQN algorithm model using a DQN algorithm model construction method, and the DQN algorithm model construction method includes:

[0028] Step C1, calculate the energy efficiency reward Rr based on the current thermal efficiency η, the target thermal efficiency ηm and the energy efficiency reward coefficient nxj1 to obtain the energy efficiency reward Rr, and set Rr = nxj1 × (η - ηm), 0 < nxj1 < 1;

[0029] Step C2, calculating the electrode life bonus Re based on the maximum expected life Lmax of the electrode, the current cumulative usage time Lc of the electrode, and the electrode life bonus coefficient nxj2 to obtain the electrode life bonus Re, setting Re = nxj2 × (Lmax - Lc), 0 < nxj2 < 1;

[0030] Step C3, calculating the final reward function R according to the electrode life reward Re and the energy efficiency reward Rr to obtain the final reward function R, and setting R = Rr + Re;

[0031] Step C4, initialize the parameters of the DQN network, obtain the initial DQN network model, calculate the target Q value y according to the final reward function R and the discount factor zky, set y = R + zky × maxa′Q2(si+1, a′, θ), 0 ≤ zky ≤ 1, and calculate the loss function L(θ) according to the number of historical experiences M, the predicted Q value Q1(si, ai; θ) in the si state, and the target Q value y in the si state, and set The loss function L(θ) is used as the loss function of the initial DQN network model to obtain the basic DQN network model. The initial DQN network model is trained through historical experience to obtain the trained DQN network model. The trained DQN network model is output as the DQN algorithm model to obtain the DQN algorithm model, where i is the order of the state and i is a positive integer.

[0032] Furthermore, when the energy efficiency balancing module obtains the optimal evaporation rate according to the DQN algorithm model, the boiler data is input into the DQN algorithm model to obtain the optimal evaporation rate output by the DQN algorithm model.

[0033] Furthermore, when the energy efficiency balancing module performs temperature adjustment on the DQN algorithm model, the ambient temperature Wh in the boiler data is compared with the preset maximum ambient temperature Whmax and the preset minimum ambient temperature Whmin, 23°C = Whmmmin, Whmax = 27°C, and the ambient temperature is judged based on the comparison result. The DQN algorithm model is temperature adjusted based on the judgment result, wherein:

[0034] When Whmin≤Wh≤Whmax, the energy efficiency balancing module determines that the ambient temperature is moderate and does not adjust the temperature of the DQN algorithm model;

[0035] When Wh<Whmin, the energy efficiency balance module determines that the ambient temperature is too low, adjusts the temperature of the DQN algorithm model, adjusts the energy efficiency reward Rr by the first temperature adjustment coefficient α, and obtains the first adjusted energy efficiency reward Rr1, adjusts the battery life reward Re by the second temperature adjustment coefficient β, and obtains the first adjusted battery life reward Re1, replaces the energy efficiency reward Rr and the battery life reward Re with the first adjusted energy efficiency reward Rr1 and the first adjusted battery life reward Re1 respectively, and recalculates the final reward function according to the first adjusted energy efficiency reward Rr1 and the first adjusted battery life reward Re1, setting α=1.34-0.34×e -(Whmin-Wh), β=0.78+0.21×e -(Whmin-Wh) , Rr1=α×Rr, Re1=Rr×β;

[0036] When Wh>Whmax, the energy efficiency balancing module determines that the ambient temperature is too high, performs temperature adjustment on the DQN algorithm model, adjusts the energy efficiency reward Rr by the third temperature adjustment coefficient γ, and obtains the second adjusted energy efficiency reward Rr2. The battery life reward Re is temperature adjusted by the second temperature adjustment coefficient μ to obtain the adjusted battery life reward Re2. The energy efficiency reward Rr and the battery life reward Re are replaced with the second adjusted energy efficiency reward Rr2 and the second adjusted battery life reward Re2, respectively. The final reward function is recalculated according to the second adjusted energy efficiency reward Rr2 and the second adjusted battery life reward Re2, and γ is set to 1.34-0.34×e -(Wh-Whmax) , μ=0.78+0.21×e -(Wh-Whmax) , Rr2=γ×Rr, Re2=Rr×μ.

[0037] Furthermore, the energy efficiency balancing module performs load correction during the temperature adjustment process, and calculates the predicted load yc n Compare with the maximum load ymax, ymax≤100MW, judge the future vacuum boiler load situation based on the comparison result, and make load correction for the temperature adjustment process based on the judgment result, where:

[0038] When yc n When ≤ymax, the energy efficiency balance module determines that the future load condition of the vacuum boiler is appropriate, and does not perform load correction in the temperature adjustment process;

[0039] When yc n >ymax, the energy efficiency balancing module determines that the future load of the vacuum boiler is inappropriate, performs load correction on the temperature adjustment process, cancels the temperature adjustment of the DQN algorithm, and uses the load correction coefficient fzb to load correct the energy efficiency reward Rr to obtain the corrected energy efficiency reward Rr3. Use the load correction coefficient fzb to load correct the battery life reward Re to obtain the corrected battery life reward Re3. The energy efficiency reward Rr and the battery life reward Re are replaced with the corrected energy efficiency reward Rr3 and the corrected battery life reward Re3 respectively, and the final reward function is recalculated based on the corrected energy efficiency reward Rr3 and the corrected battery life reward Re3, setting fzb=ymax / yc n , Rr3=fzb×Rr, Re3=fzb×Re.

[0040] Furthermore, when the coordination control module calculates the speed of the vacuum pump according to the optimal evaporation rate, it obtains the optimal evaporation rate Eopt output by the energy efficiency balance module, and calculates the target vacuum degree Vm according to the optimal evaporation rate Eopt, the first constant k and the second constant h, and sets Vm = (k / Eopt) 1 / h , calculate the future vacuum degree Vp according to the base e of the natural logarithm and the current vacuum pump speed Sc, and set The vacuum degree deviation ΔV is calculated according to the future vacuum degree Vp and the target vacuum degree Vm, and ΔV is set to Vt-Vp. The adjusted vacuum pump speed Sa is calculated according to the current vacuum pump speed Sc, the vacuum degree deviation ΔV and the vacuum proportional coefficient kp, and Sa is set to Sa=Sc+kp×ΔV. The vacuum pump is controlled according to the calculated adjusted vacuum pump speed Sa.

[0041] Compared with the existing technology, the beneficial effect of the present invention is that it realizes intelligent control of the entire process through the collaboration of multiple components. The temperature, pressure, and flow sensors collect data such as ambient temperature, steam pressure, steam flow and change in real time and transmit them to the control system. The control system dynamically adjusts the speed of the vacuum pump based on multi-source data to maintain the vacuum environment, and at the same time accurately controls the water flow rate of the water supply pipe, so that the vacuum boiler can operate efficiently under a vacuum state. The pressure sensor monitors the steam pressure in real time to ensure the safety of the system operation and avoid the risk of overpressure. The flow sensor assists in dynamically matching the evaporation rate and steam demand to reduce energy waste. The various components form a closed-loop feedback through the control system to realize real-time optimization of the vacuum boiler operating parameters, thereby achieving the comprehensive benefits of improving thermal efficiency, reducing energy consumption, adapting to load fluctuations, and extending equipment life, significantly improving the automation level, reliability and safety of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a schematic structural diagram of the energy-saving high-voltage electrode vacuum boiler of this embodiment;

[0043] Figure 2 Schematic diagram of the structure of the intelligent control system of this embodiment. DETAILED DESCRIPTION

[0044] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0045] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0046] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside", and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0047] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0048] See also Figure 1 As shown, it is a schematic structural diagram of the energy-saving high-voltage electrode vacuum boiler of this embodiment, and the device includes:

[0049] The temperature sensor 1 is connected to the vacuum boiler 5 and the control system 2 and is used to collect the ambient temperature;

[0050] A control system 2, which is connected to the temperature sensor 1, the flow sensor 7, the pressure sensor 4, and the vacuum pump 6, and is used to control the parameters of the vacuum boiler 1;

[0051] A water supply pipe 3, which is connected to the vacuum boiler 5 and is used to supply water to the vacuum boiler;

[0052] The pressure sensor 4 is connected to the vacuum boiler 5 and the control system 2 and is used to collect the steam pressure;

[0053] A vacuum boiler 5 connected to a pressure sensor 4, a water supply pipe 3, a flow sensor 7, and a vacuum pump 6;

[0054] A vacuum pump 6 connected to the vacuum boiler 5 and the control system 2, for pumping the vacuum boiler 1 into a vacuum state;

[0055] The flow sensor 7 is connected to the vacuum boiler 5 and the control system 2 and is used to collect the steam flow rate and the steam flow rate change.

[0056] Specifically, the present invention is applied to power generation in power stations, and realizes intelligent management and control of the entire process through the collaboration of multiple components. Temperature, pressure, and flow sensors collect data such as ambient temperature, steam pressure, steam flow and change in real time and transmit them to the control system. The control system dynamically adjusts the speed of the vacuum pump based on multi-source data to maintain a vacuum environment, and at the same time accurately controls the water flow rate of the water supply pipe, so that the vacuum boiler can operate efficiently under a vacuum state. The pressure sensor monitors the steam pressure in real time to ensure the safety of the system operation and avoid the risk of overpressure. The flow sensor assists in dynamically matching the evaporation rate and steam demand to reduce energy waste. Each component forms a closed-loop feedback through the control system to realize real-time optimization of the operating parameters of the vacuum boiler, thereby achieving the comprehensive benefits of improving thermal efficiency, reducing energy consumption, adapting to load fluctuations, and extending equipment life, and significantly improving the automation level, reliability and safety of the system.

[0057] See also Figure 2 As shown in FIG, which is a schematic diagram of the structure of the intelligent control system of this embodiment, the system includes:

[0058] Data acquisition module, used to acquire boiler data;

[0059] a load prediction and control module, configured to predict the predicted load using the load prediction method and based on the boiler data to obtain the predicted load, to control the vacuum boiler based on the predicted load using the vacuum boiler control method, and to calibrate the vacuum boiler control method, the load prediction and control module being connected to the data acquisition module;

[0060] An energy efficiency balancing module, configured to construct a DQN algorithm model, obtain an optimal evaporation rate based on the DQN algorithm model, perform temperature adjustment on the DQN algorithm model, and perform load correction during the temperature adjustment process. The energy efficiency balancing module is connected to the load prediction and control module;

[0061] The coordination control module is used to calculate the adjustment of the vacuum pump speed according to the optimal evaporation rate, to control the vacuum pump, to perform a first correction on the calculation process of adjusting the vacuum pump speed, and to fine-tune the correction process of the vacuum boiler control method. The coordination control module is connected to the energy efficiency balancing module.

[0062] Specifically, the intelligent control system of the present invention is applied to energy-saving high-voltage electrode vacuum boilers. By adjusting the relationship between vacuum degree and evaporation rate in real time, the safety and operational efficiency of the vacuum boiler are improved. The system obtains boiler data through the data acquisition module, and provides complete state representation for subsequent modules to avoid control deviations caused by information loss. The system also predicts load changes in advance through the load prediction and control module, avoiding the lag of traditional feedback control, allowing the boiler to adjust the electrode power and steam flow in advance, reducing energy efficiency losses during dynamic regulation, and adapting to time-varying conditions such as aging of boiler equipment and changes in fuel characteristics through an online correction mechanism. Factors ensure that the control algorithm is effective in the long term. The system also breaks through the local optimal limitations of traditional PID control through the energy efficiency balance module, and searches for the global optimal solution in the high-dimensional state space through the DQN algorithm to achieve the optimal balance between energy efficiency and electrode life, thereby avoiding the problem of excessive energy consumption of vacuum boilers. The vacuum degree and vacuum boiler parameters are adjusted in real time to ensure the adaptability of the vacuum degree and vacuum boiler parameters, thereby reducing energy loss. The system also dynamically matches the vacuum pump speed and the boiler evaporation rate through the coordinated control module to avoid the "vacuum over-regulation-evaporation rate fluctuation" problem caused by traditional independent control, thereby improving the overall efficiency of the system.

[0063] Specifically, when the data acquisition module acquires boiler data, the boiler data includes steam flow, electrode power, steam pressure, steam flow change, electrode power change, ambient temperature, current vacuum pump speed, adjusted vacuum degree and time to reach target vacuum degree. The data acquisition module acquires the steam flow and steam flow change through a flow sensor, and the data acquisition module acquires the electrode power and electrode power change through a power sensor. This embodiment does not limit the specific installation position of the power sensor. Those skilled in the art can set it according to actual needs. For example, the power sensor can be installed on the vacuum boiler. The data acquisition module acquires the current vacuum pump speed through a speed sensor. This embodiment does not limit the specific installation position of the speed sensor. The specific installation position of the vacuum sensor is not limited, and those skilled in the art can set it according to actual needs. For example, the speed sensor can be installed on the vacuum boiler, and the data acquisition module obtains the steam pressure through the pressure sensor. The data acquisition module obtains the ambient temperature through the temperature sensor. The data acquisition module obtains the adjusted vacuum degree through the vacuum sensor. This embodiment does not limit the specific installation position of the vacuum sensor. Those skilled in the art can set it according to actual needs. For example, the vacuum sensor can be installed on the vacuum boiler, and the data acquisition module obtains the time to reach the target vacuum degree through a timer. This embodiment does not limit the specific installation position of the timer. Those skilled in the art can set it according to actual needs. For example, the timer can be installed on the vacuum boiler.

[0064] Specifically, the load prediction and control module uses a load prediction method and predicts the predicted load based on boiler data to obtain the predicted load. The load prediction method includes:

[0065] Step A1, constructing a boiler load prediction model to obtain a boiler load prediction model;

[0066] Step A2: Input the load data at time t in the boiler data into the boiler load prediction model to obtain the predicted load yc at time Z output by the boiler load prediction model. n , the predicted load at Z moments is output as the predicted load.

[0067] Specifically, the load data refers to the load data of the vacuum boiler, and the load data includes steam flow, electrode power, and steam pressure. The steam flow refers to the flow of steam generated by the vacuum boiler, the electrode power refers to the power of the vacuum boiler electrode, and the steam pressure refers to the pressure of the steam generated by the vacuum boiler. This embodiment does not limit the specific construction method of constructing the boiler load prediction model. Those skilled in the art can set it according to actual needs. For example, the autoregressive integral moving average model in time series analysis can be selected to construct the boiler load prediction model. The n is the order of the time, n is a positive integer, and the maximum value of n is Z.

[0068] Specifically, the load prediction and control module controls the vacuum boiler according to the predicted load using a vacuum boiler control method, wherein the vacuum boiler control method includes:

[0069] Step B1: Based on the predicted load yc n , expected load yq, weight coefficient λj, time sequence n, total time sequence Z, steam flow change △q, electrode power change △P and change sequence j are used to construct the defined objective function J2, and set Get the objective function J2;

[0070] Step B2: solving the optimal control sequence using a quadratic programming algorithm according to the objective function J2 to obtain the optimal control sequence, and controlling the vacuum boiler according to the optimal control sequence.

[0071] Specifically, the change order refers to the change order of the steam flow rate change and the electrode power change. The expected load refers to the expected vacuum boiler load value at time n. This embodiment does not limit the specific value of the expected load. Those skilled in the art can set it according to actual needs, for example, the value of the expected load can be set according to specific parameters of the vacuum boiler. The total number of moments N refers to the total number of moments predicted by the boiler load prediction model. The steam flow rate change refers to the steam flow rate change of the vacuum boiler at time n and time n-1. The electrode power change refers to the electrode power change of the vacuum boiler at time n and time n-1. The quadratic programming algorithm refers to an algorithm for solving the optimal control sequence. This embodiment does not limit the specific control method for controlling the vacuum boiler according to the optimal control sequence. Those skilled in the art can set it according to actual needs, for example, the vacuum boiler can be controlled according to the optimal control sequence using a PID control algorithm. The weight coefficient refers to the coefficient used to construct the objective function. This embodiment does not limit the specific value of the weight coefficient. Those skilled in the art can set it according to actual needs, for example, the weight coefficient can be set according to the model of the vacuum boiler.

[0072] Specifically, when the load prediction and control module calibrates the vacuum boiler control method, the predicted load yc n Compare with the preset load yy, 50MW≤yy≤100MW, judge the predicted load demand based on the comparison result, and calibrate the vacuum boiler control method based on the judgment result, where:

[0073] When yc n When ≤yy, the load prediction and control module determines that the predicted load demand is small and does not calibrate the vacuum boiler control method;

[0074] When yc n When yy is higher, the load prediction and control module determines that the load demand is large, corrects the vacuum boiler control method, corrects the weight coefficient λj according to the demand coefficient xq, obtains the corrected weight coefficient λjz, replaces the weight coefficient λj with the corrected weight coefficient λjz, and reconstructs the objective function J2, setting e is the base of natural logarithm, λjz=λj×xq.

[0075] Specifically, the preset load refers to a preset value for judging the predicted load demand situation, and the predicted load demand situation refers to the size of the predicted load demand judged based on the predicted load and the preset load. The predicted load demand situation includes the predicted load demand situation being a small load demand and the predicted load demand situation being a large load demand.

[0076] Specifically, the load prediction and control module calibrates the objective function J2 by ​​monitoring the predicted load. When the load demand is large, the weight coefficient value is reduced to make the control of the vacuum boiler more precise, thereby improving the control accuracy of the vacuum boiler.

[0077] Specifically, when constructing the DQN algorithm model, the energy efficiency balancing module constructs the DQN algorithm model using a DQN algorithm model construction method, and the DQN algorithm model construction method includes:

[0078] Step C1, calculate the energy efficiency reward Rr based on the current thermal efficiency η, the target thermal efficiency ηm and the energy efficiency reward coefficient nxj1 to obtain the energy efficiency reward Rr, and set Rr = nxj1 × (η - ηm), 0 < nxj1 < 1;

[0079] Step C2, calculating the electrode life bonus Re based on the maximum expected life Lmax of the electrode, the current cumulative usage time Lc of the electrode, and the electrode life bonus coefficient nxj2 to obtain the electrode life bonus Re, setting Re = nxj2 × (Lmax - Lc), 0 < nxj2 < 1;

[0080] Step C3, calculating the final reward function R according to the electrode life reward Re and the energy efficiency reward Rr to obtain the final reward function R, and setting R = Rr + Re;

[0081] Step C4, initialize the parameters of the DQN network, obtain the initial DQN network model, calculate the target Q value y according to the final reward function R and the discount factor zky, set y = R + zky × maxa′Q2(si+1, a′, θ), 0 ≤ zky ≤ 1, and calculate the loss function L(θ) according to the number of historical experiences M, the predicted Q value Q1(si, ai; θ) in the si state, and the target Q value y in the si state, and set The loss function L(θ) is used as the loss function of the initial DQN network model to obtain the basic DQN network model. The initial DQN network model is trained through historical experience to obtain the trained DQN network model. The trained DQN network model is output as the DQN algorithm model to obtain the DQN algorithm model, where i is the order of the state and i is a positive integer.

[0082] Specifically, θ is the main network parameter of the initial DQN network model, and the main network parameter refers to the weight of the neural network that approximately predicts the Q value. This embodiment does not limit the specific value of θ. Those skilled in the art can set it according to actual needs, such as setting the specific value of θ according to the parameters connecting the neurons in each layer. The DQN algorithm model construction method refers to a model construction method for constructing a DQN algorithm model. The current thermal efficiency refers to the ratio of the input energy to the effective output energy of the vacuum boiler in the current operating state. This example does not limit the method for obtaining the current thermal efficiency. Those skilled in the art can set it according to actual needs, such as based on The current thermal efficiency is obtained based on the ratio of the effective output heat and the total input energy of the vacuum boiler. The target thermal efficiency refers to the energy efficiency index that the boiler should achieve under ideal working conditions. This embodiment does not limit the specific value of the target thermal efficiency. Those skilled in the art can set it according to actual needs. For example, the target thermal efficiency value can be set by the boiler design parameters or industry standards. The maximum expected life of the electrode refers to the theoretical design life of the electrode. This embodiment does not limit the specific method for obtaining the maximum expected life of the electrode. Those skilled in the art can set it according to actual needs. For example, the maximum expected life of the motor can be obtained according to the instruction manual of the vacuum boiler. The current cumulative use of the electrode The usage time refers to the cumulative running time of the electrode since it was put into use. This embodiment does not limit the specific method of obtaining the current cumulative usage time of the electrode. Those skilled in the art can set it according to actual needs, such as obtaining the current cumulative usage time of the electrode through a timer. This embodiment does not limit the specific method of initializing the parameters of the DQN network. Those skilled in the art can set it according to actual needs, such as initializing the parameters of the DQN network by random initialization. The historical experience refers to the initial DQN network model interacting with the vacuum boiler environment through the ∈-greedy strategy, randomly exploring actions with ∈ probability, selecting the action with the largest current Q value with 1-∈ probability, and executing the action. The state, action and next state are then stored to obtain the training data. The number of historical experiences refers to the number of historical experience data. The predicted Q value in the si state refers to the state-action value function predicted in the si state. The target Q value in the si state refers to the state-action value function of the target in the si state. The predicted Q value in the si state is obtained through historical experience. The maxa′Q2(si+1, a′, θ) refers to the maximum value of the Q value of all possible actions a′ in the state si+1. The all possible actions refer to all possible actions randomly explored by the DQN algorithm model with ∈ probability. The discount factor refers to the coefficient used to calculate the target Q value.

[0083] Specifically, when the energy efficiency balancing module obtains the optimal evaporation rate according to the DQN algorithm model, the boiler data is input into the DQN algorithm model to obtain the optimal evaporation rate output by the DQN algorithm model.

[0084] Specifically, when the energy efficiency balance module adjusts the temperature of the DQN algorithm model, it compares the ambient temperature Wh in the boiler data with the preset maximum ambient temperature Whmax and the preset minimum ambient temperature Whmin, 23°C = Whmmmin, Whmax = 27°C, and judges the ambient temperature based on the comparison result. The DQN algorithm model is temperature-adjusted based on the judgment result, wherein:

[0085] When Whmin≤Wh≤Whmax, the energy efficiency balancing module determines that the ambient temperature is moderate and does not adjust the temperature of the DQN algorithm model;

[0086] When Wh<Whmin, the energy efficiency balance module determines that the ambient temperature is too low, adjusts the temperature of the DQN algorithm model, adjusts the energy efficiency reward Rr by the first temperature adjustment coefficient α, and obtains the first adjusted energy efficiency reward Rr1, adjusts the battery life reward Re by the second temperature adjustment coefficient β, and obtains the first adjusted battery life reward Re1, replaces the energy efficiency reward Rr and the battery life reward Re with the first adjusted energy efficiency reward Rr1 and the first adjusted battery life reward Re1 respectively, and recalculates the final reward function according to the first adjusted energy efficiency reward Rr1 and the first adjusted battery life reward Re1, setting α=1.34-0.34×e -(Whmin-Wh) , β=0.78+0.21×e -(Whmin-Wh) , Rr1=α×Rr, Re1=Rr×β;

[0087] When Wh>Whmax, the energy efficiency balancing module determines that the ambient temperature is too high, performs temperature adjustment on the DQN algorithm model, adjusts the energy efficiency reward Rr by the third temperature adjustment coefficient γ, and obtains the second adjusted energy efficiency reward Rr2. The battery life reward Re is temperature adjusted by the second temperature adjustment coefficient μ to obtain the adjusted battery life reward Re2. The energy efficiency reward Rr and the battery life reward Re are replaced with the second adjusted energy efficiency reward Rr2 and the second adjusted battery life reward Re2, respectively. The final reward function is recalculated according to the second adjusted energy efficiency reward Rr2 and the second adjusted battery life reward Re2, and γ is set to 1.34-0.34×e -(Wh-Whmax) , μ=0.78+0.21×e -(Wh-Whmax) , Rr2=γ×Rr, Re2=Rr×μ.

[0088] Specifically, the ambient temperature refers to the temperature of the environment surrounding the vacuum boiler, the preset maximum ambient temperature refers to the maximum preset value for judging the ambient temperature condition, the preset minimum ambient temperature refers to the minimum preset value for judging the ambient temperature condition, the ambient temperature condition refers to the temperature condition of the environment surrounding the vacuum boiler judged based on the ambient temperature and the preset maximum ambient temperature and the preset minimum ambient temperature, and the ambient temperature condition includes the ambient temperature condition being moderate, the ambient temperature condition being too high, and the ambient temperature condition being too low.

[0089] Specifically, the energy efficiency balancing module monitors the ambient temperature. When the ambient temperature is too high or too low, it adjusts the energy efficiency reward and the battery life reward, increases the value of the energy efficiency reward and reduces the value of the battery life reward, so that the energy efficiency reward accounts for a higher proportion in the final reward function, and reduces the impact of temperature on the predicted Q value output, so that a stricter output control action can be achieved when the ambient temperature is not moderate.

[0090] Specifically, the energy efficiency balancing module performs load correction during the temperature adjustment process, and changes the predicted load yc n Compare with the maximum load ymax, ymax≤100MW, judge the future vacuum boiler load situation based on the comparison result, and make load correction for the temperature adjustment process based on the judgment result, where:

[0091] When yc n When ≤ymax, the energy efficiency balance module determines that the future load condition of the vacuum boiler is appropriate, and does not perform load correction in the temperature adjustment process;

[0092] When yc n >ymax, the energy efficiency balancing module determines that the future load of the vacuum boiler is inappropriate, performs load correction on the temperature adjustment process, cancels the temperature adjustment of the DQN algorithm, and uses the load correction coefficient fzb to load correct the energy efficiency reward Rr to obtain the corrected energy efficiency reward Rr3. Use the load correction coefficient fzb to load correct the battery life reward Re to obtain the corrected battery life reward Re3. The energy efficiency reward Rr and the battery life reward Re are replaced with the corrected energy efficiency reward Rr3 and the corrected battery life reward Re3 respectively, and the final reward function is recalculated based on the corrected energy efficiency reward Rr3 and the corrected battery life reward Re3, setting fzb=ymax / yc n , Rr3=fzb×Rr, Re3=fzb×Re.

[0093] Specifically, the maximum load refers to a preset value for judging the future load condition of the vacuum boiler. The future load condition of the vacuum boiler refers to the future load condition of the vacuum boiler judged by the energy efficiency balancing module based on the predicted load and the maximum load. The future load condition of the vacuum boiler includes the future load condition of the vacuum boiler being appropriate and the future load condition of the vacuum boiler being inappropriate.

[0094] Specifically, the energy efficiency balancing module monitors the predicted load. When the future vacuum boiler load is inappropriate, the module adjusts the final reward function to reduce the impact of the predicted load on the optimal evaporation rate calculation, thereby obtaining a more accurate optimal evaporation rate.

[0095] Specifically, when the coordination control module calculates the speed of the vacuum pump according to the optimal evaporation rate, it obtains the optimal evaporation rate Eopt output by the energy efficiency balance module, and calculates the target vacuum degree Vm according to the optimal evaporation rate Eopt, the first constant k and the second constant h, and sets Vm = (k / Eopt) 1 / h , calculate the future vacuum degree Vp according to the base e of the natural logarithm and the current vacuum pump speed Sc, and set The vacuum degree deviation ΔV is calculated according to the future vacuum degree Vp and the target vacuum degree Vm, and ΔV is set to Vt-Vp. The adjusted vacuum pump speed Sa is calculated according to the current vacuum pump speed Sc, the vacuum degree deviation ΔV and the vacuum proportional coefficient kp, and Sa is set to Sa=Sc+kp×ΔV. The vacuum pump is controlled according to the calculated adjusted vacuum pump speed Sa.

[0096] Specifically, the first constant refers to a constant used to calculate the target vacuum degree, and the second constant refers to a constant used to calculate the target vacuum degree. This embodiment does not limit the specific values ​​of the first constant and the second constant. Those skilled in the art may set them according to actual needs. For example, the specific values ​​of the first constant and the second constant may be set based on experiments on the relationship between vacuum degree and evaporation rate. The current vacuum pump speed refers to the real-time speed of the vacuum pump when calculating the future vacuum degree. The vacuum proportional coefficient refers to the proportional coefficient used to calculate the adjusted vacuum pump speed. This embodiment does not limit the specific value of the vacuum proportional coefficient. Those skilled in the art may set it according to actual needs. For example, the specific value of the vacuum proportional coefficient may be set based on expert experience in calculating the adjusted vacuum pump speed. This embodiment does not limit the specific control method for controlling the vacuum pump based on the calculated adjusted vacuum pump speed Sa. Those skilled in the art may set it according to actual needs. For example, a PID control algorithm may be used to control the vacuum pump based on the calculated adjusted vacuum pump speed Sa. The PID control algorithm dynamically adjusts the vacuum pump speed through a linear combination of the proportional, integral, and differential aspects of the system error, so that the vacuum pump output is as close as possible to the speed value of the adjusted vacuum pump speed.

[0097] Specifically, the coordination control module monitors the vacuum degree and rotation speed of the vacuum pump in real time, and continuously adjusts the rotation speed of the vacuum pump to make the vacuum degree of the vacuum boiler meet the demand as much as possible.

[0098] Specifically, when the coordination control module performs the first correction on the calculation process of adjusting the speed of the vacuum pump, the vacuum difference Zc is calculated according to the adjusted vacuum degree Vs and the target vacuum degree Vm in the boiler data, and Zc=|Vs-Vm| is set. The vacuum difference Zc is compared with the preset vacuum difference Zc0, 0.01MPa≤Zc0≤0.15MPa, and the vacuum adjustment error is judged according to the comparison result. The calculation process of adjusting the speed of the vacuum pump is first corrected according to the judgment result, wherein:

[0099] When Zc≤Zc0, the coordination control module determines that the vacuum adjustment error is appropriate, and does not perform the first correction on the calculation process of adjusting the vacuum pump speed;

[0100] When Zc>Zc0, the coordination control module determines that the vacuum adjustment error is inappropriate, and performs a first correction on the calculation process of adjusting the vacuum pump speed. If Vs>Vm, the target vacuum degree Vm is first corrected by the first correction coefficient Dj1 to obtain the first corrected target vacuum degree Vm1, and the target vacuum degree Vm is replaced with the first corrected target vacuum degree Vm1, and the future vacuum degree Vp is recalculated according to the first corrected target vacuum degree Vm1. If Vs<Vm, the target vacuum degree Vm is second corrected by the second correction coefficient Dj2 to obtain the second corrected target vacuum degree Vm2, and the target vacuum degree Vm is replaced with the second corrected target vacuum degree Vm2, and the future vacuum degree Vp is recalculated according to the second corrected target vacuum degree Vm2, setting Vm1=Vm×Dj1, Dj1=1-Vm / Vs, Vm2=Vm×Dj2, Dj2=Vm / Vs.

[0101] Specifically, the adjusted vacuum degree refers to the vacuum degree of the vacuum boiler measured after the vacuum pump is controlled according to the calculated adjusted vacuum pump speed Sa, the preset vacuum difference refers to the preset value for judging the vacuum adjustment error situation, the vacuum adjustment error situation refers to the error between the vacuum degree of the vacuum boiler measured after the vacuum pump is controlled according to the calculated adjusted vacuum pump speed Sa and the target vacuum degree, the vacuum adjustment error situation includes the vacuum adjustment error situation being a suitable adjustment error and the vacuum adjustment error situation being an inappropriate adjustment error.

[0102] Specifically, the coordination control module determines the error between the adjusted vacuum degree and the target vacuum degree through real-time monitoring of the adjusted vacuum degree. When the adjustment error is inappropriate, the target vacuum degree is adjusted so that the adjusted vacuum degree can be more in line with the requirements of the optimal evaporation rate. If the adjustment error is inappropriate and the adjusted vacuum degree is greater than the target vacuum degree, the target vacuum degree is reduced to make the adjusted vacuum degree more in line with the requirements of the optimal evaporation rate. If the adjustment error is inappropriate and the adjusted vacuum degree is less than the target vacuum degree, the target vacuum degree is increased to make the adjusted vacuum degree more in line with the requirements of the optimal evaporation rate, thereby achieving the purpose of intelligently controlling the vacuum boiler to save energy and have high working efficiency.

[0103] Specifically, when fine-tuning the calibration process of the vacuum boiler control method, the coordination control module compares the target vacuum degree reaching time Ty in the boiler data with the preset reaching time Ty0, 10min≤Ty0≤30min, and judges the time taken to reach the target vacuum degree based on the comparison result. The calibration process of the vacuum boiler control method is fine-tuned based on the judgment result, wherein:

[0104] When Ty≤Ty0, the coordinated control module determines that the time taken to reach the target vacuum degree is short, and does not perform fine calibration on the calibration process of the vacuum boiler control method;

[0105] When Ty>Ty0, the coordination control module determines that the time taken to reach the target vacuum degree is long, and performs fine calibration on the calibration process of the vacuum boiler control method. According to the fine calibration coefficient gj, the preset load yy is fine-calibrated to obtain the fine-calibrated preset load yy1, and the preset load yy is replaced with the fine-calibrated preset load yy1, and the predicted load yc is replaced with the fine-calibrated preset load yy1. n Re-compare with the preset load yy1 after fine calibration, and set gj = Ty / Ty0, yy1 = yy×gj.

[0106] Specifically, the time to reach the target vacuum degree refers to the time required for the vacuum boiler to reach the target vacuum degree after adjusting the vacuum pump. The preset reaching time refers to the preset value for judging the time-consuming situation of reaching the target vacuum degree. The time-consuming situation of reaching the target vacuum degree refers to the length of time consumed to reach the target vacuum degree judged based on the target vacuum degree time and the preset reaching time. The time-consuming situation of reaching the target vacuum degree includes the situation where the time-consuming situation of reaching the target vacuum degree is short and the situation where the time-consuming situation of reaching the target vacuum degree is long.

[0107] Specifically, the coordinated control module calibrates the correction process of the vacuum boiler control method by monitoring the time to reach the target vacuum degree. When the time taken to reach the target vacuum degree is long, the preset load value is increased to make it more difficult to judge the predicted load demand as a large load demand, thereby making the correction judgment of the vacuum boiler control method more accurate, thereby improving the control accuracy of the vacuum boiler.

[0108] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. An energy-saving high-voltage electrode vacuum boiler, characterized in that: include: A temperature sensor is connected to the vacuum boiler and the control system to collect the ambient temperature; A control system, which is connected to a temperature sensor, a flow sensor, a pressure sensor, and a vacuum pump, and is used to control the parameters of the vacuum boiler; a water supply pipe connected to the vacuum boiler and used for supplying water to the vacuum boiler; A pressure sensor is connected to the vacuum boiler and the control system to collect steam pressure; A vacuum boiler is connected to a pressure sensor, a water supply pipe, a flow sensor, and a vacuum pump; A vacuum pump, which is connected to the vacuum boiler and the control system and is used to pump the vacuum boiler into a vacuum state; The flow sensor is connected to the vacuum boiler and the control system and is used to collect steam flow and steam flow change.

2. An intelligent control system for the energy-saving high-voltage electrode vacuum boiler according to claim 1, characterized in that: The system comprises: Data acquisition module, used to acquire boiler data; a load prediction and control module for predicting the predicted load using a load prediction method and based on boiler data to obtain the predicted load, for controlling the vacuum boiler based on the predicted load using a vacuum boiler control method, and for calibrating the vacuum boiler control method; An energy efficiency balancing module is used to construct a DQN algorithm model, obtain an optimal evaporation rate based on the DQN algorithm model, perform temperature adjustment on the DQN algorithm model, and perform load correction on the temperature adjustment process; The coordination control module is used to calculate the adjustment of the vacuum pump speed according to the optimal evaporation rate, to control the vacuum pump, to perform a first correction on the calculation process of adjusting the vacuum pump speed, and to fine-tune the correction process of the vacuum boiler control method.

3. The energy-saving high-voltage electrode vacuum boiler and its intelligent control system according to claim 2, characterized in that: The load prediction and control module predicts the predicted load using a load prediction method and based on boiler data to obtain the predicted load. The load prediction method includes: Step A1, constructing a boiler load prediction model to obtain a boiler load prediction model; Step A2: Input the load data at time t in the boiler data into the boiler load prediction model to obtain the predicted load yc at time Z output by the boiler load prediction model. n , the predicted load at Z moments is output as the predicted load.

4. The energy-saving high-voltage electrode vacuum boiler and its intelligent control system according to claim 3 is characterized in that: The load prediction and control module controls the vacuum boiler according to the predicted load using a vacuum boiler control method, wherein the vacuum boiler control method includes: Step B1: Based on the predicted load yc n , expected load yq, weight coefficient λj, time sequence n, total time sequence Z, steam flow change △q, electrode power change △P and change sequence j are used to construct the defined objective function J2, and set Get the objective function J2; Step B2: solving the optimal control sequence using a quadratic programming algorithm according to the objective function J2 to obtain the optimal control sequence, and controlling the vacuum boiler according to the optimal control sequence.

5. The energy-saving high-voltage electrode vacuum boiler and its intelligent control system according to claim 4 is characterized in that: When the load prediction and control module calibrates the vacuum boiler control method, the predicted load yc n Compare with the preset load yy, 50MW≤yy≤100MW, judge the predicted load demand based on the comparison result, and calibrate the vacuum boiler control method based on the judgment result, where: When yc n When ≤yy, the load prediction and control module determines that the predicted load demand is small and does not calibrate the vacuum boiler control method; When yc n When yy is higher, the load prediction and control module determines that the load demand is large, corrects the vacuum boiler control method, corrects the weight coefficient λj according to the demand coefficient xq, obtains the corrected weight coefficient λjz, replaces the weight coefficient λj with the corrected weight coefficient λjz, and reconstructs the objective function J2, setting e is the base of natural logarithm, λjz=λj×xq.

6. The energy-saving high-voltage electrode vacuum boiler and its intelligent control system according to claim 5, characterized in that: When constructing the DQN algorithm model, the energy efficiency balancing module constructs the DQN algorithm model using a DQN algorithm model construction method, and the DQN algorithm model construction method includes: Step C1, calculate the energy efficiency reward Rr based on the current thermal efficiency η, the target thermal efficiency ηm and the energy efficiency reward coefficient nxj1 to obtain the energy efficiency reward Rr, and set Rr = nxj1 × (η - ηm), 0 < nxj1 < 1; Step C2, calculating the electrode life bonus Re based on the maximum expected life Lmax of the electrode, the current cumulative usage time Lc of the electrode, and the electrode life bonus coefficient nxj2 to obtain the electrode life bonus Re, setting Re = nxj2 × (Lmax - Lc), 0 < nxj2 < 1; Step C3, calculating the final reward function R according to the electrode life reward Re and the energy efficiency reward Rr to obtain the final reward function R, and setting R = Rr + Re; Step C4, initialize the parameters of the DQN network, obtain the initial DQN network model, calculate the target Q value y according to the final reward function R and the discount factor zky, set y = R + zky × maxa′Q2(si+1, a′, θ), 0 ≤ zky ≤ 1, and calculate the loss function L(θ) according to the number of historical experiences M, the predicted Q value Q1(si, ai; θ) in the si state, and the target Q value y in the si state, and set The loss function L(θ) is used as the loss function of the initial DQN network model to obtain the basic DQN network model. The initial DQN network model is trained through historical experience to obtain the trained DQN network model. The trained DQN network model is output as the DQN algorithm model to obtain the DQN algorithm model, where i is the order of the state and i is a positive integer.

7. The energy-saving high-voltage electrode vacuum boiler and its intelligent control system according to claim 6, characterized in that: When the energy efficiency balancing module obtains the optimal evaporation rate according to the DQN algorithm model, the boiler data is input into the DQN algorithm model to obtain the optimal evaporation rate output by the DQN algorithm model.

8. The energy-saving high-voltage electrode vacuum boiler and its intelligent control system according to claim 7, characterized in that: When the energy efficiency balance module adjusts the temperature of the DQN algorithm model, it compares the ambient temperature Wh in the boiler data with the preset maximum ambient temperature Whmax and the preset minimum ambient temperature Whmin, 23°C = Whmmmin, Whmax = 27°C, and judges the ambient temperature according to the comparison result. The DQN algorithm model is temperature-adjusted according to the judgment result, wherein: When Whmin≤Wh≤Whmax, the energy efficiency balancing module determines that the ambient temperature is moderate and does not adjust the temperature of the DQN algorithm model; When Wh<Whmin, the energy efficiency balance module determines that the ambient temperature is too low, adjusts the temperature of the DQN algorithm model, adjusts the energy efficiency reward Rr by the first temperature adjustment coefficient α, and obtains the first adjusted energy efficiency reward Rr1, adjusts the battery life reward Re by the second temperature adjustment coefficient β, and obtains the first adjusted battery life reward Re1, replaces the energy efficiency reward Rr and the battery life reward Re with the first adjusted energy efficiency reward Rr1 and the first adjusted battery life reward Re1 respectively, and recalculates the final reward function according to the first adjusted energy efficiency reward Rr1 and the first adjusted battery life reward Re1, setting α=1.34-0.34×e -(Whmin-Wh) , β=0.78+0.21×e -(Whmin-Wh) , Rr1=α×Rr, Re1=Rr×β; When Wh>Whmax, the energy efficiency balancing module determines that the ambient temperature is too high, performs temperature adjustment on the DQN algorithm model, adjusts the energy efficiency reward Rr by the third temperature adjustment coefficient γ, and obtains the second adjusted energy efficiency reward Rr2. The battery life reward Re is temperature adjusted by the second temperature adjustment coefficient μ to obtain the adjusted battery life reward Re2. The energy efficiency reward Rr and the battery life reward Re are replaced with the second adjusted energy efficiency reward Rr2 and the second adjusted battery life reward Re2, respectively. The final reward function is recalculated according to the second adjusted energy efficiency reward Rr2 and the second adjusted battery life reward Re2, and γ is set to 1.34-0.34×e -(Wh-Whmax) , μ=0.78+0.21×e -(Wh-Whmax) , Rr2=γ×Rr, Re2=Rr×μ.

9. The energy-saving high-voltage electrode vacuum boiler and its intelligent control system according to claim 8, characterized in that: The energy efficiency balancing module performs load correction during the temperature adjustment process, and calculates the predicted load yc n Compare with the maximum load ymax, ymax≤100MW, judge the future vacuum boiler load situation based on the comparison result, and make load correction for the temperature adjustment process based on the judgment result, where: When yc n When ≤ymax, the energy efficiency balance module determines that the future load condition of the vacuum boiler is appropriate, and does not perform load correction in the temperature adjustment process; When yc n >ymax, the energy efficiency balancing module determines that the future load of the vacuum boiler is inappropriate, performs load correction on the temperature adjustment process, cancels the temperature adjustment of the DQN algorithm, and uses the load correction coefficient fzb to load correct the energy efficiency reward Rr to obtain the corrected energy efficiency reward Rr3. Use the load correction coefficient fzb to load correct the battery life reward Re to obtain the corrected battery life reward Re3. The energy efficiency reward Rr and the battery life reward Re are replaced with the corrected energy efficiency reward Rr3 and the corrected battery life reward Re3 respectively, and the final reward function is recalculated based on the corrected energy efficiency reward Rr3 and the corrected battery life reward Re3, setting fzb=ymax / yc n , Rr3=fzb×Rr, Re3=fzb×Re.

10. The energy-saving high-voltage electrode vacuum boiler and its intelligent control system according to claim 9, characterized in that: When the coordination control module calculates the speed of the vacuum pump according to the optimal evaporation rate, it obtains the optimal evaporation rate Eopt output by the energy efficiency balance module, and calculates the target vacuum degree Vm according to the optimal evaporation rate Eopt, the first constant k and the second constant h, and sets Vm = (k / Eopt) 1 / h , calculate the future vacuum degree Vp according to the base e of the natural logarithm and the current vacuum pump speed Sc, and set The vacuum degree deviation ΔV is calculated according to the future vacuum degree Vp and the target vacuum degree Vm, and ΔV is set to Vt-Vp. The adjusted vacuum pump speed Sa is calculated according to the current vacuum pump speed Sc, the vacuum degree deviation ΔV and the vacuum proportional coefficient kp, and Sa is set to Sa=Sc+kp×ΔV. The vacuum pump is controlled according to the calculated adjusted vacuum pump speed Sa.

Images